{"forum": "maQs92RfZ8", "year": 2026, "status": "rejected", "title": "Exploring Federated Pruning for Large Language Models", "authors": ["Pengxin Guo", "Yinong Wang", "Wei Li", "Mengting Liu", "Ming Li", "Jinkai Zheng", "Liangqiong Qu"], "authorids": ["~Pengxin_Guo1", "~Yinong_Wang1", "~Wei_Li82", "~Mengting_Liu6", "~Ming_Li21", "~Jinkai_Zheng1", "~Liangqiong_Qu2"], "authors_source": "OpenReview API", "abstract": "LLM pruning has emerged as a promising technology for compressing LLMs, enabling their deployment on resource-limited devices. However, current methodologies typically require access to public calibration samples, which can be challenging to obtain in privacy-sensitive domains. To address this issue, we introduce FedPrLLM, a comprehensive federated pruning framework designed for the privacy-preserving compression of LLMs. In FedPrLLM, each client only needs to calculate a pruning mask matrix based on its local calibration data and share it with the server to prune the global model. This approach allows for collaborative pruning of the global model with the knowledge of each client while maintaining local data privacy. Additionally, we conduct extensive experiments to explore various possibilities within the FedPrLLM framework, including different comparison groups, pruning strategies, and the decision to scale weights. Our extensive evaluation reveals that one-shot pruning with layer comparison and no weight scaling is the optimal choice within the FedPrLLM framework. We hope our work will help guide future efforts in pruning LLMs in privacy-sensitive fields. Our code is available at https://anonymous.4open.science/r/FedPrLLM-15594.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=maQs92RfZ8", "arxiv_id": "2505.13547", "paper_pdf": "papers/maQs92RfZ8.pdf", "paper_pdf_sha256": "0f832bd2d83e0aa83cc5bca4f46bcf23535fe22aacba18b90418dccb761e7e30", "paper_pdf_bytes": 977965, "paper_pdf_source": "openreview", "code_url": "https://github.com/Pengxin-Guo/FedPrLLM", "code_repository": "Pengxin-Guo/FedPrLLM", "code_commit": "7e91d12e0e3250a2790e534d15140ba163cf7739", "code_archive": "repos/maQs92RfZ8.zip", "code_archive_sha256": "e199692b9595ba88bd1c0e513a67d2dcd63eaf197fe09894a93ff2ac59b8fbf3", "code_archive_bytes": 11259, "code_file_count": 5, "code_extensions": {".py": 5}, "github_disk_usage_kb": 10, "github_languages": {"Python": 48799}, "github_archived": false, "github_pushed_at": "2025-05-15T13:33:09Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/exploring-federated-pruning-for-large"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Q5CLpqbrFM", "year": 2025, "status": "rejected", "title": "Learning Representations of Instruments for Partial Identification of Treatment Effects", "authors": ["Jonas Schweisthal", "Dennis Frauen", "Maresa Schröder", "Konstantin Hess", "Niki Kilbertus", "Stefan Feuerriegel"], "authorids": ["~Jonas_Schweisthal1", "~Dennis_Frauen1", "~Maresa_Schröder1", "~Konstantin_Hess1", "~Niki_Kilbertus1", "~Stefan_Feuerriegel1"], "authors_source": "OpenReview API", "abstract": "Reliable estimation of treatment effects from observational data is important in many disciplines, such as medicine. However, estimation is challenging when unconfoundedness as a standard assumption in the causal inference literature is violated. In this work, we leverage arbitrary (potentially high-dimensional) instruments to estimate bounds on the conditional average treatment effect (CATE). Our contributions are three-fold: (1) We propose a novel approach for partial identification through a mapping of instruments to a discrete representation space so that we yield valid bounds on the CATE. This is crucial for reliable decision-making in real-world applications. (2) We derive a two-step method that learns tight bounds using a tailored neural partitioning of the latent instrument space. As a result, we avoid instability issues due to numerical approximations or adversarial training. Furthermore, our procedure aims to reduce the estimation variance in finite-sample settings to yield more reliable estimates. (3) We show theoretically that our procedure obtains valid bounds while reducing estimation variance and we perform experiments to demonstrate the effectiveness across various settings. Overall, our procedure offers a novel path for practitioners to make use of potentially high-dimensional instruments (e.g., as in Mendelian randomization).", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=Q5CLpqbrFM", "arxiv_id": "2410.08976", "paper_pdf": "papers/Q5CLpqbrFM.pdf", "paper_pdf_sha256": "78ed1fb7e0c8058d17c09de55c23dba0815eb61958e93f49b3121f194d09c6fe", "paper_pdf_bytes": 2027941, "paper_pdf_source": "openreview", "code_url": "https://github.com/JSchweisthal/ComplexPartialIdentif", "code_repository": "JSchweisthal/ComplexPartialIdentif", "code_commit": "260547060aa0f2b9e71f0179768f8a534bf6fd19", "code_archive": "repos/Q5CLpqbrFM.zip", "code_archive_sha256": "833358fb96d33775e0a634fb59489f4a4f0e0163187deb4dc1cc534cce81ddbf", "code_archive_bytes": 11992, "code_file_count": 7, "code_extensions": {".py": 7}, "github_disk_usage_kb": 10, "github_languages": {"Python": 36573}, "github_archived": false, "github_pushed_at": "2024-10-11T16:02:21Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/learning-representations-of-instruments-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "qYb0CANLGC", "year": 2024, "status": "rejected", "title": "Auto-Regressive Next-Token Predictors are Universal Learners", "authors": ["eran malach"], "authorids": ["~eran_malach1"], "authors_source": "OpenReview API", "abstract": "Large language models display remarkable capabilities in logical and mathematical reasoning, allowing them to solve complex tasks. Interestingly, these abilities emerge in networks trained on the simple task of next-token prediction. In this work, we present a theoretical framework for studying auto-regressive next-token predictors. We demonstrate that even simple models such as linear next-token predictors, trained on Chain-of-Thought (CoT) data, can approximate any function efficiently computed by a Turing machine. We introduce a new complexity measure---length complexity---which measures the number of intermediate tokens in a CoT sequence required to approximate some target function, and analyze the interplay between length complexity and other notions of complexity. Finally, we show experimentally that simple next-token predictors, such as linear networks and shallow Multi-Layer Perceptrons (MLPs), display non-trivial performance on text generation and arithmetic tasks. Our results demonstrate that the power of today's LLMs can be attributed, to a great extent, to the auto-regressive next-token training scheme, and not necessarily to a particular choice of architecture.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=qYb0CANLGC", "arxiv_id": "2309.06979", "paper_pdf": "papers/qYb0CANLGC.pdf", "paper_pdf_sha256": "ebe78d35f4d70fbb451c6e62b818867e5bc4cc9c62b6562566f0c26d4f1b2d26", "paper_pdf_bytes": 363218, "paper_pdf_source": "openreview", "code_url": "https://github.com/emalach/LinearLM", "code_repository": "emalach/LinearLM", "code_commit": "11fb93e06b9738a91b3463b7b0b7b63095928827", "code_archive": "repos/qYb0CANLGC.zip", "code_archive_sha256": "d5957a399d05f42c6edeb12417cb7a03d2c6949a12da9f08796bc5e978751112", "code_archive_bytes": 7725, "code_file_count": 3, "code_extensions": {".py": 3}, "github_disk_usage_kb": 11, "github_languages": {"Python": 14953}, "github_archived": false, "github_pushed_at": "2024-07-29T20:36:18Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/auto-regressive-next-token-predictors-are"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "LgjKqSjDzr", "year": 2022, "status": "rejected", "title": "SALT : Sharing Attention between Linear layer and Transformer for tabular dataset", "authors": ["Juseong Kim", "Jinsun Park", "Giltae Song"], "authorids": ["~Juseong_Kim2", "~Jinsun_Park1", "gsong@pusan.ac.kr"], "authors_source": "OpenReview API", "abstract": "Handling tabular data with deep learning models is a challenging problem despite their remarkable success in vision and language processing applications. Therefore, many practitioners still rely on classical models such as gradient boosting decision trees (GBDTs) rather than deep networks due to their superior performance with tabular data. In this paper, we propose a novel hybrid deep network architecture for tabular data, dubbed SALT (Sharing Attention between Linear layer and Transformer). The proposed SALT consists of two blocks: Transformers and linear layers blocks that take advantage of shared attention matrices. The shared attention matrices enable transformers and linear layers to closely cooperate with each other, and it leads to improved performance and robustness. Our algorithm outperforms tree-based ensemble models and previous deep learning methods in multiple benchmark datasets. We further demonstrate the robustness of the proposed SALT with semi-supervised learning and pre-training with small dataset scenarios.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=LgjKqSjDzr", "arxiv_id": null, "paper_pdf": "papers/LgjKqSjDzr.pdf", "paper_pdf_sha256": "e51405d2a961863fea10f81c16e63b1564adf13a5f116e713f6a9337c8e0c1a2", "paper_pdf_bytes": 999515, "paper_pdf_source": "openreview", "code_url": "https://github.com/Juseong03/SALT", "code_repository": "Juseong03/SALT", "code_commit": "8d95159be073be64d1ba1af1017ae8870a35005a", "code_archive": "repos/LgjKqSjDzr.zip", "code_archive_sha256": "1f5edbaadaeb7ad970dcf27b46a64c5559d3962ce2206a0c50a4ca9b4ecf0b04", "code_archive_bytes": 9616, "code_file_count": 4, "code_extensions": {".py": 4}, "github_disk_usage_kb": 10, "github_languages": {"Python": 35982}, "github_archived": false, "github_pushed_at": "2021-10-28T02:33:04Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/salt-sharing-attention-between-linear-layer"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "kmBFHJ5pr0o", "year": 2021, "status": "rejected", "title": "Distributed Adversarial Training to Robustify Deep Neural Networks at Scale", "authors": ["Gaoyuan Zhang", "Songtao Lu", "Sijia Liu", "Xiangyi Chen", "Pin-Yu Chen", "Lee Martie", "Lior Horesh", "Mingyi Hong"], "authorids": ["~Gaoyuan_Zhang1", "~Songtao_Lu1", "~Sijia_Liu1", "~Xiangyi_Chen1", "~Pin-Yu_Chen1", "lee.martie@ibm.com", "~Lior_Horesh1", "~Mingyi_Hong1"], "authors_source": "OpenReview API", "abstract": "Current deep neural networks are vulnerable to adversarial attacks, where adversarial perturbations to the inputs can change or manipulate classification. To defend against such attacks, an effective and popular approach, known as adversarial training, has been shown to mitigate the negative impact of adversarial attacks by virtue of a min-max robust training method. While effective, this approach is difficult to scale well to large models on large datasets (e.g., ImageNet) in general. To address this challenge, we propose distributed adversarial training (DAT), a large-batch adversarial training framework implemented over multiple machines. DAT supports one-shot and iterative attack generation methods, gradient quantization, and training over labeled and unlabeled data. Theoretically, we provide, under standard conditions in the optimization theory, the convergence rate of DAT to the first-order stationary points in general non-convex settings.  Empirically, on ResNet-18 and -50 under CIFAR-10 and ImageNet, we demonstrate that DAT either matches or outperforms state-of-the-art robust accuracies and achieves a graceful training speedup.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=kmBFHJ5pr0o", "arxiv_id": "2206.06257", "paper_pdf": "papers/kmBFHJ5pr0o.pdf", "paper_pdf_sha256": "479976bacaec51c73c97c58763ee6dad612254263f15ecbb62e51c0093883df0", "paper_pdf_bytes": 647562, "paper_pdf_source": "openreview", "code_url": "https://github.com/dat-2022/dat", "code_repository": "dat-2022/dat", "code_commit": "3fac9f60fcf0213b14cc991b3c1a9b2cf26415ba", "code_archive": "repos/kmBFHJ5pr0o.zip", "code_archive_sha256": "ce7430d997a23c756e61aae1bb1a46b4ffbe2e2e9856edc3deb76f58df45d848", "code_archive_bytes": 15125, "code_file_count": 10, "code_extensions": {".py": 10}, "github_disk_usage_kb": 12, "github_languages": {"Python": 47403}, "github_archived": false, "github_pushed_at": "2022-07-31T16:33:09Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/distributed-adversarial-training-to-robustify-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "HylNWkHtvB", "year": 2020, "status": "rejected", "title": "Domain-Independent Dominance of Adaptive Methods", "authors": ["Pedro Savarese", "David McAllester", "Sudarshan Babu", "Michael Maire"], "authorids": ["savarese@ttic.edu", "mcallester@ttic.edu", "sudarshan@ttic.edu", "mmaire@uchicago.edu"], "authors_source": "OpenReview API", "abstract": "From a simplified analysis of adaptive methods, we derive AvaGrad, a new optimizer which outperforms SGD on vision tasks when its adaptability is properly tuned. We observe that the power of our method is partially explained by a decoupling of learning rate and adaptability, greatly simplifying hyperparameter search. In light of this observation, we demonstrate that, against conventional wisdom, Adam can also outperform SGD on vision tasks, as long as the coupling between its learning rate and adaptability is taken into account. In practice, AvaGrad matches the best results, as measured by generalization accuracy, delivered by any existing optimizer (SGD or adaptive) across image classification (CIFAR, ImageNet) and character-level language modelling (Penn Treebank) tasks. This later observation, alongside of AvaGrad's decoupling of hyperparameters, could make it the preferred optimizer for deep learning, replacing both SGD and Adam.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=HylNWkHtvB", "arxiv_id": "1912.01823", "paper_pdf": "papers/HylNWkHtvB.pdf", "paper_pdf_sha256": "69ee064c63d05aba1b8e8f60d445c7df6ebbd405434135b70034287016301d43", "paper_pdf_bytes": 355911, "paper_pdf_source": "openreview", "code_url": "https://github.com/lolemacs/avagrad", "code_repository": "lolemacs/avagrad", "code_commit": "343733151f26a1fa8e504079c6d7934a6038a93b", "code_archive": "repos/HylNWkHtvB.zip", "code_archive_sha256": "4ae059a3b716f559ead8828fa2c4a0006b88eebdacaceb4066dc60c5be20255d", "code_archive_bytes": 8780, "code_file_count": 5, "code_extensions": {".py": 5}, "github_disk_usage_kb": 10, "github_languages": {"Python": 26004}, "github_archived": false, "github_pushed_at": "2020-12-15T04:01:22Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/domain-independent-dominance-of-adaptive-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "S1E64jC5tm", "year": 2019, "status": "rejected", "title": "The Forward-Backward Embedding of Directed Graphs", "authors": ["Thomas Bonald", "Nathan De Lara"], "authorids": ["thomas.bonald@telecom-paristech.fr", "nathan.delara@telecom-paristech.fr"], "authors_source": "OpenReview API", "abstract": "We introduce a novel embedding of directed graphs derived from the singular value decomposition (SVD) of the  normalized adjacency matrix. Specifically, we show that, after proper normalization of the singular vectors,  \n the  distances between vectors in  the embedding space are proportional to the mean commute times between the corresponding  nodes by a  forward-backward random walk in the graph, which follows the edges  alternately in forward and backward directions.  In particular, two nodes having many common successors in the graph tend to be represented by close vectors in the embedding space. More formally, we prove that our representation of the graph is  equivalent to the spectral embedding of some co-citation graph, where  nodes are linked with respect to their common set of successors in the original graph. The interest of our approach is that it does not require to build this co-citation graph, which is typically much denser than the original graph. Experiments  on  real datasets show the efficiency of the approach. \n", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=S1E64jC5tm", "arxiv_id": null, "paper_pdf": "papers/S1E64jC5tm.pdf", "paper_pdf_sha256": "8544849615a647b537158912f5832b63bf3b6415d01ddb6c4f9f951d494c8c04", "paper_pdf_bytes": 316920, "paper_pdf_source": "openreview", "code_url": "https://github.com/tbonald/directed", "code_repository": "tbonald/directed", "code_commit": "2d4b979f296bb0750ee1560ac07d70ff68f7c8dc", "code_archive": "repos/S1E64jC5tm.zip", "code_archive_sha256": "da3639ca90e6775705e265ab6256cb2cf997f0b08f0b91f7b7c6fb5455efc695", "code_archive_bytes": 13245, "code_file_count": 4, "code_extensions": {".py": 3, ".ipynb": 1}, "github_disk_usage_kb": 17, "github_languages": {"Python": 21945, "Jupyter Notebook": 17282}, "github_archived": false, "github_pushed_at": "2018-11-08T09:38:07Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/the-forward-backward-embedding-of-directed"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "S1EwLkW0W", "year": 2018, "status": "rejected", "title": "Dissecting Adam: The Sign, Magnitude and Variance of Stochastic Gradients", "authors": ["Lukas Balles", "Philipp Hennig"], "authorids": ["lukas.balles@tuebingen.mpg.de", "ph@tue.mpg.de"], "authors_source": "OpenReview API", "abstract": "The ADAM optimizer is exceedingly popular in the deep learning community. Often it works very well, sometimes it doesn’t. Why? We interpret ADAM as a combination of two aspects: for each weight, the update direction is determined by the sign of the stochastic gradient, whereas the update magnitude is solely determined by an estimate of its relative variance. We  disentangle these two aspects and analyze them in isolation, shedding light on ADAM ’s inner workings. Transferring the \"variance adaptation” to momentum- SGD gives rise to a novel method, completing the practitioner’s toolbox for problems where ADAM fails.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=S1EwLkW0W", "arxiv_id": "1705.07774", "paper_pdf": "papers/S1EwLkW0W.pdf", "paper_pdf_sha256": "2c820b68da0c383b2903af5ebb1dbdd750ebd91c21a5918648a4350b627fe321", "paper_pdf_bytes": 656781, "paper_pdf_source": "openreview", "code_url": "https://github.com/lballes/msvag", "code_repository": "lballes/msvag", "code_commit": "d2d467b6a9c6442d781ecb2f1d4b5c5769556363", "code_archive": "repos/S1EwLkW0W.zip", "code_archive_sha256": "cec7a33afea60c29789a2ee5ae6350da0f16dec1de490ba4dcf2a99486c1f23b", "code_archive_bytes": 8966, "code_file_count": 3, "code_extensions": {".py": 3}, "github_disk_usage_kb": 12, "github_languages": {"Python": 10920}, "github_archived": false, "github_pushed_at": "2018-05-11T14:21:18Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/dissecting-adam-the-sign-magnitude-and"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "JtX6oaaJ2d", "year": 2026, "status": "rejected", "title": "Improving LLM Unlearning Robustness via Random Perturbations", "authors": ["Dang Huu-Tien", "Hoang Thanh-Tung", "Anh Tuan Bui", "Phuong Minh Nguyen", "Le-Minh Nguyen", "Naoya Inoue"], "authorids": ["~Dang_Huu-Tien1", "~Hoang_Thanh-Tung1", "~Anh_Tuan_Bui2", "~Phuong_Minh_Nguyen3", "~Le-Minh_Nguyen1", "~Naoya_Inoue1"], "authors_source": "OpenReview API", "abstract": "Here, we show that current state-of-the-art LLM unlearning methods inherently reduce models' robustness, causing them to misbehave even when a single non-adversarial forget-token is present in the retain-query. Toward understanding underlying causes, we propose a novel theoretical framework that reframes the *unlearning process as backdoor attacks and defenses*: forget-tokens act as backdoor triggers that, when activated in retain-queries, cause disruptions in unlearned models' behaviors, similar to successful backdoor attacks. The sense that, LLM unlearning methods *themselves poison the model*, make it more vulnerable to forget-tokens, and *hide rather than erase* target knowledge, describes their true mechanism. To mitigate the vulnerability caused by the forgetting process, we reinterpret the retaining process as a backdoor defense and propose Random Noise Augmentation (RNA), a lightweight, model and method-agnostic approach with theoretical guarantees for improving the robustness of models.  Extensive experiments demonstrate that RNA significantly improves the robustness of unlearned models while preserving forget and retain performances. This backdoor attack-defense framework offers insights into the mechanisms of unlearning that can shed light on future research directions for improving unlearning robustness.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=JtX6oaaJ2d", "arxiv_id": "2501.19202", "paper_pdf": "papers/JtX6oaaJ2d.pdf", "paper_pdf_sha256": "047638bbcecfff54980a5e5c942fd1cddb60e523ef9129df9c8399213c0fa4d7", "paper_pdf_bytes": 7340032, "paper_pdf_source": "openreview", "code_url": "https://github.com/RebelsNLU-jaist/llmu-robustness", "code_repository": "RebelsNLU-jaist/llmu-robustness", "code_commit": "cdda05906725da2f6c1b7d5792eee7d780811541", "code_archive": "repos/JtX6oaaJ2d.zip", "code_archive_sha256": "2dbe319e41d72dbf0ff8662f7b10e0361be5d485a9612997ec042818202674a3", "code_archive_bytes": 25397, "code_file_count": 14, "code_extensions": {".py": 8, ".sh": 6}, "github_disk_usage_kb": 16, "github_languages": {"Python": 48251, "Shell": 6200}, "github_archived": false, "github_pushed_at": "2026-06-02T05:43:33Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/improving-the-robustness-of-representation"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "22ywev7zMt", "year": 2025, "status": "rejected", "title": "On the Out-of-Distribution Generalization of Self-Supervised Learning", "authors": ["Wenwen Qiang", "Jingyao Wang", "Zeen Song", "Jiangmeng Li", "Changwen Zheng"], "authorids": ["~Wenwen_Qiang1", "~Jingyao_Wang1", "~Zeen_Song1", "~Jiangmeng_Li1", "~Changwen_Zheng1"], "authors_source": "OpenReview API", "abstract": "In this paper, we focus on the out-of-distribution (OOD) generalization of self-supervised learning (SSL). By analyzing the mini-batch construction during SSL training phase, we first give one plausible explanation for SSL having OOD generalization. Then, from the perspective of data generation and causal inference, we analyze and conclude that SSL learns spurious correlations during the training process, which leads to a reduction in OOD generalization. To address this issue, we propose a post-intervention distribution (PID) grounded in the Structural Causal Model. PID offers a scenario where the relationships between variables are free from the influence of spurious correlations. Besides, we demonstrate that if each mini-batch during SSL training satisfies PID, the resulting SSL model can achieve optimal worst-case OOD performance. This motivates us to develop a batch sampling strategy that enforces PID constraints through the learning of a latent variable model. Through theoretical analysis, we demonstrate the identifiability of the latent variable model and validate the effectiveness of the proposed sampling strategy. Experiments conducted on various downstream OOD tasks demonstrate the effectiveness of the proposed sampling strategy.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=22ywev7zMt", "arxiv_id": "2505.16675", "paper_pdf": "papers/22ywev7zMt.pdf", "paper_pdf_sha256": "80507b1b63d95900be2c420895a12d96d9270abac2fa06b894015b0b6cb03d1f", "paper_pdf_bytes": 2564459, "paper_pdf_source": "openreview", "code_url": "https://github.com/ML-TASA/PID-SSL", "code_repository": "ML-TASA/PID-SSL", "code_commit": "095b2be7fcfe206cbf2a103039a0529d12968846", "code_archive": "repos/22ywev7zMt.zip", "code_archive_sha256": "fd1087b1fecbe86c4aeab52c00f3c95107d8f1fb6b59b21d37b5b15360c84b76", "code_archive_bytes": 7332, "code_file_count": 3, "code_extensions": {".py": 3}, "github_disk_usage_kb": 10, "github_languages": {"Python": 18876}, "github_archived": false, "github_pushed_at": "2025-06-04T08:40:48Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/on-the-out-of-distribution-generalization-of-2"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "j8s-BRxXST", "year": 2023, "status": "rejected", "title": "A Simple Contrastive Learning Objective for Alleviating Neural Text Degeneration", "authors": ["Shaojie Jiang", "Ruqing Zhang", "Svitlana Vakulenko", "Maarten de Rijke"], "authorids": ["~Shaojie_Jiang1", "~Ruqing_Zhang3", "~Svitlana_Vakulenko1", "~Maarten_de_Rijke1"], "authors_source": "OpenReview API", "abstract": "The cross-entropy objective has proved to be an all-purpose training objective for autoregressive language models (LMs). However, without distinguishing problematic tokens, LMs trained using cross-entropy exhibit text degeneration problems. To address this, unlikelihood training has been proposed to reduce the probability of unlikely tokens predicted by LMs. But unlikelihood does not explicitly consider the relationship between the label tokens and unlikely token candidates, thus showing marginal improvements in degeneration. We propose a new contrastive token learning objective that inherits the advantages of cross-entropy and unlikelihood training and avoids their limitations. The key idea is to teach a LM to generate high probabilities for label tokens and low probabilities for negative candidates. Comprehensive experiments on language modeling and open-domain dialogue generation tasks show that the proposed contrastive token objective yields much less repetitive texts, with a higher generation quality than baseline approaches, achieving the new state-of-the-art performance on text degeneration.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=j8s-BRxXST", "arxiv_id": "2205.02517", "paper_pdf": "papers/j8s-BRxXST.pdf", "paper_pdf_sha256": "13f30c6527fe9db4674103de3a836b928b2749fd60c3e6b8674717e66380bb95", "paper_pdf_bytes": 1753876, "paper_pdf_source": "openreview", "code_url": "https://github.com/ShaojieJiang/CT-Loss", "code_repository": "ShaojieJiang/CT-Loss", "code_commit": "21731b590839b31675e55333be289c7b6513ade3", "code_archive": "repos/j8s-BRxXST.zip", "code_archive_sha256": "38e85e3862162dbf45eedcd49ca03a36b7201b63f8314f378b6699d19395ea76", "code_archive_bytes": 8148, "code_file_count": 3, "code_extensions": {".py": 3}, "github_disk_usage_kb": 13, "github_languages": {"Python": 6005, "Mustache": 974}, "github_archived": false, "github_pushed_at": "2022-05-11T09:27:35Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/a-simple-contrastive-learning-objective-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "xw04RdwI2kS", "year": 2022, "status": "rejected", "title": "Inverse Contextual Bandits: Learning How Behavior Evolves over Time", "authors": ["Alihan Hüyük", "Daniel Jarrett", "Mihaela van der Schaar"], "authorids": ["~Alihan_Hüyük1", "~Daniel_Jarrett1", "~Mihaela_van_der_Schaar2"], "authors_source": "OpenReview API", "abstract": "Understanding a decision-maker's priorities by observing their behavior is critical for transparency and accountability in decision processes—such as in healthcare. Though conventional approaches to policy learning almost invariably assume stationarity in behavior, this is hardly true in practice: Medical practice is constantly evolving as clinical professionals fine-tune their knowledge over time. For instance, as the medical community's understanding of organ transplantations has progressed over the years, a pertinent question is: How have actual organ allocation policies been evolving? To give an answer, we desire a policy learning method that provides interpretable representations of decision-making, in particular capturing an agent's non-stationary knowledge of the world, as well as operating in an offline manner. First, we model the evolving behavior of decision-makers in terms of contextual bandits, and formalize the problem of Inverse Contextual Bandits (\"ICB''). Second, we propose two concrete algorithms as solutions, learning parametric and non-parametric representations of an agent's behavior. Finally, using both real and simulated data for liver transplantations, we illustrate the applicability and explainability of our method, as well as benchmarking and validating the accuracy of our algorithms.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=xw04RdwI2kS", "arxiv_id": "2107.06317", "paper_pdf": "papers/xw04RdwI2kS.pdf", "paper_pdf_sha256": "c61e0337f8e7874a6ff8e8d9ee63565d9c735628905549da16b084317611b111", "paper_pdf_bytes": 2697785, "paper_pdf_source": "openreview", "code_url": "https://github.com/alihanhyk/invconban", "code_repository": "alihanhyk/invconban", "code_commit": "b365614697053f3584915ee013cf6fee36e73de5", "code_archive": "repos/xw04RdwI2kS.zip", "code_archive_sha256": "8529981d61f2f26816a14f729e68b7d323d31a6e8c4f9518bc435c0dd9d2e4eb", "code_archive_bytes": 17069, "code_file_count": 12, "code_extensions": {".py": 11, ".sh": 1}, "github_disk_usage_kb": 12, "github_languages": {"Python": 32635, "Shell": 2089}, "github_archived": false, "github_pushed_at": "2022-06-07T11:59:01Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/inverse-contextual-bandits-learning-how"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "EsA9Nr9JHvy", "year": 2021, "status": "rejected", "title": "The Heavy-Tail Phenomenon in SGD", "authors": ["Mert Gurbuzbalaban", "Umut Simsekli", "Lingjiong Zhu"], "authorids": ["~Mert_Gurbuzbalaban1", "~Umut_Simsekli1", "~Lingjiong_Zhu1"], "authors_source": "OpenReview API", "abstract": "In recent years, various notions of capacity and complexity have been proposed for characterizing the generalization properties of stochastic gradient descent (SGD) in deep learning. Some of the popular notions that correlate well with the performance on unseen data are (i) the 'flatness' of the local minimum found by SGD, which is related to the eigenvalues of the Hessian, (ii) the ratio of the stepsize $\\eta$ to the batch size $b$, which essentially controls the magnitude of the stochastic gradient noise, and (iii) the 'tail-index', which measures the heaviness of the tails of the network weights at convergence. In this paper, we argue that these three seemingly unrelated perspectives for generalization are deeply linked to each other. We claim that depending on the structure of the Hessian of the loss at the minimum, and the choices of the algorithm parameters $\\eta$ and $b$, the SGD iterates will converge to a \\emph{heavy-tailed} stationary distribution. We rigorously prove this claim in the setting of quadratic optimization: we show that even in a simple linear regression problem with independent and identically distributed Gaussian data, the iterates can be heavy-tailed with infinite variance. We further characterize the behavior of the tails with respect to algorithm parameters, the dimension, and the curvature. We then translate our results into insights about the behavior of SGD in deep learning. We finally support our theory with experiments conducted on both synthetic data and fully connected neural networks.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=EsA9Nr9JHvy", "arxiv_id": "2006.04740", "paper_pdf": "papers/EsA9Nr9JHvy.pdf", "paper_pdf_sha256": "2b9da9bfcea610172575c5d32bd6f0bc1e57967da89c7c3f84d5efb39bac18c6", "paper_pdf_bytes": 718734, "paper_pdf_source": "openreview", "code_url": "https://github.com/umutsimsekli/sgd_ht", "code_repository": "umutsimsekli/sgd_ht", "code_commit": "67f2b51055a331f23748bd418c056de1aa7cb48f", "code_archive": "repos/EsA9Nr9JHvy.zip", "code_archive_sha256": "d88873ef01711a7064b73963ad9fb2658772af2e6e4052470c88de222f176ad2", "code_archive_bytes": 12648, "code_file_count": 6, "code_extensions": {".py": 4, ".ipynb": 2}, "github_disk_usage_kb": 12, "github_languages": {"Python": 18880, "Jupyter Notebook": 12049}, "github_archived": false, "github_pushed_at": "2022-02-15T13:12:46Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/the-heavy-tail-phenomenon-in-sgd"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ByxXZpVtPB", "year": 2020, "status": "rejected", "title": "Homogeneous Linear Inequality Constraints for Neural Network Activations", "authors": ["Thomas Frerix", "Matthias Nießner", "Daniel Cremers"], "authorids": ["thomas.frerix@tum.de", "niessner@tum.de", "cremers@tum.de"], "authors_source": "OpenReview API", "abstract": "We propose a method to impose homogeneous linear inequality constraints of the form $Ax\\leq 0$ on neural network activations. The proposed method allows a data-driven training approach to be combined with modeling prior knowledge about the task. One way to achieve this task is by means of a projection step at test time after unconstrained training.\nHowever, this is an expensive operation. By directly incorporating the constraints into the architecture, we can significantly speed-up inference at test time; for instance, our experiments show a speed-up of up to two orders of magnitude over a projection method. Our algorithm computes a suitable parameterization of the feasible set at initialization and uses standard variants of stochastic gradient descent to find solutions to the constrained network. Thus, the modeling constraints are always satisfied during training. Crucially, our approach avoids to solve an optimization problem at each training step or to manually trade-off data and constraint fidelity with additional hyperparameters. We consider constrained generative modeling as an important application domain and experimentally demonstrate the proposed method by constraining a variational autoencoder.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=ByxXZpVtPB", "arxiv_id": "1902.01785", "paper_pdf": "papers/ByxXZpVtPB.pdf", "paper_pdf_sha256": "4677f0bf54d7f1bf6d4fcbdea5d5ea6363aec79cc08e2ac1f2286647ff1ec40c", "paper_pdf_bytes": 733546, "paper_pdf_source": "openreview", "code_url": "https://github.com/tfrerix/constrained-nets", "code_repository": "tfrerix/constrained-nets", "code_commit": "1221f3fd2592056c2a00c463d65fe3afa646fd54", "code_archive": "repos/ByxXZpVtPB.zip", "code_archive_sha256": "6728797b652ebdf88e23afcf1b96053b5a62108e3c05f38349628efbfd76da9a", "code_archive_bytes": 12562, "code_file_count": 7, "code_extensions": {".py": 6, ".sh": 1}, "github_disk_usage_kb": 11, "github_languages": {"Python": 29194, "Shell": 436}, "github_archived": false, "github_pushed_at": "2021-04-08T19:17:35Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/linear-inequality-constraints-for-neural"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "rJeZS3RcYm", "year": 2019, "status": "rejected", "title": "Simple Black-box Adversarial Attacks", "authors": ["Chuan Guo", "Jacob R. Gardner", "Yurong You", "Andrew G. Wilson", "Kilian Q. Weinberger"], "authorids": ["cg563@cornell.edu", "jrg365@cornell.edu", "yy785@cornell.edu", "andrew@cornell.edu", "kqw4@cornell.edu"], "authors_source": "OpenReview API", "abstract": "The construction of adversarial images is a search problem in high dimensions within a small region around a target image. The goal is to find an imperceptibly modified image that is misclassified by a target model. In the black-box setting, only sporadic feedback is provided through occasional model evaluations. In this paper we provide a new algorithm whose search strategy is based on an intriguingly simple iterative principle: We randomly pick a low frequency component of the discrete cosine transform (DCT) and either add or subtract it to the target image. Model evaluations are only required to identify whether an operation decreases the adversarial loss. Despite its simplicity, the proposed method can be used for targeted and untargeted attacks --- resulting in previously unprecedented query efficiency in both settings. We require a median of 600 black-box model queries (ResNet-50) to produce an adversarial ImageNet image, and we successfully attack Google Cloud Vision with 2500 median queries, averaging to a cost of only $3 per image. We argue that our proposed algorithm should serve as a strong baseline for future adversarial black-box attacks, in particular because it is extremely fast and can be implemented in less than 20 lines of PyTorch code. ", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=rJeZS3RcYm", "arxiv_id": "1905.07121", "paper_pdf": "papers/rJeZS3RcYm.pdf", "paper_pdf_sha256": "fcc0747ef59f7d8c34a649fbff3282b4c583e6c42da58706e6596aa4d42b3333", "paper_pdf_bytes": 7340032, "paper_pdf_source": "openreview", "code_url": "https://github.com/cg563/simple-blackbox-attack", "code_repository": "cg563/simple-blackbox-attack", "code_commit": "3b66937945699f706918277ac55fb171f07d2bcf", "code_archive": "repos/rJeZS3RcYm.zip", "code_archive_sha256": "617afb689a318d671fbce9c38169653313174229ca1fb5e6fe025f9e86f570e0", "code_archive_bytes": 10814, "code_file_count": 4, "code_extensions": {".py": 4}, "github_disk_usage_kb": 18, "github_languages": {"Python": 28113}, "github_archived": false, "github_pushed_at": "2023-03-27T18:29:55Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/simple-black-box-adversarial-attacks-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ryykVe-0W", "year": 2018, "status": "rejected", "title": "Learning Independent Features with Adversarial Nets for Non-linear ICA", "authors": ["Philemon Brakel", "Yoshua Bengio"], "authorids": ["pbpop3@gmail.com", "yoshua.bengio@umontreal.ca"], "authors_source": "OpenReview API", "abstract": "Reliable measures of statistical dependence could potentially be useful tools for learning independent features and performing tasks like source separation using Independent Component Analysis (ICA).  Unfortunately, many of such measures, like the mutual information, are hard to estimate and optimize directly.  We propose to learn independent features with adversarial objectives (Goodfellow et al. 2014, Arjovsky et al. 2017) which optimize such measures implicitly.  These objectives compare samples from the joint distribution and the product of the marginals without the need to compute any probability densities. We also propose two methods for obtaining samples from the product of the marginals using either a simple resampling trick or a separate parametric distribution.  Our experiments show that this strategy can easily be applied to different types of model architectures and solve both linear and non-linear ICA problems.\n", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=ryykVe-0W", "arxiv_id": "1710.05050", "paper_pdf": "papers/ryykVe-0W.pdf", "paper_pdf_sha256": "a7615362a0e07145193cef06d29980c1607eb969c83baca531c847b78bc32a9c", "paper_pdf_bytes": 574546, "paper_pdf_source": "openreview", "code_url": "https://github.com/pbrakel/anica", "code_repository": "pbrakel/anica", "code_commit": "79d837addcd98ee9bab301d3966e4919cf9732f3", "code_archive": "repos/ryykVe-0W.zip", "code_archive_sha256": "a6ba4a1c3a2eb3eceae20e2aa9d3c86c4677f6883b3eb07a93d3175a5e4fd3f2", "code_archive_bytes": 21667, "code_file_count": 6, "code_extensions": {".py": 6}, "github_disk_usage_kb": 16, "github_languages": {"Python": 47021}, "github_archived": false, "github_pushed_at": "2017-10-17T13:02:32Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/learning-independent-features-with"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "LKv3bx610K", "year": 2026, "status": "rejected", "title": "Mitigating Fine-tuning Risks in LLMs via Safety-Aware Probing Optimization", "authors": ["Chengcan Wu", "Zhixin Zhang", "Zeming Wei", "Yihao Zhang", "Meng Sun"], "authorids": ["~Chengcan_Wu1", "~Zhixin_Zhang5", "~Zeming_Wei1", "~Yihao_Zhang8", "~Meng_Sun1"], "authors_source": "OpenReview API", "abstract": "The significant progress of large language models (LLMs) has led to remarkable achievements across numerous applications. However, their ability to generate harmful content has sparked substantial safety concerns. Despite the implementation of safety alignment techniques during the pre-training phase, recent research indicates that fine-tuning LLMs on adversarial or even benign data can inadvertently compromise their safety. In this paper, we re-examine the fundamental issue of why fine-tuning on non-harmful data still results in safety degradation. We introduce a safety-aware probing (SAP) optimization framework designed to mitigate the safety risks of fine-tuning LLMs. Specifically, SAP incorporates a safety-aware probe into the gradient propagation process, mitigating the model's risk of safety degradation by identifying potential pitfalls in gradient directions, thereby enhancing task-specific performance while successfully preserving model safety. Our extensive experimental results demonstrate that SAP effectively reduces harmfulness below the original fine-tuned model and achieves comparable test loss to standard fine-tuning methods. Our code is available in the supplementary materials.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=LKv3bx610K", "arxiv_id": "2505.16737", "paper_pdf": "papers/LKv3bx610K.pdf", "paper_pdf_sha256": "6825b5f34b0a8ddf68d8606fdab0ec4a6a579bef466be8f3538d9d7e7d8d0a22", "paper_pdf_bytes": 4164185, "paper_pdf_source": "openreview", "code_url": "https://github.com/ChengcanWu/SAP", "code_repository": "ChengcanWu/SAP", "code_commit": "487edc3374f24581399040beac6209c388a463ee", "code_archive": "repos/LKv3bx610K.zip", "code_archive_sha256": "19d021f6b6ea33a4069e1ae54ae35569bd44cedf368b9c92978febc6637e57ed", "code_archive_bytes": 12549, "code_file_count": 5, "code_extensions": {".py": 5}, "github_disk_usage_kb": 19, "github_languages": {"Python": 19396}, "github_archived": false, "github_pushed_at": "2026-04-23T15:02:42Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/mitigating-fine-tuning-risks-in-llms-via"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "dq3keisMjT", "year": 2025, "status": "rejected", "title": "Phase Transitions in the Output Distribution of Large Language Models", "authors": ["Julian Arnold", "Flemming Holtorf", "Frank Schäfer", "Niels Lörch"], "authorids": ["~Julian_Arnold2", "~Flemming_Holtorf1", "~Frank_Schäfer1", "~Niels_Lörch1"], "authors_source": "OpenReview API", "abstract": "In a physical system, changing parameters such as temperature can induce a phase transition: an abrupt change from one state of matter to another. Analogous phenomena have recently been observed in large language models. Typically, the task of identifying phase transitions requires human analysis and some prior understanding of the system to narrow down which low-dimensional properties to monitor and analyze. Statistical methods for the automated detection of phase transitions from data have recently been proposed within the physics community. These methods are largely system agnostic and, as shown here, can be adapted to study the behavior of large language models. In particular, we quantify distributional changes in the generated output via statistical distances, which can be efficiently estimated with access to the probability distribution over next-tokens. This versatile approach is capable of discovering new phases of behavior and unexplored transitions -- an ability that is particularly exciting in light of the rapid development of language models and their emergent capabilities.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=dq3keisMjT", "arxiv_id": "2405.17088", "paper_pdf": "papers/dq3keisMjT.pdf", "paper_pdf_sha256": "0dbb58fa3e3333bffbd88d8ebdaa8aeabedd4e59c972d818efa00c15d9ce6948", "paper_pdf_bytes": 866082, "paper_pdf_source": "openreview", "code_url": "https://github.com/llmtransitions/llmtransitions", "code_repository": "llmtransitions/llmtransitions", "code_commit": "626845657b65205672dd7c01eb2c963b5ebb53c2", "code_archive": "repos/dq3keisMjT.zip", "code_archive_sha256": "6d0508b8e098359accc7a34d53da1b4f7df2a075a3852c961b25cf15704b2d5c", "code_archive_bytes": 8937, "code_file_count": 6, "code_extensions": {".py": 3, ".ipynb": 3}, "github_disk_usage_kb": 12, "github_languages": {"Jupyter Notebook": 14391, "Python": 8484}, "github_archived": false, "github_pushed_at": "2025-08-15T13:57:56Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/phase-transitions-in-the-output-distribution"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "k3VANp85b4S", "year": 2023, "status": "rejected", "title": "On the Robustness of Randomized Ensembles to Adversarial Perturbations", "authors": ["Hassan Dbouk", "Naresh Shanbhag"], "authorids": ["~Hassan_Dbouk1", "~Naresh_Shanbhag1"], "authors_source": "OpenReview API", "abstract": "Randomized ensemble classifiers (RECs), where one classifier is randomly selected during inference, have emerged as an attractive alternative to traditional ensembling methods for realizing adversarially robust classifiers with limited compute requirements. However, recent works have shown that existing methods for constructing RECs are more vulnerable than initially claimed, casting major doubts on their efficacy and prompting fundamental questions such as: \"When are RECs useful?\", \"What are their limits?\", and \"How do we train them?\". In this work, we first demystify RECs as we derive fundamental results regarding their theoretical limits, necessary and sufficient conditions for them to be useful, and more. Leveraging this new understanding, we propose a new boosting\nalgorithm (BARRE) for training robust RECs, and empirically demonstrate its effectiveness at defending against strong $\\ell_\\infty$ norm-bounded adversaries across various network architectures and datasets. Our code is submitted as part of the supplementary material, and will be publicly released on GitHub", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=k3VANp85b4S", "arxiv_id": "2302.01375", "paper_pdf": "papers/k3VANp85b4S.pdf", "paper_pdf_sha256": "fc476d61e8af77d8b69aaa7bcdbf266b7ed3895dcff10aedaa692da3fa5515b2", "paper_pdf_bytes": 761683, "paper_pdf_source": "openreview", "code_url": "https://github.com/hsndbk4/BARRE", "code_repository": "hsndbk4/BARRE", "code_commit": "1a42d89214db58200abd73f29ae5f8ee1e3bebe6", "code_archive": "repos/k3VANp85b4S.zip", "code_archive_sha256": "97bb1b39d3c562b65753663e05ab1352faec8e3cff5ec83554c924d0abaa0f6c", "code_archive_bytes": 16547, "code_file_count": 9, "code_extensions": {".py": 9}, "github_disk_usage_kb": 15, "github_languages": {"Python": 46743}, "github_archived": false, "github_pushed_at": "2023-09-05T02:14:51Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/on-the-robustness-of-randomized-ensembles-to"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "GIEPR9OomyX", "year": 2022, "status": "rejected", "title": "Langevin Autoencoders for Learning Deep Latent Variable Models", "authors": ["Shohei Taniguchi", "Yusuke Iwasawa", "Wataru Kumagai", "Yutaka Matsuo"], "authorids": ["~Shohei_Taniguchi1", "~Yusuke_Iwasawa1", "~Wataru_Kumagai2", "~Yutaka_Matsuo1"], "authors_source": "OpenReview API", "abstract": "Markov chain Monte Carlo (MCMC), such as Langevin dynamics, is valid for approximating intractable distributions. However, its usage is limited in the context of deep latent variable models since it is not scalable to data size owing to its datapoint-wise iterations and slow convergence. This paper proposes the amortized Langevin dynamics (ALD), wherein datapoint-wise MCMC iterations are entirely replaced with updates of an inference model that maps observations into latent variables. Since it no longer depends on datapoint-wise iterations, ALD enables scalable inference from large-scale datasets. Despite its efficiency, it retains the excellent property of MCMC; we prove that ALD has the target posterior as a stationary distribution with a mild assumption. Furthermore, ALD can be extended to sampling from an unconditional distribution such as an energy-based model, enabling more flexible generative modeling by applying it to the prior distribution of the latent variable. Based on ALD, we construct a new deep latent variable model named the Langevin autoencoder (LAE). LAE uses ALD for autoencoder-like posterior inference and sampling from the latent space EBM. Using toy datasets, we empirically validate that ALD can properly obtain samples from target distributions in both conditional and unconditional cases, and ALD converges significantly faster than traditional LD. We also evaluate LAE on the image generation task using three datasets (SVHN, CIFAR-10, and CelebA-HQ). Not only can LAE be trained faster than non-amortized MCMC methods, but LAE can also generate better samples in terms of the Fréchet Inception Distance (FID) compared to AVI-based methods, such as the variational autoencoder.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=GIEPR9OomyX", "arxiv_id": "2209.07036", "paper_pdf": "papers/GIEPR9OomyX.pdf", "paper_pdf_sha256": "ec6a4622851af0812767f792117d7a2cf772caadca68c18932f84d15e4f17ad4", "paper_pdf_bytes": 2061206, "paper_pdf_source": "openreview", "code_url": "https://github.com/iShohei220/LAE", "code_repository": "iShohei220/LAE", "code_commit": "d4cfbdb8fdee520fa891610a6afb569dcb9540b6", "code_archive": "repos/GIEPR9OomyX.zip", "code_archive_sha256": "60d698fddb4b0817510629baed3f1acbf1cd844130a007b27a8315294a7f879c", "code_archive_bytes": 17968, "code_file_count": 8, "code_extensions": {".py": 8}, "github_disk_usage_kb": 13, "github_languages": {"Python": 59400}, "github_archived": false, "github_pushed_at": "2022-09-15T02:16:11Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/langevin-autoencoders-for-learning-deep-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "6VhmvP7XZue", "year": 2021, "status": "rejected", "title": "Open-world Semi-supervised Learning", "authors": ["Kaidi Cao", "Maria Brbic", "Jure Leskovec"], "authorids": ["~Kaidi_Cao1", "mbrbic@cs.stanford.edu", "~Jure_Leskovec1"], "authors_source": "OpenReview API", "abstract": "Supervised and semi-supervised learning methods have been traditionally designed for the closed-world setting which is based on the assumption that unlabeled test data contains only classes previously encountered in the labeled training data. However, the real world is often open and dynamic, and thus novel previously unseen classes may appear in the test data or during the model deployment. Here, we introduce a new open-world semi-supervised learning setting in which the model is required to recognize previously seen classes, as well as to discover novel classes never seen in the labeled dataset. To tackle the problem, we propose ORCA, an approach that jointly learns a feature representation and a classifier on the labeled and unlabeled subsets of the data. The key idea in ORCA is in introducing uncertainty based adaptive margin that effectively circumvents the bias caused by the imbalance of variance between seen and novel classes. We demonstrate that ORCA accurately discovers novel classes and assigns samples to previously seen classes on standard benchmark image classification datasets, including  CIFAR and ImageNet. Remarkably, despite solving the harder task ORCA outperforms semi-supervised methods on seen classes, as well as novel class discovery methods on unseen classes, achieving 7% and 151% improvements on seen and unseen classes of the ImageNet dataset.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=6VhmvP7XZue", "arxiv_id": "2102.03526", "paper_pdf": "papers/6VhmvP7XZue.pdf", "paper_pdf_sha256": "553138b2d0a9c4b7c990374fbb79dfb52bb5171bea9b220ebc0eab012b88d2c4", "paper_pdf_bytes": 1169175, "paper_pdf_source": "openreview", "code_url": "https://github.com/snap-stanford/orca", "code_repository": "snap-stanford/orca", "code_commit": "5f33afdeb0aa1ec51d62c9f926f265504ccf9efc", "code_archive": "repos/6VhmvP7XZue.zip", "code_archive_sha256": "ceef60e184d444cf68895244badcf770b14c988fed98df8a714eae641872efd6", "code_archive_bytes": 16865, "code_file_count": 9, "code_extensions": {".py": 9}, "github_disk_usage_kb": 13, "github_languages": {"Python": 43535}, "github_archived": false, "github_pushed_at": "2022-02-17T04:35:15Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/open-world-semi-supervised-learning-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "BJe4oxHYPB", "year": 2020, "status": "rejected", "title": "Winning the Lottery with Continuous Sparsification", "authors": ["Pedro Savarese", "Hugo Silva", "Michael Maire"], "authorids": ["savarese@ttic.edu", "hugoandradesilva664@gmail.com", "mmaire@uchicago.edu"], "authors_source": "OpenReview API", "abstract": "The Lottery Ticket Hypothesis from Frankle & Carbin (2019) conjectures that, for typically-sized neural networks, it is possible to find small sub-networks which train faster and yield superior performance than their original counterparts. The proposed algorithm to search for such sub-networks (winning tickets), Iterative Magnitude Pruning (IMP), consistently finds sub-networks with 90-95% less parameters which indeed train faster and better than the overparameterized models they were extracted from, creating potential applications to problems such as transfer learning.\n\nIn this paper, we propose a new algorithm to search for winning tickets, Continuous Sparsification, which continuously removes parameters from a network during training, and learns the sub-network's structure with gradient-based methods instead of relying on pruning strategies. We show empirically that our method is capable of finding tickets that outperforms the ones learned by Iterative Magnitude Pruning, and at the same time providing up to 5 times faster search, when measured in number of training epochs.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=BJe4oxHYPB", "arxiv_id": "1912.04427", "paper_pdf": "papers/BJe4oxHYPB.pdf", "paper_pdf_sha256": "e17cb17cf63855e0d10013ad6db1e1153c2b6dd4cf5da813959a8879c120394f", "paper_pdf_bytes": 463296, "paper_pdf_source": "openreview", "code_url": "https://github.com/lolemacs/continuous-sparsification", "code_repository": "lolemacs/continuous-sparsification", "code_commit": "5bd4039f80724bbd71a78ba45d7efd35ea2e8e20", "code_archive": "repos/BJe4oxHYPB.zip", "code_archive_sha256": "f324c6e01e535b95a3021fa560129d4d296c44f4e0cc009dd42fce19b84f37fe", "code_archive_bytes": 9023, "code_file_count": 6, "code_extensions": {".py": 6}, "github_disk_usage_kb": 17, "github_languages": {"Python": 14108}, "github_archived": false, "github_pushed_at": "2022-06-10T10:39:52Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/winning-the-lottery-with-continuous-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ryxDjjCqtQ", "year": 2019, "status": "rejected", "title": "Deconfounding Reinforcement Learning in Observational Settings", "authors": ["Chaochao Lu", "José Miguel Hernández Lobato"], "authorids": ["cl641@cam.ac.uk", "jmh233@cam.ac.uk"], "authors_source": "OpenReview API", "abstract": "In this paper, we propose a general formulation to cope with a family of reinforcement learning tasks in observational settings, that is, learning good policies solely from the historical data produced by real environments with confounders (i.e., the factors affecting both actions and rewards). Based on the proposed approach, we extend one representative of reinforcement learning algorithms: the Actor-Critic method, to its deconfounding variant, which is also straightforward to be applied to other algorithms. In addition, due to lack of datasets in this direction, a benchmark is developed for deconfounding reinforcement learning algorithms by revising OpenAI Gym and MNIST. We demonstrate that the proposed algorithms are superior to traditional reinforcement learning algorithms in confounded environments. To the best of our knowledge, this is the first time that confounders are taken into consideration for addressing full reinforcement learning problems.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=ryxDjjCqtQ", "arxiv_id": "1812.10576", "paper_pdf": "papers/ryxDjjCqtQ.pdf", "paper_pdf_sha256": "69d8396b824ab8094caee9a975aadbd99b50a68991b0c7267f3b6c845f37a7fa", "paper_pdf_bytes": 1784580, "paper_pdf_source": "openreview", "code_url": "https://github.com/CausalRL/DRL", "code_repository": "CausalRL/DRL", "code_commit": "0cd76bcbad8189ff3eb8ee750aaf6e992ce291a9", "code_archive": "repos/ryxDjjCqtQ.zip", "code_archive_sha256": "ee1b041e506020300693f36eec1b45690f4a282a573cd0953c870e4ca18aad87", "code_archive_bytes": 45636, "code_file_count": 18, "code_extensions": {".py": 18}, "github_disk_usage_kb": 24, "github_languages": {"Python": 227459}, "github_archived": false, "github_pushed_at": "2019-04-13T22:45:57Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/deconfounding-reinforcement-learning-in"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "B1bgpzZAZ", "year": 2018, "status": "rejected", "title": "ElimiNet: A Model for Eliminating Options for Reading Comprehension with Multiple Choice Questions", "authors": ["Soham Parikh", "Ananya Sai", "Preksha Nema", "Mitesh M Khapra"], "authorids": ["sohamp@cse.iitm.ac.in", "ananyasb@cse.iitm.ac.in", "preksha@cse.iitm.ac.in", "miteshk@cse.iitm.ac.in"], "authors_source": "OpenReview API", "abstract": "The task of Reading Comprehension with Multiple Choice Questions, requires a human (or machine) to read a given \\{\\textit{passage, question}\\} pair and select one of the $n$ given options. The current state of the art model for this task first computes a query-aware representation for the passage and then \\textit{selects} the option which has the maximum similarity with this representation. However, when humans perform this task they do not just focus on option selection but use a combination of \\textit{elimination} and \\textit{selection}. Specifically, a human would first try to eliminate the most irrelevant option and then read the document again in the light of this new information (and perhaps ignore portions corresponding to the eliminated option). This process could be repeated multiple times till the reader is finally ready to select the correct option. We propose \\textit{ElimiNet}, a neural network based model which tries to mimic this process. Specifically, it has gates which decide whether an option can be eliminated given the \\{\\textit{document, question}\\} pair and if so it tries to make the document representation orthogonal to this eliminatedd option (akin to ignoring portions of the document corresponding to the eliminated option). The model makes multiple rounds of partial elimination to refine the document representation and finally uses a selection module to pick the best option. We evaluate our model on the recently released large scale RACE dataset and show that it outperforms the current state of the art model on 7 out of the 13 question types in this dataset. Further we show that taking an ensemble of our \\textit{elimination-selection} based method with a \\textit{selection} based method gives us an improvement of 7\\% (relative) over the best reported performance on this dataset.    \n", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=B1bgpzZAZ", "arxiv_id": "1904.02651", "paper_pdf": "papers/B1bgpzZAZ.pdf", "paper_pdf_sha256": "b48c4c1f871aa6f700303934a8668963d33fdcfca7e1e902e04859bbb5f48ab8", "paper_pdf_bytes": 508977, "paper_pdf_source": "openreview", "code_url": "https://github.com/sohamparikh/ElimiNet", "code_repository": "sohamparikh/ElimiNet", "code_commit": "f2a9f293fa2ce5470522444c8cfa6523c848392a", "code_archive": "repos/B1bgpzZAZ.zip", "code_archive_sha256": "082dd1559dda2c785c6dafa812d121e07d489569010cc504be6e8c30ac400a0a", "code_archive_bytes": 15164, "code_file_count": 8, "code_extensions": {".py": 5, ".sh": 3}, "github_disk_usage_kb": 17, "github_languages": {"Python": 56689, "Shell": 2332}, "github_archived": false, "github_pushed_at": "2019-04-11T21:22:09Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/eliminet-a-model-for-eliminating-options-for-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "JeiaHDawhb", "year": 2025, "status": "rejected", "title": "Maximum Total Correlation Reinforcement Learning", "authors": ["Bang You", "Puze Liu", "Huaping Liu", "Jan Peters", "Oleg Arenz"], "authorids": ["~Bang_You1", "~Puze_Liu1", "~Huaping_Liu3", "~Jan_Peters3", "~Oleg_Arenz1"], "authors_source": "OpenReview API", "abstract": "Simplicity is a powerful inductive bias. In reinforcement learning, regularization is used for simpler policies, data augmentation for simpler representations, and sparse reward functions for simpler objectives, all that, with the underlying motivation to increase generalizability and robustness by focusing on the essentials. Supplementary to these techniques, we investigate how to promote simple behavior throughout the duration of the episode. To that end, we introduce a modification of the reinforcement learning problem, that additionally maximizes the total correlation within the induced trajectories. We propose a practical algorithm that optimizes all models, including policy and state representation, based on a lower bound approximation. In simulated robot locomotion environments, our method naturally generates policies that induce periodic and compressible trajectories, and that exhibit superior robustness to noise and changes in dynamics compared to baseline methods, while also improving performance in the original tasks.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=JeiaHDawhb", "arxiv_id": "2505.16734", "paper_pdf": "papers/JeiaHDawhb.pdf", "paper_pdf_sha256": "06c846adbac422813c8e2e3b292b0d92251a7c2cdbc0258157587faa17eeaaec", "paper_pdf_bytes": 454123, "paper_pdf_source": "openreview", "code_url": "https://github.com/BangYou01/MTC", "code_repository": "BangYou01/MTC", "code_commit": "7b4dfae88d6182e58fa7c7c85af5a6bfadc5f863", "code_archive": "repos/JeiaHDawhb.zip", "code_archive_sha256": "ac2fc141ce9541ce029fba1050d6a435859f1aebb0857489c4d37c9ee8e1ff90", "code_archive_bytes": 14799, "code_file_count": 6, "code_extensions": {".py": 5, ".sh": 1}, "github_disk_usage_kb": 13, "github_languages": {"Python": 50767, "Shell": 426}, "github_archived": false, "github_pushed_at": "2025-05-28T06:55:08Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/maximum-total-correlation-reinforcement"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "SHUQtRK0eU", "year": 2024, "status": "rejected", "title": "Generalized Activation via Multivariate Projection", "authors": ["Jiayun Li", "Yuxiao Cheng", "Zhuofan Xia", "Yilin Mo", "Gao Huang"], "authorids": ["~Jiayun_Li2", "~Yuxiao_Cheng1", "~Zhuofan_Xia2", "~Yilin_Mo1", "~Gao_Huang1"], "authors_source": "OpenReview API", "abstract": "Activation functions are essential to introduce nonlinearity into neural networks, with the Rectified Linear Unit (ReLU) often favored for its simplicity and effectiveness. Motivated by the structural similarity between a shallow Feedforward Neural Network (FNN) and a single iteration of the Projected Gradient Descent (PGD) algorithm, a standard approach for solving constrained optimization problems, we consider ReLU as a projection from $\\mathbb{R}$ onto the nonnegative half-line $\\mathbb{R}_+$. \nBuilding on this interpretation, we extend ReLU by substituting it with a generalized projection operator onto a convex cone, such as the Second-Order Cone (SOC) projection, thereby naturally extending it to a Multivariate Projection Unit (MPU), an activation function with multiple inputs and multiple outputs.\nWe further provide a mathematical proof establishing that FNNs activated by SOC projections outperform those utilizing ReLU in terms of expressive power. Experimental evaluations on widely-adopted architectures further corroborate MPU's effectiveness against a broader range of existing activation functions.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=SHUQtRK0eU", "arxiv_id": "2309.17194", "paper_pdf": "papers/SHUQtRK0eU.pdf", "paper_pdf_sha256": "db6c5d626825d337d5d42ef32aa6a8844a3734174b54714b3d379e7960aba569", "paper_pdf_bytes": 1459067, "paper_pdf_source": "openreview", "code_url": "https://github.com/ljy9912/mimo_nn", "code_repository": "ljy9912/mimo_nn", "code_commit": "f008dcc4ba9f1a9931d89e671654de244ede33b3", "code_archive": "repos/SHUQtRK0eU.zip", "code_archive_sha256": "5bb3a03a8c764fd117297f17e43c5257a0ccbe77928ff6e249f84921d7279bec", "code_archive_bytes": 11571, "code_file_count": 8, "code_extensions": {".py": 6, ".sh": 2}, "github_disk_usage_kb": 14, "github_languages": {"Python": 32462, "Shell": 1149}, "github_archived": false, "github_pushed_at": "2024-05-20T04:00:07Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/generalized-activation-via-multivariate"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "2_BsVZ6R-ef", "year": 2023, "status": "rejected", "title": "Analytical Composition of Differential Privacy via the Edgeworth Accountant", "authors": ["Hua Wang", "Sheng Gao", "Huanyu Zhang", "Milan Shen", "Weijie J Su"], "authorids": ["~Hua_Wang7", "~Sheng_Gao2", "~Huanyu_Zhang2", "~Milan_Shen1", "~Weijie_J_Su1"], "authors_source": "OpenReview API", "abstract": "Many modern machine learning algorithms are composed of simple private algorithms; thus, an increasingly important problem is to efficiently compute the overall privacy loss under composition. In this study, we introduce the Edgeworth Accountant, an analytical approach to composing differential privacy guarantees of private algorithms. The Edgeworth Accountant starts by losslessly tracking the privacy loss under composition using the $f$-differential privacy framework, which allows us to express the privacy guarantees using privacy-loss log-likelihood ratios (PLLRs). As the name suggests, this accountant next uses the Edgeworth expansion to the upper and lower bounds the probability distribution of the sum of the PLLRs. Moreover, by relying on a technique for approximating complex distributions using simple ones, we demonstrate that the Edgeworth Accountant can be applied to the composition of any noise-addition mechanism. Owing to certain appealing features of the Edgeworth expansion, the $(\\epsilon, \\delta)$-differential privacy bounds offered by this accountant are non-asymptotic, with essentially no extra computational cost, as opposed to the prior approaches, wherein the running times increase with the number of compositions. Finally, we demonstrate that our upper and lower $(\\epsilon, \\delta)$-differential privacy bounds are tight in federated analytics and certain regimes of training private deep learning models.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=2_BsVZ6R-ef", "arxiv_id": "2206.04236", "paper_pdf": "papers/2_BsVZ6R-ef.pdf", "paper_pdf_sha256": "e45932a3e10bf5dbb9b0605470d57980c2956ac13f84fe35762c8b1cede33e47", "paper_pdf_bytes": 394791, "paper_pdf_source": "openreview", "code_url": "https://github.com/HuaWang-wharton/EdgeworthAccountant", "code_repository": "HuaWang-wharton/EdgeworthAccountant", "code_commit": "6cd5a3eb973cfde5e734cb125119ef9ab26e0e00", "code_archive": "repos/2_BsVZ6R-ef.zip", "code_archive_sha256": "7c832955dcea2601a957eeb3684e3147df18af5e85e4621abcf493f063f38d95", "code_archive_bytes": 15139, "code_file_count": 7, "code_extensions": {".py": 7}, "github_disk_usage_kb": 15, "github_languages": {"Python": 30528}, "github_archived": false, "github_pushed_at": "2022-06-09T01:57:56Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/analytical-composition-of-differential"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "JVsvIuMDE0Z", "year": 2022, "status": "rejected", "title": "Adaptive Behavior Cloning Regularization for Stable Offline-to-Online Reinforcement Learning", "authors": ["Yi Zhao", "Rinu Boney", "Alexander Ilin", "Juho Kannala", "Joni Pajarinen"], "authorids": ["~Yi_Zhao6", "~Rinu_Boney1", "~Alexander_Ilin1", "~Juho_Kannala1", "~Joni_Pajarinen2"], "authors_source": "OpenReview API", "abstract": "Offline reinforcement learning, by learning from a fixed dataset, makes it possible to learn agent behaviors without interacting with the environment. However, depending on the quality of the offline dataset, such pre-trained agents may have limited performance and would further need to be fine-tuned online by interacting with the environment. During online fine-tuning, the performance of the pre-trained agent may collapse quickly due to the sudden distribution shift from offline to online data. While constraints enforced by offline RL methods such as a behaviour cloning loss prevent this to an extent, these constraints also significantly slow down online fine-tuning by forcing the agent to stay close to the behavior policy. We propose to adaptively weigh the behavior cloning loss during online fine-tuning based on the agent's performance and training stability. Moreover, we use a randomized ensemble of Q functions to further increase the sample efficiency of online fine-tuning by performing a large number of learning updates. Experiments show that the proposed method yields state-of-the-art offline-to-online reinforcement learning performance on the popular D4RL benchmark.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=JVsvIuMDE0Z", "arxiv_id": "2210.13846", "paper_pdf": "papers/JVsvIuMDE0Z.pdf", "paper_pdf_sha256": "70f0482bd5c20f441a0698260c63a09406237927d910e6f7782cfc896f2f46e2", "paper_pdf_bytes": 437231, "paper_pdf_source": "openreview", "code_url": "https://github.com/zhaoyi11/adaptive_bc", "code_repository": "zhaoyi11/adaptive_bc", "code_commit": "d47226d4ffc9959ac0ed1337ddd7ebb787fc05e1", "code_archive": "repos/JVsvIuMDE0Z.zip", "code_archive_sha256": "5e9adbc418534e878989df1a2cbdb26b8a0ea69264aac3dd40ff236757d8de5f", "code_archive_bytes": 13306, "code_file_count": 3, "code_extensions": {".py": 3}, "github_disk_usage_kb": 14, "github_languages": {"Python": 27310}, "github_archived": false, "github_pushed_at": "2022-07-04T08:28:37Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/adaptive-behavior-cloning-regularization-for-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "crAi7c41xTh", "year": 2021, "status": "rejected", "title": "Shape Matters: Understanding the Implicit Bias of the Noise Covariance", "authors": ["Jeff Z. HaoChen", "Colin Wei", "Jason D. Lee", "Tengyu Ma"], "authorids": ["~Jeff_Z._HaoChen1", "~Colin_Wei1", "~Jason_D._Lee1", "~Tengyu_Ma1"], "authors_source": "OpenReview API", "abstract": "The noise in stochastic gradient descent (SGD) provides a crucial implicit regularization effect for training overparameterized models. Prior theoretical work largely focuses on spherical Gaussian noise, whereas empirical studies demonstrate the phenomenon that parameter-dependent noise --- induced by mini-batches or label perturbation --- is far more effective than Gaussian noise. \nThis paper theoretically characterizes this phenomenon on a quadratically-parameterized model introduced by Vaskevicius et al. and Woodworth et al.  We show that in an over-parameterized setting, SGD with label noise recovers the sparse ground-truth with an arbitrary initialization, whereas SGD with Gaussian noise or gradient descent overfits to dense solutions with large norms. Our analysis reveals that parameter-dependent noise introduces a bias towards local minima with smaller noise variance, whereas spherical Gaussian noise does not. ", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=crAi7c41xTh", "arxiv_id": "2006.08680", "paper_pdf": "papers/crAi7c41xTh.pdf", "paper_pdf_sha256": "fd62ac31d32f0bc4744b3ce6ba516caa8054d872b808125631bbee0a69774fe3", "paper_pdf_bytes": 448050, "paper_pdf_source": "openreview", "code_url": "https://github.com/jhaochenz96/noise-implicit-bias", "code_repository": "jhaochenz96/noise-implicit-bias", "code_commit": "2b94db1672edf916851cd806c64c246000e2944e", "code_archive": "repos/crAi7c41xTh.zip", "code_archive_sha256": "ccf4a40dd299a9634fb0e5c21e326de86f0c002d39dbf6b01d5de3b498b436f7", "code_archive_bytes": 11874, "code_file_count": 8, "code_extensions": {".py": 8}, "github_disk_usage_kb": 15, "github_languages": {"Python": 27244}, "github_archived": false, "github_pushed_at": "2020-06-17T04:37:15Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/shape-matters-understanding-the-implicit-bias"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "HJlzxgBtwH", "year": 2020, "status": "rejected", "title": "Minimally distorted Adversarial Examples with a Fast Adaptive Boundary Attack", "authors": ["Francesco Croce", "Matthias Hein"], "authorids": ["francesco91.croce@gmail.com", "matthias.hein@uni-tuebingen.de"], "authors_source": "OpenReview API", "abstract": "The evaluation of robustness against adversarial manipulations of neural networks-based classifiers is mainly tested with empirical attacks as the methods for the exact computation, even when available, do not scale to large networks. We propose in this paper a new white-box adversarial attack wrt the $l_p$-norms for $p \\in \\{1,2,\\infty\\}$ aiming at finding the minimal perturbation necessary to change the class of a given input. It has an intuitive geometric meaning, yields quickly high quality results, minimizes the size of the perturbation (so that it returns the robust accuracy at every threshold with a single run). It performs better or similarly to state-of-the-art attacks which are partially specialized to one $l_p$-norm.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=HJlzxgBtwH", "arxiv_id": "1907.02044", "paper_pdf": "papers/HJlzxgBtwH.pdf", "paper_pdf_sha256": "8bc18107b29805c3f4690737aec72d218d5c1a6751e7f0ddc4b2958ad1952170", "paper_pdf_bytes": 3264275, "paper_pdf_source": "openreview", "code_url": "https://github.com/fra31/fab-attack", "code_repository": "fra31/fab-attack", "code_commit": "7c3ef4cb4cd91ade912f06c683fbed8a4660464d", "code_archive": "repos/HJlzxgBtwH.zip", "code_archive_sha256": "97d6d30d34229ac69bcb9dd17cc84210147b93fa773ed331ed9de3f0201573f2", "code_archive_bytes": 18226, "code_file_count": 7, "code_extensions": {".py": 7}, "github_disk_usage_kb": 18, "github_languages": {"Python": 47504}, "github_archived": false, "github_pushed_at": "2020-07-10T12:35:52Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/minimally-distorted-adversarial-examples-with"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "HJx7l309Fm", "year": 2019, "status": "rejected", "title": "Actor-Attention-Critic for Multi-Agent Reinforcement Learning", "authors": ["Shariq Iqbal", "Fei Sha"], "authorids": ["shariqiqbal2810@gmail.com", "feisha.work@gmail.com"], "authors_source": "OpenReview API", "abstract": "Reinforcement learning in multi-agent scenarios is important for real-world applications but presents challenges beyond those seen in single-agent settings. We present an actor-critic algorithm that trains decentralized policies in multi-agent settings, using centrally computed critics that share an attention mechanism which selects relevant information for each agent at every timestep. This attention mechanism enables more effective and scalable learning in complex multi-agent environments, when compared to recent approaches. Our approach is applicable not only to cooperative settings with shared rewards, but also individualized reward settings, including adversarial settings, and it makes no assumptions about the action spaces of the agents. As such, it is flexible enough to be applied to most multi-agent learning problems", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=HJx7l309Fm", "arxiv_id": "1810.02912", "paper_pdf": "papers/HJx7l309Fm.pdf", "paper_pdf_sha256": "dcdf7ac024d4ee05494c3f1a323a17bf4bde661b97c82292d9e5546ff9c50ccd", "paper_pdf_bytes": 739740, "paper_pdf_source": "openreview", "code_url": "https://github.com/shariqiqbal2810/MAAC", "code_repository": "shariqiqbal2810/MAAC", "code_commit": "6174a01251251e6778c4ada26bc8d9cd930e3856", "code_archive": "repos/HJx7l309Fm.zip", "code_archive_sha256": "b3ddcd48636a5fda010f64338e06e632a7c600ac3f400a04207bbec175f99636", "code_archive_bytes": 23481, "code_file_count": 12, "code_extensions": {".py": 12}, "github_disk_usage_kb": 30, "github_languages": {"Python": 66332}, "github_archived": false, "github_pushed_at": "2022-05-29T16:14:16Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/actor-attention-critic-for-multi-agent"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "SybqeKgA-", "year": 2018, "status": "rejected", "title": "On Batch Adaptive Training for Deep Learning: Lower Loss and Larger Step Size", "authors": ["Runyao Chen", "Kun Wu", "Ping Luo"], "authorids": ["chenrunyao14@mails.ucas.ac.cn", "WuKun14@mails.ucas.ac.cn", "luop@ict.ac.cn"], "authors_source": "OpenReview API", "abstract": "Mini-batch gradient descent and its variants are commonly used in deep learning. The principle of mini-batch gradient descent is to use noisy gradient calculated on a batch to estimate the real gradient, thus balancing the computation cost per iteration and the uncertainty of noisy gradient. However, its batch size is a fixed hyper-parameter requiring manual setting before training the neural network. Yin et al. (2017) proposed a batch adaptive stochastic gradient descent (BA-SGD) that can dynamically choose a proper batch size as learning proceeds. We extend the BA-SGD to momentum algorithm and evaluate both the BA-SGD and the batch adaptive momentum (BA-Momentum) on two deep learning tasks from natural language processing to image classification. Experiments confirm that batch adaptive methods can achieve a lower loss compared with mini-batch methods after scanning the same epochs of data. Furthermore, our BA-Momentum is more robust against larger step sizes, in that it can dynamically enlarge the batch size to reduce the larger uncertainty brought by larger step sizes. We also identified an interesting phenomenon, batch size boom. The code implementing batch adaptive framework is now open source, applicable to any gradient-based optimization problems.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=SybqeKgA-", "arxiv_id": null, "paper_pdf": "papers/SybqeKgA-.pdf", "paper_pdf_sha256": "e98e153f12732b25dfda3eff31c04e5ea49ba7daf66402734caf1799a3bc7ab7", "paper_pdf_bytes": 1551207, "paper_pdf_source": "openreview", "code_url": "https://github.com/thomasyao3096/Batch_Adaptive_Framework", "code_repository": "thomasyao3096/Batch_Adaptive_Framework", "code_commit": "2244eebc33938ef816120a80eb2560192ec08b41", "code_archive": "repos/SybqeKgA-.zip", "code_archive_sha256": "f53cd8052f21d0cf8cd9de2c3e605ddbd47b1f47c59cda5d3abaa5ed39d86dc9", "code_archive_bytes": 11261, "code_file_count": 4, "code_extensions": {".py": 4}, "github_disk_usage_kb": 19, "github_languages": {"Python": 32613}, "github_archived": false, "github_pushed_at": "2017-11-12T02:27:36Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/on-batch-adaptive-training-for-deep-learning"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ODibPQmeP1", "year": 2026, "status": "rejected", "title": "CHARM: Calibrating Reward Models With Chatbot Arena Scores", "authors": ["Xiao Zhu", "Chenmien Tan", "Pinzhen Chen", "Rico Sennrich", "Yanlin Zhang", "Hanxu Hu"], "authorids": ["~Xiao_Zhu4", "~Chenmien_Tan1", "~Pinzhen_Chen1", "~Rico_Sennrich1", "~Yanlin_Zhang1", "~Hanxu_Hu1"], "authors_source": "OpenReview API", "abstract": "Reward models (RMs) play a crucial role in Reinforcement Learning from Human Feedback by serving as proxies for human preferences in aligning large language models. However, they suffer from various biases which could lead to reward hacking. In this paper, we identify a model preference bias in RMs, where they systematically assign disproportionately high scores to responses from certain policy models, leading to unfair judgments. To mitigate this bias, we propose a calibration method named **CH**atbot **A**rena calibrated **R**eward **M**odeling (**CHARM**) that leverages Elo scores from the Chatbot Arena to construct debiased preference datasets and adjust reward model scoring. We conduct extensive experiments on reward model benchmarks and human preference alignment. Results demonstrate that our calibrated RMs achieve improved evaluation accuracy on RM-Bench and the Chat-Hard domain of RewardBench and exhibit a stronger correlation with human preferences by producing scores more closely aligned with Elo rankings.  Beyond this, **CHARM** enhances robustness to stylistic variations, mitigates implicit pattern bias, and generalizes to unseen models. These results demonstrate that **CHARM** provides a simple, effective, and broadly applicable approach to building more reliable and fair reward models.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=ODibPQmeP1", "arxiv_id": "2504.10045", "paper_pdf": "papers/ODibPQmeP1.pdf", "paper_pdf_sha256": "a83b6bf53aa1f23dda4b9d2b24478039d97acaeb0fbb78db1b745a65f814d236", "paper_pdf_bytes": 1281082, "paper_pdf_source": "openreview", "code_url": "https://github.com/HexagonStar/CHARM", "code_repository": "HexagonStar/CHARM", "code_commit": "719f0aef03a1eb8381bad450b0d991455b91563a", "code_archive": "repos/ODibPQmeP1.zip", "code_archive_sha256": "2629f6448f327bc04d0ff4af9b127e6069eb7b624197d5117372fc5e7dc3cb3f", "code_archive_bytes": 21492, "code_file_count": 12, "code_extensions": {".py": 7, ".sh": 5}, "github_disk_usage_kb": 42, "github_languages": {"Python": 32355, "Shell": 2532}, "github_archived": false, "github_pushed_at": "2025-04-15T05:31:31Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/charm-calibrating-reward-models-with-chatbot"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "pogJXugbN8", "year": 2024, "status": "rejected", "title": "BAFFLE: A Baseline of Backpropagation-Free Federated Learning", "authors": ["Haozhe Feng", "Tianyu Pang", "Chao Du", "Wei Chen", "Shuicheng YAN", "Min Lin"], "authorids": ["~Haozhe_Feng1", "~Tianyu_Pang1", "~Chao_Du1", "~Wei_Chen34", "~Shuicheng_YAN3", "~Min_Lin1"], "authors_source": "OpenReview API", "abstract": "Federated learning (FL) is a general principle for decentralized clients to train a server model collectively without sharing local data. FL is a promising framework with practical applications, but its standard training paradigm requires the clients to backpropagate through the model to compute gradients. Since these clients are typically edge devices and not fully trusted, executing backpropagation on them incurs computational and storage overhead as well as white-box vulnerability. In light of this, we develop backpropagation-free federated learning, dubbed BAFFLE, in which backpropagation is replaced by multiple forward processes to estimate gradients. BAFFLE is 1) memory-efficient and easily fits uploading bandwidth; 2) compatible with inference-only hardware optimization and model quantization or pruning; and 3) well-suited to trusted execution environments, because the clients in BAFFLE only execute forward propagation and return a set of scalars to the server. Empirically we use BAFFLE to train deep models from scratch or to finetune pretrained models, achieving acceptable results.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=pogJXugbN8", "arxiv_id": "2301.12195", "paper_pdf": "papers/pogJXugbN8.pdf", "paper_pdf_sha256": "a0cf61126a3f28bb76de4ec658d5bdf238284693e3dcfa510292575be86fb391", "paper_pdf_bytes": 907149, "paper_pdf_source": "openreview", "code_url": "https://github.com/FengHZ/BAFFLE", "code_repository": "FengHZ/BAFFLE", "code_commit": "b59374505eb21c6d3ae2c6a513556c05d2d1dce9", "code_archive": "repos/pogJXugbN8.zip", "code_archive_sha256": "06c2d593027c587a16f587b7a7e4aff8f309e4f23fb1e891870084d8ac0a82c9", "code_archive_bytes": 15249, "code_file_count": 8, "code_extensions": {".py": 8}, "github_disk_usage_kb": 15, "github_languages": {"Python": 45513}, "github_archived": false, "github_pushed_at": "2023-02-09T11:41:10Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/does-federated-learning-really-need"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "w1w4dGJ4qV", "year": 2023, "status": "rejected", "title": "The Benefits of Model-Based Generalization in Reinforcement Learning", "authors": ["Kenny John Young", "Aditya Ramesh", "Louis Kirsch", "Jürgen Schmidhuber"], "authorids": ["~Kenny_John_Young1", "~Aditya_Ramesh2", "~Louis_Kirsch1", "~Jürgen_Schmidhuber1"], "authors_source": "OpenReview API", "abstract": "Model-Based Reinforcement Learning (RL) is widely believed to have the potential to improve sample efficiency by allowing an agent to synthesize large amounts of imagined experience. Experience Replay (ER) can be considered a simple kind of model, which has proved extremely effective at improving the stability and efficiency of deep RL. In principle, a learned parametric model could improve on ER by generalizing from real experience to augment the dataset with additional plausible experience. However, owing to the many design choices involved in empirically successful algorithms, it can be very hard to establish where the benefits are actually coming from. Here, we provide theoretical and empirical insight into when, and how, we can expect data generated by a learned model to be useful. First, we provide a general theorem motivating how learning a model as an intermediate step can narrow down the set of possible value functions more than learning a value function directly from data using the Bellman equation.  Second, we provide an illustrative example showing empirically how a similar effect occurs in a more concrete setting with neural network function approximation.  Finally, we provide extensive experiments showing the benefit of model-based learning for online RL in environments with combinatorial complexity, but factored structure that allows a learned model to generalize. In these experiments, we take care to control for other factors in order to isolate, insofar as possible, the benefit of using experience generated by a learned model relative to ER alone.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=w1w4dGJ4qV", "arxiv_id": "2211.02222", "paper_pdf": "papers/w1w4dGJ4qV.pdf", "paper_pdf_sha256": "e2fcee91ff9096c4126dc82e6194fdf8b55a35e72f988e956fd8b47e5826df6d", "paper_pdf_bytes": 15819893, "paper_pdf_source": "openreview", "code_url": "https://github.com/kenjyoung/Model_Generalization_Code_supplement", "code_repository": "kenjyoung/Model_Generalization_Code_supplement", "code_commit": "dd3c5a950af69ea49cadbcea1315cb193a2e1c08", "code_archive": "repos/w1w4dGJ4qV.zip", "code_archive_sha256": "2c60980fe20cf29cc1e56de67e5c1c12d1187fe3d486b3cc2d9446825020f439", "code_archive_bytes": 31426, "code_file_count": 8, "code_extensions": {".py": 8}, "github_disk_usage_kb": 18, "github_languages": {"Python": 107954}, "github_archived": false, "github_pushed_at": "2023-01-13T19:03:22Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/the-benefits-of-model-based-generalization-in"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "-3Qj7Jl6UP5", "year": 2022, "status": "rejected", "title": "The magnitude vector of images", "authors": ["Michael F Adamer", "Leslie O'Bray", "Edward De Brouwer", "Bastian Rieck", "Karsten Borgwardt"], "authorids": ["~Michael_F_Adamer1", "~Leslie_O'Bray1", "~Edward_De_Brouwer1", "~Bastian_Rieck1", "~Karsten_Borgwardt2"], "authors_source": "OpenReview API", "abstract": "The magnitude of a finite metric space is a recently-introduced invariant quantity. Despite beneficial theoretical and practical properties, such as a general utility for outlier detection, and a close connection to Laplace radial basis kernels, magnitude has received little attention by the machine learning community so far. In this work, we investigate the properties of magnitude on individual images, with each image forming its own metric space. We show that the known properties of outlier detection translate to edge detection in images and we give supporting theoretical justifications. In addition, we provide a proof of concept of its utility by using a novel magnitude layer to defend against adversarial attacks. Since naive magnitude calculations may be computationally prohibitive, we introduce an algorithm that leverages the regular structure of images to dramatically reduce the computational cost.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=-3Qj7Jl6UP5", "arxiv_id": "2110.15188", "paper_pdf": "papers/-3Qj7Jl6UP5.pdf", "paper_pdf_sha256": "9d22ebc02f37c7f8999316f15db468927be2c72583fa8bfde309563ab9a6d3bc", "paper_pdf_bytes": 1553416, "paper_pdf_source": "openreview", "code_url": "https://github.com/MikeAdamer/mag-metric", "code_repository": "MikeAdamer/mag-metric", "code_commit": "1f33528faf55d32d4cd8cfa983bbc95cf9b3dce0", "code_archive": "repos/-3Qj7Jl6UP5.zip", "code_archive_sha256": "ba7b0cc320fa111f048e036608739e858eb4e488a01d2fb2f04471c6e7e19ae1", "code_archive_bytes": 21211, "code_file_count": 10, "code_extensions": {".py": 10}, "github_disk_usage_kb": 16, "github_languages": {"Python": 58499}, "github_archived": false, "github_pushed_at": "2022-09-30T14:26:52Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/the-magnitude-vector-of-images-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "BJxQxeBYwH", "year": 2020, "status": "rejected", "title": "Are Powerful Graph Neural Nets Necessary? A Dissection on Graph Classification", "authors": ["Ting Chen", "Song Bian", "Yizhou Sun"], "authorids": ["iamtingchen@gmail.com", "biansonghz@gmail.com", "yzsun@cs.ucla.edu"], "authors_source": "OpenReview API", "abstract": "Graph Neural Nets (GNNs) have received increasing attentions, partially due to their superior performance in many node and graph classification tasks. However, there is a lack of understanding on what they are learning and how sophisticated the learned graph functions are.  In this work, we propose a dissection of GNNs on graph classification into two parts: 1) the graph filtering, where graph-based neighbor aggregations are performed, and 2) the set function, where a set of hidden node features are composed for prediction. To study the importance of both parts, we propose to linearize them separately. We first linearize the graph filtering function, resulting Graph Feature Network (GFN), which is a simple lightweight neural net defined on a \\textit{set} of graph augmented features. Further linearization of GFN's set function results in Graph Linear Network (GLN), which is a linear function. Empirically we perform evaluations on common graph classification benchmarks. To our surprise, we find that, despite the simplification, GFN could match or exceed the best accuracies produced by recently proposed GNNs (with a fraction of computation cost), while GLN underperforms significantly. Our results demonstrate the importance of non-linear set function, and suggest that linear graph filtering with non-linear set function is an efficient and powerful scheme for modeling existing graph classification benchmarks.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=BJxQxeBYwH", "arxiv_id": "1905.04579", "paper_pdf": "papers/BJxQxeBYwH.pdf", "paper_pdf_sha256": "4c0c8fba22739b685bbbbe1d0504875aa20b835d46e1ff80b4aaf3696f27feba", "paper_pdf_bytes": 1431009, "paper_pdf_source": "openreview", "code_url": "https://github.com/chentingpc/gfn", "code_repository": "chentingpc/gfn", "code_commit": "b598eb77f62680e62e2adb558b158505b2a3926a", "code_archive": "repos/BJxQxeBYwH.zip", "code_archive_sha256": "e84c5442f37fd839767ca4446d3ae7bada5b171133c4c64c54094fec228983d6", "code_archive_bytes": 16195, "code_file_count": 9, "code_extensions": {".py": 9}, "github_disk_usage_kb": 21, "github_languages": {"Python": 53680}, "github_archived": false, "github_pushed_at": "2020-05-05T01:49:13Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/dissecting-graph-neural-networks-on-graph"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "S1giVsRcYm", "year": 2019, "status": "rejected", "title": "Count-Based Exploration with the Successor Representation", "authors": ["Marlos C. Machado", "Marc G. Bellemare", "Michael Bowling"], "authorids": ["machado@ualberta.ca", "bellemare@google.com", "mbowling@ualberta.ca"], "authors_source": "OpenReview API", "abstract": "The problem of exploration in reinforcement learning is well-understood in the tabular case and many sample-efficient algorithms are known. Nevertheless, it is often unclear how the algorithms in the tabular setting can be extended to tasks with large state-spaces where generalization is required. Recent promising developments generally depend on problem-specific density models or handcrafted features. In this paper we introduce a simple approach for exploration that allows us to develop theoretically justified algorithms in the tabular case but that also give us intuitions for new algorithms applicable to settings where function approximation is required. Our approach and its underlying theory is based on the substochastic successor representation, a concept we develop here. While the traditional successor representation is a representation that defines state generalization by the similarity of successor states, the substochastic successor representation is also able to implicitly count the number of times each state (or feature) has been observed. This extension connects two until now disjoint areas of research. We show in traditional tabular domains (RiverSwim and SixArms) that our algorithm empirically performs as well as other sample-efficient algorithms. We then describe a deep reinforcement learning algorithm inspired by these ideas and show that it matches the performance of recent pseudo-count-based methods in hard exploration Atari 2600 games.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=S1giVsRcYm", "arxiv_id": "1807.11622", "paper_pdf": "papers/S1giVsRcYm.pdf", "paper_pdf_sha256": "3eaf0a7a954c68f3833685770709e661b8d0aa8da03f9151a4f72dd77edb4cd0", "paper_pdf_bytes": 3941702, "paper_pdf_source": "openreview", "code_url": "https://github.com/mcmachado/count_based_exploration_sr", "code_repository": "mcmachado/count_based_exploration_sr", "code_commit": "e4d657eb498ca84a703ddd7ec426d908adc67a72", "code_archive": "repos/S1giVsRcYm.zip", "code_archive_sha256": "14563765715ec9109131eda98b3fa3c9537c66cf80d885bfb3f7e553abe5d450", "code_archive_bytes": 38266, "code_file_count": 23, "code_extensions": {".py": 23}, "github_disk_usage_kb": 31, "github_languages": {"Python": 93470}, "github_archived": false, "github_pushed_at": "2019-07-01T20:22:50Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/count-based-exploration-with-the-successor"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "B1tExikAW", "year": 2018, "status": "rejected", "title": "LatentPoison -- Adversarial Attacks On The Latent Space", "authors": ["Antonia Creswell", "Biswa Sengupta", "Anil A. Bharath"], "authorids": ["ac2211@ic.ac.uk", "b.sengupta@imperial.ac.uk", "a.bharath@imperial.ac.uk"], "authors_source": "OpenReview API", "abstract": "Robustness and security of machine learning (ML) systems are intertwined, wherein a non-robust ML system (classifiers, regressors, etc.) can be subject to attacks using a wide variety of exploits. With the advent of scalable deep learning methodologies, a lot of emphasis has been put on the robustness of supervised, unsupervised and reinforcement learning algorithms. Here, we study the robustness of the latent space of a deep variational autoencoder (dVAE), an unsupervised generative framework, to show that it is indeed possible to perturb the latent space, flip the class predictions and keep the classification probability approximately equal before and after an attack. This means that an agent that looks at the outputs of a decoder would remain oblivious to an attack.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=B1tExikAW", "arxiv_id": "1711.02879", "paper_pdf": "papers/B1tExikAW.pdf", "paper_pdf_sha256": "9eb7ea64274e6f133f2e993a4bf7b7e6dc0c50356c73ecff60f9756082ad202a", "paper_pdf_bytes": 8626408, "paper_pdf_source": "openreview", "code_url": "https://github.com/ToniCreswell/Adversarial-Attack-On-Latent-Space", "code_repository": "ToniCreswell/Adversarial-Attack-On-Latent-Space", "code_commit": "812de9a3c6a6be8ad6e2923625d09b900fe07a00", "code_archive": "repos/B1tExikAW.zip", "code_archive_sha256": "49ed91c4148515ca56b70e9c8b8ef4f686c21bcff7f6f9243919e2b4f61b9c55", "code_archive_bytes": 18206, "code_file_count": 6, "code_extensions": {".py": 5, ".ipynb": 1}, "github_disk_usage_kb": 22, "github_languages": {"Python": 37809, "Jupyter Notebook": 2909}, "github_archived": false, "github_pushed_at": "2022-12-07T23:58:00Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/latentpoison-adversarial-attacks-on-the"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Bibt0JTvpx", "year": 2026, "status": "rejected", "title": "EPiC: Towards Lossless Speedup for Reasoning Training through Edge-Preserving CoT Condensation", "authors": ["Jinghan Jia", "Hadi Reisizadeh", "Chongyu Fan", "Nathalie Baracaldo", "Mingyi Hong", "Sijia Liu"], "authorids": ["~Jinghan_Jia1", "~Hadi_Reisizadeh1", "~Chongyu_Fan1", "~Nathalie_Baracaldo1", "~Mingyi_Hong1", "~Sijia_Liu1"], "authors_source": "OpenReview API", "abstract": "Large language models (LLMs) have shown remarkable reasoning capabilities when trained with chain-of-thought (CoT) supervision. However, the long and verbose CoT traces, especially those distilled from large reasoning models (LRMs) such as DeepSeek-R1, significantly increase training costs during the distillation process, where a non-reasoning base model is taught to replicate the reasoning behavior of an LRM. In this work, we study the problem of CoT condensation for resource-efficient reasoning training, aimed at pruning intermediate reasoning steps (i.e., thoughts) in CoT traces, enabling supervised model training on length-reduced CoT data while preserving both answer accuracy and the model’s ability to generate coherent reasoning. Our rationale is that CoT traces typically follow a three-stage structure: problem understanding, exploration, and solution convergence. Through empirical analysis, we find that retaining the structure of the reasoning trace, especially the early stage of problem understanding (rich in reflective cues) and the final stage of solution convergence (which closely relates to the final answer), is sufficient to achieve lossless reasoning supervision. To this end, we propose an Edge-Preserving Condensation method, EPiC, which selectively retains only the initial and final segments of each CoT trace while discarding the middle portion. This design draws an analogy to preserving the “edge” of a reasoning trajectory, capturing both the initial problem framing and the final answer synthesis, to maintain logical continuity. Our analyses leveraging the CoT landscape and measuring the mutual information between CoT steps provide further validation for this design. Experiments across multiple model families (Qwen and LLaMA) and benchmarks show that EPiC reduces training time by over 34% while achieving lossless reasoning accuracy (e.g., on MATH500), comparable to full CoT supervision. Additionally, we show that EPiC outperforms other condensation methods, including teacher-guided regeneration of condensed CoTs.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=Bibt0JTvpx", "arxiv_id": "2506.04205", "paper_pdf": "papers/Bibt0JTvpx.pdf", "paper_pdf_sha256": "65d0f0acf1f8948c088e9aca92e717647e1cf86d6478700e52562d568a1aa382", "paper_pdf_bytes": 2033741, "paper_pdf_source": "openreview", "code_url": "https://github.com/OPTML-Group/EPiC", "code_repository": "OPTML-Group/EPiC", "code_commit": "54610896cb4659a7a05920cf8437ea070cf84d2f", "code_archive": "repos/Bibt0JTvpx.zip", "code_archive_sha256": "05d59008616105ced26ad93f3c2474f2004d00fa12c526e77dceb26cbb1b3c0a", "code_archive_bytes": 50941, "code_file_count": 20, "code_extensions": {".py": 20}, "github_disk_usage_kb": 43, "github_languages": {"Python": 96709}, "github_archived": false, "github_pushed_at": "2025-06-11T18:08:47Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/epic-towards-lossless-speedup-for-reasoning"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "yLmcYLP3Yd", "year": 2025, "status": "rejected", "title": "Discrete Neural Algorithmic Reasoning", "authors": ["Gleb Rodionov", "Liudmila Prokhorenkova"], "authorids": ["~Gleb_Rodionov1", "~Liudmila_Prokhorenkova1"], "authors_source": "OpenReview API", "abstract": "Neural algorithmic reasoning aims to capture computations with neural networks via learning the models to imitate the execution of classic algorithms. While common architectures are expressive enough to contain the correct model in the weights space, current neural reasoners are struggling to generalize well on out-of-distribution data. On the other hand, classic computations are not affected by distributional shifts as they can be described as transitions between discrete computational states. In this work, we propose to force neural reasoners to maintain the execution trajectory as a combination of finite predefined states. To achieve that, we separate discrete and continuous data flows and describe the interaction between them. Trained with supervision on the algorithm's state transitions, such models are able to perfectly align with the original algorithm. To show this, we evaluate our approach on multiple algorithmic problems and get perfect test scores both in single-task and multitask setups. Moreover, the proposed architectural choice allows us to prove the correctness of the learned algorithms for any test data.", "decision": "Reject", "meta_review": "", "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=yLmcYLP3Yd", "arxiv_id": "2402.11628", "paper_pdf": "papers/yLmcYLP3Yd.pdf", "paper_pdf_sha256": "ae3b1317f66c43efa8a65f094d15a754865be2e3bc4187f3db6892ebe110e005", "paper_pdf_bytes": 426356, "paper_pdf_source": "openreview", "code_url": "https://github.com/yandex-research/dnar", "code_repository": "yandex-research/dnar", "code_commit": "12f3f0bd0a70568386e43b2d841422f28ed53698", "code_archive": "repos/yLmcYLP3Yd.zip", "code_archive_sha256": "76cf854f1efe2f7d4ae01fb75ae2602795cd5ea6d55c58919cbecf270d66a157", "code_archive_bytes": 17641, "code_file_count": 7, "code_extensions": {".py": 7}, "github_disk_usage_kb": 15, "github_languages": {"Python": 38447}, "github_archived": false, "github_pushed_at": "2024-09-10T15:48:36Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/discrete-neural-algorithmic-reasoning"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "VLnODGVVAsL", "year": 2023, "status": "rejected", "title": "Anchor Sampling for Federated Learning with Partial Client Participation", "authors": ["Feijie Wu", "Song Guo", "Zhihao Qu", "Shiqi He", "Ziming Liu"], "authorids": ["~Feijie_Wu1", "~Song_Guo5", "~Zhihao_Qu1", "~Shiqi_He1", "~Ziming_Liu1"], "authors_source": "OpenReview API", "abstract": "In federated learning, the support of partial client participation offers a flexible training strategy, but it deteriorates the model training efficiency. In this paper, we propose a framework FedAMD to improve the convergence property and maintain flexibility. The core idea is anchor sampling, which disjoints the partial participants into anchor and miner groups. Each client in the anchor group aims at the local bullseye with the gradient computation using a large batch. Guided by the bullseyes, clients in the miner group steer multiple near-optimal local updates using small batches and update the global model. With the joint efforts from both groups, FedAMD is able to accelerate the training process as well as improve the model performance. Measured by $\\epsilon$-approximation and compared to the state-of-the-art first-order methods, FedAMD achieves the convergence by up to $O(1/\\epsilon)$ fewer communication rounds under non-convex objectives. In specific, we achieve a linear convergence rate under PL conditions. Empirical studies on real-world datasets validate the effectiveness of FedAMD and demonstrate the superiority of our proposed algorithm: Not only does it considerably save computation and communication costs, but also the test accuracy significantly improves. ", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=VLnODGVVAsL", "arxiv_id": "2206.05891", "paper_pdf": "papers/VLnODGVVAsL.pdf", "paper_pdf_sha256": "2f51defed0fe2dadd68ec7ed4dd642ce49e267a41b3effcb1ff6d0deab199e7b", "paper_pdf_bytes": 1395729, "paper_pdf_source": "openreview", "code_url": "https://github.com/HarliWu/FedAMD", "code_repository": "HarliWu/FedAMD", "code_commit": "73596d55e15a3ef6f404e5a313f50bc05a328c55", "code_archive": "repos/VLnODGVVAsL.zip", "code_archive_sha256": "c0a44531ab2903feef431e58f2eb4b87e4397d8f1933e0a8312dba08c8047aea", "code_archive_bytes": 28071, "code_file_count": 13, "code_extensions": {".py": 13}, "github_disk_usage_kb": 20, "github_languages": {"Python": 87743}, "github_archived": false, "github_pushed_at": "2024-04-09T17:55:44Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/accelerating-federated-learning-via-sampling"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "7Z7u2z1Ornl", "year": 2022, "status": "rejected", "title": "Pruning Edges and Gradients to Learn Hypergraphs from Larger Sets", "authors": ["David W Zhang", "Gertjan J. Burghouts", "Cees G. M. Snoek"], "authorids": ["~David_W_Zhang1", "~Gertjan_J._Burghouts1", "~Cees_G._M._Snoek1"], "authors_source": "OpenReview API", "abstract": "This paper aims for set-to-hypergraph prediction, where the goal is to infer the set of relations for a given set of entities. This is a common abstraction for applications in particle physics, biological systems and combinatorial optimization. We address two common scaling problems encountered in set-to-hypergraph tasks that limit the size of the input set: the exponentially growing number of hyperedges and the run-time complexity, both leading to higher memory requirements. We make three contributions. First, we propose to predict and supervise the \\emph{positive} edges only, which changes the asymptotic memory scaling from exponential to linear. Second, we introduce a training method that encourages iterative refinement of the predicted hypergraph, which allows us to skip iterations in the backward pass for improved efficiency and constant memory usage. Third, we combine both contributions in a single set-to-hypergraph model that enables us to address problems with larger input set sizes. We provide ablations for our main technical contributions and show that our model outperforms prior state-of-the-art, especially for larger sets.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=7Z7u2z1Ornl", "arxiv_id": "2106.13919", "paper_pdf": "papers/7Z7u2z1Ornl.pdf", "paper_pdf_sha256": "7cb5550039f5258201fc6903989fd4ca391dba4a9ac50056e3e05d832f07af94", "paper_pdf_bytes": 497702, "paper_pdf_source": "openreview", "code_url": "https://github.com/davzha/recurrently_predicting_hypergraphs", "code_repository": "davzha/recurrently_predicting_hypergraphs", "code_commit": "c2b9a4959a1b9e3bac05ad3e5264029ccad2447d", "code_archive": "repos/7Z7u2z1Ornl.zip", "code_archive_sha256": "f924f1cdd07be8dfb94f21a4f4cb4c71a681edf06fd4a3f32775c1ff327bc15e", "code_archive_bytes": 25112, "code_file_count": 15, "code_extensions": {".py": 15}, "github_disk_usage_kb": 17, "github_languages": {"Python": 60006}, "github_archived": false, "github_pushed_at": "2021-09-21T09:02:49Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/recurrently-predicting-hypergraphs"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "yeeS_HULL7Z", "year": 2021, "status": "rejected", "title": "Attention-Based Clustering: Learning a Kernel from Context", "authors": ["Samuel Coward", "Erik Visse-Martindale", "Chithrupa Ramesh"], "authorids": ["~Samuel_Coward1", "erik.visse-martindale@uk.zuken.com", "~Chithrupa_Ramesh1"], "authors_source": "OpenReview API", "abstract": "In machine learning, no data point stands alone. We believe that context is an underappreciated concept in many machine learning methods. We propose Attention-Based Clustering (ABC), a neural architecture based on the attention mechanism, which is designed to learn latent representations that adapt to context within an input set, and which is inherently agnostic to input sizes and number of clusters. By learning a similarity kernel, our method directly combines with any out-of-the-box kernel-based clustering approach. We present competitive results for clustering Omniglot characters and include analytical evidence of the effectiveness of an attention-based approach for clustering. ", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=yeeS_HULL7Z", "arxiv_id": "2010.01040", "paper_pdf": "papers/yeeS_HULL7Z.pdf", "paper_pdf_sha256": "fa9d14e53527c1987340d84f5e913c8bf4ca4f66bf236df811c54929caf461b7", "paper_pdf_bytes": 396620, "paper_pdf_source": "openreview", "code_url": "https://github.com/DramaCow/ABC", "code_repository": "DramaCow/ABC", "code_commit": "20e084d2b4d9b3fb0456c9927cc96174c0fd1657", "code_archive": "repos/yeeS_HULL7Z.zip", "code_archive_sha256": "8bd8de7b6c6e52f4da379206ba570ff77d0ddd8316c02319b1424d7c7a184066", "code_archive_bytes": 23618, "code_file_count": 16, "code_extensions": {".py": 15, ".sh": 1}, "github_disk_usage_kb": 20, "github_languages": {"Python": 64158, "Shell": 194}, "github_archived": false, "github_pushed_at": "2024-09-13T11:36:50Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/attention-based-clustering-learning-a-kernel"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "S1lF8xHYwS", "year": 2020, "status": "rejected", "title": "Unsupervised Domain Adaptation through Self-Supervision", "authors": ["Yu Sun", "Eric Tzeng", "Trevor Darrell", "Alexei A. Efros"], "authorids": ["yusun@berkeley.edu", "etzeng@eecs.berkeley.edu", "trevor@eecs.berkeley.edu", "efros@eecs.berkeley.edu"], "authors_source": "OpenReview API", "abstract": "This paper addresses unsupervised domain adaptation, the setting where labeled training data is available on a source domain, but the goal is to have good performance on a target domain with only unlabeled data. Like much of previous work, we seek to align the learned representations of the source and target domains while preserving discriminability. The way we accomplish alignment is by learning to perform auxiliary self-supervised task(s) on both domains simultaneously.  Each self-supervised task brings the two domains closer together along the direction relevant to that task. Training this jointly with the main task classifier on the source domain is shown to successfully generalize to the unlabeled target domain.  The presented objective is straightforward to implement and easy to optimize. We achieve state-of-the-art results on four out of seven standard benchmarks, and competitive results on segmentation adaptation. We also demonstrate that our method composes well with another popular pixel-level adaptation method.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=S1lF8xHYwS", "arxiv_id": "1909.11825", "paper_pdf": "papers/S1lF8xHYwS.pdf", "paper_pdf_sha256": "1c40f9ea0702869dd2834daf98891e8a8ae20111d5eab9e25410d219ff355f19", "paper_pdf_bytes": 6346299, "paper_pdf_source": "openreview", "code_url": "https://github.com/yueatsprograms/uda_release", "code_repository": "yueatsprograms/uda_release", "code_commit": "b256316283e74b5d1f16777f029c384ee9b6e2e7", "code_archive": "repos/S1lF8xHYwS.zip", "code_archive_sha256": "116a62eb06d838bea684a05e1410d7bf7ad8ab1b8ca1b06b54ef3cadc4c12a33", "code_archive_bytes": 24041, "code_file_count": 29, "code_extensions": {".py": 22, ".sh": 7}, "github_disk_usage_kb": 21, "github_languages": {"Python": 43943, "Shell": 4095}, "github_archived": false, "github_pushed_at": "2021-10-06T17:41:03Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/unsupervised-domain-adaptation-through-self-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "H1lnJ2Rqt7", "year": 2019, "status": "rejected", "title": "LARGE BATCH SIZE TRAINING OF NEURAL NETWORKS WITH ADVERSARIAL TRAINING AND SECOND-ORDER INFORMATION", "authors": ["Zhewei Yao", "Amir Gholami", "Kurt Keutzer", "Michael Mahoney"], "authorids": ["zheweiy@berkeley.edu", "amirgh@berkeley.edu", "keutzer@berkeley.edu", "mmahoney@stat.berkeley.edu"], "authors_source": "OpenReview API", "abstract": "Stochastic Gradient Descent (SGD) methods using randomly selected batches are widely-used to train neural network (NN) models. Performing design exploration to find the best NN for a particular task often requires extensive training with different models on a large dataset,  which is very computationally expensive. The most straightforward method to accelerate this computation is to distribute the batch of SGD over multiple processors. However, large batch training often times leads to degradation in accuracy, poor generalization, and even poor robustness to adversarial attacks.  Existing solutions for large batch training either do not work or require massive hyper-parameter tuning. To address this issue, we propose a novel large batch training method which combines recent results in adversarial training (to regularize against ``sharp minima'') and second order optimization (to use curvature information to change batch size adaptively during training). We extensively evaluate our method on Cifar-10/100, SVHN, TinyImageNet, and ImageNet datasets, using multiple NNs, including residual networks as well as compressed networks such as SqueezeNext.  Our new approach exceeds the performance of the existing solutions in terms of both accuracy and the number of SGD iterations (up to 1\\% and $3\\times$, respectively). We emphasize that this is achieved without any additional hyper-parameter tuning to tailor our method to any of these experiments.\n", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=H1lnJ2Rqt7", "arxiv_id": "1810.01021", "paper_pdf": "papers/H1lnJ2Rqt7.pdf", "paper_pdf_sha256": "eeb0d26bc21ceac77f642114a6564744a91686ec797a5a1131febc95691a8143", "paper_pdf_bytes": 721965, "paper_pdf_source": "openreview", "code_url": "https://github.com/amirgholami/HessianFlow", "code_repository": "amirgholami/HessianFlow", "code_commit": "644fc8a57472f3895fc21ba68357e46ad723beec", "code_archive": "repos/H1lnJ2Rqt7.zip", "code_archive_sha256": "74e706620eb7f530638b28ad68b540f83bfa1bd8d8d434b824782a1901960504", "code_archive_bytes": 32302, "code_file_count": 14, "code_extensions": {".py": 14}, "github_disk_usage_kb": 35, "github_languages": {"Python": 44843}, "github_archived": false, "github_pushed_at": "2020-01-15T06:36:26Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/large-batch-size-training-of-neural-networks"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "rJ7yZ2P6-", "year": 2018, "status": "rejected", "title": "Enhance Word Representation for Out-of-Vocabulary on Ubuntu Dialogue Corpus", "authors": ["JIANXIONG DONG", "Jim Huang"], "authorids": ["jdongca2003@gmail.com", "ccjimhuang@gmail.com"], "authors_source": "OpenReview API", "abstract": "Ubuntu dialogue corpus is the largest public available dialogue corpus to make it feasible to build end-to-end\ndeep neural network models directly from the conversation data. One challenge of Ubuntu dialogue corpus is \nthe large number of out-of-vocabulary words. In this paper we proposed an algorithm which combines the general pre-trained word embedding vectors with those  generated on the task-specific training set to address this issue.  We integrated character embedding into Chen et al's Enhanced LSTM method (ESIM) and used it to evaluate the effectiveness of our proposed method. For the task of next utterance selection, the proposed method has demonstrated a significant performance improvement against original ESIM and the new model has achieved state-of-the-art results on both Ubuntu dialogue corpus and Douban conversation corpus. In addition, we investigated the performance impact of end-of-utterance and end-of-turn token tags. ", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=rJ7yZ2P6-", "arxiv_id": "1802.02614", "paper_pdf": "papers/rJ7yZ2P6-.pdf", "paper_pdf_sha256": "c5060bc86dca3475ba979043a06f7dd348e6c5ce568df078379ebbd8a515259f", "paper_pdf_bytes": 327399, "paper_pdf_source": "openreview", "code_url": "https://github.com/jdongca2003/next_utterance_selection", "code_repository": "jdongca2003/next_utterance_selection", "code_commit": "7491e972d58412c175166f6564d77b7436e71a87", "code_archive": "repos/rJ7yZ2P6-.zip", "code_archive_sha256": "773b6feb8049ca8dc1e0ce3d2fb4bc77dd3a4dd2142e680369b67f9d14437d9f", "code_archive_bytes": 23916, "code_file_count": 10, "code_extensions": {".py": 7, ".sh": 3}, "github_disk_usage_kb": 24, "github_languages": {"Python": 46161, "Shell": 2365}, "github_archived": false, "github_pushed_at": "2018-05-07T05:07:38Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/enhance-word-representation-for-out-of"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "bDee2EgvWJ", "year": 2026, "status": "rejected", "title": "Approximate Message Passing for Bayesian Neural Networks", "authors": ["Romeo Sommerfeld", "Christian Helms", "Jan Niklas Groeneveld", "Rainer Schlosser", "Ralf Herbrich"], "authorids": ["~Romeo_Sommerfeld1", "~Christian_Helms1", "~Jan_Niklas_Groeneveld1", "~Rainer_Schlosser1", "~Ralf_Herbrich1"], "authors_source": "OpenReview API", "abstract": "Bayesian methods for learning predictive models have the ability to consider both sources of uncertainty (i.e., data and model uncertainty) within a single framework and thereby provide a powerful tool for decision-making. Bayesian neural networks (BNNs) hold great potential for  training data efficiency due to full uncertainty quantification, making them promising candidates for more data-efficient AI in data-constrained settings such as reinforcement learning in the physical world.  However, current computational approaches for learning BNNs often face limitations such as overconfidence, sensitivity to hyperparameters, and posterior collapse, highlighting the need for alternative computational approaches. In this paper, we introduce a novel method that leverages approximate message passing (MP) of a full factorized neural network model using mixed approximations to overcome these problems while maintaining data efficiency. Our framework supports convolutional neural networks while addressing the issue of double-counting training data, which has been a key source of overconfidence in prior work. We demonstrate the data-efficiency of our method on multiple benchmark datasets in comparison to state-of-the-art methods for learning neural networks.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=bDee2EgvWJ", "arxiv_id": "2501.15573", "paper_pdf": "papers/bDee2EgvWJ.pdf", "paper_pdf_sha256": "d8c03d50a4202cf7a5f7fddd5623d4d2fc085b6c0cecea79b7135e578188bd61", "paper_pdf_bytes": 2803386, "paper_pdf_source": "openreview", "code_url": "https://github.com/christian-helms/mpbnns", "code_repository": "christian-helms/mpbnns", "code_commit": "85f12084c7ab9d89acb9628b02413e841c7346f7", "code_archive": "repos/bDee2EgvWJ.zip", "code_archive_sha256": "ebbf2273e0467ed6483c8c79ce6b024267b2bfa88e09167dff7e7eeac3efe2fc", "code_archive_bytes": 57358, "code_file_count": 21, "code_extensions": {".jl": 17, ".py": 3, ".sh": 1}, "github_disk_usage_kb": 54, "github_languages": {"Julia": 180263, "Python": 19933, "Shell": 388}, "github_archived": false, "github_pushed_at": "2025-01-26T17:12:50Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/approximate-message-passing-for-bayesian"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "lAXlDAdan5", "year": 2025, "status": "rejected", "title": "Accelerating Error Correction Code Transformers", "authors": ["Matan Levy", "Yoni Choukroun", "Lior Wolf"], "authorids": ["~Matan_Levy2", "~Yoni_Choukroun1", "~Lior_Wolf1"], "authors_source": "OpenReview API", "abstract": "Error correction codes (ECC) are crucial for ensuring reliable information transmission in communication systems. Choukroun & Wolf (2022b) recently introduced the Error Correction Code Transformer (ECCT), which has demonstrated promising performance across various transmission channels and families of codes. However, its high computational and memory demands limit its practical applications compared to traditional decoding algorithms. Achieving effective quantization of the ECCT presents significant challenges due to its inherently small architecture, since existing, very low-precision quantization techniques often lead to performance degradation in compact neural networks. In this paper, we introduce a novel acceleration method for transformer-based decoders. We first propose a ternary weight quantization method specifically designed for the ECCT, inducing a decoder with multiplication-free linear layers. We present an\noptimized self-attention mechanism to reduce computational complexity via codeaware multi-heads processing. Finally, we provide positional encoding via the Tanner graph eigendecomposition, enabling a richer representation of the graph connectivity. The approach not only matches or surpasses ECCT’s performance but also significantly reduces energy consumption, memory footprint, and computational complexity. Our method brings transformer-based error correction closer to practical implementation in resource-constrained environments, achieving a 90% compression ratio and reducing arithmetic operation energy consumption by at least 224 times on modern hardware.", "decision": "Reject", "meta_review": "", "num_reviews": 6, "openreview_url": "https://openreview.net/forum?id=lAXlDAdan5", "arxiv_id": "2410.05911", "paper_pdf": "papers/lAXlDAdan5.pdf", "paper_pdf_sha256": "521ebc1740dc0f78a95b3742ca1a5e88013508acbf748a9022196483c13eacc7", "paper_pdf_bytes": 2465131, "paper_pdf_source": "openreview", "code_url": "https://github.com/mlaetvayn/AECCT", "code_repository": "mlaetvayn/AECCT", "code_commit": "d66d5dc9312bde0484c126eba0a1dd857fcd2418", "code_archive": "repos/lAXlDAdan5.zip", "code_archive_sha256": "4457907076cdc95c4a02315d7a8ec7a510ec7ee532a3c5c40bec084d086239eb", "code_archive_bytes": 17429, "code_file_count": 5, "code_extensions": {".py": 5}, "github_disk_usage_kb": 16, "github_languages": {"Python": 32650}, "github_archived": false, "github_pushed_at": "2024-10-16T09:16:41Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/accelerating-error-correction-code"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "YkEW5TabYN", "year": 2024, "status": "rejected", "title": "Perturbed examples reveal invariances shared by language models", "authors": ["Ruchit Rawal", "Mariya Toneva"], "authorids": ["~Ruchit_Rawal1", "~Mariya_Toneva1"], "authors_source": "OpenReview API", "abstract": "An explosion of work in language is leading to ever-increasing numbers of available natural language processing models, with little understanding of how new models compare to better-understood models. One major reason for this difficulty is saturating benchmark datasets, which may not reflect well differences in model performance in the wild. In this work, we propose a novel framework for comparing two natural language processing models by revealing their shared invariance to interpretable input perturbations that are designed to target a specific linguistic capability (e.g., Synonym-Invariance, Typo-Invariance). Via experiments on models from within the same and across different architecture families, this framework offers a number of insights about how changes in models (e.g. distillation, increase in size, amount of pre-training) affect multiple well-defined linguistic capabilities. Furthermore, we also demonstrate how our framework can enable evaluation of the invariances shared between models that are available as commercial black-box APIs (e.g., InstructGPT family) and models that are relatively better understood (e.g., GPT-2). Across several experiments, we observe that large language models share many of the invariances encoded by models of various sizes, whereas the invariances encoded by large language models are only shared by other large models. Possessing a wide variety of invariances may be a key reason for the recent successes of large language models, and our framework can shed light on the types of invariances that are retained by or emerge in new models.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=YkEW5TabYN", "arxiv_id": "2311.04166", "paper_pdf": "papers/YkEW5TabYN.pdf", "paper_pdf_sha256": "f61900781619f6513465cbf29a7de809e536e6853c69af1c4a73fcf82e6d072e", "paper_pdf_bytes": 1219604, "paper_pdf_source": "openreview", "code_url": "https://github.com/bridge-ai-neuro/shared_invariances_acl", "code_repository": "bridge-ai-neuro/shared_invariances_acl", "code_commit": "f2615b69f213931488074d842514cef4ee139327", "code_archive": "repos/YkEW5TabYN.zip", "code_archive_sha256": "77c3c5ec76609da9bca135acb87c8580e7f0ef043a45bc89d0635c2857a56e55", "code_archive_bytes": 20792, "code_file_count": 14, "code_extensions": {".py": 11, ".sh": 3}, "github_disk_usage_kb": 16, "github_languages": {"Python": 48260, "Shell": 1750}, "github_archived": false, "github_pushed_at": "2024-06-09T10:05:31Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/perturbed-examples-reveal-invariances-shared"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "CEhy-i7_KfC", "year": 2023, "status": "rejected", "title": "Pretraining the Vision Transformer using self-supervised methods for vision based Deep Reinforcement Learning", "authors": ["Manuel Goulão", "Arlindo L. Oliveira"], "authorids": ["~Manuel_Goulão1", "~Arlindo_L._Oliveira1"], "authors_source": "OpenReview API", "abstract": "The Vision Transformer architecture has shown to be competitive in the computer vision (CV) space where it has dethroned convolution-based networks in several benchmarks. Nevertheless, Convolutional Neural Networks (CNN) remain the preferential architecture for the representation module in Reinforcement Learning. In this work, we study pretraining a Vision Transformer using several state-of-the-art self-supervised methods and assess data-efficiency gains from this training framework. We propose a new self-supervised learning method called TOV-VICReg that extends VICReg to better capture temporal relations between observations by adding a temporal order verification task. Furthermore, we evaluate the resultant encoders with Atari games in a sample-efficiency regime. Our results show that the vision transformer, when pretrained with TOV-VICReg, outperforms the other self-supervised methods but still struggles to overcome a CNN. Nevertheless, we were able to outperform a CNN in two of the ten games where we perform a 100k steps evaluation. Ultimately, we believe that such approaches in Deep Reinforcement Learning (DRL) might be the key to achieving new levels of performance as seen in natural language processing and computer vision.", "decision": "Reject", "meta_review": null, "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=CEhy-i7_KfC", "arxiv_id": "2209.10901", "paper_pdf": "papers/CEhy-i7_KfC.pdf", "paper_pdf_sha256": "70eb5a4040f93871f4aca16d008325a6d8980b5e0c8683a4e040275976a0cb68", "paper_pdf_bytes": 393689, "paper_pdf_source": "openreview", "code_url": "https://github.com/mgoulao/TOV-VICReg", "code_repository": "mgoulao/TOV-VICReg", "code_commit": "535b1643134b407ca40ca2c6553e04b50d3968ae", "code_archive": "repos/CEhy-i7_KfC.zip", "code_archive_sha256": "ca3a98c07a3ef237691d3f0301ed573eb8785863038ba5bf1659cc8c10ab99b9", "code_archive_bytes": 24578, "code_file_count": 13, "code_extensions": {".py": 11, ".sh": 2}, "github_disk_usage_kb": 20, "github_languages": {"Python": 59289, "Shell": 2501}, "github_archived": false, "github_pushed_at": "2022-09-22T10:33:19Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/pretraining-the-vision-transformer-using-self"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "FKotzp6PZJw", "year": 2021, "status": "rejected", "title": "On the Estimation Bias in Double Q-Learning", "authors": ["Zhizhou Ren", "Guangxiang Zhu", "Beining Han", "Jianglun Chen", "Chongjie Zhang"], "authorids": ["~Zhizhou_Ren1", "~Guangxiang_Zhu1", "~Beining_Han1", "~Jianglun_Chen2", "~Chongjie_Zhang1"], "authors_source": "OpenReview API", "abstract": "Double Q-learning is a classical method for reducing overestimation bias, which is caused by taking maximum estimated values in the Bellman operator. Its variants in the deep Q-learning paradigm have shown great promise in producing reliable value prediction and improving learning performance. However, as shown by prior work, double Q-learning is not fully unbiased and still suffers from underestimation bias. In this paper, we show that such underestimation bias may lead to multiple non-optimal fixed points under an approximated Bellman operation. To address the concerns of converging to non-optimal stationary solutions, we propose a simple and effective approach as a partial fix for underestimation bias in double Q-learning. This approach leverages real returns to bound the target value. We extensively evaluate the proposed method in the Atari benchmark tasks and demonstrate its significant improvement over baseline algorithms.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=FKotzp6PZJw", "arxiv_id": "2109.14419", "paper_pdf": "papers/FKotzp6PZJw.pdf", "paper_pdf_sha256": "48483f43dd599eae7d07bd52ed6c742e38f3bfde6167ff91e1549abd6e1e6547", "paper_pdf_bytes": 1085232, "paper_pdf_source": "openreview", "code_url": "https://github.com/Stilwell-Git/Doubly-Bounded-Q-Learning", "code_repository": "Stilwell-Git/Doubly-Bounded-Q-Learning", "code_commit": "028ee629781f3fc7daa149b5b838884cfa09fcd1", "code_archive": "repos/FKotzp6PZJw.zip", "code_archive_sha256": "b7abd2079ad9c96c5b328240dc13250ed8b887a6d2cdf02d67d12f834847bf37", "code_archive_bytes": 22158, "code_file_count": 18, "code_extensions": {".py": 17, ".cpp": 1}, "github_disk_usage_kb": 20, "github_languages": {"Python": 40506, "C++": 4793}, "github_archived": false, "github_pushed_at": "2022-11-14T14:00:54Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/on-the-estimation-bias-in-double-q-learning-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "H1lKd6NYPS", "year": 2020, "status": "rejected", "title": "Online Meta-Critic Learning for Off-Policy Actor-Critic Methods", "authors": ["Wei Zhou", "Yiying Li", "Yongxin Yang", "Huaimin Wang", "Timothy M. Hospedales"], "authorids": ["zhouwei14@nudt.edu.cn", "liyiying10@nudt.edu.cn", "yongxin.yang@ed.ac.uk", "hmwang@nudt.edu.cn", "t.hospedales@ed.ac.uk"], "authors_source": "OpenReview API", "abstract": "Off-Policy Actor-Critic (Off-PAC) methods have proven successful in a variety of continuous control tasks. Normally, the critic’s action-value function is updated using temporal-difference, and the critic in turn provides a loss for the actor that trains it to take actions with higher expected return. In this paper, we introduce a novel and flexible meta-critic that observes the learning process and meta-learns an additional loss for the actor that accelerates and improves actor-critic learning. Compared to the vanilla critic, the meta-critic network is explicitly trained to accelerate the learning process; and compared to existing meta-learning algorithms, meta-critic is rapidly learned online for a single task, rather than slowly over a family of tasks. Crucially, our meta-critic framework is designed for off-policy based learners, which currently provide state-of-the-art reinforcement learning sample efficiency. We demonstrate that online meta-critic learning leads to improvements in a variety of continuous control environments when combined with contemporary Off-PAC methods DDPG, TD3 and the state-of-the-art SAC. ", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=H1lKd6NYPS", "arxiv_id": "2003.05334", "paper_pdf": "papers/H1lKd6NYPS.pdf", "paper_pdf_sha256": "3f73a188b1d752d13af721ddf2d41d99a57316b24523da888689f411b6a5ad89", "paper_pdf_bytes": 7208362, "paper_pdf_source": "openreview", "code_url": "https://github.com/zwfightzw/Meta-Critic", "code_repository": "zwfightzw/Meta-Critic", "code_commit": "d8045d66fa82d5035868b82e6bd9cbdbe6fa5955", "code_archive": "repos/H1lKd6NYPS.zip", "code_archive_sha256": "4ec08efa2248202c6989da1920d4263836bab039a868c110ccb47ac8068e5ed6", "code_archive_bytes": 26608, "code_file_count": 12, "code_extensions": {".py": 12}, "github_disk_usage_kb": 22, "github_languages": {"Python": 78982}, "github_archived": false, "github_pushed_at": "2020-10-19T06:25:17Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/online-meta-critic-learning-for-off-policy-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Sygx4305KQ", "year": 2019, "status": "rejected", "title": "Small steps and giant leaps: Minimal Newton solvers for Deep Learning", "authors": ["Joao Henriques", "Sebastien Ehrhardt", "Samuel Albanie", "Andrea Vedaldi"], "authorids": ["joao@robots.ox.ac.uk", "hyenal@robots.ox.ac.uk", "albanie@robots.ox.ac.uk", "vedali@robots.ox.ac.uk"], "authors_source": "OpenReview API", "abstract": "We propose a fast second-order method that can be used as a drop-in replacement for current deep learning solvers. Compared to stochastic gradient descent (SGD), it only requires two additional forward-mode automatic differentiation operations per iteration, which has a computational cost comparable to two standard forward passes and is easy to implement. Our method addresses long-standing issues with current second-order solvers, which invert an approximate Hessian matrix every iteration exactly or by conjugate-gradient methods, procedures that are much slower than a SGD step. Instead, we propose to keep a single estimate of the gradient projected by the inverse Hessian matrix, and update it once per iteration with just two passes over the network. This estimate has the same size and is similar to the momentum variable that is commonly used in SGD. No estimate of the Hessian is maintained.\nWe first validate our method, called CurveBall, on small problems with known solutions (noisy Rosenbrock function and degenerate 2-layer linear networks), where current deep learning solvers struggle. We then train several large models on CIFAR and ImageNet, including ResNet and VGG-f networks, where we demonstrate faster convergence with no hyperparameter tuning. We also show our optimiser's generality by testing on a large set of randomly-generated architectures.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=Sygx4305KQ", "arxiv_id": "1805.08095", "paper_pdf": "papers/Sygx4305KQ.pdf", "paper_pdf_sha256": "d41e9bfffb0357512e05446983925195dd5f02a8abd22512172c6a81369dcbb5", "paper_pdf_bytes": 2346524, "paper_pdf_source": "openreview", "code_url": "https://github.com/jotaf98/curveball", "code_repository": "jotaf98/curveball", "code_commit": "1dc37325382c12e3fc9b2e7e27c47e6d7a17021a", "code_archive": "repos/Sygx4305KQ.zip", "code_archive_sha256": "7d7433458348ca7b73a461edfc7b926a18453def06cd9543c8227492bb79c16a", "code_archive_bytes": 43815, "code_file_count": 35, "code_extensions": {".m": 21, ".cu": 10, ".cpp": 2, ".sh": 1, ".hpp": 1}, "github_disk_usage_kb": 37, "github_languages": {"MATLAB": 58871, "Cuda": 42788, "C++": 1212, "Shell": 548}, "github_archived": false, "github_pushed_at": "2018-10-29T13:59:54Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/small-steps-and-giant-leaps-minimal-newton"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "HyEi7bWR-", "year": 2018, "status": "rejected", "title": "Orthogonal Recurrent Neural Networks with Scaled Cayley Transform", "authors": ["Kyle Helfrich", "Devin Willmott", "Qiang Ye"], "authorids": ["kyle.helfrich@uky.edu", "devin.willmott@uky.edu", "qiang.ye@uky.edu"], "authors_source": "OpenReview API", "abstract": "Recurrent Neural Networks (RNNs) are designed to handle sequential data but suffer from vanishing or exploding gradients.  Recent work on Unitary Recurrent Neural Networks (uRNNs) have been used to address this issue and in some cases, exceed the capabilities of Long Short-Term Memory networks (LSTMs).  We propose a simpler and novel update scheme to maintain orthogonal recurrent weight matrices without using complex valued matrices. This is done by parametrizing with a skew-symmetric matrix using the Cayley transform. Such a parametrization is unable to represent matrices with negative one eigenvalues, but this limitation is overcome by scaling the recurrent weight matrix by a diagonal matrix consisting of ones and negative ones.  The proposed training scheme involves a straightforward gradient calculation and update step. In several experiments, the proposed scaled Cayley orthogonal recurrent neural network (scoRNN) achieves superior results with fewer trainable parameters than other unitary RNNs.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=HyEi7bWR-", "arxiv_id": "1707.09520", "paper_pdf": "papers/HyEi7bWR-.pdf", "paper_pdf_sha256": "31ec6aad5b3148eb8a25485574a1b299fcf6cfc7bd774161abf6fd7fcfdbc71d", "paper_pdf_bytes": 676036, "paper_pdf_source": "openreview", "code_url": "https://github.com/SpartinStuff/scoRNN", "code_repository": "SpartinStuff/scoRNN", "code_commit": "d390c9bac62c510963ff90d386fb02beccff0a1e", "code_archive": "repos/HyEi7bWR-.zip", "code_archive_sha256": "c89c8a0d7597fa0f6d04b4f31b3a0d3f1ba9f5b19be80c929c69ba2c42de1726", "code_archive_bytes": 12462, "code_file_count": 4, "code_extensions": {".py": 4}, "github_disk_usage_kb": 26, "github_languages": {"Python": 32173}, "github_archived": false, "github_pushed_at": "2024-05-24T09:14:20Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/orthogonal-recurrent-neural-networks-with"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "oVd1y7ilTk", "year": 2026, "status": "rejected", "title": "On Designing Diffusion Autoencoders for Efficient Generation and Representation Learning", "authors": ["Magdalena Proszewska", "Nikolay Malkin", "Siddharth N"], "authorids": ["~Magdalena_Proszewska1", "~Nikolay_Malkin1", "~Siddharth_N1"], "authors_source": "OpenReview API", "abstract": "Diffusion autoencoders (DAs) are variants of diffusion generative models that use an input-dependent latent variable to capture representations alongside the diffusion process. These representations can be used for tasks such as downstream classification, controllable generation, and interpolation. However, the generative performance of DAs relies heavily on how well the prior distribution over the latent variables can be modelled and subsequently sampled from. Better generative modelling is also the goal of another class of diffusion models—those that learn their forward (noising) process. While effective at adjusting the noise process in an input-dependent manner, they must satisfy additional constraints derived from the terminal conditions of the diffusion process. Here, we draw a connection between these two classes of models and show that certain design decisions (latent variable choice, conditioning method, etc.) in the DA framework—leading to a model we term DMZ—enable effective representations as evaluated on downstream tasks, including domain transfer, as well as more efficient modelling and generation with fewer denoising steps compared to standard diffusion models.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=oVd1y7ilTk", "arxiv_id": "2506.00136", "paper_pdf": "papers/oVd1y7ilTk.pdf", "paper_pdf_sha256": "79bed6630dc80a05b2da429c077a3f2632d1fffe35ab7f3b45ed02b0f854630f", "paper_pdf_bytes": 9032494, "paper_pdf_source": "openreview", "code_url": "https://github.com/exlab-research/dmz", "code_repository": "exlab-research/dmz", "code_commit": "1d740e18e68348e3dd43acfda37508d9aed67a8a", "code_archive": "repos/oVd1y7ilTk.zip", "code_archive_sha256": "8f75859b96c80c6a3819d8c6925123ec98fcbe1ae043b67712f445b2d4492e23", "code_archive_bytes": 47223, "code_file_count": 20, "code_extensions": {".py": 20}, "github_disk_usage_kb": 56, "github_languages": {"Python": 149055}, "github_archived": false, "github_pushed_at": "2025-09-25T09:44:51Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/on-designing-diffusion-autoencoders-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "zCncHdGsOa", "year": 2025, "status": "rejected", "title": "Efficient optimization with orthogonality constraint: a randomized Riemannian submanifold method", "authors": ["Andi Han", "Pierre-Louis Poirion", "Akiko Takeda"], "authorids": ["~Andi_Han1", "~Pierre-Louis_Poirion1", "~Akiko_Takeda2"], "authors_source": "OpenReview API", "abstract": "Optimization with orthogonality constraints frequently arise in various fields such as machine learning, signal processing and computer vision. Riemannian optimization offers a powerful framework for solving these problems by equipping the constraint set with a Riemannian manifold structure and performing optimization intrinsically on the manifold. This approach typically involves computing a search direction in the tangent space and updating variables via a retraction operation. However, as the size of the variables increases, the computational cost of the retraction can become prohibitively high, limiting the applicability of Riemannian optimization to large-scale problems.  To address this challenge and enhance scalability, we propose a novel approach that restricts each update on a random submanifold, thereby significantly reducing the per-iteration complexity. We introduce two sampling strategies for selecting the random submanifold and theoretically analyze the convergence of the proposed method. We provide convergence results for general nonconvex functions and functions that satisfy Riemannian Polyak–Łojasiewicz condition as well as for stochastic optimization settings. Extensive experiments verify the benefits of the proposed method, showcasing its effectiveness across a wide variety of problem instances.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=zCncHdGsOa", "arxiv_id": "2505.12378", "paper_pdf": "papers/zCncHdGsOa.pdf", "paper_pdf_sha256": "b3a8931dc232e0663dd4d50438dd76c65caeea97600ddff6b216e203c4791663", "paper_pdf_bytes": 1255804, "paper_pdf_source": "openreview", "code_url": "https://github.com/andyjm3/RSDM", "code_repository": "andyjm3/RSDM", "code_commit": "ac14994038a100b1047a324c7931b529ac0397f5", "code_archive": "repos/zCncHdGsOa.zip", "code_archive_sha256": "6503c40e738e953eea87e17120ce9996b1accc5ceba55efd8636e57300993994", "code_archive_bytes": 15170, "code_file_count": 6, "code_extensions": {".py": 6}, "github_disk_usage_kb": 17, "github_languages": {"Python": 30936}, "github_archived": false, "github_pushed_at": "2025-05-21T10:10:56Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/efficient-optimization-with-orthogonality"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "tB7p0SM5TH", "year": 2024, "status": "rejected", "title": "GraSP: Simple yet Effective Graph Similarity Predictions", "authors": ["Haoran Zheng", "Jieming Shi"], "authorids": ["~Haoran_Zheng1", "~Jieming_Shi1"], "authors_source": "OpenReview API", "abstract": "Graph similarity computation (GSC) is considered one of the essential operations because of its wide range of applications in various fields. Graph Edit Distance (GED) and Maximum Common Subgraph (MCS) are the most popular graph similarity metrics. However, calculating exact GED and MCS is a complex task that falls under the category of NP-hard problems. Consequently, state-of-the-art methodologies learn data-driven models leveraging graph neural networks (GNNs) for estimating GED and MCS values. A perceived limitation of these approaches includes reliance on computationally expensive cross-graph node-level interaction components but to little avail. Instead of building up complicated components, we aim to make the complicated simple and present GraSP, a simple yet highly effective approach for GSC.  In particular, to achieve higher expressiveness,  we design techniques to enhance node features via positional encoding, employ a graph neural network backbone with a gating mechanism and residual connections, and develop a multi-scale pooling technique to generate meaningful representations. We theoretically prove that our method is more expressive and passes 1-WL test performance capabilities. Notably, GraSP is versatile in accurately predicting GED and MCS metrics. In extensive experiments against numerous competitors on real-world datasets, we demonstrate the superiority of GraSP over prior arts regarding effectiveness and efficiency. The source code is available at https://anonymous.4open.science/r/GraSP.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=tB7p0SM5TH", "arxiv_id": "2412.09968", "paper_pdf": "papers/tB7p0SM5TH.pdf", "paper_pdf_sha256": "707198cc4b4e9ccf36a0238d3a3f2ca62b0a5fe09b5767e57ec4428d1d4df89e", "paper_pdf_bytes": 1295848, "paper_pdf_source": "openreview", "code_url": "https://github.com/HaoranZ99/GraSP", "code_repository": "HaoranZ99/GraSP", "code_commit": "d7d89bf1197ec3fb8f4c9ab578ab4ebace757767", "code_archive": "repos/tB7p0SM5TH.zip", "code_archive_sha256": "b597a61fdfa6fc6a6b97ae7b47a99dbe750f4b99084eaeb3eb4eefebb41a7d02", "code_archive_bytes": 16412, "code_file_count": 8, "code_extensions": {".py": 7, ".sh": 1}, "github_disk_usage_kb": 16, "github_languages": {"Python": 47638, "Shell": 867}, "github_archived": false, "github_pushed_at": "2026-02-25T12:42:37Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/grasp-simple-yet-effective-graph-similarity"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "OYKIo3ySkxA", "year": 2023, "status": "rejected", "title": "DIGEST: FAST AND COMMUNICATION EFFICIENT DECENTRALIZED LEARNING WITH LOCAL UPDATES", "authors": ["Peyman Gholami", "Hulya Seferoglu"], "authorids": ["~Peyman_Gholami1", "~Hulya_Seferoglu1"], "authors_source": "OpenReview API", "abstract": "Decentralized learning advocates the elimination of centralized parameter servers\n(aggregation points) for potentially better utilization of underlying resources, de-\nlay reduction, and resiliency against parameter server unavailability and catas-\ntrophic failures. Gossip based decentralized algorithms, where each node in a net-\nwork has its own locally kept model on which it effectuates the learning by talking\nto its neighbors, received a lot of attention recently. Despite their potential, Gossip\nalgorithms introduce huge communication costs. In this work, we show that nodes\ndo not need to communicate as frequently as in Gossip for fast convergence; in\nfact, a sporadic exchange of a digest of a trained model is sufficient. Thus, we\ndesign a fast and communication-efficient decentralized learning mechanism; DI-\nGEST by particularly focusing on stochastic gradient descent (SGD). DIGEST is\na decentralized algorithm building on local-SGD algorithms, which are originally\ndesigned for communication efficient centralized learning. We show through anal-\nysis and experiments that DIGEST significantly reduces the communication cost\nwithout hurting convergence time for both iid and non-iid data.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=OYKIo3ySkxA", "arxiv_id": "2307.07652", "paper_pdf": "papers/OYKIo3ySkxA.pdf", "paper_pdf_sha256": "3518c5921b252f6f3b450ae28a83dd746017e2a794fdaadf63f0511f7b4a98bb", "paper_pdf_bytes": 1272206, "paper_pdf_source": "openreview", "code_url": "https://github.com/Anonymous404404/DigestCode", "code_repository": "Anonymous404404/DigestCode", "code_commit": "9499a2d81eaf1380a823cdb73436af6f417bf161", "code_archive": "repos/OYKIo3ySkxA.zip", "code_archive_sha256": "8ebf4c976ce8d08dbb8928312017dbe24dba1eea86b253056a2cd2424bde237f", "code_archive_bytes": 23552, "code_file_count": 7, "code_extensions": {".py": 7}, "github_disk_usage_kb": 21, "github_languages": {"Python": 41988}, "github_archived": false, "github_pushed_at": "2023-01-26T19:34:46Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/digest-fast-and-communication-efficient"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "r1lUl6NFDH", "year": 2020, "status": "rejected", "title": "Mirror Descent View For Neural Network Quantization", "authors": ["Thalaiyasingam Ajanthan", "Kartik Gupta", "Philip H. S. Torr", "Richard Hartley", "Puneet K. Dokania"], "authorids": ["thalaiyasingam.ajanthan@anu.edu.au", "kartik.gupta@anu.edu.au", "phst@robots.ox.ac.uk", "richard.hartley@anu.edu.au", "puneet@robots.ox.ac.uk"], "authors_source": "OpenReview API", "abstract": "Quantizing large Neural Networks (NN) while maintaining the performance is highly desirable for resource-limited devices due to reduced memory and time complexity. NN quantization is usually formulated as a constrained optimization problem and optimized via a modified version of gradient descent. In this work, by interpreting the continuous parameters (unconstrained) as the dual of the quantized ones, we introduce a Mirror Descent (MD) framework (Bubeck (2015)) for NN quantization. Specifically, we provide conditions on the projections (i.e., mapping from continuous to quantized ones) which would enable us to derive valid mirror maps and in turn the respective MD updates. Furthermore, we discuss a numerically stable implementation of MD by storing an additional set of auxiliary dual variables (continuous). This update is strikingly analogous to the popular Straight Through Estimator (STE) based method which is typically viewed as a “trick” to avoid vanishing gradients issue but here we show that it is an implementation method for MD for certain projections. Our experiments on standard classification datasets (CIFAR-10/100, TinyImageNet) with convolutional and residual architectures show that our MD variants obtain fully-quantized networks with accuracies very close to the floating-point networks.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=r1lUl6NFDH", "arxiv_id": "1910.08237", "paper_pdf": "papers/r1lUl6NFDH.pdf", "paper_pdf_sha256": "c8e50b0d3dc1dde1a52f8f4dd4dab3a037acf02fad32172b251e0edaf7ce4525", "paper_pdf_bytes": 564973, "paper_pdf_source": "openreview", "code_url": "https://github.com/kartikgupta-at-anu/md-bnn", "code_repository": "kartikgupta-at-anu/md-bnn", "code_commit": "b42d5aabd78b73b2baee0858b7f2b8ee72e36d6c", "code_archive": "repos/r1lUl6NFDH.zip", "code_archive_sha256": "bbe81658a01507d01fb7f3e7ebba485364d55da038f32cff708027df4f17772b", "code_archive_bytes": 49542, "code_file_count": 30, "code_extensions": {".py": 20, ".sh": 10}, "github_disk_usage_kb": 24, "github_languages": {"Python": 131342, "Shell": 16704}, "github_archived": false, "github_pushed_at": "2021-02-19T00:59:07Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/mirror-descent-view-for-neural-network"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "By41BjA9YQ", "year": 2019, "status": "rejected", "title": "Laplacian Smoothing Gradient Descent", "authors": ["Stanley J. Osher", "Bao Wang", "Penghang Yin", "Xiyang Luo", "Minh Pham", "Alex T. Lin"], "authorids": ["sjo@math.ucla.edu", "wangbaonj@gmail.com", "yph@g.ucla.edu", "xylmath@gmail.com", "minhrose@ucla.edu", "atlin@math.ucla.edu"], "authors_source": "OpenReview API", "abstract": "We propose a class of very simple modifications of gradient descent and stochastic gradient descent. We show that when applied to a large variety of machine learning problems, ranging from softmax regression to deep neural nets, the proposed surrogates can dramatically reduce the variance and improve the generalization accuracy. The methods only involve multiplying the usual (stochastic) gradient by the inverse of a positive definitive matrix coming from the discrete Laplacian or its high order generalizations. The theory of Hamilton-Jacobi partial differential equations demonstrates that the implicit version of new algorithm is almost the same as doing gradient descent on a new function which (i) has the same global minima as the original function and (ii) is ``more convex\". We show that optimization algorithms with these surrogates converge uniformly in the discrete Sobolev $H_\\sigma^p$ sense and reduce the optimality gap for convex optimization problems. We implement our algorithm into both PyTorch and Tensorflow platforms which only involves changing of a few lines of code. The code will be available on Github.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=By41BjA9YQ", "arxiv_id": "1806.06317", "paper_pdf": "papers/By41BjA9YQ.pdf", "paper_pdf_sha256": "ee79d1e9ddd55db98ff61e5f20c598b8ffe83d3ba0f44c3a48a09629784f58ac", "paper_pdf_bytes": 1165708, "paper_pdf_source": "openreview", "code_url": "https://github.com/BaoWangMath/LaplacianSmoothing-GradientDescent", "code_repository": "BaoWangMath/LaplacianSmoothing-GradientDescent", "code_commit": "ba394837a6857677c1e243a375f495be32da8a5e", "code_archive": "repos/By41BjA9YQ.zip", "code_archive_sha256": "5264e98d4795aa28d5bab59df6cc05595d5eed7d1372c1daaf4d5d04c450fd08", "code_archive_bytes": 28493, "code_file_count": 15, "code_extensions": {".py": 15}, "github_disk_usage_kb": 50, "github_languages": {"Python": 77387}, "github_archived": false, "github_pushed_at": "2019-05-26T18:29:37Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/laplacian-smoothing-gradient-descent"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "HJSA_e1AW", "year": 2018, "status": "rejected", "title": "Normalized Direction-preserving Adam", "authors": ["Zijun Zhang", "Lin Ma", "Zongpeng Li", "Chuan Wu"], "authorids": ["zijun.zhang@ucalgary.ca", "linmawhu@gmail.com", "zongpeng@ucalgary.ca", "cwu@cs.hku.hk"], "authors_source": "OpenReview API", "abstract": "Optimization algorithms for training deep models not only affects the convergence rate and stability of the training process, but are also highly related to the generalization performance of trained models. While adaptive algorithms, such as Adam and RMSprop, have shown better optimization performance than stochastic gradient descent (SGD) in many scenarios, they often lead to worse generalization performance than SGD, when used for training deep neural networks (DNNs). In this work, we identify two problems regarding the direction and step size for updating the weight vectors of hidden units, which may degrade the generalization performance of Adam. As a solution, we propose the normalized direction-preserving Adam (ND-Adam) algorithm, which controls the update direction and step size more precisely, and thus bridges the generalization gap between Adam and SGD. Following a similar rationale, we further improve the generalization performance in classification tasks by regularizing the softmax logits. By bridging the gap between SGD and Adam, we also shed some light on why certain optimization algorithms generalize better than others.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=HJSA_e1AW", "arxiv_id": "1709.04546", "paper_pdf": "papers/HJSA_e1AW.pdf", "paper_pdf_sha256": "a9e704f314545c18f8c31037845072a1b898f859041161991bab73471c4d3cda", "paper_pdf_bytes": 650436, "paper_pdf_source": "openreview", "code_url": "https://github.com/zj10/ND-Adam", "code_repository": "zj10/ND-Adam", "code_commit": "b4a9d59b9c3607bbf734dc8c893fdbdce7bd1bef", "code_archive": "repos/HJSA_e1AW.zip", "code_archive_sha256": "98ecfe34d94e0eab54ded44bf322641d1206e2d6dcf213eaa722f66bfa2ab467", "code_archive_bytes": 19208, "code_file_count": 9, "code_extensions": {".py": 9}, "github_disk_usage_kb": 26, "github_languages": {"Python": 51045}, "github_archived": false, "github_pushed_at": "2018-09-19T04:58:16Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/normalized-direction-preserving-adam"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "LJ6AvummWu", "year": 2026, "status": "rejected", "title": "SOReL and TOReL: Two Methods for Fully Offline Reinforcement Learning", "authors": ["Mattie Fellows", "Clarisse Wibault", "Uljad Berdica", "Johannes Forkel", "Michael A Osborne", "Jakob Nicolaus Foerster"], "authorids": ["~Mattie_Fellows1", "~Clarisse_Wibault1", "~Uljad_Berdica1", "~Johannes_Forkel1", "~Michael_A_Osborne1", "~Jakob_Nicolaus_Foerster1"], "authors_source": "OpenReview API", "abstract": "Sample efficiency remains a major obstacle for real world adoption of reinforcement learning (RL): success has been limited to settings where simulators provide access to essentially unlimited environment interactions, which in reality are typically costly or dangerous to obtain. Offline RL in principle offers a solution by exploiting offline data to learn a near-optimal policy before deployment. In practice, however, current offline RL methods rely on extensive online interactions for hyperparameter tuning, and have no reliable bound on their initial online performance. To address these two issues, we introduce two algorithms. Firstly, SOReL: an algorithm for safe offline reinforcement learning. Using only offline data our Bayesian approach infers a posterior over environment dynamics to obtain a reliable estimate of the online performance via the posterior predictive uncertainty. Crucially, all hyperparameters are also tuned fully offline. Secondly, we introduce TOReL: a tuning for offline reinforcement learning algorithm that extends our information rate based offline hyperparameter tuning methods to general offline RL approaches. Our empirical evaluation confirms SOReL's ability to accurately estimate regret in the Bayesian setting whilst TOReL's offline hyperparameter tuning achieves competitive performance with the best online hyperparameter tuning methods using only offline data. Thus, SOReL and TOReL make a significant step towards safe and reliable offline RL, unlocking the potential for RL in the real world.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=LJ6AvummWu", "arxiv_id": "2505.22442", "paper_pdf": "papers/LJ6AvummWu.pdf", "paper_pdf_sha256": "101a7e7b321d394a985c9b6d12e48597c5aff6718f7d07974400677c95105b82", "paper_pdf_bytes": 6541472, "paper_pdf_source": "openreview", "code_url": "https://github.com/CWibault/sorel_torel", "code_repository": "CWibault/sorel_torel", "code_commit": "8babfc8950e8acbdbf252b1e6cd3f34048c17793", "code_archive": "repos/LJ6AvummWu.zip", "code_archive_sha256": "3c9fe3661e799299cff2a9b26a5a4e6c2b8cba127db80381da91ee2359f64844", "code_archive_bytes": 81895, "code_file_count": 41, "code_extensions": {".py": 37, ".sh": 4}, "github_disk_usage_kb": 60, "github_languages": {"Python": 216922, "Shell": 998}, "github_archived": false, "github_pushed_at": "2025-05-27T22:26:15Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/sorel-and-torel-two-methods-for-fully-offline"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "cWrqs2lwCJ", "year": 2025, "status": "rejected", "title": "Thinking Forward and Backward: Effective Backward Planning with Large Language Models", "authors": ["Allen Z. Ren", "brian ichter", "Anirudha Majumdar"], "authorids": ["~Allen_Z._Ren1", "~brian_ichter1", "~Anirudha_Majumdar1"], "authors_source": "OpenReview API", "abstract": "Large language models (LLMs) have exhibited remarkable reasoning and planning capabilities. Most prior work in this area has used LLMs to reason through steps from an initial to a goal state or criterion, thereby effectively reasoning in a forward direction. Nonetheless, many planning problems exhibit an inherent asymmetry such that planning backward from the goal is significantly easier --- for example, if there are bottlenecks close to the goal. We take inspiration from this observation and demonstrate that this bias holds for LLM planning as well: planning performance in one direction correlates with the planning complexity of the problem in that direction. However, our experiments also reveal systematic biases which lead to poor planning in the backward direction. With this knowledge, we propose a backward planning algorithm for LLMs that first flips the problem and then plans forward in the flipped problem. This helps avoid the backward bias, generate more diverse candidate plans, and exploit asymmetries between the forward and backward directions in planning problems --- we find that combining planning in both directions with self-verification improves the overall planning success rates by 4-24% in three planning domains.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=cWrqs2lwCJ", "arxiv_id": "2411.01790", "paper_pdf": "papers/cWrqs2lwCJ.pdf", "paper_pdf_sha256": "4ee1a03edcc2ea1b74f12996f37a6e59ab01cbbb52cc4fde1e03f95ff03a6da9", "paper_pdf_bytes": 4620505, "paper_pdf_source": "openreview", "code_url": "https://github.com/irom-princeton/llm-backward", "code_repository": "irom-princeton/llm-backward", "code_commit": "b215c6785a18803a4b116f06e6850a19cc7a6048", "code_archive": "repos/cWrqs2lwCJ.zip", "code_archive_sha256": "bb0ad5cafa0bcf8812a3d139d30e81825a5104a49e0703a4b8ebe82c9f120f19", "code_archive_bytes": 22707, "code_file_count": 12, "code_extensions": {".py": 12}, "github_disk_usage_kb": 18, "github_languages": {"Python": 78294}, "github_archived": false, "github_pushed_at": "2024-11-05T03:32:56Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/thinking-forward-and-backward-effective"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "53kW6e1uNN", "year": 2024, "status": "rejected", "title": "AFDGCF: Adaptive Feature De-correlation Graph Collaborative Filtering for Recommendations", "authors": ["Wei Wu", "Chao Wang", "Dazhong Shen", "Chuan Qin", "Hui Xiong"], "authorids": ["~Wei_Wu25", "~Chao_Wang14", "~Dazhong_Shen1", "~Chuan_Qin1", "~Hui_Xiong1"], "authors_source": "OpenReview API", "abstract": "Collaborative filtering methods based on graph neural networks (GNNs) have witnessed significant success in recommender systems (RS), capitalizing on their ability to capture collaborative signals within intricate user-item relationships via message-passing mechanisms. However, these GNN-based RS inadvertently introduce a linear correlation between user and item embeddings, contradicting the goal of providing personalized recommendations. While existing research predominantly ascribes this flaw to the over-smoothing problem, this paper underscores the critical, often overlooked role of the over-correlation issue in diminishing the effectiveness of GNN representations and subsequent recommendation performance. The unclear relationship between over-correlation and over-smoothing in RS, coupled with the challenge of adaptively minimizing the impact of over-correlation while preserving collaborative filtering signals, is quite challenging. To this end, this paper aims to address the aforementioned gap by undertaking a comprehensive study of the over-correlation issue in graph collaborative filtering models. Empirical evidence substantiates the widespread prevalence of over-correlation in these models. Furthermore, a theoretical analysis establishes a pivotal connection between the over-correlation and over-smoothing predicaments. Leveraging these insights, we introduce the Adaptive Feature De-correlation Graph Collaborative Filtering (AFDGCF) Framework, which dynamically applies correlation penalties to the feature dimensions of the representation matrix, effectively alleviating both over-correlation and over-smoothing challenges. The efficacy of the proposed framework is corroborated through extensive experiments conducted with four different graph collaborative filtering models across four publicly available datasets, demonstrating the superiority of AFDGCF in enhancing the performance landscape of graph collaborative filtering models.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=53kW6e1uNN", "arxiv_id": "2403.17416", "paper_pdf": "papers/53kW6e1uNN.pdf", "paper_pdf_sha256": "f2c133df5a7f07873788c4cdf3ae3e1e2e8354091dca97c5e6ac53b1ed79e345", "paper_pdf_bytes": 6094555, "paper_pdf_source": "openreview", "code_url": "https://github.com/U-rara/AFDGCF", "code_repository": "U-rara/AFDGCF", "code_commit": "0eba7b6dbf7cfd2c1a0dcebd9e2f0742a68b89aa", "code_archive": "repos/53kW6e1uNN.zip", "code_archive_sha256": "38919b617762d350bf5da4ffcd8222d0bf84bfcab807a169ba094aec4410d806", "code_archive_bytes": 29630, "code_file_count": 11, "code_extensions": {".py": 11}, "github_disk_usage_kb": 16, "github_languages": {"Python": 81015}, "github_archived": false, "github_pushed_at": "2024-01-23T13:47:58Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/afdgcf-adaptive-feature-de-correlation-graph"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "t-hNmA0cVSW", "year": 2023, "status": "rejected", "title": "Semi-supervised Counting via Pixel-by-pixel Density Distribution Modelling", "authors": ["Hui LIN", "Zhiheng Ma", "Rongrong Ji", "Yaowei Wang", "su zhou", "Xiaopeng Hong"], "authorids": ["~Hui_LIN4", "~Zhiheng_Ma1", "~Rongrong_Ji5", "~Yaowei_Wang1", "~su_zhou1", "~Xiaopeng_Hong4"], "authors_source": "OpenReview API", "abstract": "This paper focuses on semi-supervised crowd counting, where only a small portion of the training data are labeled. We formulate the pixel-wise density value to regress as a probability distribution, instead of a single deterministic value, and utilize a dual-branch structure to model the corresponding discrete form of the distribution function. On the basis, we propose a semi-supervised crowd counting model. Firstly, we enhance the transformer decoder by usingdensity tokens to specialize the forwards of decoders w.r.t. different density intervals; Secondly, we design a pixel-wise distribution matching loss to measure the differences in the pixel-wise density distributions between the prediction and the ground-truth; Thirdly, we propose an interleaving consistency regularization term to align the prediction of two branches and make them consistent. Extensive experiments on four datasets are performed to show that our method clearly outperforms the competitors by a large margin under various labeled ratio settings.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=t-hNmA0cVSW", "arxiv_id": "2402.15297", "paper_pdf": "papers/t-hNmA0cVSW.pdf", "paper_pdf_sha256": "a9f514d0bce69c0ecbdd86e9de62c489df551d4d369dc5cd0f5b309098caf099", "paper_pdf_bytes": 1719675, "paper_pdf_source": "openreview", "code_url": "https://github.com/LoraLinH/Semi-supervised-Counting-via-Pixel-by-pixel-Density-Distribution-Modelling", "code_repository": "LoraLinH/Semi-supervised-Counting-via-Pixel-by-pixel-Density-Distribution-Modelling", "code_commit": "66034010cc7134fa5265b282e606cdcb7beb3399", "code_archive": "repos/t-hNmA0cVSW.zip", "code_archive_sha256": "2ab266684d36250a645091676a1e02f61111bfd66817cb416102b131df813996", "code_archive_bytes": 19965, "code_file_count": 11, "code_extensions": {".py": 11}, "github_disk_usage_kb": 21, "github_languages": {"Python": 56116}, "github_archived": false, "github_pushed_at": "2025-04-02T11:55:14Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/semi-supervised-counting-via-pixel-by-pixel"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "gxk4-rVATDA", "year": 2022, "status": "rejected", "title": "Bit-wise Training of Neural Network Weights", "authors": ["Cristian Ivan"], "authorids": ["~Cristian_Ivan1"], "authors_source": "OpenReview API", "abstract": "We propose an algorithm where the individual bits representing the weights of a neural network are learned. This method allows training weights with integer values on arbitrary bit-depths and naturally uncovers sparse networks, without additional constraints or regularization techniques. We show better results than the standard training technique with fully connected networks and similar performance as compared to standard training for residual networks. By training bits in a selective manner we found that the biggest contribution to achieving high accuracy is given by the first three most significant bits, while the rest provide an intrinsic regularization. As a consequence we show that more than 90% of a network can be used to store arbitrary codes without affecting the its accuracy. These codes can be random noise, binary files or even the weights of previously trained networks.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=gxk4-rVATDA", "arxiv_id": "2202.09571", "paper_pdf": "papers/gxk4-rVATDA.pdf", "paper_pdf_sha256": "525dfaab09c5bd0aaf95bff67a39eb0a9c58b57103475c37b2bf7086db9e04ed", "paper_pdf_bytes": 2383381, "paper_pdf_source": "openreview", "code_url": "https://github.com/iclr2022-2798/bit-wise-training", "code_repository": "iclr2022-2798/bit-wise-training", "code_commit": "39fdc97caa080bd3b09ae7abcef75f79ca6dac5d", "code_archive": "repos/gxk4-rVATDA.zip", "code_archive_sha256": "5131a7aabdd85cade759642bde287a67de69543b370984bf71e1e753f8197425", "code_archive_bytes": 12573, "code_file_count": 5, "code_extensions": {".py": 5}, "github_disk_usage_kb": 20, "github_languages": {"Python": 47611}, "github_archived": false, "github_pushed_at": "2021-10-11T20:45:43Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/bit-wise-training-of-neural-network-weights-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ztMLindFLWR", "year": 2021, "status": "rejected", "title": "Breaking the Expressive Bottlenecks of Graph Neural Networks", "authors": ["Mingqi Yang", "Yanming Shen", "Heng Qi", "Baocai Yin"], "authorids": ["~Mingqi_Yang1", "shen@dlut.edu.cn", "hengqi@dlut.edu.cn", "~Baocai_Yin1"], "authors_source": "OpenReview API", "abstract": "Recently, the Weisfeiler-Lehman (WL) graph isomorphism test was used to measure the expressiveness of graph neural networks (GNNs), showing that the neighborhood aggregation GNNs were at most as powerful as 1-WL test in distinguishing graph structures. There were also improvements proposed in analogy to $k$-WL test ($k>1$). However, the aggregators in these GNNs are far from injective as required by the WL test, and suffer from weak distinguishing strength, making it become expressive bottlenecks. In this paper, we improve the expressiveness by exploring powerful aggregators. We reformulate aggregation with the corresponding aggregation coefficient matrix, and then systematically analyze the requirements of the aggregation coefficient matrix for building more powerful aggregators and even injective aggregators. It can also be viewed as the strategy for preserving the rank of hidden features, and implies that basic aggregators correspond to a special case of low-rank transformations. We also show the necessity of applying nonlinear units ahead of aggregation, which is different from most aggregation-based GNNs. Based on our theoretical analysis, we develop two GNN layers, ExpandingConv and CombConv. Experimental results show that our models significantly boost performance, especially for large and densely connected graphs.", "decision": "Reject", "meta_review": null, "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=ztMLindFLWR", "arxiv_id": "2012.07219", "paper_pdf": "papers/ztMLindFLWR.pdf", "paper_pdf_sha256": "676865eab7cb0510e8c452632a31dfb4db77efee5a9020e0ca4ba0758ee19194", "paper_pdf_bytes": 1365702, "paper_pdf_source": "openreview", "code_url": "https://github.com/qslim/epcb-gnns", "code_repository": "qslim/epcb-gnns", "code_commit": "87bca5bb371ec15fe0fb68f7a93ae9933ef0e76a", "code_archive": "repos/ztMLindFLWR.zip", "code_archive_sha256": "a35d57fdb76fe8cf73bf1e332a2b5b0fd2ca66da11ef37c997592de85ba0f047", "code_archive_bytes": 36353, "code_file_count": 24, "code_extensions": {".py": 19, ".sh": 5}, "github_disk_usage_kb": 27, "github_languages": {"Python": 99910, "Shell": 3373}, "github_archived": false, "github_pushed_at": "2022-06-21T08:29:30Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/breaking-the-expressive-bottlenecks-of-graph-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "H1xTup4KPr", "year": 2020, "status": "rejected", "title": "Needles in Haystacks: On Classifying Tiny Objects in Large Images", "authors": ["Nick Pawlowski", "Suvrat Bhooshan", "Nicolas Ballas", "Francesco Ciompi", "Ben Glocker", "Michal Drozdzal"], "authorids": ["pawlowski.nick@gmail.com", "sbh@fb.com", "ballasn@fb.com", "f.ciompi@gmail.com", "b.glocker@imperial.ac.uk", "mdrozdzal@fb.com"], "authors_source": "OpenReview API", "abstract": "In some important computer vision domains, such as medical or hyperspectral imaging, we care about the classification of tiny objects in large images. However, most Convolutional Neural Networks (CNNs) for image classification were developed using biased datasets that contain large objects, in mostly central image positions. To assess whether classical CNN architectures work well for tiny object classification we build a comprehensive testbed containing two datasets: one derived from MNIST digits and one from histopathology images. This testbed allows controlled experiments to stress-test CNN architectures with a broad spectrum of signal-to-noise ratios. Our observations indicate that: (1) There exists a limit to signal-to-noise below which CNNs fail to generalize and that this limit is affected by dataset size - more data leading to better performances; however, the amount of training data required for the model to generalize scales rapidly with the inverse of the object-to-image ratio (2) in general, higher capacity models exhibit better generalization; (3) when knowing the approximate object sizes, adapting receptive field is beneficial; and (4) for very small signal-to-noise ratio the choice of global pooling operation affects optimization, whereas for relatively large signal-to-noise values, all tested global pooling operations exhibit similar performance.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=H1xTup4KPr", "arxiv_id": "1908.06037", "paper_pdf": "papers/H1xTup4KPr.pdf", "paper_pdf_sha256": "7da3f4a4f9ca3f738bbf5e6132d49c609df7a82a032badf85147fc6afc45cc3a", "paper_pdf_bytes": 13200937, "paper_pdf_source": "openreview", "code_url": "https://github.com/facebookresearch/Needles-in-Haystacks", "code_repository": "facebookresearch/Needles-in-Haystacks", "code_commit": "bd08ae43b8174f9e5cda8763d99ba70a26a5b0c6", "code_archive": "repos/H1xTup4KPr.zip", "code_archive_sha256": "c1803276f7adfe6b7477887cf4442a0fc16eaf26f99b9bf1e8c8e81943af993c", "code_archive_bytes": 32871, "code_file_count": 17, "code_extensions": {".py": 17}, "github_disk_usage_kb": 25, "github_languages": {"Python": 58499}, "github_archived": true, "github_pushed_at": "2019-06-28T12:18:01Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/needles-in-haystacks-on-classifying-tiny"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Bkf1tjR9KQ", "year": 2019, "status": "rejected", "title": "DVOLVER: Efficient Pareto-Optimal Neural Network Architecture Search", "authors": ["Guillaume Michel", "Mohammed Amine Alaoui", "Alice Lebois", "Amal Feriani", "Mehdi Felhi"], "authorids": ["guillaume.michel@netatmo.com", "mohammed-amine.alaoui@netatmo.com", "alice.lebois@netatmo.com", "amal.feriani@netatmo.com", "mehdi.felhi@netatmo.com"], "authors_source": "OpenReview API", "abstract": "Automatic search of neural network architectures is a standing research topic. In addition to the fact that it presents a faster alternative to hand-designed architectures, it can improve their efficiency and for instance generate Convolutional Neural Networks (CNN) adapted for mobile devices. In this paper, we present a multi-objective neural architecture search method to find a family of CNN models with the best accuracy and computational resources tradeoffs, in a search space inspired by the state-of-the-art findings in neural search. Our work, called Dvolver, evolves a population of architectures and iteratively improves an approximation of the optimal Pareto front. Applying Dvolver on the model accuracy and on the number of floating points operations as objective functions, we are able to find, in only 2.5 days 1 , a set of competitive mobile models on ImageNet. Amongst these models one architecture has the same Top-1 accuracy on ImageNet as NASNet-A mobile with 8% less floating point operations and another one has a Top-1 accuracy of 75.28% on ImageNet exceeding by 0.28% the best MobileNetV2 model for the same computational resources.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=Bkf1tjR9KQ", "arxiv_id": "1902.01654", "paper_pdf": "papers/Bkf1tjR9KQ.pdf", "paper_pdf_sha256": "64c31a0ea1fa3bb2fdcfb0e2b8a41bee8f6e93f094ec74d634c03301deb5df7b", "paper_pdf_bytes": 376334, "paper_pdf_source": "openreview", "code_url": "https://github.com/guillaume-michel/dvolver", "code_repository": "guillaume-michel/dvolver", "code_commit": "b1301c2790b91172fc3322b1836cb71e55afe7b9", "code_archive": "repos/Bkf1tjR9KQ.zip", "code_archive_sha256": "7760873ac19c77427b0984c6b4dd01022f3300df9dfe73ff128e2931e60291c8", "code_archive_bytes": 65891, "code_file_count": 29, "code_extensions": {".py": 27, ".sh": 2}, "github_disk_usage_kb": 54, "github_languages": {"Python": 198562, "Shell": 4794}, "github_archived": false, "github_pushed_at": "2019-02-04T12:39:58Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/dvolver-efficient-pareto-optimal-neural"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "SJmAXkgCb", "year": 2018, "status": "rejected", "title": "DNN Feature Map Compression using Learned Representation over GF(2)", "authors": ["Denis A. Gudovskiy", "Alec Hodgkinson", "Luca Rigazio"], "authorids": ["denis.gudovskiy@us.panasonic.com", "alec.hodgkinson@us.panasonic.com", "luca.rigazio@us.panasonic.com"], "authors_source": "OpenReview API", "abstract": "In this paper, we introduce a method to compress intermediate feature maps of deep neural networks (DNNs) to decrease memory storage and bandwidth requirements during inference. Unlike previous works, the proposed method is based on converting fixed-point activations into vectors over the smallest GF(2) finite field followed by nonlinear dimensionality reduction (NDR) layers embedded into a DNN. Such an end-to-end learned representation finds more compact feature maps by exploiting quantization redundancies within the fixed-point activations along the channel or spatial dimensions. We apply the proposed network architecture to the tasks of ImageNet classification and PASCAL VOC object detection. Compared to prior approaches, the conducted experiments show a factor of 2 decrease in memory requirements with minor degradation in accuracy while adding only bitwise computations.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=SJmAXkgCb", "arxiv_id": "1808.05285", "paper_pdf": "papers/SJmAXkgCb.pdf", "paper_pdf_sha256": "55d99b9ba1df096bd61823c91e709a819b825bff577e7b60fc196caa026789d4", "paper_pdf_bytes": 280243, "paper_pdf_source": "openreview", "code_url": "https://github.com/gudovskiy/fmap_compression", "code_repository": "gudovskiy/fmap_compression", "code_commit": "0e60097613b7fd7eac1d72775b202fc7a9bbbcea", "code_archive": "repos/SJmAXkgCb.zip", "code_archive_sha256": "eb371b7d36a9d4aa2b7ac6a2bc66806b25eb170273d7ba63fdda38dde43d6f23", "code_archive_bytes": 75728, "code_file_count": 17, "code_extensions": {".cpp": 7, ".cu": 6, ".hpp": 3, ".cuh": 1}, "github_disk_usage_kb": 29, "github_languages": {"C++": 72058, "Cuda": 28024}, "github_archived": false, "github_pushed_at": "2018-11-21T01:40:02Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/dnn-feature-map-compression-using-learned"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Am95bfE207", "year": 2026, "status": "rejected", "title": "From Low Intrinsic Dimensionality to Non-Vacuous Generalization Bounds in Deep Multi-Task Learning", "authors": ["Hossein Zakerinia", "Dorsa Ghobadi", "Christoph H. Lampert"], "authorids": ["~Hossein_Zakerinia1", "~Dorsa_Ghobadi2", "~Christoph_H._Lampert6"], "authors_source": "OpenReview API", "abstract": "Deep learning methods are known to generalize well from training to future data, even in an overparametrized regime, where they could easily overfit. One explanation for this phenomenon is that even when their ambient dimensionality, (i.e. the number of parameters) is large, the models’ intrinsic dimensionality is small; specifically, their learning takes place in a small subspace of all possible weight configurations. In this work, we confirm this phenomenon in the setting of deep multi-task learning. We introduce a method to parametrize multi-task network directly in the low-dimensional space, facilitated by the use of random expansions techniques. We then show that high-accuracy multi-task solutions can be found with much smaller intrinsic dimensionality (fewer free parameters) than what single-task learning requires. Subsequently, we show that the low-dimensional representations in combination with weight compression and PAC-Bayesian reasoning lead to the first non-vacuous generalization bounds for deep multi-task networks.", "decision": "Reject", "meta_review": "", "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=Am95bfE207", "arxiv_id": "2501.19067", "paper_pdf": "papers/Am95bfE207.pdf", "paper_pdf_sha256": "afc7be806de1d88d2201446c496281181c0d0472e4f4299f9035659d7286d605", "paper_pdf_bytes": 350604, "paper_pdf_source": "openreview", "code_url": "https://github.com/hzakerinia/MTL", "code_repository": "hzakerinia/MTL", "code_commit": "b20ca886303c8cfe0a067422e653373c5c239c0a", "code_archive": "repos/Am95bfE207.zip", "code_archive_sha256": "a01c1af3089054a0e9a9933edc755f63dadf690ee359023d4041ab5bed4e8eff", "code_archive_bytes": 73462, "code_file_count": 32, "code_extensions": {".py": 32}, "github_disk_usage_kb": 60, "github_languages": {"Python": 257002}, "github_archived": false, "github_pushed_at": "2025-05-14T12:33:35Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/deep-multi-task-learning-has-low-amortized"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "fvo6q86NKG", "year": 2025, "status": "rejected", "title": "CBF-LLM: Safe Control for LLM Alignment", "authors": ["Yuya Miyaoka", "Masaki Inoue"], "authorids": ["~Yuya_Miyaoka1", "~Masaki_Inoue1"], "authors_source": "OpenReview API", "abstract": "This paper proposes a control-based framework for aligning large language models (LLMs) by leveraging a control barrier function (CBF) to ensure user-desirable text generation. \nThe presented framework applies the CBF safety filter to the predicted token generated from the baseline LLM, to intervene in the generated text.\nThe safety filter includes two significant advantages:\nthis safety filter is an add-on type, allowing it to be used for alignment purposes without fine-tuning the baseline LLM,\nand if there is an evaluation model regarding the desired alignment, it can be directly applied to the filter design.\nThe overall text-generation system is implemented with Llama 3 and a BERT model, aiming to generate positive text.\nFinally, further applications and limitations of the CBF-LLM for other alignment tasks, including topic-keeping and hallucination mitigating, are discussed.", "decision": "Reject", "meta_review": "", "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=fvo6q86NKG", "arxiv_id": "2408.15625", "paper_pdf": "papers/fvo6q86NKG.pdf", "paper_pdf_sha256": "f1b3100272c76f672f0a93be47f34dc94fe7aefab5b3bb1c3597e8bb4b0dd4c9", "paper_pdf_bytes": 486425, "paper_pdf_source": "openreview", "code_url": "https://github.com/Mya-Mya/CBF-LLM", "code_repository": "Mya-Mya/CBF-LLM", "code_commit": "27cc8388f0f7e6f463848c2ecc0ca7989fc8b579", "code_archive": "repos/fvo6q86NKG.zip", "code_archive_sha256": "d84b0d6dd7e378e0029816e21fdaf39b5350d31504eedef9116ce29e21a5165d", "code_archive_bytes": 8167, "code_file_count": 9, "code_extensions": {".py": 9}, "github_disk_usage_kb": 20, "github_languages": {"Python": 13744}, "github_archived": false, "github_pushed_at": "2024-12-10T11:15:13Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/cbf-llm-safe-control-for-llm-alignment"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "mOTiVzTgF2", "year": 2024, "status": "rejected", "title": "ResiDual: Transformer with Dual Residual Connections", "authors": ["Shufang Xie", "Huishuai Zhang", "Junliang Guo", "Xu Tan", "Jiang Bian", "Hany Hassan Awadalla", "Arul Menezes", "Tao Qin", "Rui Yan"], "authorids": ["~Shufang_Xie1", "~Huishuai_Zhang3", "~Junliang_Guo1", "~Xu_Tan1", "~Jiang_Bian1", "~Hany_Hassan_Awadalla1", "~Arul_Menezes1", "~Tao_Qin1", "~Rui_Yan2"], "authors_source": "OpenReview API", "abstract": "Transformer networks have become the preferred architecture for many tasks due to their state-of-the-art performance. However, the optimal way to implement residual connections in Transformer, which are essential for effective training, is still debated. Two widely used variants are the Post-Layer-Normalization (Post-LN) and Pre-Layer-Normalization (Pre-LN) Transformers, which apply layer normalization after each residual block's output or before each residual block's input, respectively. While both variants enjoy their advantages, they also suffer from severe limitations: Post-LN causes gradient vanishing issue that hinders training deep Transformers, and Pre-LN causes representation collapse issue that limits model capacity. In this paper, we propose ResiDual, a novel Transformer architecture with Pre-Post-LN (PPLN), which fuses the connections in Post-LN and Pre-LN together, and inherits their advantages while avoids their limitations. We conduct both theoretical analyses and empirical experiments to verify the effectiveness of ResiDual. Theoretically, we prove that ResiDual has a lower bound on the gradient to avoid the vanishing issue due to the residual connection from Pre-LN. Moreover, ResiDual also has diverse model representations to avoid the collapse issue due to the residual connection from Post-LN. Empirically, ResiDual outperforms both Post-LN and Pre-LN on several machine translation benchmarks across different network depths and data sizes.", "decision": "Reject", "meta_review": "", "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=mOTiVzTgF2", "arxiv_id": "2304.14802", "paper_pdf": "papers/mOTiVzTgF2.pdf", "paper_pdf_sha256": "b74039d1cd2b681315278bc56f8d4dfc03d1617ebc4765548fc803e82aaa5b7b", "paper_pdf_bytes": 607065, "paper_pdf_source": "openreview", "code_url": "https://github.com/microsoft/ResiDual", "code_repository": "microsoft/ResiDual", "code_commit": "8682f7510be105a0caf5b98a021e712f44f90ace", "code_archive": "repos/mOTiVzTgF2.zip", "code_archive_sha256": "cec65c73f8e80c77f8450dcd56aefe54692dbfcb4702ca6c61522f17e9994aa7", "code_archive_bytes": 15699, "code_file_count": 3, "code_extensions": {".py": 3}, "github_disk_usage_kb": 17, "github_languages": {"Python": 20837}, "github_archived": true, "github_pushed_at": "2023-08-18T18:23:56Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/residual-transformer-with-dual-residual"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "4nrZXPFN1c4", "year": 2023, "status": "rejected", "title": "Energy Transformer", "authors": ["Benjamin Hoover", "Yuchen Liang", "Bao Pham", "Rameswar Panda", "Hendrik Strobelt", "Duen Horng Chau", "Mohammed J Zaki", "Dmitry Krotov"], "authorids": ["~Benjamin_Hoover1", "~Yuchen_Liang2", "~Bao_Pham1", "~Rameswar_Panda1", "~Hendrik_Strobelt1", "~Duen_Horng_Chau1", "~Mohammed_J_Zaki1", "~Dmitry_Krotov2"], "authors_source": "OpenReview API", "abstract": "Transformers have become the de facto  models of choice in machine learning, typically leading to impressive performance on many applications. At the same time, the architectural development in the transformer world is mostly driven by empirical findings, and  the theoretical understanding of their architectural building blocks is rather limited. In contrast, Dense Associative Memory models or Modern Hopfield Networks have a well-established theoretical foundation, but have not yet demonstrated truly impressive practical results. We propose a transformer architecture that replaces the sequence of feedforward transformer blocks with a single large Associative Memory model. Our novel architecture, called Energy Transformer (or ET for short), has many of the familiar architectural primitives that are often used in the current generation of transformers. However, it is not identical to the existing architectures. The sequence of transformer layers in ET is purposely designed to minimize a specifically engineered energy function, which is responsible for representing the relationships between the tokens. As a consequence of this computational principle, the attention in ET is different from the conventional attention mechanism.  In this work, we introduce the theoretical foundations of ET, explore it's empirical capabilities using the image completion task, and obtain strong quantitative results on the graph anomaly detection task.", "decision": "Reject", "meta_review": null, "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=4nrZXPFN1c4", "arxiv_id": "2302.07253", "paper_pdf": "papers/4nrZXPFN1c4.pdf", "paper_pdf_sha256": "4583760c02fa5b8a3975c53b295601c92b33a05e9c07e98fad9260c38a7d17cc", "paper_pdf_bytes": 19682945, "paper_pdf_source": "openreview", "code_url": "https://github.com/zhuergou/Energy-Transformer-for-Graph-Anomaly-Detection", "code_repository": "zhuergou/Energy-Transformer-for-Graph-Anomaly-Detection", "code_commit": "958cb1db8a2a1eae106ba2a179c952e38cab179e", "code_archive": "repos/4nrZXPFN1c4.zip", "code_archive_sha256": "9ee7efb5b8d5824211794a9fe2c79c43429f044c2db429d225a56900cd9fd0bf", "code_archive_bytes": 19192, "code_file_count": 9, "code_extensions": {".py": 9}, "github_disk_usage_kb": 25, "github_languages": {"Python": 57204}, "github_archived": false, "github_pushed_at": "2023-05-21T03:57:58Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/energy-transformer"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "IEsx-jwFk3g", "year": 2022, "status": "rejected", "title": "Deep Representations for Time-varying Brain Datasets", "authors": ["Sikun Lin", "Shuyun Tang", "Ambuj Singh"], "authorids": ["~Sikun_Lin1", "~Shuyun_Tang1", "~Ambuj_Singh1"], "authors_source": "OpenReview API", "abstract": "Finding an appropriate representation of dynamic activities in the brain is crucial for many downstream applications. Due to its highly dynamic nature, temporally averaged fMRI (functional magnetic resonance imaging) cannot capture the whole picture of underlying brain activities, and previous works lack the ability to learn and interpret the latent dynamics in brain architectures. In this paper, we build an efficient graph neural network model that incorporates both region-mapped fMRI sequences and structural connectivities obtained from DWI (diffusion-weighted imaging) as inputs. Through novel sample-level adaptive adjacency matrix learning and multi-resolution inner cluster smoothing, we find good representations of the latent brain dynamics. We also attribute inputs with integrated gradients, which enables us to infer (1) highly involved brain connections and subnetworks for each task (2) keyframes of imaging sequences along the temporal axis, and (3) subnetworks that discriminate between individual subjects. This ability to identify critical subnetworks that characterize brain states across heterogeneous tasks and individuals is of great importance to neuroscience research. Extensive experiments and ablation studies demonstrate our proposed method's superiority and efficiency in spatial-temporal graph signal modeling with insightful interpretations of brain dynamics.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=IEsx-jwFk3g", "arxiv_id": "2205.11648", "paper_pdf": "papers/IEsx-jwFk3g.pdf", "paper_pdf_sha256": "e45af310eaeeb14566f45d6e65cfa34cac9477d05d84831e4c79dc7f0139c109", "paper_pdf_bytes": 34179969, "paper_pdf_source": "openreview", "code_url": "https://github.com/sklin93/ReBraiD", "code_repository": "sklin93/ReBraiD", "code_commit": "7fbecd68e0fba5b86670916441dc509098417b42", "code_archive": "repos/IEsx-jwFk3g.zip", "code_archive_sha256": "f3a2ddf49f894acec99d6785488dae4eaa60e6033cbdd2a8cdf4821190f8fbc5", "code_archive_bytes": 18930, "code_file_count": 5, "code_extensions": {".py": 5}, "github_disk_usage_kb": 22, "github_languages": {"Python": 77048}, "github_archived": false, "github_pushed_at": "2022-08-16T19:31:48Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/deep-representations-for-time-varying-brain-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "RGeQOjc58d", "year": 2021, "status": "rejected", "title": "Improved Gradient based Adversarial Attacks for Quantized Networks", "authors": ["Kartik Gupta", "Thalaiyasingam Ajanthan"], "authorids": ["~Kartik_Gupta2", "~Thalaiyasingam_Ajanthan1"], "authors_source": "OpenReview API", "abstract": "Neural network quantization has become increasingly popular due to efficient memory consumption and faster computation resulting from bitwise operations on the quantized networks. Even though they exhibit excellent generalization capabilities, their robustness properties are not well-understood. In this work, we systematically study the robustness of quantized networks against gradient based adversarial attacks and demonstrate that these quantized models suffer from gradient vanishing issues and show a fake sense of robustness. By attributing gradient vanishing to poor forward-backward signal propagation in the trained network, we introduce a simple temperature scaling approach to mitigate this issue while preserving the decision boundary. Despite being a simple modification to existing gradient based adversarial attacks, experiments on CIFAR-10/100 datasets with multiple network architectures demonstrate that our temperature scaled attacks obtain near-perfect success rate on quantized networks while outperforming original attacks on adversarially trained models as well as floating point networks.\n", "decision": "Reject", "meta_review": null, "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=RGeQOjc58d", "arxiv_id": "2003.13511", "paper_pdf": "papers/RGeQOjc58d.pdf", "paper_pdf_sha256": "c13eafc3dddfaebeae3649727ec54e41cb61d874d173bc4ac9a31dc71090386e", "paper_pdf_bytes": 455817, "paper_pdf_source": "openreview", "code_url": "https://github.com/kartikgupta-at-anu/attack-bnn", "code_repository": "kartikgupta-at-anu/attack-bnn", "code_commit": "79c60f64b9ced3084571c805404876475f2bf1bd", "code_archive": "repos/RGeQOjc58d.zip", "code_archive_sha256": "66abc5e1ddba9997967ba2a07f88f849a1510a09070e72843bbb04b229823d3b", "code_archive_bytes": 40718, "code_file_count": 26, "code_extensions": {".py": 20, ".sh": 6}, "github_disk_usage_kb": 28, "github_languages": {"Python": 111733, "Shell": 30760}, "github_archived": false, "github_pushed_at": "2022-05-06T13:58:39Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/improved-gradient-based-adversarial-attacks"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "SylUzpNFDS", "year": 2020, "status": "rejected", "title": "SoftLoc: Robust Temporal Localization under Label Misalignment", "authors": ["Julien Schroeter", "Kirill Sidorov", "Dave Marshall"], "authorids": ["schroeterj1@cardiff.ac.uk", "sidorovk@cardiff.ac.uk", "marshallad@cardiff.ac.uk"], "authors_source": "OpenReview API", "abstract": "This work addresses the long-standing problem of robust event localization in the presence of temporally of misaligned labels in the training data. We propose a novel versatile loss function that generalizes a number of training regimes from standard fully-supervised cross-entropy to count-based weakly-supervised learning. Unlike classical models which are constrained to strictly fit the annotations during training, our soft localization learning approach relaxes the reliance on the exact position of labels instead. Training with this new loss function exhibits strong robustness to temporal misalignment of labels, thus alleviating the burden of precise annotation of temporal sequences. We demonstrate state-of-the-art performance against standard benchmarks in a number of challenging experiments and further show that robustness to label noise is not achieved at the expense of raw performance. ", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=SylUzpNFDS", "arxiv_id": null, "paper_pdf": "papers/SylUzpNFDS.pdf", "paper_pdf_sha256": "e29061a3b055fb66bfb6dbc3dcc1cad4df2748cc151d5544853debbf84cdc2c8", "paper_pdf_bytes": 1232214, "paper_pdf_source": "openreview", "code_url": "https://github.com/SoftLocNIPS/submission", "code_repository": "SoftLocNIPS/submission", "code_commit": "c5a5106c68070cd8fb037f2940e885132e58f463", "code_archive": "repos/SylUzpNFDS.zip", "code_archive_sha256": "db19d79331b9dbeb14a099eab4f7402570adf7b023b482f76c5d02378f130f9d", "code_archive_bytes": 28830, "code_file_count": 12, "code_extensions": {".py": 11, ".sh": 1}, "github_disk_usage_kb": 25, "github_languages": {"Python": 75271, "Shell": 324}, "github_archived": false, "github_pushed_at": "2019-08-12T13:18:36Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/softloc-robust-temporal-localization-under"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "rkxjnjA5KQ", "year": 2019, "status": "rejected", "title": "Transfer Learning for Related Reinforcement Learning Tasks via Image-to-Image Translation", "authors": ["Shani Gamrian", "Yoav Goldberg"], "authorids": ["gamrianshani@gmail.com", "yoav.goldberg@gmail.com"], "authors_source": "OpenReview API", "abstract": "Deep Reinforcement Learning has managed to achieve state-of-the-art results in learning control policies directly from raw pixels. However, despite its remarkable success, it fails to generalize, a fundamental component required in a stable Artificial Intelligence system. Using the Atari game Breakout, we demonstrate the difficulty of a trained agent in adjusting to simple modifications in the raw image, ones that a human could adapt to trivially. In transfer learning, the goal is to use the knowledge gained from the source task to make the training of the target task faster and better. We show that using various forms of fine-tuning, a common method for transfer learning, is not effective for adapting to such small visual changes. In fact, it is often easier to re-train the agent from scratch than to fine-tune a trained agent. We suggest that in some cases transfer learning can be improved by adding a dedicated component whose goal is to learn to visually map between the known domain and the new one. Concretely, we use Unaligned Generative Adversarial Networks (GANs) to create a mapping function to translate images in the target task to corresponding images in the source task. These mapping functions allow us to transform between various variations of the Breakout game, as well as between different levels of a Nintendo game, Road Fighter. We show that learning this mapping is substantially more efficient than re-training. A visualization of a trained agent playing Breakout and Road Fighter, with and without the GAN transfer, can be seen in \\url{https://streamable.com/msgtm} and \\url{https://streamable.com/5e2ka}.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=rkxjnjA5KQ", "arxiv_id": "1806.07377", "paper_pdf": "papers/rkxjnjA5KQ.pdf", "paper_pdf_sha256": "ddf056ac04908711dd3f5205bc275217b564ac6218eaecd8c2738f4d56b60b5b", "paper_pdf_bytes": 928773, "paper_pdf_source": "openreview", "code_url": "https://github.com/ShaniGam/RL-GAN", "code_repository": "ShaniGam/RL-GAN", "code_commit": "6c4e5f95826b2b99a893e66380050ff38b1d0cf5", "code_archive": "repos/rkxjnjA5KQ.zip", "code_archive_sha256": "4c66f9ec7d6373bd2d651ab858d9607e6a180170597f8c9e34d39b2270e559c8", "code_archive_bytes": 53667, "code_file_count": 24, "code_extensions": {".py": 24}, "github_disk_usage_kb": 55, "github_languages": {"Python": 147451}, "github_archived": false, "github_pushed_at": "2020-03-22T17:20:07Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/transfer-learning-for-related-reinforcement"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "SyL9u-WA-", "year": 2018, "status": "rejected", "title": "Stabilizing Gradients for Deep Neural Networks via Efficient SVD Parameterization", "authors": ["Jiong Zhang", "Qi Lei", "Inderjit S. Dhillon"], "authorids": ["zhangjiong724@utexas.edu", "leiqi@ices.utexas.edu", "inderjit@cs.utexas.edu"], "authors_source": "OpenReview API", "abstract": "Vanishing and exploding gradients are two of the main obstacles in training deep neural networks, especially in capturing long range dependencies in recurrent neural networks (RNNs). In this paper, we present an efficient parametrization of the transition matrix of an RNN that allows us to stabilize the gradients that arise in its training. Specifically, we parameterize the transition matrix by its singular value decomposition (SVD), which allows us to explicitly track and control its singular values. We attain efficiency by using tools that are common in numerical linear algebra, namely Householder reflectors for representing the orthogonal matrices that arise in the SVD. By explicitly controlling the singular values, our proposed svdRNN method allows us to easily solve the exploding gradient problem and we observe that it empirically solves the vanishing gradient issue to a large extent. We note that the SVD parameterization can be used for any rectangular weight matrix, hence it can be easily extended to any deep neural network, such as a multi-layer perceptron. Theoretically, we demonstrate that our parameterization does not lose any expressive power, and show how it potentially makes the optimization process easier. Our extensive  experimental results also demonstrate that the proposed framework converges faster, and has good generalization, especially when the depth is large. \n", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=SyL9u-WA-", "arxiv_id": "1803.09327", "paper_pdf": "papers/SyL9u-WA-.pdf", "paper_pdf_sha256": "fcca7d93b35e55d0b202f55be69405cfd6d0b95252b4207da358946bedc07260", "paper_pdf_bytes": 2266663, "paper_pdf_source": "openreview", "code_url": "https://github.com/zhangjiong724/spectral-RNN", "code_repository": "zhangjiong724/spectral-RNN", "code_commit": "c15407a84de5a0fe9244ccbde4de8b351ae18162", "code_archive": "repos/SyL9u-WA-.zip", "code_archive_sha256": "81c35bb8b80e387262f17cecd33f010be4ad9e60a1ffe014440fda53b1029dc7", "code_archive_bytes": 45387, "code_file_count": 22, "code_extensions": {".cc": 12, ".py": 10}, "github_disk_usage_kb": 33, "github_languages": {"C++": 73502, "Python": 52179, "Makefile": 4469}, "github_archived": false, "github_pushed_at": "2018-06-05T22:55:24Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/stabilizing-gradients-for-deep-neural"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "uDmxJ6133n", "year": 2026, "status": "rejected", "title": "TOWARD MEMORY-AIDED WORLD MODELS: BENCHMARKING VIA SPATIAL CONSISTENCY", "authors": ["Kewei Lian", "Shaofei Cai", "Yilun Du", "Yitao Liang"], "authorids": ["~Kewei_Lian1", "~Shaofei_Cai2", "~Yilun_Du1", "~Yitao_Liang1"], "authors_source": "OpenReview API", "abstract": "The ability to simulate the world in a spatially consistent manner is a crucial requirements for effective world models. Such a model enables high-quality visual generation, and also ensures the reliability of world models for downstream tasks such as simulation and planning. Designing a memory module is a crucial component for addressing spatial consistency: such a model must not only retain long-horizon observational information, but also enables the construction of explicit or implicit internal spatial representations. However, there are no dataset designed to promote the development of memory modules by explicitly enforcing spatial consistency constraints. Furthermore, most existing benchmarks primarily emphasize visual coherence or generation quality, neglecting the requirement of long-range spatial consistency. To bridge this gap, we construct a dataset and corresponding benchmark by sampling 150 distinct locations within the open-world environment of Minecraft, collecting about 250 hours (20 million frames) of loop-based navigation videos with actions. Our dataset follows a curriculum design of sequence lengths, allowing models to learn spatial consistency on increasingly complex navigation trajectories. Furthermore, our data collection pipeline is easily extensible to new Minecraft environments and modules. Four representative world model baselines are evaluated on our benchmark. Dataset, benchmark, and code are open-sourced to support future research.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=uDmxJ6133n", "arxiv_id": "2505.22976", "paper_pdf": "papers/uDmxJ6133n.pdf", "paper_pdf_sha256": "41b5769420a7277901ee5195af95f916da274dafbbe6f985c8a651bfa0415920", "paper_pdf_bytes": 18188318, "paper_pdf_source": "openreview", "code_url": "https://github.com/Kevin-lkw/LoopNav", "code_repository": "Kevin-lkw/LoopNav", "code_commit": "d60a052ecf7dd8f7ff176e34f5065ccb0d9cdfc8", "code_archive": "repos/uDmxJ6133n.zip", "code_archive_sha256": "8eb651772287b4f9c0fa4ec70bcbbb20f07990a33e334b1d4b816d1271ac6c82", "code_archive_bytes": 61598, "code_file_count": 17, "code_extensions": {".py": 14, ".js": 3}, "github_disk_usage_kb": 62, "github_languages": {"Python": 157556, "JavaScript": 17112}, "github_archived": false, "github_pushed_at": "2026-05-08T03:56:56Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/toward-memory-aided-world-models-benchmarking"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "o2uHg0Skil", "year": 2025, "status": "rejected", "title": "RL, but don't do anything I wouldn't do", "authors": ["Michael K. Cohen", "Marcus Hutter", "Yoshua Bengio", "Stuart Russell"], "authorids": ["~Michael_K._Cohen1", "~Marcus_Hutter1", "~Yoshua_Bengio1", "~Stuart_Russell1"], "authors_source": "OpenReview API", "abstract": "In reinforcement learning, if the agent's reward differs from the designers' true utility, even only rarely, the state distribution resulting from the agent's policy can be very bad, in theory and in practice. When RL policies would devolve into undesired behavior, a common countermeasure is KL regularization to a trusted policy (\"Don't do anything I wouldn't do\"). All current cutting-edge language models are RL agents that are KL-regularized to a \"base policy\" that is purely predictive. Unfortunately, we demonstrate that when this base policy is a Bayesian predictive model of a trusted policy, the KL constraint is no longer reliable for controlling the behavior of an advanced RL agent. We demonstrate this theoretically using algorithmic information theory, and while systems today are too weak to exhibit this theorized failure precisely, we RL-finetune a language model and find evidence that our formal results are plausibly relevant in practice. We also propose a theoretical alternative that avoids this problem by replacing the \"Don't do anything I wouldn't do\" principle with \"Don't do anything I mightn't do\".", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=o2uHg0Skil", "arxiv_id": "2410.06213", "paper_pdf": "papers/o2uHg0Skil.pdf", "paper_pdf_sha256": "0a6fdb60e3ba466a31a1978c9b7289c81868236a42cf42fcb9339bfbc2bccf81", "paper_pdf_bytes": 1964952, "paper_pdf_source": "openreview", "code_url": "https://github.com/mkc1000/kl-fixed-mixture", "code_repository": "mkc1000/kl-fixed-mixture", "code_commit": "494011e784f97b9501584a9289c0c8428b1dba64", "code_archive": "repos/o2uHg0Skil.zip", "code_archive_sha256": "c845cfdbde0394e436d6de3a211a84d8b241a566717862834d831f3cc151a782", "code_archive_bytes": 5717, "code_file_count": 3, "code_extensions": {".py": 3}, "github_disk_usage_kb": 22, "github_languages": {"Python": 6995}, "github_archived": false, "github_pushed_at": "2024-10-11T05:58:05Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/rl-but-don-t-do-anything-i-wouldn-t-do"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ZA9XUTseA9", "year": 2024, "status": "rejected", "title": "On the Implicit Bias of Adam", "authors": ["Matias D. Cattaneo", "Jason Matthew Klusowski", "Boris Shigida"], "authorids": ["~Matias_D._Cattaneo1", "~Jason_Matthew_Klusowski1", "~Boris_Shigida1"], "authors_source": "OpenReview API", "abstract": "In previous literature, backward error analysis was used to find ordinary differential equations (ODEs) approximating the gradient descent trajectory. It was found that finite step sizes implicitly regularize solutions because terms appearing in the ODEs penalize the two-norm of the loss gradients. We prove that the exis- tence of similar implicit regularization in RMSProp and Adam depends on their hyperparameters and the training stage, but with a different “norm” involved: the corresponding ODE terms either penalize the (perturbed) one-norm of the loss gradients or, on the contrary, hinder its decrease (the latter case being typical). We also conduct numerical experiments and discuss how the proven facts can influence generalization.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=ZA9XUTseA9", "arxiv_id": "2309.00079", "paper_pdf": "papers/ZA9XUTseA9.pdf", "paper_pdf_sha256": "cd10ecbd46a1d312d3e340bca44707eb5e40480e65b84af53a2c48e1a5954bf2", "paper_pdf_bytes": 629521, "paper_pdf_source": "openreview", "code_url": "https://github.com/borshigida/implicit-bias-of-adam", "code_repository": "borshigida/implicit-bias-of-adam", "code_commit": "e25ea9a22085331bc7c2ef7b98752ffc8cf1ac5c", "code_archive": "repos/ZA9XUTseA9.zip", "code_archive_sha256": "18b6ad2865c7045d94589be466a317f049c615549dcc67a4829b20f3f51e1b73", "code_archive_bytes": 12609, "code_file_count": 5, "code_extensions": {".py": 5}, "github_disk_usage_kb": 18, "github_languages": {"Python": 32552}, "github_archived": false, "github_pushed_at": "2024-07-29T13:01:08Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/on-the-implicit-bias-of-adam"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "hTCBqt7pgxf", "year": 2023, "status": "rejected", "title": "Efficient block contrastive learning via parameter-free meta-node approximation", "authors": ["Gayan K Kulatilleke", "Marius Portmann", "Shekhar S. Chandra"], "authorids": ["~Gayan_K_Kulatilleke1", "~Marius_Portmann1", "~Shekhar_S._Chandra1"], "authors_source": "OpenReview API", "abstract": "Contrastive learning has recently achieved remarkable success in many domains including graphs. However contrastive loss, especially for graphs, requires a large number of negative samples which is unscalable and computationally prohibitive with a quadratic time complexity. Sub-sampling is not optimal and incorrect negative sampling leads to sampling bias. In this work, we propose a meta-node based approximation technique that can (a) proxy all negative combinations (b) in quadratic cluster size time complexity, (c) at graph level, not node level, and (d) exploit graph sparsity. By replacing node-pairs with additive cluster-pairs, we compute the negatives in cluster-time at graph level. The resulting Proxy approximated meta-node Contrastive (PamC) loss, based on simple optimized GPU operations, captures the full set of negatives, yet is efficient with a linear time complexity. By avoiding sampling, we effectively eliminate sample bias. We meet the criterion for larger number of samples, thus achieving block-contrastiveness, which is proven to outperform pair-wise losses. We use learnt soft cluster assignments for the meta-node constriction, and avoid possible heterophily and noise added during edge creation. Theoretically, we show that real world graphs easily satisfy conditions necessary for our approximation. Empirically, we show promising accuracy gains over state-of-the-art graph clustering on 6 benchmarks. Importantly, we gain substantially in efficiency; up to 3x in training time, 1.8x in inference time and over 5x in GPU memory reduction.\n", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=hTCBqt7pgxf", "arxiv_id": "2209.14067", "paper_pdf": "papers/hTCBqt7pgxf.pdf", "paper_pdf_sha256": "1faa3aa876f62b796fbe66fb364b6c80dddda2ebac6d7be7e6616e7db8fe4563", "paper_pdf_bytes": 4049478, "paper_pdf_source": "openreview", "code_url": "https://github.com/gayanku/PAMC", "code_repository": "gayanku/PAMC", "code_commit": "f33a8b347b23945194f1425227243fe6b8745ffc", "code_archive": "repos/hTCBqt7pgxf.zip", "code_archive_sha256": "277b9d70042e5802907bd4703d261c6b737c1422c49272547198123618e364c1", "code_archive_bytes": 13311, "code_file_count": 5, "code_extensions": {".py": 5}, "github_disk_usage_kb": 28, "github_languages": {"Python": 49547}, "github_archived": false, "github_pushed_at": "2022-09-29T00:46:41Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/efficient-block-contrastive-learning-via"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "gccdzDu5Ur", "year": 2022, "status": "rejected", "title": "Combining Diverse Feature Priors", "authors": ["Saachi Jain", "Dimitris Tsipras", "Aleksander Madry"], "authorids": ["~Saachi_Jain1", "~Dimitris_Tsipras1", "~Aleksander_Madry1"], "authors_source": "OpenReview API", "abstract": "To improve model generalization, model designers often restrict the features that their models use, either implicitly or explicitly. In this work, we explore the design space of leveraging such feature priors by viewing them as distinct perspectives on the data. Specifically, we find that models trained with diverse sets of explicit feature priors have less overlapping failure modes, and can thus be combined more effectively. Moreover, we demonstrate that jointly training such models on additional (unlabeled) data allows them to correct each other's mistakes, which, in turn, leads to better generalization and resilience to spurious correlations.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=gccdzDu5Ur", "arxiv_id": "2110.08220", "paper_pdf": "papers/gccdzDu5Ur.pdf", "paper_pdf_sha256": "4907d8e26920118f3ea6199aa60ee36ad71a3609225cfe748fe0ef4469163a54", "paper_pdf_bytes": 3702503, "paper_pdf_source": "openreview", "code_url": "https://github.com/MadryLab/copriors", "code_repository": "MadryLab/copriors", "code_commit": "da5146bc3d02e7cfe3e48cde70f0d3b8eac536de", "code_archive": "repos/gccdzDu5Ur.zip", "code_archive_sha256": "32dc2a2c5a37aeb7c6df414973f2c0f9ef88eb60a274ff958919f3ac549a857c", "code_archive_bytes": 24795, "code_file_count": 12, "code_extensions": {".py": 12}, "github_disk_usage_kb": 25, "github_languages": {"Python": 67395}, "github_archived": false, "github_pushed_at": "2021-10-18T01:35:53Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/combining-diverse-feature-priors-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "a5KvtsZ14ev", "year": 2021, "status": "rejected", "title": "SLAPS: Self-Supervision Improves Structure Learning for Graph Neural Networks", "authors": ["Bahare Fatemi", "Seyed Mehran Kazemi", "Layla El Asri"], "authorids": ["~Bahare_Fatemi1", "~Seyed_Mehran_Kazemi1", "~Layla_El_Asri2"], "authors_source": "OpenReview API", "abstract": "Graph neural networks (GNNs) work well when the graph structure is provided. However, this structure may not always be available in real-world applications. One solution to this problem is to infer the latent structure and then apply a GNN to the inferred graph. Unfortunately, the space of possible graph structures grows super-exponentially with the number of nodes and so the available node labels may be insufficient for learning both the structure and the GNN parameters. In this work, we propose the Simultaneous Learning of Adjacency and GNN Parameters with Self-supervision, or SLAPS, a method that provides more supervision for inferring a graph structure. This approach consists of training a denoising autoencoder GNN in parallel with the task-specific GNN. The autoencoder is trained to reconstruct the initial node features given noisy node features as well as a structure provided by a learnable graph generator. We explore the design space of SLAPS by comparing different graph generation and symmetrization approaches. A comprehensive experimental study demonstrates that SLAPS scales to large graphs with hundreds of thousands of nodes and outperforms several models that have been proposed to learn a task-specific graph structure on established benchmarks.", "decision": "Reject", "meta_review": null, "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=a5KvtsZ14ev", "arxiv_id": "2102.05034", "paper_pdf": "papers/a5KvtsZ14ev.pdf", "paper_pdf_sha256": "bb9e2f129bf24419b4ead3ef650b6c585b0bcee101cc9fefd5558f2bc3ffce93", "paper_pdf_bytes": 634660, "paper_pdf_source": "openreview", "code_url": "https://github.com/BorealisAI/SLAPS-GNN", "code_repository": "BorealisAI/SLAPS-GNN", "code_commit": "489481fd3fa1ba0be6b1d42b40acec8b8858b7ec", "code_archive": "repos/a5KvtsZ14ev.zip", "code_archive_sha256": "e11aeddc36869c9f631c3b1f4479aea731b9754a62bb65295b5934113067dd5d", "code_archive_bytes": 22056, "code_file_count": 7, "code_extensions": {".py": 7}, "github_disk_usage_kb": 29, "github_languages": {"Python": 41538}, "github_archived": false, "github_pushed_at": "2021-10-25T22:12:41Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/slaps-self-supervision-improves-structure-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "rkxWpCNKvS", "year": 2020, "status": "rejected", "title": "Improved Image Augmentation for Convolutional Neural Networks by Copyout and CopyPairing", "authors": ["Philip May"], "authorids": ["eniak.info@gmail.com"], "authors_source": "OpenReview API", "abstract": "Image augmentation is a widely used technique to improve the performance of convolutional neural networks (CNNs). In common image shifting, cropping, flipping, shearing and rotating are used for augmentation. But there are more advanced techniques like Cutout and SamplePairing.\n\nIn this work we present two improvements of the state-of-the-art Cutout and SamplePairing techniques. Our new method called Copyout takes a square patch of another random training image and copies it onto a random location of each image used for training. The second technique we discovered is called CopyPairing. It combines Copyout and SamplePairing for further augmentation and even better performance.\n\nWe apply different experiments with these augmentation techniques on the CIFAR-10 dataset to evaluate and compare them under different configurations. In our experiments we show that Copyout reduces the test error rate by 8.18% compared with Cutout and 4.27% compared with SamplePairing. CopyPairing reduces the test error rate by 11.97% compared with Cutout and 8.21% compared with SamplePairing.\n\nCopyout and CopyPairing implementations are available at https://github.com/anonym/anonym.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=rkxWpCNKvS", "arxiv_id": "1909.00390", "paper_pdf": "papers/rkxWpCNKvS.pdf", "paper_pdf_sha256": "ff24017a0a807d407b345a9fc932691e5194fdc0fb7eb2437e3b12665362a49a", "paper_pdf_bytes": 146220, "paper_pdf_source": "openreview", "code_url": "https://github.com/t-systems-on-site-services-gmbh/coocop", "code_repository": "t-systems-on-site-services-gmbh/coocop", "code_commit": "95f844404e22a99cc93a058fa2bd085685dc88c7", "code_archive": "repos/rkxWpCNKvS.zip", "code_archive_sha256": "41be181c2fc2c6ddd2f1d1305314665e4518ae3ab8f12268748bb888e7980987", "code_archive_bytes": 6027, "code_file_count": 5, "code_extensions": {".py": 4, ".sh": 1}, "github_disk_usage_kb": 25, "github_languages": {"Python": 9552, "Shell": 201}, "github_archived": false, "github_pushed_at": "2019-10-11T19:13:12Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/improved-image-augmentation-for-convolutional"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "SJNRHiAcYX", "year": 2019, "status": "rejected", "title": "Boosting Trust Region Policy Optimization by Normalizing flows Policy", "authors": ["Yunhao Tang", "Shipra Agrawal"], "authorids": ["yt2541@columbia.edu", "sa3305@columbia.edu"], "authors_source": "OpenReview API", "abstract": "We propose to improve trust region policy search with normalizing flows policy. We illustrate that when the trust region is constructed by KL divergence constraint, normalizing flows policy can generate samples far from the 'center' of the previous policy iterate, which potentially enables better exploration and helps avoid bad local optima. We show that normalizing flows policy significantly improves upon factorized Gaussian policy baseline, with both TRPO and ACKTR, especially on tasks with complex dynamics such as Humanoid.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=SJNRHiAcYX", "arxiv_id": "1809.10326", "paper_pdf": "papers/SJNRHiAcYX.pdf", "paper_pdf_sha256": "91db1cb255b3600befff0f4db9014ede4425171c84b729f9bcd92e7dc5065c87", "paper_pdf_bytes": 4355503, "paper_pdf_source": "openreview", "code_url": "https://github.com/robintyh1/onpolicybaselines", "code_repository": "robintyh1/onpolicybaselines", "code_commit": "58d401622b38a9127437cf48b1b62b27028d694a", "code_archive": "repos/SJNRHiAcYX.zip", "code_archive_sha256": "c421d56b8f02615e28b2c5eafb81d3ac91f10049375defedaa9c3da8a85b1e19", "code_archive_bytes": 138276, "code_file_count": 61, "code_extensions": {".py": 61}, "github_disk_usage_kb": 56, "github_languages": {"Python": 370267}, "github_archived": false, "github_pushed_at": "2020-04-03T18:54:27Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/boosting-trust-region-policy-optimization-by"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Hk2MHt-3-", "year": 2018, "status": "rejected", "title": "Coupled Ensembles of Neural Networks", "authors": ["Anuvabh Dutt", "Denis Pellerin", "Georges Quénot"], "authorids": ["anuvabh.dutt@univ-grenoble-alpes.fr", "denis.pellerin@gipsa-lab.grenoble-inp.fr", "georges.quenot@imag.fr"], "authors_source": "OpenReview API", "abstract": "We investigate in this paper the architecture of deep convolutional networks. Building on existing state of the art models, we propose a reconfiguration of the model parameters into several parallel branches at the global network level, with each branch being a standalone CNN. We show that this arrangement is an efficient way to significantly reduce the number of parameters while at the same time improving the performance. The use of branches brings an additional form of regularization. In addition to splitting the parameters into parallel branches, we propose a tighter coupling of these branches by averaging their log-probabilities. The tighter coupling favours the learning of better representations, even at the level of the individual branches, as compared to when each branch is trained independently. We refer to this branched architecture as \"coupled ensembles\". The approach is very generic and can be applied with almost any neural network architecture. With coupled ensembles of DenseNet-BC and parameter budget of 25M, we obtain error rates of 2.92%, 15.68% and 1.50% respectively on CIFAR-10, CIFAR-100 and SVHN tasks. For the same parameter budget, DenseNet-BC has an error rate of 3.46%, 17.18%, and 1.8% respectively.  With ensembles of coupled ensembles, of DenseNet-BC networks, with 50M total parameters, we obtain error rates of 2.72%, 15.13% and 1.42% respectively on these tasks.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=Hk2MHt-3-", "arxiv_id": "1709.06053", "paper_pdf": "papers/Hk2MHt-3-.pdf", "paper_pdf_sha256": "e231be8aef025a2c6b1fdec34ece99bb2749ce5048791eb7174ed2e86fd6fe15", "paper_pdf_bytes": 987876, "paper_pdf_source": "openreview", "code_url": "https://github.com/vabh/coupled_ensembles", "code_repository": "vabh/coupled_ensembles", "code_commit": "c4ced6a13189e6f8b5420917509c3873e5ed6fb1", "code_archive": "repos/Hk2MHt-3-.zip", "code_archive_sha256": "2853f7669f0426629bf49825eec6639330c4c1743066495371c117bf7522cf00", "code_archive_bytes": 27692, "code_file_count": 10, "code_extensions": {".py": 9, ".sh": 1}, "github_disk_usage_kb": 35, "github_languages": {"Python": 46710, "Shell": 761}, "github_archived": false, "github_pushed_at": "2019-02-25T12:59:07Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/coupled-ensembles-of-neural-networks"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ZkiVWvWwic", "year": 2026, "status": "rejected", "title": "Dynadiff: Single-stage Decoding of Images from Continuously Evolving fMRI", "authors": ["Marlene Careil", "Yohann Benchetrit", "Jean-Remi King"], "authorids": ["~Marlene_Careil1", "~Yohann_Benchetrit1", "~Jean-Remi_King1"], "authors_source": "OpenReview API", "abstract": "Brain-to-image decoding has been recently propelled by the progress in generative AI models and the availability of large ultra-high field functional Magnetic Resonance Imaging (fMRI). However, current approaches depend on complicated multi-stage pipelines and preprocessing steps that typically collapse the temporal dimension of brain recordings, thereby limiting time-resolved brain decoders. Here, we introduce Dynadiff (Dynamic Neural Activity Diffusion for Image Reconstruction), a new single-stage diffusion model designed for reconstructing images from dynamically evolving fMRI recordings. Our approach offers three main contributions. First, Dynadiff simplifies training as compared to existing approaches. Second, our model outperforms state-of-the-art models on time-resolved fMRI signals, especially on high-level semantic image reconstruction metrics, while remaining competitive on preprocessed fMRI data that collapse time. Third, this approach allows a precise characterization of the evolution of image representations in brain activity.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=ZkiVWvWwic", "arxiv_id": "2505.14556", "paper_pdf": "papers/ZkiVWvWwic.pdf", "paper_pdf_sha256": "b062985363f375d3641bfe3d25cfeb9abbc93dfa97de654d5c6509b1a4f02cf9", "paper_pdf_bytes": 16270608, "paper_pdf_source": "openreview", "code_url": "https://github.com/facebookresearch/dynadiff", "code_repository": "facebookresearch/dynadiff", "code_commit": "b8d1f84a054826c2cea28a44beb2aac5a1839b8c", "code_archive": "repos/ZkiVWvWwic.zip", "code_archive_sha256": "7e328e7c2a86be382d2cdb3c929d8cbc160377d4e65a1c519c930abde59973d4", "code_archive_bytes": 71422, "code_file_count": 16, "code_extensions": {".py": 13, ".sh": 3}, "github_disk_usage_kb": 65, "github_languages": {"Python": 95810, "Shell": 2753}, "github_archived": false, "github_pushed_at": "2025-05-13T21:27:55Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/dynadiff-single-stage-decoding-of-images-from"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "N5ID99rsUq", "year": 2024, "status": "rejected", "title": "Stability and Generalization in Free Adversarial Training", "authors": ["Xiwei Cheng", "Kexin Fu", "Farzan Farnia"], "authorids": ["~Xiwei_Cheng2", "~Kexin_Fu2", "~Farzan_Farnia1"], "authors_source": "OpenReview API", "abstract": "While adversarial training methods have resulted in significant improvements in the deep neural nets' robustness against norm-bounded adversarial perturbations, their generalization performance from training samples to test data has been shown to be considerably worse than standard empirical risk minimization methods. Several recent studies seek to connect the generalization behavior of adversarially trained classifiers to various gradient-based min-max optimization algorithms used for their training. In this work, we study the generalization performance of adversarial training methods using the algorithmic stability framework. Specifically, our goal is to compare the generalization performance of vanilla adversarial training scheme fully optimizing the perturbations at every iteration vs. the free adversarial training simultaneously optimizing the norm-bounded perturbations and classifier parameters. Our proven generalization bounds indicate that the free adversarial training method could enjoy a lower generalization gap between training and test samples due to the simultaneous nature of its min-max optimization algorithm. We perform several numerical experiments to evaluate the generalization performance of vanilla, fast, and free adversarial training methods. Our empirical findings also show the improved generalization performance of the free adversarial training method and further demonstrate that the better generalization result could translate to greater robustness against black-box attack schemes and higher transferability of the adversarial examples designed for free adversarially trained neural networks.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=N5ID99rsUq", "arxiv_id": "2404.08980", "paper_pdf": "papers/N5ID99rsUq.pdf", "paper_pdf_sha256": "245600e9c70dccdcdf487b8ce159d9a0b14950e2e5725971c34cf582e28acf86", "paper_pdf_bytes": 1917096, "paper_pdf_source": "openreview", "code_url": "https://github.com/Xiwei-Cheng/Stability_FreeAT", "code_repository": "Xiwei-Cheng/Stability_FreeAT", "code_commit": "aecc74df99c4874a23d1e3a1b6a7c84b682604a8", "code_archive": "repos/N5ID99rsUq.zip", "code_archive_sha256": "95d6bea8379d5e0b920490bd1414b8e62b8befe85c9c9e0d8b40aee18ba96af7", "code_archive_bytes": 17848, "code_file_count": 7, "code_extensions": {".py": 7}, "github_disk_usage_kb": 19, "github_languages": {"Python": 67532}, "github_archived": false, "github_pushed_at": "2024-04-08T09:28:46Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/stability-and-generalization-in-free"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "gLl0fZQo6Vu", "year": 2023, "status": "rejected", "title": "Agent-Controller Representations: Principled Offline RL with Rich Exogenous Information", "authors": ["Riashat Islam", "Manan Tomar", "Alex Lamb", "Hongyu Zang", "Yonathan Efroni", "Dipendra Misra", "Xin Li", "Harm van Seijen", "Remi Tachet des Combes", "John Langford"], "authorids": ["~Riashat_Islam1", "~Manan_Tomar1", "~Alex_Lamb1", "~Hongyu_Zang1", "~Yonathan_Efroni2", "~Dipendra_Misra1", "~Xin_Li31", "~Harm_van_Seijen1", "~Remi_Tachet_des_Combes1", "~John_Langford1"], "authors_source": "OpenReview API", "abstract": "Learning to control an agent from data collected offline in a rich pixel-based visual observation space is vital for real-world applications of reinforcement learning (RL). A major challenge in this setting is the presence of input information that is hard to model and irrelevant to controlling the agent. This problem has been approached by the theoretical RL community through the lens of exogenous information, i.e, any control-irrelevant information contained in observations. For example, a robot navigating in busy streets needs to ignore irrelevant information, such as other people walking in the background, textures of objects, or birds in the sky. In this paper, we focus on the setting with visually detailed exogenous information, and introduce new offline RL benchmarks offering the ability to study this problem. We find that contemporary representation learning techniques can fail on datasets where the noise is a complex and time dependent process, which is prevalent in practical applications. To address these, we propose to use multi-step inverse models, which have seen a great deal of interest in the RL theory community, to learn Agent-Controller Representations for Offline-RL (ACRO). Despite being simple and requiring no reward, we show theoretically and empirically that the representation created by this objective greatly outperforms baselines.  ", "decision": "Reject", "meta_review": null, "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=gLl0fZQo6Vu", "arxiv_id": "2211.00164", "paper_pdf": "papers/gLl0fZQo6Vu.pdf", "paper_pdf_sha256": "f38065170ab60adab869938b5b866e802caaafefa4b1e60722c8486a605e63db", "paper_pdf_bytes": 18339013, "paper_pdf_source": "openreview", "code_url": "https://github.com/manantomar/agent-centric-representations", "code_repository": "manantomar/agent-centric-representations", "code_commit": "f4622fadcb1fa2b6724dbab21ee62d1ac250bfe4", "code_archive": "repos/gLl0fZQo6Vu.zip", "code_archive_sha256": "aea44ce95fee289f1a7a9434acae24387e48459e73863ead03950895890a7926", "code_archive_bytes": 36760, "code_file_count": 9, "code_extensions": {".py": 9}, "github_disk_usage_kb": 28, "github_languages": {"Python": 86139}, "github_archived": false, "github_pushed_at": "2023-08-14T01:00:34Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/agent-controller-representations-principled"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "MDT30TEtaVY", "year": 2022, "status": "rejected", "title": "Set Norm and Equivariant Skip Connections: Putting the Deep in Deep Sets", "authors": ["Lily H Zhang", "Veronica Tozzo", "John M. Higgins", "Rajesh Ranganath"], "authorids": ["~Lily_H_Zhang1", "~Veronica_Tozzo2", "~John_M._Higgins1", "~Rajesh_Ranganath2"], "authors_source": "OpenReview API", "abstract": "Permutation invariant neural networks are a promising tool for predictive modeling of set data. We show, however, that existing architectures struggle to perform well when they are deep. In this work, we address this issue for the two most widely used permutation invariant networks, Deep Sets and its transformer analogue Set Transformer. We take inspiration from previous efforts to scale neural network architectures by incorporating normalization layers and skip connections that work for sets. First, we motivate and develop set norm, a normalization tailored for sets. Then, we employ equivariant residual connections and introduce the ``clean path principle'' for their placement. With these changes, our many-layer Deep Sets++ and Set Transformer++ models reach comparable or better performance than their original counterparts on a diverse suite of tasks, from point cloud classification to regression on sets of images. We additionally introduce Flow-RBC, a new single-cell dataset and real-world application of permutation invariant prediction. On this task, our new models outperform existing methods as well as a clinical baseline. We open-source our data and code here: link-omitted-for-anonymity.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=MDT30TEtaVY", "arxiv_id": "2206.11925", "paper_pdf": "papers/MDT30TEtaVY.pdf", "paper_pdf_sha256": "e3a789479d8691761dcb0d4e57be72c22934034e3f297dabafceabc6e904ea9e", "paper_pdf_bytes": 461720, "paper_pdf_source": "openreview", "code_url": "https://github.com/rajesh-lab/deep_permutation_invariant", "code_repository": "rajesh-lab/deep_permutation_invariant", "code_commit": "7d25da12329d3d89a69c2f5333f94234248f01a5", "code_archive": "repos/MDT30TEtaVY.zip", "code_archive_sha256": "a9e89578b413ce3fbe4bbac92580dd6b282418379548c6def7601e9949c456dc", "code_archive_bytes": 35988, "code_file_count": 23, "code_extensions": {".py": 23}, "github_disk_usage_kb": 25, "github_languages": {"Python": 98456}, "github_archived": false, "github_pushed_at": "2022-10-11T20:48:51Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/set-norm-and-equivariant-skip-connections-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "jpm1AfJucwt", "year": 2021, "status": "rejected", "title": "Revisiting Loss Modelling for Unstructured Pruning", "authors": ["César Laurent", "Camille Ballas", "Thomas George", "Pascal Vincent", "Nicolas Ballas"], "authorids": ["~César_Laurent1", "~Camille_Ballas1", "~Thomas_George2", "~Pascal_Vincent1", "~Nicolas_Ballas1"], "authors_source": "OpenReview API", "abstract": "By removing parameters from deep neural networks, unstructured pruning methods aim at cutting down memory footprint and computational cost, while maintaining prediction accuracy. In order to tackle this otherwise intractable problem, many of these methods model the loss landscape using first or second order Taylor expansions to identify which parameters can be discarded. We revisit loss modelling for unstructured pruning: we show the importance of ensuring locality of the pruning steps, and systematically compare first and second order Taylor expansions. Finally, we show that better preserving the original network function does not necessarily transfer to better performing networks after fine-tuning, suggesting that only considering the impact of pruning on the loss might not be a sufficient objective to design good pruning criteria.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=jpm1AfJucwt", "arxiv_id": "2006.12279", "paper_pdf": "papers/jpm1AfJucwt.pdf", "paper_pdf_sha256": "fb73c1aabdd4e9dcc51ba19a09b9c9c23079b20ba77f0f70a24a0135ef7a81d2", "paper_pdf_bytes": 1425439, "paper_pdf_source": "openreview", "code_url": "https://github.com/Thrandis/loss-models-pruning", "code_repository": "Thrandis/loss-models-pruning", "code_commit": "b784b84cd2494e59673849dfd3b3e45e996a7e7a", "code_archive": "repos/jpm1AfJucwt.zip", "code_archive_sha256": "b474a4e926a41e698873d8c0deb6a703d8a079f61daaa2c59588c02954d86286", "code_archive_bytes": 14683, "code_file_count": 8, "code_extensions": {".py": 8}, "github_disk_usage_kb": 29, "github_languages": {"Python": 36312, "Dockerfile": 194}, "github_archived": false, "github_pushed_at": "2020-09-17T19:48:09Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/revisiting-loss-modelling-for-unstructured"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "rJggX0EKwS", "year": 2020, "status": "rejected", "title": "The Benefits of Over-parameterization at Initialization in Deep ReLU Networks", "authors": ["Devansh Arpit", "Yoshua Bengio"], "authorids": ["devansharpit@gmail.com", "yoshua.bengio@mila.quebec"], "authors_source": "OpenReview API", "abstract": "It has been noted in existing literature that over-parameterization in ReLU networks generally improves performance. While there could be several factors involved behind this, we prove some desirable theoretical properties at initialization which may be enjoyed by ReLU networks. Specifically, it is known that He initialization in deep ReLU networks asymptotically preserves variance of activations in the forward pass and variance of gradients in the backward pass for infinitely wide networks, thus preserving the flow of information in both directions. Our paper goes beyond these results and shows novel properties that hold under He initialization: i) the norm of hidden activation of each layer is equal to the norm of the input, and, ii) the norm of weight gradient of each layer is equal to the product of norm of the input vector and the error at output layer. These results are derived using the PAC analysis framework, and hold true for finitely sized datasets such that the width of the ReLU network only needs to be larger than a certain finite lower bound. As we show, this lower bound depends on the depth of the network and the number of samples, and by the virtue of being a lower bound, over-parameterized ReLU networks are endowed with these desirable properties. For the aforementioned hidden activation norm property under He initialization, we further extend our theory and show that this property holds for a finite width network even when the number of data samples is infinite. Thus we overcome several limitations of existing papers, and show new properties of deep ReLU networks at initialization.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=rJggX0EKwS", "arxiv_id": "1901.03611", "paper_pdf": "papers/rJggX0EKwS.pdf", "paper_pdf_sha256": "7fb0a3faf914f6f556351d6d4aa8222e6c0053195d847f89ae11fcd60cdb4d8f", "paper_pdf_bytes": 622695, "paper_pdf_source": "openreview", "code_url": "https://github.com/devansharpit/overparametrization_benefits", "code_repository": "devansharpit/overparametrization_benefits", "code_commit": "9711531e87e51712eca8bb625e1a801534ef8676", "code_archive": "repos/rJggX0EKwS.zip", "code_archive_sha256": "c324a1c08ac36e17ffcfb8832943bc1250ace9be4e36553d524828f93d9a6921", "code_archive_bytes": 5091, "code_file_count": 3, "code_extensions": {".py": 3}, "github_disk_usage_kb": 28, "github_languages": {"Python": 6016}, "github_archived": false, "github_pushed_at": "2019-08-20T02:44:58Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/the-benefits-of-over-parameterization-at"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "HJl0jiRqtX", "year": 2019, "status": "rejected", "title": "EDDI: Efficient Dynamic Discovery of High-Value Information with Partial VAE", "authors": ["Chao Ma", "Sebastian Tschiatschek", "Konstantina Palla", "Jose Miguel Hernandez Lobato", "Sebastian Nowozin", "Cheng Zhang"], "authorids": ["cm905@cam.ac.uk", "sebastian.tschiatschek@microsoft.com", "konstantina.palla@microsoft.com", "jmh233@cam.ac.uk", "sebastian.nowozin@microsoft.com", "cheng.zhang@microsoft.com"], "authors_source": "OpenReview API", "abstract": "Making decisions requires information relevant to the task at hand. Many real-life decision-making situations allow acquiring further relevant information at a specific cost. For example, in assessing the health status of a patient we may decide to take additional measurements such as diagnostic tests or imaging scans before making a final assessment. More information that is relevant allows for better decisions but it may be costly to acquire all of this information.  How can we trade off the desire to make good decisions with the option to acquire further information at a cost? To this end, we propose a principled framework, named EDDI (Efficient Dynamic Discovery of high-value Information), based on the theory of Bayesian experimental design. In EDDI we propose a novel partial variational autoencoder (Partial VAE), to efficiently handle missing data over varying subsets of known information. EDDI combines this Partial VAE with an acquisition function that maximizes expected information gain on a set of target variables. EDDI is efficient and demonstrates that dynamic discovery of high-value information is possible; we show cost reduction at the same decision quality and improved decision quality at the same cost in benchmarks and in two health-care applications.. We believe there is great potential for realizing these gains in real-world decision support systems.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=HJl0jiRqtX", "arxiv_id": "1809.11142", "paper_pdf": "papers/HJl0jiRqtX.pdf", "paper_pdf_sha256": "4d9e5643c6d5f5bc86c5b9094b9a748e53f881cb5b97ede8a555e1fcd5ccfee3", "paper_pdf_bytes": 3555859, "paper_pdf_source": "openreview", "code_url": "https://github.com/microsoft/EDDI", "code_repository": "microsoft/EDDI", "code_commit": "835f39e13371a9727b8946e6fa5dcb21eaf28e14", "code_archive": "repos/HJl0jiRqtX.zip", "code_archive_sha256": "bff214ea9fdd435670872fc0a62c9551552fa7bea831480501a68460ca0d9a99", "code_archive_bytes": 38451, "code_file_count": 7, "code_extensions": {".py": 7}, "github_disk_usage_kb": 82, "github_languages": {"Python": 68526}, "github_archived": true, "github_pushed_at": "2023-06-12T18:56:30Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/eddi-efficient-dynamic-discovery-of-high"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "SJzMATlAZ", "year": 2018, "status": "rejected", "title": "Deep Continuous Clustering", "authors": ["Sohil Atul Shah", "Vladlen Koltun"], "authorids": ["sohilas@umd.edu", "vkoltun@gmail.com"], "authors_source": "OpenReview API", "abstract": "Clustering high-dimensional datasets is hard because interpoint distances become less informative in high-dimensional spaces. We present a clustering algorithm that performs nonlinear dimensionality reduction and clustering jointly. The data is embedded into a lower-dimensional space by a deep autoencoder. The autoencoder is optimized as part of the clustering process. The resulting network produces clustered data. The presented approach does not rely on prior knowledge of the number of ground-truth clusters. Joint nonlinear dimensionality reduction and clustering are formulated as optimization of a global continuous objective. We thus avoid discrete reconfigurations of the objective that characterize prior clustering algorithms. Experiments on datasets from multiple domains demonstrate that the presented algorithm outperforms state-of-the-art clustering schemes, including recent methods that use deep networks.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=SJzMATlAZ", "arxiv_id": "1803.01449", "paper_pdf": "papers/SJzMATlAZ.pdf", "paper_pdf_sha256": "f7c1c1ff97a746db6ca181809bf69b5c4c78d4f5abd5d6c2dc2d7494ac699021", "paper_pdf_bytes": 991706, "paper_pdf_source": "openreview", "code_url": "https://github.com/shahsohil/DCC", "code_repository": "shahsohil/DCC", "code_commit": "775e07e17547df2804fd9eead3f1b128d303b577", "code_archive": "repos/SJzMATlAZ.zip", "code_archive_sha256": "74c957409776942d41486910683790491c7a3ec12fb53b4c59bdf5666bc04029", "code_archive_bytes": 35209, "code_file_count": 16, "code_extensions": {".py": 16}, "github_disk_usage_kb": 52, "github_languages": {"Python": 77928}, "github_archived": false, "github_pushed_at": "2021-07-14T09:52:55Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/deep-continuous-clustering"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "PZQHihJlfm", "year": 2026, "status": "rejected", "title": "Next-Scale Autoregressive Models are Zero-Shot Single-Image Object View Synthesizers", "authors": ["Shiran Yuan", "Hao Zhao"], "authorids": ["~Shiran_Yuan1", "~Hao_Zhao1"], "authors_source": "OpenReview API", "abstract": "Learning to synthesize novel views without explicit 3D representations or hand-crafted 3D inductive bias has recently gained attention: it is simpler, more formally direct, and better aligned with the lesson that scalable learning paradigms with less assumptions built into architectural design (e.g., regarding geometry) often win. However, the current dominant solutions are diffusion-based, which typically suffer from problems like slow inference. We introduce ArchonView, the first autoregressive model for zero-shot single-image, object-centric novel view synthesis (NVS), achieving substantially faster inference, higher accuracies, and notably not relying on fine-tuning of 2D generative checkpoints (challenging the common assumption that 2D priors are required in diffusion-based NVS). We design innovative methods of both global and local conditioning to suit characteristics of the NVS task. Crucially, a naïve application of next-scale autoregression fails; we identify two design choices that unlock performance: local conditioning pre-filling, and removing global AdaLN at the classifier head. ArchonView delivers state-of-the-art zero-shot results across six standard benchmarks (GSO, ABO, OmniObject3D, RTMV, NeRF-Synthetic, ShapeNet), while being several times faster than diffusion baselines (e.g., 0.22s v.s. 1.7–1.8s per view at matched parameter count). It consistently improves synthesis accuracy, and scales predictably with both model size (135M–2B) and data size, exhibiting clear scaling-law-like trends. Our findings suggest a paradigm shift and challenge an existing assumption: first, for object-centric NVS, next-scale autoregression can be faster, simpler, and more accurate than diffusion; and second, priors obtained from fine-tuning 2D-pretrained models may not be necessary for generative NVS. Our code is open-sourced at https://anonymous.4open.science/r/ArchonView/.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=PZQHihJlfm", "arxiv_id": "2503.13588", "paper_pdf": "papers/PZQHihJlfm.pdf", "paper_pdf_sha256": "d0ae1551e84c4a5257f6aed353bf338866ade9f622998a2f9f330faee372a698", "paper_pdf_bytes": 4537378, "paper_pdf_source": "openreview", "code_url": "https://github.com/Shiran-Yuan/ArchonView", "code_repository": "Shiran-Yuan/ArchonView", "code_commit": "07be09646242b4625653c9d2ff1ec14330415655", "code_archive": "repos/PZQHihJlfm.zip", "code_archive_sha256": "3616eeed0dee05f406ada0d19b40b12ddc1ba5d2a28fecccb6948fffb737a2c7", "code_archive_bytes": 47624, "code_file_count": 17, "code_extensions": {".py": 17}, "github_disk_usage_kb": 74, "github_languages": {"Python": 141757}, "github_archived": false, "github_pushed_at": "2025-03-19T01:41:30Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/next-scale-autoregressive-models-are-zero"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "3SMBSTG3qN", "year": 2025, "status": "rejected", "title": "Beyond CVaR: Leveraging Static Spectral Risk Measures for Enhanced Decision-Making in Distributional Reinforcement Learning", "authors": ["Mehrdad Moghimi", "Hyejin Ku"], "authorids": ["~Mehrdad_Moghimi1", "~Hyejin_Ku1"], "authors_source": "OpenReview API", "abstract": "In domains such as finance, healthcare, and robotics, managing worst-case scenarios is critical, as failure to do so can lead to catastrophic outcomes. Distributional Reinforcement Learning (DRL) provides a natural framework to incorporate risk sensitivity into decision-making processes. However, existing approaches face two key limitations: (1) the use of fixed risk measures at each decision step often results in overly conservative policies, and (2) the interpretation and theoretical properties of the learned policies remain unclear. While optimizing a static risk measure addresses these issues, its use in the DRL framework has been limited to the simple static CVaR risk measure. In this paper, we present a novel DRL algorithm with convergence guarantees that optimizes for a broader class of static Spectral Risk Measures (SRM). Additionally, we provide a clear interpretation of the learned policy by leveraging the distribution of returns in DRL and the decomposition of static coherent risk measures. Extensive experiments demonstrate that our model learns policies aligned with the SRM objective, and outperforms existing risk-neutral and risk-sensitive DRL models in various settings.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=3SMBSTG3qN", "arxiv_id": "2501.02087", "paper_pdf": "papers/3SMBSTG3qN.pdf", "paper_pdf_sha256": "e907f006291434dc5a3999005872480a65130785f13be089928628de07bfde18", "paper_pdf_bytes": 933625, "paper_pdf_source": "openreview", "code_url": "https://github.com/MehrdadMoghimi/QRSRM", "code_repository": "MehrdadMoghimi/QRSRM", "code_commit": "ed267453bc326f0ff79f4a530480ffc6232906ba", "code_archive": "repos/3SMBSTG3qN.zip", "code_archive_sha256": "0a5dba171dbb8a64b2aa347e50a68c341bae53204c01d8923370e72bffa9ad99", "code_archive_bytes": 36988, "code_file_count": 8, "code_extensions": {".py": 8}, "github_disk_usage_kb": 25, "github_languages": {"Python": 132192}, "github_archived": false, "github_pushed_at": "2025-12-30T04:17:41Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/beyond-cvar-leveraging-static-spectral-risk"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "cSSHiLnjsJ", "year": 2024, "status": "rejected", "title": "Traveling Words: A Geometric Interpretation of Transformers", "authors": ["Raul Molina"], "authorids": ["~Raul_Molina1"], "authors_source": "OpenReview API", "abstract": "Transformers have significantly advanced the field of natural language processing, but comprehending their internal mechanisms remains a challenge. In this paper, we introduce a novel geometric perspective that elucidates the inner mechanisms of transformer operations. Our primary contribution is illustrating how layer normalization confines the latent features to a hyper-sphere, subsequently enabling attention to mold the semantic representation of words on this surface. This geometric viewpoint seamlessly connects established properties such as iterative refinement and contextual embeddings. We validate our insights by probing a pre-trained 124M parameter GPT-2 model. Our findings reveal clear query-key attention patterns in early layers and build upon prior observations regarding the subject-specific nature of attention heads at deeper layers. Harnessing these geometric insights, we present an intuitive understanding of transformers, depicting them as processes that model the trajectory of word particles along the hyper-sphere.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=cSSHiLnjsJ", "arxiv_id": "2309.07315", "paper_pdf": "papers/cSSHiLnjsJ.pdf", "paper_pdf_sha256": "8b7b86a04e6113f4d813ae3107bb40f8e1dcd4980ab7416e88021927f4faae3a", "paper_pdf_bytes": 1715297, "paper_pdf_source": "openreview", "code_url": "https://github.com/santiag0m/traveling-words", "code_repository": "santiag0m/traveling-words", "code_commit": "137aa6090aef2d3d954998bb3557f9e4a0b1ab41", "code_archive": "repos/cSSHiLnjsJ.zip", "code_archive_sha256": "b73afe4e30e50ce5d2e2305b7d07a95b99b6710c3bf67e4354ea2bc611de4084", "code_archive_bytes": 18095, "code_file_count": 9, "code_extensions": {".py": 9}, "github_disk_usage_kb": 20, "github_languages": {"Python": 42512}, "github_archived": false, "github_pushed_at": "2023-12-20T22:13:49Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/traveling-words-a-geometric-interpretation-of"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "1jDN-RfQfrb", "year": 2023, "status": "rejected", "title": "Unveiling Transformers with LEGO: A Synthetic Reasoning Task", "authors": ["Yi Zhang", "Arturs Backurs", "Sebastien Bubeck", "Ronen Eldan", "Suriya Gunasekar", "Tal Wagner"], "authorids": ["~Yi_Zhang1", "~Arturs_Backurs1", "~Sebastien_Bubeck1", "~Ronen_Eldan1", "~Suriya_Gunasekar1", "~Tal_Wagner1"], "authors_source": "OpenReview API", "abstract": "We propose a synthetic reasoning task, LEGO (Learning Equality and Group Operations), that encapsulates the problem of following a chain of reasoning, and we study how the Transformer architectures learn this task. We pay special attention to data effects such as pretraining (on seemingly unrelated NLP tasks) and dataset composition (e.g., differing chain length at training and test time), as well as architectural variants such as weight-tied layers or adding convolutional components. We study how the trained models eventually succeed at the task, and in particular, we are able to understand (to some extent) some of the attention heads as well as how the information flows in the network. Based on these observations we propose a hypothesis that here pretraining helps for LEGO tasks due to certain structured attention patterns, and we experimentally verify this hypothesis. We also observe that in some data regimes the trained transformer finds ``shortcut\" solutions to follow the chain of reasoning, which impedes the model's robustness, and moreover we propose ways to prevent it. Motivated by our findings on structured attention patterns, we propose to replace certain attention heads with hardcoded patterns. This architectural change significantly reduces Flops and maintains or even improves the model's performance at large-scale pretraining.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=1jDN-RfQfrb", "arxiv_id": "2206.04301", "paper_pdf": "papers/1jDN-RfQfrb.pdf", "paper_pdf_sha256": "6ede72542471c625623656f177678287fca4a23c99163a8efcf86005b4cc0601", "paper_pdf_bytes": 1216225, "paper_pdf_source": "openreview", "code_url": "https://github.com/yizhangzzz/transformers-lego", "code_repository": "yizhangzzz/transformers-lego", "code_commit": "6b12f63450155b8e0612891f03a613e1f067563f", "code_archive": "repos/1jDN-RfQfrb.zip", "code_archive_sha256": "a205ef188847f3160719ae836c4a4b2441b249f72223164dd676722c2af2712c", "code_archive_bytes": 13633, "code_file_count": 5, "code_extensions": {".py": 4, ".ipynb": 1}, "github_disk_usage_kb": 28, "github_languages": {"Python": 14011, "Jupyter Notebook": 7693}, "github_archived": false, "github_pushed_at": "2022-12-15T22:16:33Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/unveiling-transformers-with-lego-a-synthetic"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "O9DAoNnYVlM", "year": 2022, "status": "rejected", "title": "Federated Learning via Plurality Vote", "authors": ["Kai Yue", "Richeng Jin", "Chau-Wai Wong", "Huaiyu Dai"], "authorids": ["~Kai_Yue1", "rjin2@ncsu.edu", "~Chau-Wai_Wong1", "~Huaiyu_Dai1"], "authors_source": "OpenReview API", "abstract": "Federated learning allows collaborative workers to solve a machine learning problem while preserving data privacy. Recent studies have tackled various challenges in federated learning, but the joint optimization of communication overhead, learning reliability, and deployment efficiency is still an open problem. To this end, we propose a new scheme named federated learning via plurality vote (FedVote). In each communication round of FedVote, workers transmit binary or ternary weights to the server with low communication overhead. The model parameters are aggregated via weighted voting to enhance the resilience against Byzantine attacks. When deployed for inference, the model with binary or ternary weights is resource-friendly to edge devices. We show that our proposed method can reduce quantization error and converges faster compared with the methods directly quantizing the model updates.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=O9DAoNnYVlM", "arxiv_id": "2110.02998", "paper_pdf": "papers/O9DAoNnYVlM.pdf", "paper_pdf_sha256": "140e2042c9151a91368b6725a90fbaa6bd8c516b2e0b49cc12e54f8598682ee4", "paper_pdf_bytes": 538435, "paper_pdf_source": "openreview", "code_url": "https://github.com/KAI-YUE/fedvote", "code_repository": "KAI-YUE/fedvote", "code_commit": "454b61d14c2e55aebe445cc40929041d89db3edc", "code_archive": "repos/O9DAoNnYVlM.zip", "code_archive_sha256": "31bdf65ccf2915073526e1b5d6d16a424487aaa84c0bcc0dd35069764dd25a0a", "code_archive_bytes": 25996, "code_file_count": 16, "code_extensions": {".py": 16}, "github_disk_usage_kb": 26, "github_languages": {"Python": 58859}, "github_archived": false, "github_pushed_at": "2023-06-01T23:24:49Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/federated-learning-via-plurality-vote"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "-NEXDKk8gZ", "year": 2021, "status": "rejected", "title": "Improved Denoising Diffusion Probabilistic Models", "authors": ["Alexander Quinn Nichol", "Prafulla Dhariwal"], "authorids": ["~Alexander_Quinn_Nichol1", "~Prafulla_Dhariwal1"], "authors_source": "OpenReview API", "abstract": "We explore denoising diffusion probabilistic models, a class of generative models which have recently been shown to produce excellent samples in the image and audio domains. While these models produce excellent samples, it has yet to be shown that they can achieve competitive log-likelihoods. We show that, with several small modifications, diffusion models can achieve competitive log-likelihoods in the image domain while maintaining high sample quality. Additionally, our models allow for sampling with an order of magnitude fewer diffusion steps with only a modest difference in sample quality. Finally, we explore how sample quality and log-likelihood scale with the number of diffusion steps and the amount of model capacity. We conclude that denoising diffusion probabilistic models are a promising class of generative models with excellent scaling properties and sample quality.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=-NEXDKk8gZ", "arxiv_id": "2102.09672", "paper_pdf": "papers/-NEXDKk8gZ.pdf", "paper_pdf_sha256": "fc2553e36d03bed548970db5ce78e17041965ac3e232ddf1b0b461b7e13bc3dc", "paper_pdf_bytes": 8094735, "paper_pdf_source": "openreview", "code_url": "https://github.com/openai/improved-diffusion", "code_repository": "openai/improved-diffusion", "code_commit": "1bc7bbbdc414d83d4abf2ad8cc1446dc36c4e4d5", "code_archive": "repos/-NEXDKk8gZ.zip", "code_archive_sha256": "ca9e7beb876fdeaea50f514df3a33183828e62e9233a179b2dda2a60a2bbff15", "code_archive_bytes": 46777, "code_file_count": 21, "code_extensions": {".py": 21}, "github_disk_usage_kb": 30, "github_languages": {"Python": 129652}, "github_archived": false, "github_pushed_at": "2024-07-18T02:45:29Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/improved-denoising-diffusion-probabilistic-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ryljMpNtwr", "year": 2020, "status": "rejected", "title": "Benchmarking Robustness in Object Detection: Autonomous Driving when Winter is Coming", "authors": ["Claudio Michaelis", "Benjamin Mitzkus", "Robert Geirhos", "Evgenia Rusak", "Oliver Bringmann", "Alexander S. Ecker", "Matthias Bethge", "Wieland Brendel"], "authorids": ["claudio.michaelis@uni-tuebingen.de", "benjamin.mitzkus@uni-tuebingen.de", "robert@geirhos.de", "evgenia.rusak@bethgelab.org", "oliver.bringmann@uni-tuebingen.de", "alexander.ecker@uni-tuebingen.de", "matthias@bethgelab.org", "wieland.brendel@bethgelab.org"], "authors_source": "OpenReview API", "abstract": "The ability to detect objects regardless of image distortions or weather conditions is crucial for real-world applications of deep learning like autonomous driving. We here provide an easy-to-use benchmark to assess how object detection models perform when image quality degrades. The three resulting benchmark datasets, termed PASCAL-C, COCO-C and Cityscapes-C, contain a large variety of image corruptions. We show that a range of standard object detection models suffer a severe performance loss on corrupted images (down to 30-60% of the original performance). However, a simple data augmentation trick - stylizing the training images - leads to a substantial increase in robustness across corruption type, severity and dataset. We envision our comprehensive benchmark to track future progress towards building robust object detection models. Benchmark, code and data are available at: (hidden for double blind review)", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=ryljMpNtwr", "arxiv_id": "1907.07484", "paper_pdf": "papers/ryljMpNtwr.pdf", "paper_pdf_sha256": "1b731982ccdb0e246ea6a9c3a8df69dc5e828644eeb8ada3f6711e30ce1dc2a0", "paper_pdf_bytes": 7695760, "paper_pdf_source": "openreview", "code_url": "https://github.com/bethgelab/stylize-datasets", "code_repository": "bethgelab/stylize-datasets", "code_commit": "71def61edba22bfec690eb461e2bb80a55e326f7", "code_archive": "repos/ryljMpNtwr.zip", "code_archive_sha256": "6a7798366da6adf178fd77033d931a2f0cd7adc2a0509edbf08b1cf6d31798cc", "code_archive_bytes": 8351, "code_file_count": 4, "code_extensions": {".py": 4}, "github_disk_usage_kb": 29, "github_languages": {"Python": 14900}, "github_archived": false, "github_pushed_at": "2021-07-05T17:29:09Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/benchmarking-robustness-in-object-detection"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "SkVRTj0cYQ", "year": 2019, "status": "rejected", "title": "Differentially Private Federated Learning: A Client Level Perspective", "authors": ["Robin C. Geyer", "Tassilo J. Klein", "Moin Nabi"], "authorids": ["geyerr@ethz.ch", "tassilo.klein@sap.com", "moin.nabi@sap.com"], "authors_source": "OpenReview API", "abstract": "Federated learning is a recent advance in privacy protection. \nIn this context, a trusted curator aggregates parameters optimized in decentralized fashion by multiple clients. The resulting model is then distributed back to all clients, ultimately converging to a joint representative model without explicitly having to share the data. \nHowever, the protocol is vulnerable to differential attacks, which could originate from any party contributing during federated optimization. In such an attack, a client's contribution during training and information about their data set is revealed through analyzing the distributed model. \nWe tackle this problem and propose an algorithm for client sided differential privacy preserving federated optimization. The aim is to hide clients' contributions during training, balancing the trade-off between privacy loss and model performance. \nEmpirical studies suggest that given a sufficiently large number of participating clients, our proposed procedure can maintain client-level differential privacy at only a minor cost in model performance. ", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=SkVRTj0cYQ", "arxiv_id": "1712.07557", "paper_pdf": "papers/SkVRTj0cYQ.pdf", "paper_pdf_sha256": "d4c626e4614a1e561a6c9911687592fb0cd1beb3917d623357cbe5d3eceb842c", "paper_pdf_bytes": 615031, "paper_pdf_source": "openreview", "code_url": "https://github.com/SAP-archive/machine-learning-diff-private-federated-learning", "code_repository": "SAP-archive/machine-learning-diff-private-federated-learning", "code_commit": "915c2fa63751403cbfcb910c101eb95377f65f19", "code_archive": "repos/SkVRTj0cYQ.zip", "code_archive_sha256": "c3f25c1b799b20511d0392678d0fa49deddc0861c11583bbd352dc88a2d1818d", "code_archive_bytes": 45061, "code_file_count": 10, "code_extensions": {".py": 9, ".sh": 1}, "github_disk_usage_kb": 85, "github_languages": {"Python": 89358, "Shell": 1104}, "github_archived": true, "github_pushed_at": "2025-03-07T12:53:08Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/differentially-private-federated-learning-a"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "HknbyQbC-", "year": 2018, "status": "rejected", "title": "Generating Adversarial Examples with Adversarial Networks", "authors": ["Chaowei Xiao", "Bo Li", "Jun-Yan Zhu", "Warren He", "Mingyan Liu", "Dawn Song"], "authorids": ["xiaocw@umich.edu", "lxbosky@gmail.com", "junyanz@berkeley.edu", "_w@eecs.berkeley.edu", "mingyan@umich.edu", "dawnsong.travel@gmail.com"], "authors_source": "OpenReview API", "abstract": "Deep neural networks (DNNs) have been found to be vulnerable to adversarial examples resulting from adding small-magnitude perturbations to inputs. Such adversarial examples can mislead DNNs to produce adversary-selected results.\nDifferent attack strategies have been proposed to generate adversarial examples, but how to produce them with high perceptual quality and more efficiently requires more research efforts. \nIn this paper, we propose AdvGAN to generate adversarial examples with generative adversarial networks (GANs), which can learn and approximate the distribution of original instances. \nFor AdvGAN, once the generator is trained, it can generate adversarial perturbations efficiently for any instance, so as to potentially accelerate adversarial training as defenses.  \nWe apply AdvGAN in both semi-whitebox and black-box attack settings. In semi-whitebox attacks, there is no need to access the original target model after the generator is trained, in contrast to traditional white-box attacks. In black-box attacks, we dynamically train a distilled model for the black-box model and optimize the generator accordingly.\nAdversarial examples generated by AdvGAN on different target models have high attack success rate under state-of-the-art defenses compared to other attacks. Our attack  has placed the first with 92.76% accuracy on a public MNIST black-box attack challenge. ", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=HknbyQbC-", "arxiv_id": "1801.02610", "paper_pdf": "papers/HknbyQbC-.pdf", "paper_pdf_sha256": "01826ca59d53cf9ea824554bb01a1c58a1323299e41066502d8813c45472ede3", "paper_pdf_bytes": 8716232, "paper_pdf_source": "openreview", "code_url": "https://github.com/MadryLab/mnist_challenge", "code_repository": "MadryLab/mnist_challenge", "code_commit": "3ee3643c4a8c59458d8c191b84027f4a6cbd9580", "code_archive": "repos/HknbyQbC-.zip", "code_archive_sha256": "024f20d6670726da3170f4af9a1ff05e79c1f98296c97e711840c29ee86cfe4f", "code_archive_bytes": 15328, "code_file_count": 7, "code_extensions": {".py": 7}, "github_disk_usage_kb": 57, "github_languages": {"Python": 21165}, "github_archived": false, "github_pushed_at": "2022-05-03T17:21:40Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/generating-adversarial-examples-with"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "1KUMxnrhnH", "year": 2026, "status": "rejected", "title": "A Separable Self-attention Inspired by the State Space Model for Computer Vision", "authors": ["Juntao Zhang", "Jun Zhou", "Kun Bian", "zhou you", "Jianning Liu"], "authorids": ["~Juntao_Zhang1", "~Jun_Zhou16", "~Kun_Bian1", "~zhou_you1", "~Jianning_Liu1"], "authors_source": "OpenReview API", "abstract": "Separable self-attention is an early attention mechanism with linear complexity. When parameters and FLOPs are comparable, lightweight networks built upon separable self-attention and its variants underperform the recent Vision Mamba (ViM). By analyzing the strengths and weaknesses of separable self-attention, we distill four design principles and, inspired by the State Space Model (SSM) serving as the core of ViM, propose a novel separable self-attention termed Vision Mamba Inspired Separable self-Attention (VMI-SA). Notably, VMI-SA does not incorporate any SSM blocks, and its attention computation process differs from all existing attention mechanisms to the best of our knowledge. We introduce proof-of-concept networks, VMINet and VMIFormer, enabling fair comparisons with ViMs through deliberate control of parameters, FLOPs, and encoder numbers. Compared to state-of-the-art Transformers, CNNs, and ViMs, VMINet and VMIFormer achieve competitive results in image classification and high-resolution dense prediction tasks.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=1KUMxnrhnH", "arxiv_id": "2501.02040", "paper_pdf": "papers/1KUMxnrhnH.pdf", "paper_pdf_sha256": "59e664034feb379394158560b5b8499353969dd6839b8700b1316067d387b9b2", "paper_pdf_bytes": 821704, "paper_pdf_source": "openreview", "code_url": "https://github.com/yws-wxs/VMINet", "code_repository": "yws-wxs/VMINet", "code_commit": "740b9beda36168983776c5bc0ef638539e077160", "code_archive": "repos/1KUMxnrhnH.zip", "code_archive_sha256": "44b31951bf4751c63594c4816702cd4cda2e8d66c48b2c9e6dec344810110525", "code_archive_bytes": 36014, "code_file_count": 4, "code_extensions": {".py": 4}, "github_disk_usage_kb": 85, "github_languages": {"Python": 75236}, "github_archived": false, "github_pushed_at": "2025-04-22T11:28:23Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/a-separable-self-attention-inspired-by-the"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "7S1xDos9pH", "year": 2025, "status": "rejected", "title": "Generalized Gaussian Temporal Difference Error for Uncertainty-aware Reinforcement Learning", "authors": ["Seyeon Kim", "Joonhun Lee", "Namhoon Cho", "Sungjun Han", "Wooseop Hwang"], "authorids": ["~Seyeon_Kim5", "~Joonhun_Lee1", "~Namhoon_Cho2", "~Sungjun_Han2", "~Wooseop_Hwang1"], "authors_source": "OpenReview API", "abstract": "Conventional uncertainty-aware temporal difference (TD) learning methods often rely on simplistic assumptions, typically including a zero-mean Gaussian distribution for TD errors. Such oversimplification can lead to inaccurate error representations and compromised uncertainty estimation. In this paper, we introduce a novel framework for generalized Gaussian error modeling in deep reinforcement learning, applicable to both discrete and continuous control settings. Our framework enhances the flexibility of error distribution modeling by incorporating additional higher-order moment, particularly kurtosis, thereby improving the estimation and mitigation of data-dependent noise, i.e., aleatoric uncertainty. We examine the influence of the shape parameter of the generalized Gaussian distribution (GGD) on aleatoric uncertainty and provide a closed-form expression that demonstrates an inverse relationship between uncertainty and the shape parameter. Additionally, we propose a theoretically grounded weighting scheme to fully leverage the GGD. To address epistemic uncertainty, we enhance the batch inverse variance weighting by incorporating bias reduction and kurtosis considerations, resulting in improved robustness. Extensive experimental evaluations using policy gradient algorithms demonstrate the consistent efficacy of our method, showcasing significant performance improvements.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=7S1xDos9pH", "arxiv_id": "2408.02295", "paper_pdf": "papers/7S1xDos9pH.pdf", "paper_pdf_sha256": "386b26c58b03fe5decfb4d3fd288d447649bb8057edbcbee76b2b98776c791a2", "paper_pdf_bytes": 16992959, "paper_pdf_source": "openreview", "code_url": "https://github.com/ait-lab/ggtde", "code_repository": "ait-lab/ggtde", "code_commit": "c2881d8d669469b2490e47b679c088b67f52bb6f", "code_archive": "repos/7S1xDos9pH.zip", "code_archive_sha256": "bbb7a95076fe2ae99002f239eafb97aca84ccb7cdea9e8b211ca8572102bc8e7", "code_archive_bytes": 33588, "code_file_count": 27, "code_extensions": {".py": 27}, "github_disk_usage_kb": 27, "github_languages": {"Python": 88286}, "github_archived": false, "github_pushed_at": "2026-07-23T08:29:48Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/generalized-gaussian-temporal-difference"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "9FXGX00iMF", "year": 2024, "status": "rejected", "title": "BWS: Best Window Selection Based on Sample Scores for Data Pruning across Broad Ranges", "authors": ["Hoyong Choi", "Nohyun Ki", "Hye Won Chung"], "authorids": ["~Hoyong_Choi1", "~Nohyun_Ki1", "~Hye_Won_Chung2"], "authors_source": "OpenReview API", "abstract": "Data subset selection aims to find a smaller yet informative subset of a large dataset that can approximate the full-dataset training, addressing challenges associated with training neural networks on large-scale datasets. However, existing methods tend to specialize in either high or low selection ratio regimes, lacking a universal approach that consistently achieves competitive performance across a broad range of selection ratios. We introduce a universal and efficient data subset selection method, Best Window Selection (BWS), by proposing a method to choose the best window subset from samples ordered based on their difficulty scores. This approach offers flexibility by allowing the choice of window intervals that span from easy to difficult samples. Furthermore, we provide an efficient mechanism for selecting the best window subset by evaluating its quality using kernel ridge regression. Our experimental results demonstrate the superior performance of BWS compared to other baselines across a broad range of selection ratios over datasets, including CIFAR-10/100 and ImageNet, and the scenarios involving training from random initialization or fine-tuning of pre-trained models.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=9FXGX00iMF", "arxiv_id": "2406.03057", "paper_pdf": "papers/9FXGX00iMF.pdf", "paper_pdf_sha256": "0da5ce530cf3a945c7fa30a003acffe8ad9b37bbe424bd786d4af5af01e55ab7", "paper_pdf_bytes": 1927725, "paper_pdf_source": "openreview", "code_url": "https://github.com/NohyunKi/BWS", "code_repository": "NohyunKi/BWS", "code_commit": "be73a6f64c57ebc12fb0ff57320cde4a82d6c47f", "code_archive": "repos/9FXGX00iMF.zip", "code_archive_sha256": "5b11e6606f2a3a33a8cb988025519dfbbfa74f7229caaf50bea8a1a31ffd2af0", "code_archive_bytes": 31806, "code_file_count": 20, "code_extensions": {".py": 16, ".sh": 4}, "github_disk_usage_kb": 23, "github_languages": {"Python": 102610, "Shell": 6051}, "github_archived": false, "github_pushed_at": "2024-05-23T15:48:39Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/bws-best-window-selection-based-on-sample"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "B7HJ9KLFV9U", "year": 2023, "status": "rejected", "title": "Thinking Two Moves Ahead: Anticipating Other Users Improves Backdoor Attacks in Federated Learning", "authors": ["Yuxin Wen", "Jonas Geiping", "Liam H Fowl", "Hossein Souri", "Rama Chellappa", "Micah Goldblum", "Tom Goldstein"], "authorids": ["~Yuxin_Wen2", "~Jonas_Geiping1", "~Liam_H_Fowl1", "~Hossein_Souri1", "~Rama_Chellappa1", "~Micah_Goldblum1", "~Tom_Goldstein1"], "authors_source": "OpenReview API", "abstract": "Federated learning is particularly susceptible to model poisoning and backdoor attacks because individual users have direct control over the training data and model updates.  At the same time, the attack power of an individual user is limited because their updates are quickly drowned out by those of many other users. Existing attacks do not account for future behaviors of other users, and thus require many sequential updates and their effects are quickly erased. We propose an attack that anticipates and accounts for the entire federated learning pipeline, including behaviors of other clients, and ensures that backdoors are effective quickly and persist even after multiple rounds of community updates. We show that this new attack is effective in realistic scenarios where the attacker only contributes to a small fraction of randomly sampled rounds and demonstrate this attack on image classification, next-word prediction, and sentiment analysis.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=B7HJ9KLFV9U", "arxiv_id": "2210.09305", "paper_pdf": "papers/B7HJ9KLFV9U.pdf", "paper_pdf_sha256": "6c38085c5a6c64e5bb66eca85ef429e62c799332d52a72ae6099faf74ac743b2", "paper_pdf_bytes": 1590399, "paper_pdf_source": "openreview", "code_url": "https://github.com/YuxinWenRick/thinking-two-moves-ahead", "code_repository": "YuxinWenRick/thinking-two-moves-ahead", "code_commit": "091f08b2af80e3454eb11fb0207cdba77a2fd902", "code_archive": "repos/B7HJ9KLFV9U.zip", "code_archive_sha256": "0fb3a04665df509fce3d40c7f9d46003ffb5d3c96a54f4b3804a1793a0a91f99", "code_archive_bytes": 35960, "code_file_count": 21, "code_extensions": {".py": 21}, "github_disk_usage_kb": 28, "github_languages": {"Python": 101761}, "github_archived": false, "github_pushed_at": "2022-10-18T01:02:31Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/thinking-two-moves-ahead-anticipating-other"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "YDqIYJBQTQs", "year": 2022, "status": "rejected", "title": "Unsupervised Object Learning via Common Fate", "authors": ["Matthias Tangemann", "Steffen Schneider", "Julius Von Kügelgen", "Francesco Locatello", "Peter Vincent Gehler", "Thomas Brox", "Matthias Kuemmerer", "Matthias Bethge", "Bernhard Schölkopf"], "authorids": ["~Matthias_Tangemann1", "~Steffen_Schneider1", "~Julius_Von_Kügelgen1", "~Francesco_Locatello1", "~Peter_Vincent_Gehler1", "~Thomas_Brox1", "~Matthias_Kuemmerer1", "~Matthias_Bethge1", "~Bernhard_Schölkopf1"], "authors_source": "OpenReview API", "abstract": "Learning generative object models from unlabelled videos is a long standing problem and is required for causal scene modeling. We decompose this problem into three easier subtasks, and provide candidate solutions for each of them. Inspired by the Common Fate Principle of Gestalt Psychology, we first extract (noisy) masks of moving objects via unsupervised motion segmentation. Second, generative models are trained on the masks of the background and the moving objects, respectively. Third, background and foreground models are combined in a conditional ``dead leaves scene model to sample novel scene configurations where occlusions and depth layering  arise naturally. To evaluate the individual stages, we introduce the Fishbowl dataset positioned between complex real-world scenes and common object-centric benchmarks of simplistic objects. We show that our approach allows learning generative models that generalize beyond the occlusions present in the input videos, and represent scenes in a modular fashion that allows sampling plausible scenes outside the training distribution by permitting, for instance, object numbers or densities not observed in the training set.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=YDqIYJBQTQs", "arxiv_id": "2110.06562", "paper_pdf": "papers/YDqIYJBQTQs.pdf", "paper_pdf_sha256": "1f85961f59d9f5cac26cda30e6b2999bc30f1c7d73f183869378cdbb70cee170", "paper_pdf_bytes": 19945939, "paper_pdf_source": "openreview", "code_url": "https://github.com/mtangemann/common_fate_object_learning", "code_repository": "mtangemann/common_fate_object_learning", "code_commit": "9c14ac4aa151504e7d9f0a7be1810fee506eadf2", "code_archive": "repos/YDqIYJBQTQs.zip", "code_archive_sha256": "a612f84fe86220ca37621b3e9b1cd2af0efdd0a98b882ed76b53abd6a287c2de", "code_archive_bytes": 30770, "code_file_count": 8, "code_extensions": {".py": 8}, "github_disk_usage_kb": 26, "github_languages": {"Python": 74727}, "github_archived": false, "github_pushed_at": "2023-04-14T07:05:02Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/unsupervised-object-learning-via-common-fate-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Svfh1_hYEtF", "year": 2021, "status": "rejected", "title": "Federated Continual Learning with Weighted Inter-client Transfer", "authors": ["Jaehong Yoon", "Wonyong Jeong", "Giwoong Lee", "Eunho Yang", "Sung Ju Hwang"], "authorids": ["~Jaehong_Yoon1", "~Wonyong_Jeong1", "~Giwoong_Lee1", "~Eunho_Yang1", "~Sung_Ju_Hwang1"], "authors_source": "OpenReview API", "abstract": "There has been a surge of interest in continual learning and federated learning, both of which are important in deep neural networks in real-world scenarios. Yet little research has been done regarding the scenario where each client learns on a sequence of tasks from a private local data stream.  This problem of federated continual learning poses new challenges to continual learning, such as utilizing knowledge from other clients, while preventing interference from irrelevant knowledge.  To resolve these issues, we propose a novel federated continual learning framework, Federated Weighted Inter-client Transfer (FedWeIT), which decomposes the network weights into global federated parameters and sparse task-specific parameters, and each client receives selective knowledge from other clients by taking a weighted combination of their task-specific parameters.FedWeITminimizes interference between incompatible tasks, and also allows positive knowledge transfer across clients during learning. We validate ourFedWeITagainst existing federated learning and continual learning methods under varying degrees of task similarity across clients, and our model significantly outperforms them with a large reduction in the communication cost.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=Svfh1_hYEtF", "arxiv_id": "2003.03196", "paper_pdf": "papers/Svfh1_hYEtF.pdf", "paper_pdf_sha256": "82e13bf797edce2ab40efdf4256086e031b35ff223468c09ab46f50b7c5166cd", "paper_pdf_bytes": 2101073, "paper_pdf_source": "openreview", "code_url": "https://github.com/wyjeong/FedWeIT", "code_repository": "wyjeong/FedWeIT", "code_commit": "83c0bcd10b4934abe3af3a9b6d2e5b77566b9567", "code_archive": "repos/Svfh1_hYEtF.zip", "code_archive_sha256": "7fcd6881ceb2f2209f35dc0127c326927fdb63bfafe6eccbb1d353d4fae5851a", "code_archive_bytes": 29425, "code_file_count": 15, "code_extensions": {".py": 13, ".sh": 2}, "github_disk_usage_kb": 30, "github_languages": {"Python": 66487, "Shell": 770}, "github_archived": false, "github_pushed_at": "2021-06-02T05:12:26Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/federated-continual-learning-with-adaptive"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "rJe9fTNtPS", "year": 2020, "status": "rejected", "title": "AHash: A Load-Balanced One Permutation Hash", "authors": ["Chenxingyu Zhao", "Jie Gui", "Yixiao Guo", "Jie Jiang", "Tong Yang", "Bin Cui", "Gong Zhang"], "authorids": ["dkzcxy@pku.edu.cn", "guisj2017@pku.edu.cn", "1700016637@pku.edu.cn", "jie.jiang@pku.edu.cn", "yangtongemail@gmail.com", "bin.cui@pku.edu.cn", "nicholas.zhang@huawei.com"], "authors_source": "OpenReview API", "abstract": "Minwise Hashing (MinHash) is a fundamental method to compute set similarities and compact high-dimensional data for efficient learning and searching. The bottleneck of MinHash is computing k (usually hundreds) MinHash values. One Permutation Hashing (OPH) only requires one permutation (hash function)  to get k MinHash values by dividing elements into k bins.  One drawback of OPH is that the load of the bins (the number of elements in a bin) could be unbalanced, which leads to the existence of empty bins and false similarity computation. Several strategies for densification, that is, filling empty bins, have been proposed. However, the densification is just a remedial strategy and cannot eliminate the error incurred by the unbalanced load. Unlike the densification to fill the empty bins after they undesirably occur, our design goal is to balance the load so as to reduce the empty bins in advance.  In this paper, we propose a load-balanced hashing, Amortization Hashing (AHash), which can generate as few empty bins as possible. Therefore, AHash is more load-balanced and accurate without hurting runtime efficiency compared with OPH and densification strategies. Our experiments on real datasets validate the claim. All source codes and datasets have been provided as Supplementary Materials and released on GitHub anonymously.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=rJe9fTNtPS", "arxiv_id": null, "paper_pdf": "papers/rJe9fTNtPS.pdf", "paper_pdf_sha256": "5ddec72d536b9df6913d1d2a641b3dd1eea670d01b7082c35df378e8c83bd73b", "paper_pdf_bytes": 997501, "paper_pdf_source": "openreview", "code_url": "https://github.com/AHashCodes/AHash", "code_repository": "AHashCodes/AHash", "code_commit": "6d3fcaedb10ae0f03b28f3e6127f3cdc6db5c463", "code_archive": "repos/rJe9fTNtPS.zip", "code_archive_sha256": "adc1d7c36b0f78176a096c288a66daf12a0fd08a1291bfe3b315a55e238b267e", "code_archive_bytes": 37242, "code_file_count": 22, "code_extensions": {".h": 16, ".cpp": 4, ".sh": 2}, "github_disk_usage_kb": 34, "github_languages": {"C++": 85815, "Shell": 1902, "Makefile": 393, "CMake": 374}, "github_archived": false, "github_pushed_at": "2019-06-08T12:09:00Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/ahash-a-load-balanced-one-permutation-hash"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "HyfyN30qt7", "year": 2019, "status": "rejected", "title": "NICE: noise injection and clamping estimation for neural network quantization", "authors": ["Chaim Baskin", "Natan Liss", "Yoav Chai", "Evgenii Zheltonozhskii", "Eli Schwartz", "Raja Girayes", "Avi Mendelson", "Alexander M.Bronstein"], "authorids": ["chaimbaskin@cs.technion.ac.il", "lissnatan@campus.technion.ac.il", "yoavchai1@mail.tau.ac.il", "evgeniizh@campus.technion.ac.il", "eli.shw@gmail.com", "raja@tauex.tau.ac.il", "avi.mendelson@tce.technion.ac.il", "bron@cs.technion.ac.il"], "authors_source": "OpenReview API", "abstract": "Convolutional Neural Networks (CNN) are very popular in many fields including computer vision, speech recognition, natural language processing, to name a few. Though deep learning leads to groundbreaking performance in these domains, the networks used are very demanding computationally and are far from real-time even on a GPU, which is not power efficient and therefore does not suit low power systems such as mobile devices. To overcome this challenge, some solutions have been proposed for quantizing the weights and activations of these networks, which accelerate the runtime significantly. Yet, this acceleration comes at the cost of a larger error. The NICE method proposed in this work trains quantized neural networks by noise injection and a learned clamping, which improve the accuracy. This leads to state-of-the-art results on various regression and classification tasks, e.g., ImageNet classification with architectures such as ResNet-18/34/50 with low as 3-bit weights and 3 -bit activations. We implement the proposed solution on an FPGA to demonstrate its applicability for low power real-time applications.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=HyfyN30qt7", "arxiv_id": "1810.00162", "paper_pdf": "papers/HyfyN30qt7.pdf", "paper_pdf_sha256": "30eef30e6bed8b39487d0e58273ad09bd345b1cd5322d7212a4db197ca032ab2", "paper_pdf_bytes": 336064, "paper_pdf_source": "openreview", "code_url": "https://github.com/LsNatan/NICE", "code_repository": "LsNatan/NICE", "code_commit": "1f5ef43c9442931dd26e05a0dd68e6602f28f569", "code_archive": "repos/HyfyN30qt7.zip", "code_archive_sha256": "7e41888fb55c7235eec2a40649c05afbef40b58ecc971157b49abf523694dede", "code_archive_bytes": 97409, "code_file_count": 31, "code_extensions": {".py": 31}, "github_disk_usage_kb": 89, "github_languages": {"Python": 246385}, "github_archived": false, "github_pushed_at": "2018-10-01T11:15:23Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/nice-noise-injection-and-clamping-estimation"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "HkepKG-Rb", "year": 2018, "status": "rejected", "title": "A Semantic Loss Function for Deep Learning with Symbolic Knowledge", "authors": ["Jingyi Xu", "Zilu Zhang", "Tal Friedman", "Yitao Liang", "Guy Van den Broeck"], "authorids": ["jixu@g.ucla.edu", "zhangzilu@pku.edu.cn", "tal@cs.ucla.edu", "yliang@cs.ucla.edu", "guyvdb@cs.ucla.edu"], "authors_source": "OpenReview API", "abstract": "This paper develops a novel methodology for using symbolic knowledge in deep learning. From first principles, we derive a semantic loss function that bridges between neural output vectors and logical constraints. This loss function captures how close the neural network is to satisfying the constraints on its output. An experimental evaluation shows that our semantic loss function effectively guides the learner to achieve (near-)state-of-the-art results on semi-supervised multi-class classification. Moreover, it significantly increases the ability of the neural network to predict structured objects, such as rankings and shortest paths. These discrete concepts are tremendously difficult to learn, and benefit from a tight integration of deep learning and symbolic reasoning methods.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=HkepKG-Rb", "arxiv_id": "1711.11157", "paper_pdf": "papers/HkepKG-Rb.pdf", "paper_pdf_sha256": "aa75cbe8d14bacab4747aae480f95c1794c40bb291d07b52c5d21a7c264996d3", "paper_pdf_bytes": 345574, "paper_pdf_source": "openreview", "code_url": "https://github.com/UCLA-StarAI/Semantic-Loss", "code_repository": "UCLA-StarAI/Semantic-Loss", "code_commit": "86764c2294b89fc36fac8979e46bb0a6e062f1ec", "code_archive": "repos/HkepKG-Rb.zip", "code_archive_sha256": "77d89c6a94c73a90bb0ab2ba00f4fb4aeaf396929234f220e0fe22558c95f72b", "code_archive_bytes": 82493, "code_file_count": 8, "code_extensions": {".py": 8}, "github_disk_usage_kb": 86, "github_languages": {"Python": 48636}, "github_archived": false, "github_pushed_at": "2019-05-01T22:29:17Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/a-semantic-loss-function-for-deep-learning"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "nEcs0Duf86", "year": 2026, "status": "rejected", "title": "From Invariant Representations to Invariant Data: Provable Robustness to Spurious Correlations via Noisy Counterfactual Matching", "authors": ["Ruqi Bai", "Yao Ji", "Zeyu Zhou", "David I. Inouye"], "authorids": ["~Ruqi_Bai1", "~Yao_Ji1", "~Zeyu_Zhou1", "~David_I._Inouye1"], "authors_source": "OpenReview API", "abstract": "Models that learn spurious correlations from training data often fail when deployed in new environments. While many methods aim to learn invariant representations to address this, they often underperform standard empirical risk minimization (ERM). We propose a data-centric alternative that shifts the focus from learning invariant representations to leveraging invariant data pairs---pairs of samples that should have the same prediction. We prove that certain counterfactuals naturally satisfy this invariance property. Based on this, we introduce Noisy Counterfactual Matching (NCM), a simple constraint-based method that improves robustness by leveraging even a small number of \\emph{noisy} counterfactual pairs---improving upon prior works that do not explicitly consider noise. For linear causal models, we prove that NCM's test-domain error is bounded by its in-domain error plus a term dependent on the counterfactuals' quality and diversity. Experiments on synthetic data validate our theory, and we demonstrate NCM's effectiveness on real-world datasets.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=nEcs0Duf86", "arxiv_id": "2505.24843", "paper_pdf": "papers/nEcs0Duf86.pdf", "paper_pdf_sha256": "f6adf367f49b825ae7d97327031ae2108828d79ee1f34c1d47f11a0b78889f70", "paper_pdf_bytes": 2075695, "paper_pdf_source": "openreview", "code_url": "https://github.com/inouye-lab/GRIT", "code_repository": "inouye-lab/GRIT", "code_commit": "da418a4ab8fc91403267ff44fc7369a51fed6d41", "code_archive": "repos/nEcs0Duf86.zip", "code_archive_sha256": "7a82ff7d55555ad9fbd93c7e55e6e53182239f94f1c42d7a249e018f875be632", "code_archive_bytes": 134396, "code_file_count": 92, "code_extensions": {".py": 92}, "github_disk_usage_kb": 86, "github_languages": {"Python": 174170}, "github_archived": false, "github_pushed_at": "2026-04-13T13:59:50Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/from-invariant-representations-to-invariant"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "lK0WxHeups", "year": 2024, "status": "rejected", "title": "Iteration and Stochastic First-order Oracle Complexities of Stochastic Gradient Descent using Constant and Decaying Learning Rates", "authors": ["Kento Imaizumi", "Hideaki Iiduka"], "authorids": ["~Kento_Imaizumi1", "~Hideaki_Iiduka1"], "authors_source": "OpenReview API", "abstract": "The performance of stochastic gradient descent (SGD), which is the simplest first-order optimizer for training deep neural networks, depends on not only the learning rate but also the batch size. They both affect the number of iterations and the stochastic first-order oracle (SFO) complexity needed for training. In particular, the previous numerical results indicated that, for SGD using a constant learning rate, the number of iterations needed for training decreases when the batch size increases, and the SFO complexity needed for training is minimized at a critical batch size and increases once the batch size exceeds that size. This paper studies the relationship between batch size and the iteration and the SFO complexities needed for nonconvex optimization in deep learning with SGD using constant/decay learning rates. We show that SGD using a step-decay learning rate and a small batch size reduces the SFO complexity to find a local minimizer of a loss function. We also provide numerical comparisons of SGD with the existing first-order optimizers and show the usefulness of SGD using a step-decay learning rate and a small batch size.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=lK0WxHeups", "arxiv_id": "2402.15344", "paper_pdf": "papers/lK0WxHeups.pdf", "paper_pdf_sha256": "47d36e70caf8e9455c6e6fa7be6f3bb98b8227dc969db534c02ab0f4fb6bf147", "paper_pdf_bytes": 414298, "paper_pdf_source": "openreview", "code_url": "https://github.com/imakn0907/SGD_using_decaying", "code_repository": "imakn0907/SGD_using_decaying", "code_commit": "26927a50592eb9db787bd972a7cd12b6fe3040e4", "code_archive": "repos/lK0WxHeups.zip", "code_archive_sha256": "905564fccb1082c81e06d1e3d76fb83edd19a8159dc7b23f838f752f7e01e815", "code_archive_bytes": 9837, "code_file_count": 5, "code_extensions": {".py": 5}, "github_disk_usage_kb": 24, "github_languages": {"Python": 27913}, "github_archived": false, "github_pushed_at": "2024-07-19T02:07:48Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/iteration-and-stochastic-first-order-oracle"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Z-aIURmBbBk", "year": 2023, "status": "rejected", "title": "Multimodal Masked Autoencoders Learn Transferable Representations", "authors": ["Xinyang Geng", "Hao Liu", "Lisa Lee", "Dale Schuurmans", "Sergey Levine", "Pieter Abbeel"], "authorids": ["~Xinyang_Geng1", "~Hao_Liu1", "~Lisa_Lee1", "~Dale_Schuurmans1", "~Sergey_Levine1", "~Pieter_Abbeel2"], "authors_source": "OpenReview API", "abstract": "Building scalable models to learn from diverse, multimodal data remains an open challenge.\nFor vision-language data, the dominant approaches are based on contrastive learning objectives that train a separate encoder for each modality. While effective, contrastive learning approaches introduce sampling bias depending on the data augmentations used, which can degrade performance on downstream tasks. Moreover, these methods are limited to paired image-text data, and cannot leverage widely-available unpaired data. In this paper, we investigate whether a large multimodal model trained purely via masked token prediction, without using modality-specific encoders or contrastive learning, can learn transferable representations for downstream tasks. We propose a simple and scalable network architecture, the Multimodal Masked Autoencoder (M3AE), which learns a unified encoder for both vision and language data via masked token prediction. We provide an empirical study of M3AE trained on a large-scale image-text dataset, and find that M3AE is able to learn generalizable representations that transfer well to downstream tasks. Surprisingly, we find that M3AE benefits from a higher text mask ratio (50-90%), in contrast to BERT whose standard masking ratio is 15%, due to the joint training of two data modalities. We also provide qualitative analysis showing that the learned representation incorporates meaningful information from both image and language. Lastly, we demonstrate the scalability of M3AE with larger model size and training time, and its flexibility to train on both paired image-text data as well as unpaired data.\n", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=Z-aIURmBbBk", "arxiv_id": "2205.14204", "paper_pdf": "papers/Z-aIURmBbBk.pdf", "paper_pdf_sha256": "66784235da8e99eb1e93821727d5a2b4d847295c5cccd1a57cf0994c4132b5ed", "paper_pdf_bytes": 8448643, "paper_pdf_source": "openreview", "code_url": "https://github.com/young-geng/m3ae_public", "code_repository": "young-geng/m3ae_public", "code_commit": "a5eb14f851a2ad9b4fe84768af9d853b1e10436e", "code_archive": "repos/Z-aIURmBbBk.zip", "code_archive_sha256": "e623194547da52deef192200b3bfdedfaa3e7969a4f4df9dcf7a7494f1bf1e11", "code_archive_bytes": 40863, "code_file_count": 9, "code_extensions": {".py": 9}, "github_disk_usage_kb": 29, "github_languages": {"Python": 138734}, "github_archived": false, "github_pushed_at": "2025-02-26T19:05:32Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/multimodal-masked-autoencoders-learn"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "5_zwnS5oJDp", "year": 2022, "status": "rejected", "title": "Bayesian Learning with Information Gain Provably Bounds Risk for a Robust Adversarial Defense", "authors": ["Bao Gia Doan", "Ehsan M Abbasnejad", "Damith C. Ranasinghe"], "authorids": ["~Bao_Gia_Doan1", "~Ehsan_M_Abbasnejad1", "damith.ranasinghe@adelaide.edu.au"], "authors_source": "OpenReview API", "abstract": "In this paper, we present a novel method to learn a Bayesian neural network robust against adversarial attacks. Previous algorithms have shown an adversarially trained Bayesian Neural Network (BNN) provides improved robustness against attacks. However, the learning approach for approximating the multi-modal Bayesian posterior leads to mode collapse with consequential sub-par robustness and under performance of an adversarially trained BNN. Instead, we propose approximating the multi-modal posterior of a BNN to prevent mode collapse and encourage diversity over learned posterior distributions of models to develop a novel adversarial training method for BNNs.  Importantly, we conceptualize and formulate information gain (IG) in the adversarial Bayesian learning context and prove, training a BNN with IG bounds the difference between the conventional empirical risk with the risk obtained from adversarial training---our intuition is that information gain from benign and adversarial examples should be the same for a robust BNN. Extensive experimental results demonstrate our proposed algorithm to achieve state-of-the-art performance under strong adversarial attacks.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=5_zwnS5oJDp", "arxiv_id": "2212.02003", "paper_pdf": "papers/5_zwnS5oJDp.pdf", "paper_pdf_sha256": "72193fa422207159a2e9e73cded029e1b15b05ea040551fff0a164edb77dbd4c", "paper_pdf_bytes": 1049078, "paper_pdf_source": "openreview", "code_url": "https://github.com/baogiadoan/IG-BNN", "code_repository": "baogiadoan/IG-BNN", "code_commit": "1976cce0d0d4a7242f83fcfee825862cb53aea6b", "code_archive": "repos/5_zwnS5oJDp.zip", "code_archive_sha256": "11ccdc3d8099cc916ac27e51fe53a7d462088ce35bf726f56b344de7b4faf474", "code_archive_bytes": 29436, "code_file_count": 24, "code_extensions": {".py": 22, ".sh": 2}, "github_disk_usage_kb": 27, "github_languages": {"Python": 83703, "Shell": 879}, "github_archived": false, "github_pushed_at": "2024-12-05T07:23:39Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/bayesian-learning-with-information-gain-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "HmAhqnu3qu", "year": 2021, "status": "rejected", "title": "Graph Representation Learning for Multi-Task Settings: a Meta-Learning Approach", "authors": ["Davide Buffelli", "Fabio Vandin"], "authorids": ["~Davide_Buffelli1", "~Fabio_Vandin2"], "authors_source": "OpenReview API", "abstract": "Graph Neural Networks (GNNs) have become the state-of-the-art method for many applications on graph structured data. GNNs are a framework for graph representation learning, where a model learns to generate low dimensional node embeddings that encapsulate structural and feature-related information. GNNs are usually trained in an end-to-end fashion, leading to highly specialized node embeddings. While this approach achieves great results in the single-task setting, generating node embeddings that can be used to perform multiple tasks (with performance comparable to single-task models) is an open problem. We propose a novel representation learning strategy, based on meta-learning, capable of producing multi-task node embeddings. Our method avoids the difficulties arising when learning to perform multiple tasks concurrently by, instead, learning to quickly (i.e. with a few steps of gradient descent) adapt to multiple tasks singularly.  We show that the embeddings produced by our method can be used to perform multiple tasks with comparable or higher performance than both single-task and multi-task end-to-end models. Our method is model-agnostic and task-agnostic and can hence be applied to a wide variety of multi-task domains.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=HmAhqnu3qu", "arxiv_id": "2201.03326", "paper_pdf": "papers/HmAhqnu3qu.pdf", "paper_pdf_sha256": "f55b84171afba0b57895775543088267a6852eef1abb73461ecc475d8cbc0119", "paper_pdf_bytes": 452963, "paper_pdf_source": "openreview", "code_url": "https://github.com/DavideBuffelli/SAME", "code_repository": "DavideBuffelli/SAME", "code_commit": "cfbdda3a4b63bb4a5b92ec6cd21a37fcede47bad", "code_archive": "repos/HmAhqnu3qu.zip", "code_archive_sha256": "d45745192255214352d324a0524e6526fca71a07a4bbe06917a1b63d7184bea6", "code_archive_bytes": 31523, "code_file_count": 13, "code_extensions": {".py": 13}, "github_disk_usage_kb": 31, "github_languages": {"Python": 118158}, "github_archived": false, "github_pushed_at": "2022-04-26T18:48:13Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/graph-representation-learning-for-multi-task-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "B1xfElrKPr", "year": 2020, "status": "rejected", "title": "Enhancing the Transformer with explicit relational encoding for math problem solving", "authors": ["Imanol Schlag", "Paul Smolensky", "Roland Fernandez", "Nebojsa Jojic", "Jürgen Schmidhuber", "Jianfeng Gao"], "authorids": ["imanol@idsia.ch", "paul.smolensky@gmail.com", "rfernand@microsoft.com", "jojic@microsoft.com", "juergen@idsia.ch", "jfgao@microsoft.com"], "authors_source": "OpenReview API", "abstract": "We incorporate Tensor-Product Representations within the Transformer in order to better support the explicit representation of relation structure.\nOur Tensor-Product Transformer (TP-Transformer) sets a new state of the art on the recently-introduced Mathematics Dataset containing 56 categories of free-form math word-problems.\nThe essential component of the model is a novel attention mechanism, called TP-Attention, which explicitly encodes the relations between each Transformer cell and the other cells from which values have been retrieved by attention. TP-Attention goes beyond linear combination of retrieved values, strengthening representation-building and resolving ambiguities introduced by multiple layers of regular attention.\nThe TP-Transformer's attention maps give better insights into how it is capable of solving the Mathematics Dataset's challenging problems.\nPretrained models and code will be made available after publication.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=B1xfElrKPr", "arxiv_id": "1910.06611", "paper_pdf": "papers/B1xfElrKPr.pdf", "paper_pdf_sha256": "9dfa6beed6a16cafe2a20a45c151a4684205a18265a702466642c231ca15203c", "paper_pdf_bytes": 992772, "paper_pdf_source": "openreview", "code_url": "https://github.com/ischlag/TP-Transformer", "code_repository": "ischlag/TP-Transformer", "code_commit": "62aa2f63cd1195b1bf42052a4d8e7d8696b6a566", "code_archive": "repos/B1xfElrKPr.zip", "code_archive_sha256": "4cd20e676a1d9dc66b81c9fc6665138919a21c3b7e01c5a44c880495e6333952", "code_archive_bytes": 29967, "code_file_count": 14, "code_extensions": {".py": 10, ".sh": 3, ".ipynb": 1}, "github_disk_usage_kb": 34, "github_languages": {"Python": 77241, "Jupyter Notebook": 14318, "Shell": 847}, "github_archived": false, "github_pushed_at": "2021-03-08T08:37:40Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/enhancing-the-transformer-with-explicit-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Hyg74h05tX", "year": 2019, "status": "rejected", "title": "Flow++: Improving Flow-Based Generative Models  with  Variational Dequantization and Architecture Design  ", "authors": ["Jonathan Ho", "Xi Chen", "Aravind Srinivas", "Yan Duan", "Pieter Abbeel"], "authorids": ["jonathanho@berkeley.edu", "peter@covariant.ai", "aravind_srinivas@berkeley.edu", "dementrock@gmail.com", "pabbeel@cs.berkeley.edu"], "authors_source": "OpenReview API", "abstract": "Flow-based generative models are powerful exact likelihood models with efficient sampling and inference. \nDespite their computational efficiency, flow-based models generally have much worse density modeling performance compared to state-of-the-art autoregressive models. In this paper, we investigate and improve upon three limiting design choices employed by flow-based models in prior work: the use of uniform noise for dequantization, the use of inexpressive affine flows, and the use of purely convolutional conditioning networks in coupling layers. Based on our findings, we propose Flow++, a new flow-based model that is now the state-of-the-art non-autoregressive model for unconditional density estimation on standard image benchmarks. Our work has begun to close the significant performance gap that has so far existed between autoregressive models and flow-based models.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=Hyg74h05tX", "arxiv_id": "1902.00275", "paper_pdf": "papers/Hyg74h05tX.pdf", "paper_pdf_sha256": "3bfa1a50d178268cc3f7c1de62bcdd7614ac678e0bf5ad4ae85efe23a3dae12e", "paper_pdf_bytes": 8888869, "paper_pdf_source": "openreview", "code_url": "https://github.com/aravindsrinivas/flowpp", "code_repository": "aravindsrinivas/flowpp", "code_commit": "737fadb2218c1e2810a91b523498f97def2c30de", "code_archive": "repos/Hyg74h05tX.zip", "code_archive_sha256": "b3556a3586397611298f9cea1385d7e269a8ce215de1066f0889d0a028ca8fcf", "code_archive_bytes": 78475, "code_file_count": 19, "code_extensions": {".py": 19}, "github_disk_usage_kb": 105, "github_languages": {"Python": 305193}, "github_archived": false, "github_pushed_at": "2019-04-27T21:40:11Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/flow-improving-flow-based-generative-models"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "B1nxTzbRZ", "year": 2018, "status": "rejected", "title": "Forward Modeling for Partial Observation Strategy Games - A StarCraft Defogger", "authors": ["Gabriel Synnaeve", "Zeming Lin", "Jonas Gehring", "Vasil Khalidov", "Nicolas Carion", "Nicolas Usunier"], "authorids": ["gab@fb.com", "zlin@fb.com", "jgehring@fb.com", "vkhalidov@fb.com", "alcinos@fb.com", "usunier@fb.com"], "authors_source": "OpenReview API", "abstract": "This paper we present a defogger, a model that learns to predict future hidden information from partial observations. We formulate this model in the context of forward modeling and leverage spatial and sequential constraints and correlations via convolutional neural networks and long short-term memory networks, respectively. We evaluate our approach on a large dataset of human games of StarCraft: Brood War, a real-time strategy video game. Our models consistently beat strong rule-based baselines and qualitatively produce sensible future game states.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=B1nxTzbRZ", "arxiv_id": "1812.00054", "paper_pdf": "papers/B1nxTzbRZ.pdf", "paper_pdf_sha256": "7f25eaf03da1e0078db5859e2bdca93b1746695d2b6f2d43166d26803688de28", "paper_pdf_bytes": 284097, "paper_pdf_source": "openreview", "code_url": "https://github.com/facebookresearch/starcraft_defogger", "code_repository": "facebookresearch/starcraft_defogger", "code_commit": "092e7accbc25794bd66eaa716d5ea7bfff67ac15", "code_archive": "repos/B1nxTzbRZ.zip", "code_archive_sha256": "05195c11fdfc60eedff61b0013f90d22a9856fbf791366d1d8bc87fdcb20b4cc", "code_archive_bytes": 95956, "code_file_count": 25, "code_extensions": {".py": 10, ".h": 8, ".cpp": 7}, "github_disk_usage_kb": 90, "github_languages": {"Python": 142076, "C++": 139691, "Makefile": 79}, "github_archived": true, "github_pushed_at": "2021-08-30T19:19:49Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/forward-modeling-for-partial-observation"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Ef0nUFg3Jw", "year": 2026, "status": "rejected", "title": "LoRA Users Beware: A Few Spurious Tokens Can Manipulate Your Finetuned Model", "authors": ["Marcel Mateos Salles", "Praney Goyal", "Pradyut Sekhsaria", "Hai Huang", "Randall Balestriero"], "authorids": ["~Marcel_Mateos_Salles1", "~Praney_Goyal1", "~Pradyut_Sekhsaria1", "~Hai_Huang5", "~Randall_Balestriero1"], "authors_source": "OpenReview API", "abstract": "Large Language Models (LLMs) are commonly finetuned for a variety of use cases and domains. A common approach is to leverage Low-Rank Adaptation (LoRA)--known to provide strong performance at low resource costs. In this study, we demonstrate that LoRA actually opens the door to short-cut vulnerabilities--and the more resource efficient is the LoRA setup, the more vulnerable will be the finetuned model to aggressive attacks. To measure that vulnerability, we introduce Seamless Spurious Token Injection (SSTI), where we find that LoRA exclusively focuses on even just a single token that is spuriously correlated with downstream labels. In short, injection of that spurious token during finetuning ensure that the model’s prediction at test-time can be manipulated on-demand. We conducted experiments across model families and datasets to evaluate the impact of SSTI during LoRA finetuning while providing possible mitigations. Our experiments conclude that none of the existing checkers and preprocessors can sanitize a dataset raising new concerns for data quality and AI safety.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=Ef0nUFg3Jw", "arxiv_id": "2506.11402", "paper_pdf": "papers/Ef0nUFg3Jw.pdf", "paper_pdf_sha256": "2edb10ae1847fdd8680a7b5968d039ff4f5dfe1d6d6e9fcd9634ed31fa000108", "paper_pdf_bytes": 1913566, "paper_pdf_source": "openreview", "code_url": "https://github.com/pradyut3501/spurious_corr", "code_repository": "pradyut3501/spurious_corr", "code_commit": "8feaa005775642368ab52c4d500cb1d36aabcc08", "code_archive": "repos/Ef0nUFg3Jw.zip", "code_archive_sha256": "23b79fb6b0638d393b1bace23639082b045207b2e3ee6791e555c069d3b7a9a0", "code_archive_bytes": 26740, "code_file_count": 13, "code_extensions": {".py": 13}, "github_disk_usage_kb": 96, "github_languages": {"Python": 48306}, "github_archived": false, "github_pushed_at": "2025-06-17T10:23:39Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/lora-users-beware-a-few-spurious-tokens-can"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "xlxGsX1pc7", "year": 2025, "status": "rejected", "title": "U-MATH: A University-Level Benchmark for Evaluating Mathematical Skills in LLMs", "authors": ["Konstantin Chernyshev", "Vitaliy Polshkov", "Ekaterina Artemova", "Alex Myasnikov", "Vlad Stepanov", "Alexei Miasnikov", "Sergei Tilga"], "authorids": ["~Konstantin_Chernyshev1", "~Vitaliy_Polshkov1", "~Ekaterina_Artemova2", "~Alex_Myasnikov1", "~Vlad_Stepanov1", "~Alexei_Miasnikov1", "~Sergei_Tilga1"], "authors_source": "OpenReview API", "abstract": "The current evaluation of mathematical skills in LLMs is limited, as existing benchmarks are relatively small, primarily focus on elementary and high-school problems, or lack diversity in topics. Additionally, the inclusion of visual elements in tasks remains largely under-explored. \n\nTo address these gaps, we introduce **U-MATH**, a novel benchmark of \\textbf{1,100} unpublished open-ended university-level problems sourced from teaching materials. It is balanced across six core subjects, with  \\textbf{20\\% of multimodal problems}. Given the open-ended nature of U-MATH problems, we employ an LLM to judge the correctness of generated solutions. To this end, we release **$\\boldsymbol\\mu$-MATH**, an dataset to evaluate the LLMs' capabilities in judging solutions.\n\nThe evaluation of general domain, math-specific, and multimodal LLMs highlights the challenges presented by U-MATH. Our findings reveal that LLMs achieve a maximum accuracy of only 63\\% on text-based tasks, with even lower 45\\% on visual problems. The solution assessment proves challenging for LLMs, with the best LLM judge having an F1-score of 80\\% on $\\mu$-MATH.\n    \nWe open-source U-MATH, $\\mu$-MATH, and evaluation code on GitHub.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=xlxGsX1pc7", "arxiv_id": "2412.03205", "paper_pdf": "papers/xlxGsX1pc7.pdf", "paper_pdf_sha256": "90189bcd89f414cba00affeb2aab331515920a52ff2b85d66c3ad6891356f1f6", "paper_pdf_bytes": 1851859, "paper_pdf_source": "openreview", "code_url": "https://github.com/Toloka/u-math", "code_repository": "Toloka/u-math", "code_commit": "7c23fea9823f26aa3f8736d47fc339ebb3404fcc", "code_archive": "repos/xlxGsX1pc7.zip", "code_archive_sha256": "c7049b98b49b49a2511e6f1fda12f8db4d8e9f88fedae4f5e3702239ae0d2d31", "code_archive_bytes": 9245, "code_file_count": 4, "code_extensions": {".py": 4}, "github_disk_usage_kb": 27, "github_languages": {"Python": 16037}, "github_archived": false, "github_pushed_at": "2026-01-30T19:24:40Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/u-math-a-university-level-benchmark-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "WReszdNNdP", "year": 2024, "status": "rejected", "title": "BOWLL: A DECEPTIVELY SIMPLE OPEN WORLD LIFELONG LEARNER", "authors": ["Roshni Ramanna Kamath", "Rupert Mitchell", "Subarnaduti Paul", "Kristian Kersting", "Martin Mundt"], "authorids": ["~Roshni_Ramanna_Kamath1", "~Rupert_Mitchell1", "~Subarnaduti_Paul1", "~Kristian_Kersting1", "~Martin_Mundt1"], "authors_source": "OpenReview API", "abstract": "The quest to improve scalar performance numbers on predetermined benchmarks seems to be deeply engraved in deep learning. However, the real world is seldom carefully curated and applications are seldom limited to excelling on test sets. A practical system is generally required to recognize novel concepts, refrain from actively including uninformative data, and retain previously acquired knowledge throughout its lifetime. Despite these key elements being rigorously researched individually, the study of their conjunction, open world lifelong learning, is only a recent trend. To accelerate this multifaceted field’s exploration, we introduce its first monolithic and much-needed baseline. Leveraging the ubiquitous use of batch normalization across deep neural networks, we propose a deceptively simple yet highly effective way to repurpose standard models for open world lifelong learning. Through extensive empirical evaluation, we highlight why our approach should serve as a future standard for models that are able to effectively maintain their knowledge, selectively focus on informative data, and accelerate future learning.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=WReszdNNdP", "arxiv_id": "2402.04814", "paper_pdf": "papers/WReszdNNdP.pdf", "paper_pdf_sha256": "6ea0d9c6979c33abf45afc588d51047a233b2b96086f40ff9292ff7c3c82008b", "paper_pdf_bytes": 547197, "paper_pdf_source": "openreview", "code_url": "https://github.com/ml-research/bowll", "code_repository": "ml-research/bowll", "code_commit": "8eb70c82d91c8fd134ce5b60401fbe55cb2621e7", "code_archive": "repos/WReszdNNdP.zip", "code_archive_sha256": "dc4c41cddc8254ea59f16bca587e46253e5091863233ea8cdeaa4fd7f0d829fb", "code_archive_bytes": 29447, "code_file_count": 8, "code_extensions": {".py": 8}, "github_disk_usage_kb": 24, "github_languages": {"Python": 86831}, "github_archived": false, "github_pushed_at": "2024-02-08T09:06:16Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/bowll-a-deceptively-simple-open-world"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "sNKZaNkyi7Q", "year": 2023, "status": "rejected", "title": "Provable Benefits of Representational Transfer in Reinforcement Learning", "authors": ["Alekh Agarwal", "Yuda Song", "Wen Sun", "Kaiwen Wang", "Mengdi Wang", "Xuezhou Zhang"], "authorids": ["~Alekh_Agarwal2", "~Yuda_Song2", "~Wen_Sun1", "~Kaiwen_Wang1", "~Mengdi_Wang1", "~Xuezhou_Zhang2"], "authors_source": "OpenReview API", "abstract": "We study the problem of representational transfer in RL, where an agent first pretrains in a number of source tasks to discover a shared representation, which is subsequently used to learn a good policy in a target task. We propose a new notion of task relatedness between source and target tasks, and develop a novel approach for representational transfer under this assumption. Concretely, we show that given a generative access to source tasks, we can discover a representation, using which subsequent linear RL techniques quickly converge to a near-optimal policy, with only online access to the target task. The sample complexity is close to knowing the ground truth features in the target task, and comparable to prior representation learning results in the source tasks. We complement our positive results with lower bounds without generative access, and validate our findings with empirical evaluation on rich observation MDPs that require deep exploration.\n", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=sNKZaNkyi7Q", "arxiv_id": "2205.14571", "paper_pdf": "papers/sNKZaNkyi7Q.pdf", "paper_pdf_sha256": "4b98e962d4a9160750d84c584806c7c6c2ca8bfb257d7df3dd0c3d78a4352f91", "paper_pdf_bytes": 1212001, "paper_pdf_source": "openreview", "code_url": "https://github.com/yudasong/briee", "code_repository": "yudasong/briee", "code_commit": "c6920ff1806e636a4d806ffd2a6a7d26db7bab6c", "code_archive": "repos/sNKZaNkyi7Q.zip", "code_archive_sha256": "cfd411dc86197fab28ea5c52dbbf4bd2b05f66bb492036a23783f0ad9ee9952c", "code_archive_bytes": 14609, "code_file_count": 9, "code_extensions": {".py": 6, ".sh": 3}, "github_disk_usage_kb": 30, "github_languages": {"Python": 41358, "Shell": 585}, "github_archived": false, "github_pushed_at": "2022-06-01T19:02:12Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/provable-benefits-of-representational"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ExJ4lMbZcqa", "year": 2022, "status": "rejected", "title": "Learning Audio-Visual Dereverberation", "authors": ["Changan Chen", "Wei Sun", "David Harwath", "Kristen Grauman"], "authorids": ["~Changan_Chen2", "~Wei_Sun10", "~David_Harwath1", "~Kristen_Grauman1"], "authors_source": "OpenReview API", "abstract": "Reverberation from audio reflecting off surfaces and objects in the environment not only degrades the quality of speech for human perception, but also severely impacts the accuracy of automatic speech recognition. Prior work attempts to remove reverberation based on the audio modality only. Our idea is to learn to dereverberate speech from audio-visual observations. The visual environment surrounding a human speaker reveals important cues about the room geometry, materials, and speaker location, all of which influence the precise reverberation effects in the audio stream. We introduce Visually-Informed Dereverberation of Audio (VIDA), an end-to-end approach that learns to remove reverberation based on both the observed sounds and visual scene. In support of this new task, we develop a large-scale dataset that uses realistic acoustic renderings of speech in real-world 3D scans of homes offering a variety of room acoustics. Demonstrating our approach on both simulated and real imagery for speech enhancement, speech recognition, and speaker identification, we show it achieves state-of-the-art performance and substantially improves over traditional audio-only methods.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=ExJ4lMbZcqa", "arxiv_id": "2106.07732", "paper_pdf": "papers/ExJ4lMbZcqa.pdf", "paper_pdf_sha256": "f69623afbc0ef2b2e257786589839f887b557e29bf135a2ee591c8d1a31116f1", "paper_pdf_bytes": 22695264, "paper_pdf_source": "openreview", "code_url": "https://github.com/facebookresearch/learning-audio-visual-dereverberation", "code_repository": "facebookresearch/learning-audio-visual-dereverberation", "code_commit": "ef5ceb4b8fe59d10798628b7a4498c686473fd3d", "code_archive": "repos/ExJ4lMbZcqa.zip", "code_archive_sha256": "18b866747ed1236c55e01b1d02d8d6c93c8ef49bfa1c73f049e467a17519e3f0", "code_archive_bytes": 29334, "code_file_count": 6, "code_extensions": {".py": 6}, "github_disk_usage_kb": 27, "github_languages": {"Python": 60359}, "github_archived": true, "github_pushed_at": "2022-08-10T02:04:01Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/learning-audio-visual-dereverberation"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "lSijhyKKsct", "year": 2021, "status": "rejected", "title": "Reinforcement Learning with Latent Flow", "authors": ["Wenling Shang", "Xiaofei Wang", "Aravind Rajeswaran", "Aravind Srinivas", "Yang Gao", "Pieter Abbeel", "Michael Laskin"], "authorids": ["~Wenling_Shang1", "w.xf@berkeley.edu", "~Aravind_Rajeswaran1", "~Aravind_Srinivas1", "~Yang_Gao1", "~Pieter_Abbeel2", "~Michael_Laskin1"], "authors_source": "OpenReview API", "abstract": "Temporal information is essential to learning effective policies with Reinforcement Learning (RL). However, current state-of-the-art RL algorithms either assume that such information is given as part of the state space or, when learning from pixels, use the simple heuristic of frame-stacking to implicitly capture temporal information present in the image observations. This heuristic is in contrast to the current paradigm in video classification architectures, which utilize explicit encodings of temporal information through methods such as optical flow and two-stream architectures to achieve state-of-the-art performance. Inspired by leading video classification architectures, we introduce the Flow of Latents for Reinforcement Learning Flare, a network architecture for RL that explicitly encodes temporal information through latent vector differences. We show that Flare (i) recovers optimal performance in state-based RL without explicit access to the state velocity, solely with positional state information, (ii) achieves state-of-the-art performance on pixel-based continuous control tasks within the DeepMind control benchmark suite, (iii) is the most sample efficient model-free pixel-based RL algorithm on challenging environments in the DeepMind control suite such as quadruped walk, hopper hop, finger turn hard, pendulum swing, and walker run, outperforming the prior model-free state-of-the-art by 1.9 and 1.5 on the 500k and 1M step benchmarks, respectively, and (iv), when augmented over rainbow DQN, outperforms or matches the baseline on a diversity of challenging Atari games at 50M time step benchmark.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=lSijhyKKsct", "arxiv_id": "2101.01857", "paper_pdf": "papers/lSijhyKKsct.pdf", "paper_pdf_sha256": "70d550e0b1b25c8e21133bd3eff77ccfbeba1142358c5a978d939116b5fb0930", "paper_pdf_bytes": 3646630, "paper_pdf_source": "openreview", "code_url": "https://github.com/WendyShang/flare", "code_repository": "WendyShang/flare", "code_commit": "ef12a7f4b2fc639d491ca63648ea1e8c58d07263", "code_archive": "repos/lSijhyKKsct.zip", "code_archive_sha256": "6e9ab16c900e97ffcb6a910988b27dd95e36ba28741ee096defb9bf3b355b2ee", "code_archive_bytes": 26182, "code_file_count": 12, "code_extensions": {".py": 8, ".sh": 4}, "github_disk_usage_kb": 31, "github_languages": {"Python": 78060, "Shell": 2529}, "github_archived": false, "github_pushed_at": "2021-03-25T20:56:51Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/reinforcement-learning-with-latent-flow-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "r1l1myStwr", "year": 2020, "status": "rejected", "title": "Continuous Meta-Learning without Tasks", "authors": ["James Harrison", "Apoorva Sharma", "Chelsea Finn", "Marco Pavone"], "authorids": ["jharrison@stanford.edu", "apoorva@stanford.edu", "cbfinn@cs.stanford.edu", "pavone@stanford.edu"], "authors_source": "OpenReview API", "abstract": "Meta-learning is a promising strategy for learning to efficiently learn within new tasks, using data gathered from a distribution of tasks. However, the meta-learning literature thus far has focused on the task segmented setting, where at train-time, offline data is assumed to be split according to the underlying task, and at test-time, the algorithms are optimized to learn in a single task. In this work, we enable the application of generic meta-learning algorithms to settings where this task segmentation is unavailable, such as continual online learning with a time-varying task. We present meta-learning via online changepoint analysis (MOCA), an approach which augments a meta-learning algorithm with a differentiable Bayesian changepoint detection scheme. The framework allows both training and testing directly on time series data without segmenting it into discrete tasks. We demonstrate the utility of this approach on a nonlinear meta-regression benchmark as well as two meta-image-classification benchmarks.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=r1l1myStwr", "arxiv_id": "1912.08866", "paper_pdf": "papers/r1l1myStwr.pdf", "paper_pdf_sha256": "ed3077a4b4326f6ee496d6bf9e93a6af5679dc3899969895f02e3d55e19b6b28", "paper_pdf_bytes": 1479739, "paper_pdf_source": "openreview", "code_url": "https://github.com/StanfordASL/moca", "code_repository": "StanfordASL/moca", "code_commit": "a4eb2d950b9a54e221786734665ed5d7d9e34646", "code_archive": "repos/r1l1myStwr.zip", "code_archive_sha256": "33922c7fadac4a01fc61b9c5c2e34c4e9fcd6786565e555f57fb22c72258b83b", "code_archive_bytes": 19385, "code_file_count": 11, "code_extensions": {".py": 11}, "github_disk_usage_kb": 35, "github_languages": {"Python": 60493}, "github_archived": false, "github_pushed_at": "2022-11-22T04:38:05Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/continuous-meta-learning-without-tasks-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "S1gd7nCcF7", "year": 2019, "status": "rejected", "title": "Self-Supervised Generalisation with Meta Auxiliary Learning", "authors": ["Shikun Liu", "Edward Johns", "Andrew Davison"], "authorids": ["shikun.liu17@imperial.ac.uk", "e.johns@imperial.ac.uk", "a.davison@imperial.ac.uk"], "authors_source": "OpenReview API", "abstract": "Auxiliary learning has been shown to improve the generalisation performance of a principal task. But typically, this requires manually-defined auxiliary tasks based on domain knowledge. In this paper, we consider that it may be possible to automatically learn these auxiliary tasks to best suit the principal task, towards optimum auxiliary tasks without any human knowledge. We propose a novel method, Meta Auxiliary Learning (MAXL), which we design for the task of image classification, where the auxiliary task is hierarchical sub-class image classification. The role of the meta learner is to determine sub-class target labels to train a multi-task evaluator, such that these labels improve the generalisation performance on the principal task. Experiments on three different CIFAR datasets show that MAXL outperforms baseline auxiliary learning methods, and is competitive even with a method which uses human-defined sub-class hierarchies. MAXL is self-supervised and general, and therefore offers a promising new direction towards automated generalisation.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=S1gd7nCcF7", "arxiv_id": "1901.08933", "paper_pdf": "papers/S1gd7nCcF7.pdf", "paper_pdf_sha256": "56be89bca044ac69b543e91267024ab3b9d39fb5e4968b2ea8e740f9459c0370", "paper_pdf_bytes": 17204922, "paper_pdf_source": "openreview", "code_url": "https://github.com/lorenmt/maxl", "code_repository": "lorenmt/maxl", "code_commit": "266f5b68df0ecdffc73da40e003fc6d58cceef96", "code_archive": "repos/S1gd7nCcF7.zip", "code_archive_sha256": "1e5e13b5ce774cea22679d9a3d5cd47639e4bc7c64d1ff18548461619e89c083", "code_archive_bytes": 18093, "code_file_count": 5, "code_extensions": {".py": 5}, "github_disk_usage_kb": 111, "github_languages": {"Python": 59824}, "github_archived": false, "github_pushed_at": "2021-12-19T20:46:37Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/self-supervised-generalisation-with-meta"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "HJUOHGWRb", "year": 2018, "status": "rejected", "title": "Contextual Explanation Networks", "authors": ["Maruan Al-Shedivat", "Avinava Dubey", "Eric P. Xing"], "authorids": ["alshedivat@cs.cmu.edu", "akdubey@cs.cmu.edu", "epxing@cs.cmu.edu"], "authors_source": "OpenReview API", "abstract": "We introduce contextual explanation networks (CENs)---a class of models that learn to predict by generating and leveraging intermediate explanations. CENs are deep networks that generate parameters for context-specific probabilistic graphical models which are further used for prediction and play the role of explanations. Contrary to the existing post-hoc model-explanation tools, CENs learn to predict and to explain jointly. Our approach offers two major advantages: (i) for each prediction, valid instance-specific explanations are generated with no computational overhead and (ii) prediction via explanation acts as a regularization and boosts performance in low-resource settings. We prove that local approximations to the decision boundary of our networks are consistent with the generated explanations. Our results on image and text classification and survival analysis tasks demonstrate that CENs are competitive with the state-of-the-art while offering additional insights behind each prediction, valuable for decision support.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=HJUOHGWRb", "arxiv_id": "1705.10301", "paper_pdf": "papers/HJUOHGWRb.pdf", "paper_pdf_sha256": "55db7c1e0c7c66a1b9f6d26c21df49f3f91d30e709bbee5f47c252fe28511bc3", "paper_pdf_bytes": 1881482, "paper_pdf_source": "openreview", "code_url": "https://github.com/alshedivat/cen", "code_repository": "alshedivat/cen", "code_commit": "f03397a0bf4ac24162e270907d623f8658179e88", "code_archive": "repos/HJUOHGWRb.zip", "code_archive_sha256": "b17324a3cfc8d36f3e5c0138da1fb5b99b7710ee3e259b3617d1cbfb8835366f", "code_archive_bytes": 100856, "code_file_count": 44, "code_extensions": {".py": 38, ".sh": 6}, "github_disk_usage_kb": 96, "github_languages": {"Python": 139132}, "github_archived": false, "github_pushed_at": "2020-03-19T20:57:37Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/contextual-explanation-networks"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "vu0wCqW51g", "year": 2026, "status": "rejected", "title": "Content-Rich AIGC Video Quality Assessment via Intricate Text Alignment and Motion-Aware Consistency", "authors": ["Shangkun Sun", "Xiaoyu Liang", "Bowen Qu", "Ge Li", "Wei Gao"], "authorids": ["~Shangkun_Sun3", "~Xiaoyu_Liang2", "~Bowen_Qu1", "~Ge_Li2", "~Wei_Gao12"], "authors_source": "OpenReview API", "abstract": "The advent of next-generation video generation models like Sora poses challenges for AI-generated content (AIGC) video quality assessment (VQA). These models substantially mitigate flickering artifacts prevalent in prior models, enable longer and complex text prompts and generate longer videos with intricate, diverse motion patterns. Conventional VQA methods designed for simple text and basic motion patterns struggle to evaluate these content-rich videos. To this end, we propose **CRAVE** (Content-Rich AIGC Video Evaluator), specifically for the evaluation of Sora-era AIGC videos. CRAVE proposes the multi-granularity text-temporal fusion that aligns long-form complex textual semantics with video dynamics. Additionally, CRAVE leverages the hybrid motion-fidelity modeling to assess temporal artifacts. Furthermore, given the straightforward prompts and content in current AIGC VQA datasets, we introduce **CRAVE-DB**, a benchmark featuring content-rich videos from next-generation models paired with elaborate prompts. Extensive experiments have shown that the proposed CRAVE achieves excellent results on multiple AIGC VQA benchmarks, demonstrating a high degree of alignment with human perception. All data and code will be publicly available.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=vu0wCqW51g", "arxiv_id": "2502.04076", "paper_pdf": "papers/vu0wCqW51g.pdf", "paper_pdf_sha256": "0800beab20ba09644b519f7f8b0293689aa1db38d874c941d0508e86f07f0e3c", "paper_pdf_bytes": 5070168, "paper_pdf_source": "openreview", "code_url": "https://github.com/littlespray/CRAVE", "code_repository": "littlespray/CRAVE", "code_commit": "2c4973b73dadef6eb6bcccf50e272c24df8f9c11", "code_archive": "repos/vu0wCqW51g.zip", "code_archive_sha256": "c450ca4d9a6967cbef35f0f599e443fab8015637ba0ea151d197e760e51128b8", "code_archive_bytes": 109603, "code_file_count": 22, "code_extensions": {".py": 22}, "github_disk_usage_kb": 103, "github_languages": {"Python": 356310}, "github_archived": false, "github_pushed_at": "2025-07-31T16:11:42Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/content-rich-aigc-video-quality-assessment"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "bBUhlynfRX", "year": 2025, "status": "rejected", "title": "A Brain-Inspired Regularizer for Adversarial Robustness", "authors": ["Elie Attias", "Cengiz Pehlevan", "Dina Obeid"], "authorids": ["~Elie_Attias1", "~Cengiz_Pehlevan2", "~Dina_Obeid1"], "authors_source": "OpenReview API", "abstract": "Convolutional Neural Networks (CNNs) excel in many visual tasks, but they tend to be sensitive to slight input perturbations that are imperceptible to the human eye, often resulting in task failures. Recent studies indicate that training CNNs with regularizers that promote brain-like representations, using neural recordings, can improve model robustness. However, the requirement to use neural data severely restricts the utility of these methods. Is it possible to develop regularizers that mimic the computational function of neural regularizers without the need for neural recordings, thereby expanding the usability and effectiveness of these techniques? In this work, we inspect a neural regularizer introduced in Li et al. to extract its underlying strength. The regularizer uses neural representational similarities, which we find also correlate with pixel similarities. Motivated by this finding, we introduce a new regularizer that retains the essence of the original but is computed using image pixel similarities, eliminating the need for neural recordings.  We show that our regularization method 1) significantly increases model robustness against a variety of black box attacks, 2) relies only on original, unaugmented datasets and 3) is computationally inexpensive. Our work explores how biologically motivated loss functions can be used to drive the performance of artificial neural networks.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=bBUhlynfRX", "arxiv_id": "2410.03952", "paper_pdf": "papers/bBUhlynfRX.pdf", "paper_pdf_sha256": "ac7e90280194abfa4e7222c109fbc276b93c20a497da35528fd1801e203c8fc0", "paper_pdf_bytes": 3582587, "paper_pdf_source": "openreview", "code_url": "https://github.com/elieattias1/pixel-reg", "code_repository": "elieattias1/pixel-reg", "code_commit": "69d13eb5efa3a70eebcc38705c2fd0ab663e8d1a", "code_archive": "repos/bBUhlynfRX.zip", "code_archive_sha256": "cd8a51bbcf8153bf92929aa6352aa76885862e3362147dae7662faac69b49c11", "code_archive_bytes": 29683, "code_file_count": 15, "code_extensions": {".py": 15}, "github_disk_usage_kb": 28, "github_languages": {"Python": 95335}, "github_archived": false, "github_pushed_at": "2024-10-16T01:18:15Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/a-brain-inspired-regularizer-for-adversarial"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "6rEcB9m9AI", "year": 2024, "status": "rejected", "title": "Promoting Exploration in Memory-Augmented Adam using Critical Momenta", "authors": ["Pranshu Malviya", "Goncalo Mordido", "Aristide Baratin", "Reza Babanezhad Harikandeh", "Jerry Huang", "Simon Lacoste-Julien", "Razvan Pascanu", "Sarath Chandar"], "authorids": ["~Pranshu_Malviya1", "~Goncalo_Mordido1", "~Aristide_Baratin1", "~Reza_Babanezhad_Harikandeh1", "~Jerry_Huang1", "~Simon_Lacoste-Julien1", "~Razvan_Pascanu1", "~Sarath_Chandar1"], "authors_source": "OpenReview API", "abstract": "Adaptive gradient-based optimizers, particularly Adam, have left their mark in training large-scale deep learning models. The strength of such optimizers is that they exhibit fast convergence while being more robust to hyperparameter choice. However, they often generalize worse than non-adaptive methods. Recent studies have tied this performance gap to flat minima selection: adaptive methods tend to find solutions in sharper basins of the loss landscape, which in turn hurts generalization. To overcome this issue, we propose a new memory-augmented version of Adam that promotes {exploration} towards flatter minima by using a buffer of critical momentum terms during training. Intuitively, the use of the buffer makes the optimizer overshoot outside the basin of attraction if it is not wide enough. We empirically show that our method improves model performance on standard supervised and online learning tasks.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=6rEcB9m9AI", "arxiv_id": "2307.09638", "paper_pdf": "papers/6rEcB9m9AI.pdf", "paper_pdf_sha256": "063650ba28cb26183d2bd9e4899e90b7fa64428f374488cb4ad356f55e449245", "paper_pdf_bytes": 5510489, "paper_pdf_source": "openreview", "code_url": "https://github.com/chandar-lab/CMOptimizer", "code_repository": "chandar-lab/CMOptimizer", "code_commit": "bdc77310624345ce88e299137b88e746c079b607", "code_archive": "repos/6rEcB9m9AI.zip", "code_archive_sha256": "99ce7fdb578157596c765c3748a57271c18665e47b05d432df9e09f7917d545c", "code_archive_bytes": 8731, "code_file_count": 3, "code_extensions": {".py": 3}, "github_disk_usage_kb": 25, "github_languages": {"Python": 33464}, "github_archived": false, "github_pushed_at": "2024-06-18T13:34:56Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/promoting-exploration-in-memory-augmented"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "5DkfiQPy9A", "year": 2023, "status": "rejected", "title": "ACAT: Adversarial Counterfactual Attention for Classification and Detection in Medical Imaging", "authors": ["Alessandro Fontanella", "Antreas Antoniou", "Wenwen Li", "Joanna Wardlaw", "Grant Mair", "Emanuele Trucco", "Amos Storkey"], "authorids": ["~Alessandro_Fontanella1", "~Antreas_Antoniou3", "~Wenwen_Li2", "~Joanna_Wardlaw1", "~Grant_Mair1", "~Emanuele_Trucco1", "~Amos_Storkey1"], "authors_source": "OpenReview API", "abstract": "In some medical imaging tasks and other settings where only small parts of the image are informative for the classification task, traditional CNNs can sometimes struggle to generalise. Manually annotated Regions of Interest (ROI) are sometimes used to isolate the most informative parts of the image. However, these are expensive to collect and may vary significantly across annotators. To overcome these issues, we propose a method to generate saliency maps, obtained from adversarially generated counterfactual images. With this method, we are able to isolate the area of interest in brain and lung CT scans without using any manual annotations. Our saliency maps, in the task of localising the lesion location out of 6 possible regions, obtain a score of $65.05 \\%$ on brain CT scans, improving the score of $61.29 \\%$ obtained with the best competing method. We also employ the saliency maps in a framework that refines a classifier pipeline. In particular, the saliency maps are used to obtain soft spatial attention masks that modulate the image features at different scales. We refer to our method as \\emph{Adversarial Counterfactual Attention} (ACAT). ACAT increases the baseline classification accuracy of lesions in brain CT scans from $71.39 \\%$ to $72.55 \\%$ and of COVID-19 related findings in lung CT scans from $67.71 \\%$ to $70.84 \\%$ and exceeds the performance of competing methods.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=5DkfiQPy9A", "arxiv_id": "2303.15421", "paper_pdf": "papers/5DkfiQPy9A.pdf", "paper_pdf_sha256": "18258ee1fc1fcb3d74c83d2c759ee43d6491a3cea2db0026234c2f5d6935a20a", "paper_pdf_bytes": 8985424, "paper_pdf_source": "openreview", "code_url": "https://github.com/alessandro-f/ACAT", "code_repository": "alessandro-f/ACAT", "code_commit": "7e4f289e768a41fe24abcf87ff05fc591ea439b0", "code_archive": "repos/5DkfiQPy9A.zip", "code_archive_sha256": "3cb26c3a88531af496a447b75ade008ee579d4c3f1da662d47e95a676c4568bd", "code_archive_bytes": 38592, "code_file_count": 11, "code_extensions": {".py": 11}, "github_disk_usage_kb": 31, "github_languages": {"Python": 301499}, "github_archived": false, "github_pushed_at": "2023-03-24T21:11:42Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/acat-adversarial-counterfactual-attention-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "zyrhwrd9EYs", "year": 2022, "status": "rejected", "title": "To Impute or Not To Impute? Missing Data in Treatment Effect Estimation", "authors": ["Jeroen Berrevoets", "Fergus Imrie", "Trent Kyono", "James Jordon", "Mihaela van der Schaar"], "authorids": ["~Jeroen_Berrevoets1", "~Fergus_Imrie1", "~Trent_Kyono1", "~James_Jordon1", "~Mihaela_van_der_Schaar2"], "authors_source": "OpenReview API", "abstract": "Missing data is a systemic problem in practical scenarios that causes noise and bias when estimating treatment effects. This makes treatment effect estimation from data with missingness a particularly tricky endeavour. A key reason for this is that standard assumptions on missingness are rendered insufficient due to the presence of an additional variable, treatment, besides the individual and the outcome.  Having a treatment variable introduces additional complexity with respect to why some variables are missing that is overlooked by previous work. In our work we identify a new missingness mechanism, which we term mixed confounded missingness (MCM), where some missingness determines treatment selection and other missingness is determined by treatment selection. Given MCM, we show that naively imputing all data leads to poor performing treatment effects models, as the act of imputation effectively removes information necessary to provide unbiased estimates. However, no imputation at all also leads to biased estimates, as missingness determined by treatment divides the population in distinct subpopulations, where estimates across these populations will be biased. Our solution is selective imputation, where we use insights from MCM to inform precisely which variables should be imputed and which should not. We empirically demonstrate how various learners benefit from selective imputation compared to other solutions for missing data.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=zyrhwrd9EYs", "arxiv_id": "2202.02096", "paper_pdf": "papers/zyrhwrd9EYs.pdf", "paper_pdf_sha256": "6f3d75e7d95ac655638164da8a3d65078101c60416d50dfefd8298a06afd3b5e", "paper_pdf_bytes": 306015, "paper_pdf_source": "openreview", "code_url": "https://github.com/jeroenbe/mcm", "code_repository": "jeroenbe/mcm", "code_commit": "41688030608ed8cbd3c9a58169bf3ff03ddf1e54", "code_archive": "repos/zyrhwrd9EYs.zip", "code_archive_sha256": "19fcdfc14b247cf4dcb495b4d9e93019ad18d424def0c05ebd74994eba00324e", "code_archive_bytes": 29813, "code_file_count": 9, "code_extensions": {".ipynb": 6, ".py": 3}, "github_disk_usage_kb": 27, "github_languages": {"Jupyter Notebook": 84481, "Python": 9308}, "github_archived": false, "github_pushed_at": "2023-04-26T22:10:25Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/to-impute-or-not-to-impute-missing-data-in-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "JAlqRs9duhz", "year": 2021, "status": "rejected", "title": "Straight to the Gradient: Learning to Use Novel Tokens for Neural Text Generation", "authors": ["Xiang Lin", "SIMENG HAN", "Shafiq Joty"], "authorids": ["~Xiang_Lin2", "~SIMENG_HAN1", "~Shafiq_Joty1"], "authors_source": "OpenReview API", "abstract": "Advanced large-scale neural language models have led to significant success in many natural language generation tasks. However, the most commonly used training objective, Maximum Likelihood Estimation (MLE), has been shown to be problematic, where the trained model prefers using dull and repetitive phrases. In this work, we introduce ScaleGrad, a modification straight to the gradient of the loss function, to remedy the degeneration issues of the standard MLE objective. By directly maneuvering the gradient information, ScaleGrad makes the model learn to use novel tokens during training. Empirical results show the effectiveness of our method not only in open-ended generation, but also in directed generation. With the simplicity in architecture, our method can serve as a general training objective that is applicable to most of the neural text generation tasks.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=JAlqRs9duhz", "arxiv_id": "2106.07207", "paper_pdf": "papers/JAlqRs9duhz.pdf", "paper_pdf_sha256": "f6d87ae4824d8d8062607554e7d2b1e9be55d6e97cfe036c93535cf1cc15de91", "paper_pdf_bytes": 779546, "paper_pdf_source": "openreview", "code_url": "https://github.com/shawnlimn/ScaleGrad", "code_repository": "shawnlimn/ScaleGrad", "code_commit": "b2685f9c8680e731316ca6149c1171fda7af5ead", "code_archive": "repos/JAlqRs9duhz.zip", "code_archive_sha256": "2fc69b03145935cc9d1c0a973293a658b520b90c0dd895bf3282d3dea4cf0d6d", "code_archive_bytes": 30116, "code_file_count": 12, "code_extensions": {".py": 12}, "github_disk_usage_kb": 32, "github_languages": {"Python": 81982}, "github_archived": false, "github_pushed_at": "2021-12-28T04:15:40Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/straight-to-the-gradient-learning-to-use-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Hkl_bCVKDr", "year": 2020, "status": "rejected", "title": "Scaleable input gradient regularization for adversarial robustness", "authors": ["Chris Finlay", "Adam M Oberman"], "authorids": ["christopher.finlay@mail.mcgill.ca", "adam.oberman@mcgill.ca"], "authors_source": "OpenReview API", "abstract": "In this work we revisit gradient regularization for adversarial robustness with some new ingredients.  First, we derive new per-image theoretical robustness bounds based on local gradient information. These bounds strongly motivate input gradient regularization.  Second, we implement a scaleable version of input gradient regularization which avoids double backpropagation: adversarially robust ImageNet models are trained in 33 hours on four consumer grade GPUs.  Finally, we show experimentally and through theoretical certification that input gradient regularization is competitive with adversarial training. Moreover we demonstrate that gradient regularization does not lead to gradient obfuscation or gradient masking.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=Hkl_bCVKDr", "arxiv_id": "1905.11468", "paper_pdf": "papers/Hkl_bCVKDr.pdf", "paper_pdf_sha256": "758646b3250b5685d366c1be1740e45e665cd1162580a93e10bf05584c061e61", "paper_pdf_bytes": 357781, "paper_pdf_source": "openreview", "code_url": "https://github.com/cfinlay/tulip", "code_repository": "cfinlay/tulip", "code_commit": "92c83e794127f0feb2770a8887ee9ca61364c194", "code_archive": "repos/Hkl_bCVKDr.zip", "code_archive_sha256": "41ccb795e140c7e3bc51f4de0274e79cc78eb0164fddcdb34665add3e4fdc5ad", "code_archive_bytes": 42298, "code_file_count": 21, "code_extensions": {".py": 19, ".sh": 2}, "github_disk_usage_kb": 36, "github_languages": {"Python": 105919, "Shell": 2040}, "github_archived": false, "github_pushed_at": "2019-07-08T18:38:16Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/scaleable-input-gradient-regularization-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "HJMghjA9YX", "year": 2019, "status": "rejected", "title": "Model Comparison for Semantic Grouping", "authors": ["Francisco Vargas", "Kamen Brestnichki", "Nils Hammerla"], "authorids": ["francisco.vargas@babylonhealth.com", "kamen.brestnichki@babylonhealth.com", "nils.hammerla@babylonhealth.com"], "authors_source": "OpenReview API", "abstract": "We introduce a probabilistic framework for quantifying the semantic similarity between two groups of embeddings. We formulate the task of semantic similarity as a model comparison task in which we contrast a generative model which jointly models two sentences versus one that does not. We illustrate how this framework can be used for the Semantic Textual Similarity tasks using clear assumptions about how the embeddings of words are generated. We apply information criteria based model comparison to overcome the shortcomings of Bayesian model comparison, whilst still penalising model complexity. We achieve competitive results by applying the proposed framework with an appropriate choice of likelihood on the STS datasets.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=HJMghjA9YX", "arxiv_id": "1904.13323", "paper_pdf": "papers/HJMghjA9YX.pdf", "paper_pdf_sha256": "9bdee0d1695d25160f4210512ad49e3b321949804e6ad7ce6d4d14b16ca67493", "paper_pdf_bytes": 406144, "paper_pdf_source": "openreview", "code_url": "https://github.com/babylonhealth/MCSG", "code_repository": "babylonhealth/MCSG", "code_commit": "f445fbba676f84ec590b5d13493dca16bd0641d8", "code_archive": "repos/HJMghjA9YX.zip", "code_archive_sha256": "99485f3fc27e5c8fa62f6735a28f9ddd10ad5eaa8d596600079d057bbc53913a", "code_archive_bytes": 61313, "code_file_count": 31, "code_extensions": {".py": 28, ".bash": 3}, "github_disk_usage_kb": 149, "github_languages": {"Python": 101171, "Shell": 9785, "sed": 2192}, "github_archived": false, "github_pushed_at": "2022-07-13T15:19:44Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/model-comparison-for-semantic-grouping"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "SkJKHMW0Z", "year": 2018, "status": "rejected", "title": "Recurrent Relational Networks for complex relational reasoning", "authors": ["Rasmus Berg Palm", "Ulrich Paquet", "Ole Winther"], "authorids": ["rasmusbergpalm@gmail.com", "upaq@google.com", "olwi@dtu.dk"], "authors_source": "OpenReview API", "abstract": "Humans possess an ability to abstractly reason about objects and their interactions, an ability not shared with state-of-the-art deep learning models. Relational networks, introduced by Santoro et al. (2017), add the capacity for relational reasoning to deep neural networks, but are limited in the complexity of the reasoning tasks they can address. We introduce recurrent relational networks which increase the suite of solvable tasks to those that require an order of magnitude more steps of relational reasoning. We use recurrent relational networks to solve Sudoku puzzles and achieve state-of-the-art results by solving 96.6% of the hardest Sudoku puzzles, where relational networks fail to solve any. We also apply our model to the BaBi textual QA dataset solving 19/20 tasks which is competitive with state-of-the-art sparse differentiable neural computers. The recurrent relational network is a general purpose module that can augment any neural network model with the capacity to do many-step relational reasoning.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=SkJKHMW0Z", "arxiv_id": null, "paper_pdf": "papers/SkJKHMW0Z.pdf", "paper_pdf_sha256": "d950d569ecc1b1115ca03ae059a72f1d8cb415f8c1d2a1a1618a2579c8ede64c", "paper_pdf_bytes": 425384, "paper_pdf_source": "openreview", "code_url": "https://github.com/Kyubyong/sudoku", "code_repository": "Kyubyong/sudoku", "code_commit": "cb611230ad602aafebe6914c3e53c81671bba605", "code_archive": "repos/SkJKHMW0Z.zip", "code_archive_sha256": "6a16f3941c49bd63ce45c541d4f95e5f40d8aa0e282111ce00e361d6fb2e8b1d", "code_archive_bytes": 52938, "code_file_count": 6, "code_extensions": {".py": 6}, "github_disk_usage_kb": 100, "github_languages": {"Python": 23747}, "github_archived": false, "github_pushed_at": "2023-02-17T05:22:35Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/recurrent-relational-networks-for-complex"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "LDGAjUuEVZ", "year": 2026, "status": "rejected", "title": "BrowseComp-ZH: Benchmarking Web Browsing Ability of Large Language Models in Chinese", "authors": ["Peilin Zhou", "Junling Liu", "Xiang Ying", "Can Zhang", "Yifan Shao", "Qichen Ye", "Dading Chong", "Zhiling Jin", "Chenxuan Xie", "Meng Cao", "Yuxin Gu", "Sixin Hong", "Jing Ren", "Jian Chen", "Chao Liu", "Yining Hua", "Sunghun Kim"], "authorids": ["~Peilin_Zhou1", "~Junling_Liu1", "~Xiang_Ying3", "~Can_Zhang2", "~Yifan_Shao3", "~Qichen_Ye1", "~Dading_Chong1", "~Zhiling_Jin3", "~Chenxuan_Xie3", "~Meng_Cao1", "~Yuxin_Gu1", "~Sixin_Hong1", "~Jing_Ren2", "~Jian_Chen18", "~Chao_Liu8", "~Yining_Hua1", "~Sunghun_Kim1"], "authors_source": "OpenReview API", "abstract": "As large language models (LLMs) evolve into web-interacting agents, their ability to retrieve and reason over real-time information has become a crucial benchmark for general intelligence. However, existing benchmarks such as BrowseComp focus solely on English, neglecting the linguistic, infrastructural, and retrieval-specific challenges posed by other information ecosystems—particularly the Chinese web. We present BrowseComp-ZH, a high-difficulty, natively-constructed benchmark designed to assess LLM agents’ web browsing abilities in Chinese. Rather than translating from English, all questions in BrowseComp-ZH are written from scratch by native speakers to reflect authentic information-seeking behaviors and cultural contexts. The dataset comprises 289 multi-hop questions across 11 diverse domains, each reverse-engineered from a short, verifiable answer and filtered through a twostage quality control pipeline to ensure retrieval hardness and answer uniqueness. We evaluate over 20 leading LLMs and agentic search systems. Despite strong language and retrieval abilities, most models perform poorly: many score below 10% accuracy, and only a few exceed 20%. Even the best system achieves just 42.9% accuracy. These results highlight the considerable difficulty of BrowseComp-ZH, where success requires not only robust retrieval strategies but also advanced multihop reasoning and information reconciliation—abilities that remain challenging for current models. BrowseComp-ZH thus serves as a stress test for web-interactive LLMs beyond English, offering a rigorous and linguistically diverse evaluation framework to guide future research on multilingual agent capabilities.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=LDGAjUuEVZ", "arxiv_id": "2504.19314", "paper_pdf": "papers/LDGAjUuEVZ.pdf", "paper_pdf_sha256": "4e9829515cc04f79f052e5748ac09ee69bd03e467b9862cf902210a43dadf7b2", "paper_pdf_bytes": 3669748, "paper_pdf_source": "openreview", "code_url": "https://github.com/PALIN2018/BrowseComp-ZH", "code_repository": "PALIN2018/BrowseComp-ZH", "code_commit": "86abe635e7deef89ec00c68ff1c2588f0e2f2099", "code_archive": "repos/LDGAjUuEVZ.zip", "code_archive_sha256": "e6902b2303ea2b55b592e8efd96f3d893cd3510bf5340714bfaf191537ecbce2", "code_archive_bytes": 104576, "code_file_count": 5, "code_extensions": {".py": 4, ".sh": 1}, "github_disk_usage_kb": 107, "github_languages": {"Python": 27501, "Shell": 291}, "github_archived": false, "github_pushed_at": "2025-05-14T07:18:36Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/browsecomp-zh-benchmarking-web-browsing"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "DRKkO2Tejc", "year": 2025, "status": "rejected", "title": "Label Privacy in Split Learning for Large Models with Parameter-Efficient Training", "authors": ["Philip Zmushko", "Marat Mansurov", "Ruslan Svirschevski", "Denis Kuznedelev", "Max Ryabinin", "Aleksandr Beznosikov"], "authorids": ["~Philip_Zmushko1", "~Marat_Mansurov1", "~Ruslan_Svirschevski1", "~Denis_Kuznedelev1", "~Max_Ryabinin1", "~Aleksandr_Beznosikov1"], "authors_source": "OpenReview API", "abstract": "As deep learning models become larger and more expensive, many practitioners turn to fine-tuning APIs. \nThese web services allow fine-tuning a model between two parties: the client that provides the data, and the server that hosts the model.\nWhile convenient, these APIs raise a new concern: the data of the client is at risk of privacy breach during the training procedure.\nThis challenge presents an important practical case of vertical federated learning, where the two parties perform parameter-efficient fine-tuning (PEFT) of a large model.\nIn this study, we systematically search for a way to fine-tune models over an API *while keeping the labels private*.\nWe analyze the privacy of LoRA, a popular approach for parameter-efficient fine-tuning when training over an API.\nUsing this analysis, we propose P$^3$EFT, a multi-party  split learning algorithm that takes advantage of existing PEFT properties to maintain privacy at a lower performance overhead.\nTo validate our algorithm, we fine-tune DeBERTa-v2-XXLarge, Flan-T5 Large and LLaMA-2 7B using LoRA adapters on a range of NLP tasks. We find that P$^3$EFT is competitive with existing privacy-preserving methods in multi-party and  two-party setups while having higher accuracy.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=DRKkO2Tejc", "arxiv_id": "2412.16669", "paper_pdf": "papers/DRKkO2Tejc.pdf", "paper_pdf_sha256": "18e48ac9b8651fbf00287200e4f3cc4419eadd22eacb4901e9b4a08145f0b4c2", "paper_pdf_bytes": 740933, "paper_pdf_source": "openreview", "code_url": "https://github.com/p3eft-iclr-2025/p3eft-iclr-2025", "code_repository": "p3eft-iclr-2025/p3eft-iclr-2025", "code_commit": "419418877dfeb07bc2dd2bd4da609c5ef4d7c207", "code_archive": "repos/DRKkO2Tejc.zip", "code_archive_sha256": "7e66ca80a35f04f11702fc20c8c000122d7ff147d09cf4b4ace6c78994003404", "code_archive_bytes": 36871, "code_file_count": 15, "code_extensions": {".sh": 8, ".py": 5, ".ipynb": 2}, "github_disk_usage_kb": 28, "github_languages": {"Python": 72124, "Jupyter Notebook": 28708, "Shell": 14928}, "github_archived": false, "github_pushed_at": "2024-10-08T14:55:14Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/label-privacy-in-split-learning-for-large"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "WSzRdcOkEx", "year": 2024, "status": "rejected", "title": "GREAT Score: Global Robustness Evaluation of Adversarial Perturbation using Generative Models", "authors": ["ZAITANG LI", "Pin-Yu Chen", "Tsung-Yi Ho"], "authorids": ["~ZAITANG_LI1", "~Pin-Yu_Chen1", "~Tsung-Yi_Ho2"], "authors_source": "OpenReview API", "abstract": "Current studies on adversarial robustness mainly focus on aggregating \\textit{local} robustness results from a set of data samples to evaluate and rank different models. However, the local statistics may not well represent the true \\textit{global} robustness of the underlying unknown data distribution. To address this challenge, this paper makes the first attempt to present a new framework, called \\textit{GREAT Score}, for global robustness evaluation of adversarial perturbation using generative models. Formally, GREAT Score carries the physical meaning of  a global statistic capturing a mean certified attack-proof perturbation level over all samples drawn from a generative model. For finite-sample evaluation, we also derive a probabilistic guarantee on the sample complexity and the difference between the sample mean and the true mean. GREAT Score has several advantages: (1) Robustness evaluations using GREAT Score are efficient and scalable to large models, by sparing the need of running adversarial attacks. In particular, we show high correlation and significantly reduced computation cost of GREAT Score when compared to the attack-based model ranking on RobustBench \\cite{croce2021robustbench}.\n(2) The use of generative models facilitates the approximation of the unknown data distribution. In our ablation study with different generative adversarial networks (GANs), we observe consistency between global robustness evaluation and the quality of GANs. (3) GREAT Score can be used for remote auditing of privacy-sensitive black-box models, as demonstrated by our robustness evaluation on several online facial recognition services.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=WSzRdcOkEx", "arxiv_id": "2304.09875", "paper_pdf": "papers/WSzRdcOkEx.pdf", "paper_pdf_sha256": "315c7b539a7d6c3f04d2d0947a586563550c787d8293353c984b7a9f7ecc0800", "paper_pdf_bytes": 7561210, "paper_pdf_source": "openreview", "code_url": "https://github.com/IBM/GREAT-Score", "code_repository": "IBM/GREAT-Score", "code_commit": "beadb769ac59f434b42e6a40e08ba87c758444ff", "code_archive": "repos/WSzRdcOkEx.zip", "code_archive_sha256": "511178ac754305ca0a9f7e07b532de0930a234cca339e9f6baf468daf48f9b1c", "code_archive_bytes": 28372, "code_file_count": 16, "code_extensions": {".py": 16}, "github_disk_usage_kb": 26, "github_languages": {"Python": 63514}, "github_archived": false, "github_pushed_at": "2025-09-18T00:02:59Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/great-score-global-robustness-evaluation-of"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "nG08xiRT2As", "year": 2023, "status": "rejected", "title": "InfoOT: Information Maximizing Optimal Transport", "authors": ["Ching-Yao Chuang", "Stefanie Jegelka", "David Alvarez-Melis"], "authorids": ["~Ching-Yao_Chuang1", "~Stefanie_Jegelka3", "~David_Alvarez-Melis1"], "authors_source": "OpenReview API", "abstract": "Optimal transport aligns samples across distributions by minimizing the transportation cost between them, e.g., the geometric distances. Yet, it ignores coherence structure in the data such as clusters, does not handle outliers well, and cannot integrate new data points. To address these drawbacks, we propose InfoOT, an information-theoretic extension of optimal transport that maximizes the mutual information between domains while minimizing geometric distances. The resulting objective can still be formulated as a (generalized) optimal transport problem, and can be efficiently solved by projected gradient descent. This formulation yields a new projection method that is robust to outliers and generalizes to unseen samples. Empirically, InfoOT improves the quality of alignments across benchmarks in domain adaptation, cross-domain retrieval, and single-cell alignment.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=nG08xiRT2As", "arxiv_id": "2210.03164", "paper_pdf": "papers/nG08xiRT2As.pdf", "paper_pdf_sha256": "df0584ccfdfb10292719d82b782e1bda65dd907883144e0b376a90c06ab06b8f", "paper_pdf_bytes": 11806626, "paper_pdf_source": "openreview", "code_url": "https://github.com/chingyaoc/InfoOT", "code_repository": "chingyaoc/InfoOT", "code_commit": "352efd202f5b475dc170a8d08a99049689d5ee1a", "code_archive": "repos/nG08xiRT2As.zip", "code_archive_sha256": "0e0a6202890a82aee9e74349efb75e671a42ed5a825943bd6b1d8cc770bcb3f4", "code_archive_bytes": 6228, "code_file_count": 3, "code_extensions": {".py": 3}, "github_disk_usage_kb": 32, "github_languages": {"Python": 11361}, "github_archived": false, "github_pushed_at": "2023-04-27T03:49:02Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/infoot-information-maximizing-optimal"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "dYUdt59fJ0e", "year": 2022, "status": "rejected", "title": "Yformer: U-Net Inspired Transformer Architecture for Far Horizon Time Series Forecasting", "authors": ["Kiran Madhusudhanan", "Johannes Burchert", "Nghia Duong-Trung", "Stefan Born", "Lars Schmidt-Thieme"], "authorids": ["~Kiran_Madhusudhanan1", "burchert@ismll.uni-hildesheim.de", "nghia.duong-trung@tu-berlin.de", "born@math.tu-berlin.de", "~Lars_Schmidt-Thieme1"], "authors_source": "OpenReview API", "abstract": "Time series data is ubiquitous in research as well as in a wide variety of industrial applications. Effectively analyzing the available historical data and providing insights into the far future allows us to make effective decisions. Recent research has witnessed the superior performance of transformer-based architectures, especially in the regime of far horizon time series forecasting. However, the current state of the art sparse Transformer architectures fail to couple down- and upsampling procedures to produce outputs in a similar resolution as the input. We propose the Yformer model, based on a novel Y-shaped encoder-decoder architecture that (1) uses direct connection from the downscaled encoder layer to the corresponding upsampled decoder layer in a U-Net inspired architecture, (2) Combines the downscaling/upsampling with sparse attention to capture long-range effects, and (3) stabilizes the encoder-decoder stacks with the addition of an auxiliary reconstruction loss. Extensive experiments have been conducted with relevant baselines on four benchmark datasets, demonstrating an average improvement of 19.82, 18.41 percentage MSE and 13.62, 11.85 percentage MAE in comparison to the current state of the art for the univariate and the multivariate settings respectively.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=dYUdt59fJ0e", "arxiv_id": "2110.08255", "paper_pdf": "papers/dYUdt59fJ0e.pdf", "paper_pdf_sha256": "799459e97228541446cb56257f2b0b786c26f54be7b5fccb0029dff53e67b378", "paper_pdf_bytes": 750164, "paper_pdf_source": "openreview", "code_url": "https://github.com/18kiran12/Yformer-Time-Series-Forecasting", "code_repository": "18kiran12/Yformer-Time-Series-Forecasting", "code_commit": "65783ab25ad350d55bc12938630565727eb03b2b", "code_archive": "repos/dYUdt59fJ0e.zip", "code_archive_sha256": "09f688a196f54fd5ca348575574ef522255341bb2318150010b49206daee31f9", "code_archive_bytes": 26078, "code_file_count": 14, "code_extensions": {".py": 14}, "github_disk_usage_kb": 29, "github_languages": {"Python": 90622}, "github_archived": false, "github_pushed_at": "2022-08-25T15:15:10Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/yformer-u-net-inspired-transformer-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "iMKvxHlrZb3", "year": 2021, "status": "rejected", "title": "Scalable Graph Neural Networks for Heterogeneous Graphs", "authors": ["Lingfan Yu", "Jiajun Shen", "Jinyang Li", "Adam Lerer"], "authorids": ["~Lingfan_Yu1", "~Jiajun_Shen1", "~Jinyang_Li1", "~Adam_Lerer1"], "authors_source": "OpenReview API", "abstract": "Graph neural networks (GNNs) are a popular class of parametric model for learning over graph-structured data. Recent work has argued that GNNs primarily use the graph for feature smoothing, and have shown competitive results on benchmark tasks by simply operating on graph-smoothed node features, rather than using end-to-end learned feature hierarchies that are challenging to scale to large graphs. In this work, we ask whether these results can be extended to heterogeneous graphs, which encode multiple types of relationship between different entities. We propose Neighbor Averaging over Relation Subgraphs (NARS), which trains a classifier on neighbor-averaged features for randomly-sampled subgraphs of the ‘metagraph‘ of relations. We describe optimizations to allow these sets of node features to be computed in a memory-efficient way, both at training and inference time. NARS achieves a new state of the art accuracy on several benchmark datasets, outperforming more expensive GNN-based methods.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=iMKvxHlrZb3", "arxiv_id": "2011.09679", "paper_pdf": "papers/iMKvxHlrZb3.pdf", "paper_pdf_sha256": "8f537614d0cc5169b90c72da6a4458446bd153f9c641e438c0f9989b1707f73d", "paper_pdf_bytes": 544130, "paper_pdf_source": "openreview", "code_url": "https://github.com/facebookresearch/NARS", "code_repository": "facebookresearch/NARS", "code_commit": "27850d3812e04d2057ba6b53aae02fd35a0e180c", "code_archive": "repos/iMKvxHlrZb3.zip", "code_archive_sha256": "5b3c04190d6b98de1bf59d42f925c53226f0a5709aa3ef619ac69c6988c86818", "code_archive_bytes": 39672, "code_file_count": 16, "code_extensions": {".py": 12, ".sh": 4}, "github_disk_usage_kb": 34, "github_languages": {"Python": 54277, "Shell": 2468, "Dockerfile": 772}, "github_archived": true, "github_pushed_at": "2020-11-24T18:24:06Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/scalable-graph-neural-networks-for-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "BylB4kBtwB", "year": 2020, "status": "rejected", "title": "Retrieving Signals in the Frequency Domain with Deep Complex Extractors", "authors": ["Chiheb Trabelsi", "Olexa Bilaniuk", "Ousmane Dia", "Ying Zhang", "Mirco Ravanelli", "Jonathan Binas", "Negar Rostamzadeh", "Christopher  J Pal"], "authorids": ["chiheb.trabelsi@polymtl.ca", "olexa.bilaniuk@umontreal.ca", "ousmane@elementai.com", "ying@elementai.com", "mirco.ravanelli@gmail.com", "jbinas@gmail.com", "negar@elementai.com", "christopher.pal@elementai.com"], "authors_source": "OpenReview API", "abstract": "Recent advances have made it possible to create deep complex-valued neural networks. Despite this progress, the potential power of fully complex intermediate computations and representations has not yet been explored for many challenging learning problems. Building on recent advances, we propose a novel mechanism for extracting signals in the frequency domain. As a case study, we perform audio source separation in the Fourier domain. Our extraction mechanism could be regarded as a local ensembling method that combines a complex-valued convolutional version of Feature-Wise Linear Modulation (FiLM) and a signal averaging operation. We also introduce a new explicit amplitude and phase-aware loss, which is scale and time invariant, taking into account the complex-valued components of the spectrogram. Using the Wall Street Journal Dataset, we compare our phase-aware loss to several others that operate both in the time and frequency domains and demonstrate the effectiveness of our proposed signal extraction method and proposed loss. When operating in the complex-valued frequency domain, our deep complex-valued network substantially outperforms its real-valued counterparts even with half the depth and a third of the parameters. Our proposed mechanism improves significantly deep complex-valued networks' performance and we demonstrate the usefulness of its regularizing effect.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=BylB4kBtwB", "arxiv_id": null, "paper_pdf": "papers/BylB4kBtwB.pdf", "paper_pdf_sha256": "4dc98b6bf38ee2e87ac865673422291b6bbd436a423ccf04793601a86104c7e0", "paper_pdf_bytes": 763456, "paper_pdf_source": "openreview", "code_url": "https://github.com/FourierSignalRetrievalICLR2020/FourierExtraction", "code_repository": "FourierSignalRetrievalICLR2020/FourierExtraction", "code_commit": "7dab26ced2f47c509f82963b5bb7ab459f535d55", "code_archive": "repos/BylB4kBtwB.zip", "code_archive_sha256": "c8b235b6f107088825f31fb58b816516fedd2c3ce93c82f6919579cedd61cd2f", "code_archive_bytes": 48797, "code_file_count": 16, "code_extensions": {".py": 16}, "github_disk_usage_kb": 37, "github_languages": {"Python": 220455}, "github_archived": false, "github_pushed_at": "2019-09-24T18:38:53Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/retrieving-signals-in-the-frequency-domain"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ryfaViR9YX", "year": 2019, "status": "rejected", "title": "Variation Network: Learning High-level Attributes for Controlled Input Manipulation", "authors": ["Gaëtan Hadjeres"], "authorids": ["hadjeres.g@gmail.com"], "authors_source": "OpenReview API", "abstract": "This paper presents the Variation Network (VarNet), a  generative model providing means to manipulate the high-level attributes of a given input. The originality of our approach is that VarNet is not only capable of handling pre-defined attributes but can also learn the relevant attributes of the dataset by itself.  These two settings can be easily combined  which makes VarNet applicable for a wide variety of tasks. Further, VarNet has a sound probabilistic interpretation which grants us with  a novel way to navigate in the latent spaces as well as means to control how the  attributes are learned. We demonstrate  experimentally that this model is capable of performing interesting input manipulation  and that the learned attributes are relevant and interpretable.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=ryfaViR9YX", "arxiv_id": "1901.03634", "paper_pdf": "papers/ryfaViR9YX.pdf", "paper_pdf_sha256": "0bea74c62322db9de1f98d183bade4b31a9f07521b051e2fa1ed140f92be8999", "paper_pdf_bytes": 2706922, "paper_pdf_source": "openreview", "code_url": "https://github.com/Ghadjeres/VarNet", "code_repository": "Ghadjeres/VarNet", "code_commit": "08a76075b833fbf71ddc9c73598ca9e7398a480d", "code_archive": "repos/ryfaViR9YX.zip", "code_archive_sha256": "9c89bc83f39bc86375180e3508f3cd3747cf17484ea093d7f08e52b124f1be22", "code_archive_bytes": 157988, "code_file_count": 31, "code_extensions": {".py": 31}, "github_disk_usage_kb": 151, "github_languages": {"Python": 181929}, "github_archived": false, "github_pushed_at": "2019-03-12T15:18:57Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/variation-network-learning-high-level"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "S1fduCl0b", "year": 2018, "status": "rejected", "title": "Lifelong Generative Modeling", "authors": ["Jason Ramapuram", "Magda Gregorova", "Alexandros Kalousis"], "authorids": ["jason.ramapuram@etu.unige.ch", "magda.gregorova@unige.ch", "alexandros.kalousis@hesge.ch"], "authors_source": "OpenReview API", "abstract": "Lifelong learning is the problem of learning multiple consecutive tasks in a sequential manner where knowledge gained from previous tasks is retained and used for future learning. It is essential towards the development of intelligent machines that can adapt to their surroundings. In this work we focus on a lifelong learning approach to generative modeling where we continuously incorporate newly observed streaming distributions into our learnt model. We do so through a student-teacher architecture which allows us to learn and preserve all the distributions seen so far without the need to retain the past data nor the past models. Through the introduction of a novel cross-model regularizer, the student model leverages the information learnt by the teacher, which acts as a summary of everything seen till now. The regularizer has the additional benefit of reducing the effect of catastrophic interference that appears when we learn over streaming data. We demonstrate its efficacy on streaming distributions as well as its ability to learn a common latent representation across a complex transfer learning scenario.\n", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=S1fduCl0b", "arxiv_id": "1705.09847", "paper_pdf": "papers/S1fduCl0b.pdf", "paper_pdf_sha256": "d0c222940124836a05fec8fabc888bdb3a74bd363bff16fdfad1e0cd66dda26a", "paper_pdf_bytes": 3621573, "paper_pdf_source": "openreview", "code_url": "https://github.com/jramapuram/LifelongVAE_pytorch", "code_repository": "jramapuram/LifelongVAE_pytorch", "code_commit": "0199cb3c9659c4e22f20e1cd4af8cf1c8f8a1e8d", "code_archive": "repos/S1fduCl0b.zip", "code_archive_sha256": "cb80cc0aa709616d74b712fb1b21f0eec7f1f9fad1f33d3a1427c0b745a0c55d", "code_archive_bytes": 36322, "code_file_count": 14, "code_extensions": {".py": 13, ".sh": 1}, "github_disk_usage_kb": 133, "github_languages": {"Python": 107264, "Shell": 1251}, "github_archived": false, "github_pushed_at": "2019-12-03T23:44:13Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/lifelong-generative-modeling"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "skIx8CgP8V", "year": 2026, "status": "rejected", "title": "Noise Augmented Fine Tuning for Mitigating Hallucinations in Large Language Models", "authors": ["Afshin Khadangi", "Amir Sartipi", "Igor Tchappi", "Ramin Bahmani"], "authorids": ["~Afshin_Khadangi1", "~Amir_Sartipi2", "~Igor_Tchappi1", "~Ramin_Bahmani1"], "authors_source": "OpenReview API", "abstract": "Large language models (LLMs) often produce inaccurate or misleading content—hallucinations. To address this challenge, we introduce Noise-Augmented Fine-Tuning (NoiseFiT), a novel framework that leverages adaptive noise injection based on the signal-to-noise ratio (SNR) to enhance model robustness. Our contribution is threefold. First, NoiseFiT selectively perturbs layers identified as either high-SNR (more robust) or low-SNR (potentially under-regularized) using a dynamically scaled Gaussian noise. Second, we further propose a hybrid loss that combines standard cross-entropy, soft cross-entropy, and consistency regularization to ensure stable and accurate outputs under noisy training conditions. Third, a theoretical analysis proposed shows that adaptive noise injection is both unbiased and variance-preserving, providing strong guarantees for convergence in expectation. Moreover, empirical results on multiple test and benchmark datasets demonstrate that NoiseFiT significantly reduces hallucination rates, often improving or matching baseline performance in key tasks. These findings highlight the promise of noise-driven strategies for achieving robust, trustworthy language modeling without incurring prohibitive computational overhead. We have publicly released the fine-tuning logs, benchmark evaluation artifacts, and source code online at W&B, Hugging Face, and GitHub, respectively, to foster further research, accessibility and reproducibility.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=skIx8CgP8V", "arxiv_id": "2504.03302", "paper_pdf": "papers/skIx8CgP8V.pdf", "paper_pdf_sha256": "7eb8e2c4e8aeb3db330e4cd55761efe02500825b10e4f854b57ca1e78703e775", "paper_pdf_bytes": 23393123, "paper_pdf_source": "openreview", "code_url": "https://github.com/akhadangi/NoiseFiT", "code_repository": "akhadangi/NoiseFiT", "code_commit": "48c61c88a1a9bb808a8091276473d7770ea33ef9", "code_archive": "repos/skIx8CgP8V.zip", "code_archive_sha256": "5be9d9e9258bbf4839be9b0cd8a7a0b1224caa648f685ffccaa49d807e388b70", "code_archive_bytes": 61203, "code_file_count": 5, "code_extensions": {".py": 5}, "github_disk_usage_kb": 108, "github_languages": {"Python": 135894}, "github_archived": false, "github_pushed_at": "2026-05-27T18:21:31Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/noise-augmented-fine-tuning-for-mitigating"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ptTt8mhS7n", "year": 2025, "status": "rejected", "title": "In-Context Transfer Learning: Demonstration Synthesis by Transferring Similar Tasks", "authors": ["Dingzirui Wang", "Xuanliang Zhang", "Qiguang Chen", "Longxu Dou", "Xiao Xu", "Rongyu Cao", "YINGWEI MA", "Qingfu Zhu", "Wanxiang Che", "Binhua Li", "Fei Huang", "Yongbin Li"], "authorids": ["~Dingzirui_Wang1", "~Xuanliang_Zhang1", "~Qiguang_Chen1", "~Longxu_Dou1", "~Xiao_Xu1", "~Rongyu_Cao2", "~YINGWEI_MA2", "~Qingfu_Zhu1", "~Wanxiang_Che1", "~Binhua_Li1", "~Fei_Huang2", "~Yongbin_Li2"], "authors_source": "OpenReview API", "abstract": "In-context learning (ICL) is an effective approach to help large language models (LLMs) adapt to various tasks by providing demonstrations of the target task. Considering the high cost of labeling demonstrations, many methods propose synthesizing demonstrations from scratch using LLMs. However, the quality of the demonstrations synthesized from scratch is limited by the capabilities and knowledge of LLMs. To address this, inspired by transfer learning, we propose In-Context Transfer Learning (ICTL), which synthesizes target task demonstrations by transferring labeled demonstrations from similar source tasks. ICTL consists of two steps: source sampling and target transfer. First, we define an optimization objective, which minimizes transfer error to sample source demonstrations similar to the target task.  Then, we employ LLMs to transfer the sampled source demonstrations to match the definition and format of the target task. Experiments on Super-NI show that ICTL outperforms synthesis from scratch by 2.0% on average, demonstrating the effectiveness of our method.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=ptTt8mhS7n", "arxiv_id": "2410.01548", "paper_pdf": "papers/ptTt8mhS7n.pdf", "paper_pdf_sha256": "a1cb4e1d424362746e7bd3c5fb4d3d81e656fbf1707aab06cf5388bcb39a90dd", "paper_pdf_bytes": 2458199, "paper_pdf_source": "openreview", "code_url": "https://github.com/zirui-HIT/ICTL", "code_repository": "zirui-HIT/ICTL", "code_commit": "e6a407345276fe999eb1de9f04de3a1957f91cb0", "code_archive": "repos/ptTt8mhS7n.zip", "code_archive_sha256": "a54341c1cc69880b1d313cd936efed7069965af038756b5b18edc8b7c8f62ce2", "code_archive_bytes": 34994, "code_file_count": 27, "code_extensions": {".py": 18, ".sh": 9}, "github_disk_usage_kb": 28, "github_languages": {"Python": 68716, "Shell": 6238}, "github_archived": false, "github_pushed_at": "2024-10-01T15:26:29Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/in-context-transfer-learning-demonstration"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "JB3lbDtsFS", "year": 2024, "status": "rejected", "title": "It HAS to be Subjective: Human Annotator Simulation via Zero-shot Density Estimation", "authors": ["Wen Wu", "Wenlin Chen", "Chao Zhang", "Phil Woodland"], "authorids": ["~Wen_Wu2", "~Wenlin_Chen2", "~Chao_Zhang20", "~Phil_Woodland1"], "authors_source": "OpenReview API", "abstract": "Human annotator simulation (HAS) serves as a cost-effective substitute for human evaluation such as data annotation and system assessment. Human perception and behaviour during human evaluation exhibit inherent variability due to diverse cognitive processes and subjective interpretations, which should be taken into account in modelling to better mimic the way people perceive and interact with the world. This paper introduces a novel meta-learning framework that treats HAS as a zero-shot density estimation problem, which incorporates human variability and allows for the efficient generation of human-like annotations for unlabelled test inputs. Under this framework, we propose two new model classes, conditional integer flows and conditional softmax flows, to account for ordinal and categorical annotations, respectively. The proposed method is evaluated on three real-world human evaluation tasks and shows superior capability and efficiency to predict the aggregated behaviours of human annotators, match the distribution of human annotations, and simulate the inter-annotator disagreements.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=JB3lbDtsFS", "arxiv_id": "2310.00486", "paper_pdf": "papers/JB3lbDtsFS.pdf", "paper_pdf_sha256": "7bb2763b53709194c082f23e57d1b0260f0ae9bd17ae2250872093c9922e2c81", "paper_pdf_bytes": 1378326, "paper_pdf_source": "openreview", "code_url": "https://github.com/W-Wu/HAS_CNF", "code_repository": "W-Wu/HAS_CNF", "code_commit": "62b5e149a8a27f1ebfa6bd03d46dfa44fab844b8", "code_archive": "repos/JB3lbDtsFS.zip", "code_archive_sha256": "e8a3936386c1cf9be63f1575e36b71df8afd7f5ec9017d78695b26c19ebe045e", "code_archive_bytes": 20152, "code_file_count": 9, "code_extensions": {".py": 9}, "github_disk_usage_kb": 27, "github_languages": {"Python": 44520}, "github_archived": false, "github_pushed_at": "2024-08-24T12:30:16Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/it-has-to-be-subjective-human-annotator"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "NZZoABNZECq", "year": 2023, "status": "rejected", "title": "Mechanistic Mode Connectivity", "authors": ["Ekdeep Singh Lubana", "Eric J Bigelow", "Robert P. Dick", "David Krueger", "Hidenori Tanaka"], "authorids": ["~Ekdeep_Singh_Lubana1", "~Eric_J_Bigelow1", "~Robert_P._Dick1", "~David_Krueger1", "~Hidenori_Tanaka1"], "authors_source": "OpenReview API", "abstract": "With the rise of pretrained models, fine-tuning has become of central importance in deep learning. However, unlike retraining from scratch, fine-tuning can fail to qualitatively change the behavior of a pre-trained network. For instance, we find in practice that naive fine-tuning does not eliminate a model’s sensitivity to spurious features. To understand and address this limitation, we study the geometry of neural network loss landscapes through the lens of mode-connectivity. Our work addresses two questions about mode-connectivity: 1) Are models trained on different data distributions mode-connected? 2) Can we fine tune a pre-trained model to switch modes? We define a notion of mechanistic mode-connectivity, and find that only models that already share the same invariances (which we call “mechanistically similar”) are mechanistically mode-connected. We hypothesize this property explains inability of naive fine-tuning methods to induce invariance to spurious features. Based on our analysis, we propose and validate a method of “mechanistic fine-tuning” called connectivity-based fine-tuning (CBFT)", "decision": "Reject", "meta_review": null, "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=NZZoABNZECq", "arxiv_id": "2211.08422", "paper_pdf": "papers/NZZoABNZECq.pdf", "paper_pdf_sha256": "f6f9f6400447f0b46a4ff0c23e06618788c881eaa879848ea590487db1b54620", "paper_pdf_bytes": 5743017, "paper_pdf_source": "openreview", "code_url": "https://github.com/EkdeepSLubana/MMC", "code_repository": "EkdeepSLubana/MMC", "code_commit": "f884ca780441dbd1b7cee0393c9df13950e7f13b", "code_archive": "repos/NZZoABNZECq.zip", "code_archive_sha256": "b8bf5354a3d3ee2013fedd231aa43a9771eb3e3864728830b0d5435ead8d78fb", "code_archive_bytes": 21316, "code_file_count": 8, "code_extensions": {".py": 8}, "github_disk_usage_kb": 32, "github_languages": {"Python": 84265}, "github_archived": false, "github_pushed_at": "2023-07-14T16:22:14Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/mechanistic-mode-connectivity"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "_ysluXvD1M", "year": 2022, "status": "rejected", "title": "Equal Experience in Recommender Systems", "authors": ["Jaewoong Cho", "Moonseok Choi", "Changho Suh"], "authorids": ["~Jaewoong_Cho1", "~Moonseok_Choi1", "~Changho_Suh1"], "authors_source": "OpenReview API", "abstract": "We explore the fairness issue that arises in recommender systems. Biased data due to inherent stereotypes of particular groups (e.g., male students' average rating on mathematics is often higher than that on humanities, and vice versa for females) may yield a limited scope of suggested items to a certain group of users. Our main contribution lies in the introduction of a novel fairness notion (that we call equal experience), which can serve to regulate such unfairness in the presence of biased data. The notion captures the degree of the equal experience of item recommendations across distinct groups. We propose an optimization framework that incorporates the fairness notion as a regularization term, as well as introduce computationally-efficient algorithms that solve the optimization. Experiments on synthetic and benchmark real datasets demonstrate that the proposed framework can indeed mitigate such unfairness while exhibiting a minor degradation of recommendation accuracy.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=_ysluXvD1M", "arxiv_id": "2210.05936", "paper_pdf": "papers/_ysluXvD1M.pdf", "paper_pdf_sha256": "978977a30f6dfa2eb2cea598118e7975e05f8fd4120b70a4db01f0f51cae1e90", "paper_pdf_bytes": 902580, "paper_pdf_source": "openreview", "code_url": "https://github.com/cjw2525/FairRec", "code_repository": "cjw2525/FairRec", "code_commit": "c57677441f9c0e2b7ab215d0b30a41fcc872fc47", "code_archive": "repos/_ysluXvD1M.zip", "code_archive_sha256": "70bf6fa32747f4465f9ffe59b1c46d9626133201de18be653c4ccc7a4f7e998b", "code_archive_bytes": 35852, "code_file_count": 12, "code_extensions": {".py": 11, ".ipynb": 1}, "github_disk_usage_kb": 30, "github_languages": {"Python": 20930, "Jupyter Notebook": 4490}, "github_archived": false, "github_pushed_at": "2021-10-05T22:57:50Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/equal-experience-in-recommender-systems-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "-Qaj4_O3cO", "year": 2021, "status": "rejected", "title": "DCT-SNN: Using DCT to Distribute Spatial Information over Time for Learning Low-Latency Spiking Neural Networks", "authors": ["Isha Garg", "Sayeed Shafayet Chowdhury", "Kaushik Roy"], "authorids": ["~Isha_Garg1", "~Sayeed_Shafayet_Chowdhury3", "~Kaushik_Roy1"], "authors_source": "OpenReview API", "abstract": "Spiking Neural Networks (SNNs) offer a promising alternative to traditional deep learning frameworks, since they provide higher \ncomputational efficiency due to event-driven information processing. SNNs distribute the analog values of pixel intensities into binary spikes over time. However, the most widely used input coding schemes, such as Poisson based rate-coding, do not leverage the additional temporal learning capability of SNNs effectively. Moreover, these SNNs suffer from high inference latency which is a major bottleneck to their deployment. To overcome this, we propose a scalable time-based encoding scheme that utilizes the Discrete Cosine Transform (DCT) to reduce the number of timesteps required for inference. DCT decomposes an image into a weighted sum of sinusoidal basis images. At each time step, a single frequency base, taken in order and modulated\nby its corresponding DCT coefficient, is input to an accumulator that generates spikes upon crossing a threshold. We use the proposed scheme to learn DCT-SNN, a low-latency deep SNN with leaky-integrate-and-fire neurons, trained using surrogate gradient descent based backpropagation. We achieve top-1 accuracy of 89.94%, 68.3% and 52.43% on CIFAR-10, CIFAR-100 and TinyImageNet, respectively using VGG architectures. Notably, DCT-SNN performs inference with 2-14X reduced latency compared to other state-of-the-art SNNs, while achieving comparable accuracy to their standard deep learning counterparts. The dimension of the transform allows us to control the number of timesteps required for inference. Additionally, we can trade-off accuracy with latency in a principled manner by dropping the highest frequency components during inference.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=-Qaj4_O3cO", "arxiv_id": "2010.01795", "paper_pdf": "papers/-Qaj4_O3cO.pdf", "paper_pdf_sha256": "bec6fc58978939a8874f86faeab53a3bb0e5c50afe3d1f9890b10ceb6adde686", "paper_pdf_bytes": 1575748, "paper_pdf_source": "openreview", "code_url": "https://github.com/SayeedChowdhury/dct-snn", "code_repository": "SayeedChowdhury/dct-snn", "code_commit": "c0ef166a8e7df4185c7774900f381d23e87e54ca", "code_archive": "repos/-Qaj4_O3cO.zip", "code_archive_sha256": "9d8f51c6a8d81a8b062bac6d242c20b94dd76c189a3388ddbc903de20843e193", "code_archive_bytes": 25656, "code_file_count": 6, "code_extensions": {".py": 6}, "github_disk_usage_kb": 34, "github_languages": {"Python": 91108}, "github_archived": false, "github_pushed_at": "2021-09-30T16:30:41Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/dct-snn-using-dct-to-distribute-spatial-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Bkf4XgrKvS", "year": 2020, "status": "rejected", "title": "Unsupervised Learning of Graph Hierarchical Abstractions with Differentiable Coarsening and Optimal Transport", "authors": ["Tengfei Ma", "Jie Chen"], "authorids": ["tengfei.ma1@ibm.com", "chenjie@us.ibm.com"], "authors_source": "OpenReview API", "abstract": "Hierarchical abstractions are a methodology for solving large-scale graph problems in various disciplines. Coarsening is one such approach: it generates a pyramid of graphs whereby the one in the next level is a structural summary of the prior one. With a long history in scientific computing, many coarsening strategies were developed based on mathematically driven heuristics. Recently, resurgent interests exist in deep learning to design hierarchical methods learnable through differentiable parameterization. These approaches are paired with downstream tasks for supervised learning. In this work, we propose an unsupervised approach, coined \\textsc{OTCoarsening}, with the use of optimal transport. Both the coarsening matrix and the transport cost matrix are parameterized, so that an optimal coarsening strategy can be learned and tailored for a given set of graphs. We demonstrate that the proposed approach produces meaningful coarse graphs and yields competitive performance compared with supervised methods for graph classification.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=Bkf4XgrKvS", "arxiv_id": "1912.11176", "paper_pdf": "papers/Bkf4XgrKvS.pdf", "paper_pdf_sha256": "55e4f939f73ecba04e48928a4cee48ae8c3d0b5a21ff929385f4df07214a8647", "paper_pdf_bytes": 449208, "paper_pdf_source": "openreview", "code_url": "https://github.com/matenure/OTCoarsening", "code_repository": "matenure/OTCoarsening", "code_commit": "ec047ed7a98087e81fcb6b8e7d3035fc4764e113", "code_archive": "repos/Bkf4XgrKvS.zip", "code_archive_sha256": "2d85e265cb0b6653605a13fbe5cf54b8e36d1b14c33d300814b55ca975203c47", "code_archive_bytes": 30304, "code_file_count": 10, "code_extensions": {".py": 10}, "github_disk_usage_kb": 39, "github_languages": {"Python": 52286}, "github_archived": false, "github_pushed_at": "2021-03-02T19:41:44Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/unsupervised-learning-of-graph-hierarchical-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Bklzkh0qFm", "year": 2019, "status": "rejected", "title": "Relational Graph Attention Networks", "authors": ["Dan Busbridge", "Dane Sherburn", "Pietro Cavallo", "Nils Y. Hammerla"], "authorids": ["dan.busbridge@gmail.com", "danesherbs@gmail.com", "p.cavallo85@gmail.com", "nils.hammerla@babylonhealth.com"], "authors_source": "OpenReview API", "abstract": "We investigate Relational Graph Attention Networks, a class of models that extends non-relational graph attention mechanisms to incorporate relational information, opening up these methods to a wider variety of problems. A thorough evaluation of these models is performed, and comparisons are made against established benchmarks. To provide a meaningful comparison, we retrain Relational Graph Convolutional Networks, the spectral counterpart of Relational Graph Attention Networks, and evaluate them under the same conditions. We find that Relational Graph Attention Networks perform worse than anticipated, although some configurations are marginally beneficial for modelling molecular properties. We provide insights as to why this may be, and suggest both modifications to evaluation strategies, as well as directions to investigate for future work.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=Bklzkh0qFm", "arxiv_id": "1904.05811", "paper_pdf": "papers/Bklzkh0qFm.pdf", "paper_pdf_sha256": "975457c8dc4e7f6c8fe7d0219928033c7bc4b3be6349f9b95d46503613c46d4b", "paper_pdf_bytes": 2240922, "paper_pdf_source": "openreview", "code_url": "https://github.com/babylonhealth/rgat", "code_repository": "babylonhealth/rgat", "code_commit": "79a49a123a3ac0c25436504182b16d91dac8eb60", "code_archive": "repos/Bklzkh0qFm.zip", "code_archive_sha256": "cd8a29ed6969e3ddd0153aa345ea85195d329d51688d1e8c01b5d570b823e812", "code_archive_bytes": 173971, "code_file_count": 15, "code_extensions": {".py": 15}, "github_disk_usage_kb": 163, "github_languages": {"Python": 76946}, "github_archived": false, "github_pushed_at": "2022-07-13T16:27:30Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/relational-graph-attention-networks-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "WqSEVQRlUB", "year": 2026, "status": "rejected", "title": "Tokenize Image as a Set", "authors": ["Zigang Geng", "Mengde Xu", "Han Hu", "Shuyang Gu"], "authorids": ["~Zigang_Geng1", "~Mengde_Xu1", "~Han_Hu1", "~Shuyang_Gu1"], "authors_source": "OpenReview API", "abstract": "This paper proposes a new paradigm for image generation through set-based tokenization and modeling. Unlike conventional methods that serialize images into fixed-position latent codes with a uniform compression ratio, we introduce an unordered token set representation to dynamically allocate coding capacity based on regional semantic complexity. This TokenSet enhances global context aggregation and improves robustness against local perturbations. To address the critical challenge of modeling discrete sets, we devise a dual transformation mechanism that bijectively converts sets into fixed-length integer sequences while preserving summation constraints. Further, we propose Fixed-Sum Discrete Diffusion—the first framework to simultaneously handle discrete values, fixed sequence length, and summation invariance—enabling effective set distribution modeling. Experiments demonstrate our method's superiority in semantic-aware representation and generation quality. Our innovations, spanning novel representation and modeling strategies, advance visual generation beyond traditional sequential token paradigms.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=WqSEVQRlUB", "arxiv_id": "2503.16425", "paper_pdf": "papers/WqSEVQRlUB.pdf", "paper_pdf_sha256": "48b0779bdeaab0c61de87de7c8642ae1df082d14ec689ef9c15abfe518fadc19", "paper_pdf_bytes": 12069518, "paper_pdf_source": "openreview", "code_url": "https://github.com/Gengzigang/TokenSet", "code_repository": "Gengzigang/TokenSet", "code_commit": "6d0fc53e9aa39dcb2f19cbbe2b5646b39b9b43f7", "code_archive": "repos/WqSEVQRlUB.zip", "code_archive_sha256": "e49bda4ebcd29837884bcfefff4a2273a4702a72caa055ace0d5dad5ec509401", "code_archive_bytes": 136893, "code_file_count": 9, "code_extensions": {".py": 9}, "github_disk_usage_kb": 131, "github_languages": {"Python": 82313}, "github_archived": false, "github_pushed_at": "2025-03-21T02:11:29Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/tokenize-image-as-a-set"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "0xUEBQV54B", "year": 2025, "status": "rejected", "title": "Large Language Monkeys: Scaling Inference Compute with Repeated Sampling", "authors": ["Bradley Brown", "Jordan Juravsky", "Ryan Saul Ehrlich", "Ronald Clark", "Quoc V Le", "Christopher Re", "Azalia Mirhoseini"], "authorids": ["~Bradley_Brown1", "~Jordan_Juravsky1", "~Ryan_Saul_Ehrlich1", "~Ronald_Clark2", "~Quoc_V_Le1", "~Christopher_Re1", "~Azalia_Mirhoseini3"], "authors_source": "OpenReview API", "abstract": "Scaling the amount of compute used to train language models has dramatically improved their capabilities. However, when it comes to inference, we often limit the amount of compute to only one attempt per problem. Here, we explore inference compute as another axis for scaling, using the simple technique of repeatedly sampling candidate solutions from a model. Across multiple tasks and models, we observe that coverage – the fraction of problems that are solved by any generated sample – scales with the number of samples over four orders of magnitude. Interestingly, the relationship between coverage and the number of samples is often log-linear and can be modelled with an exponentiated power law, suggesting the existence of inference-time scaling laws. In domains like coding and formal proofs, where answers can be automatically verified, these increases in coverage directly translate into improved performance. When we apply repeated sampling to SWE-bench Lite, the fraction of issues solved with DeepSeek-Coder-V2-Instruct increases from 15.9% with one sample to 56% with 250 samples, outperforming the single-sample state-of-the-art of 43%. In domains without automatic verifiers, we find that common methods for picking from a sample collection (majority voting and reward models) plateau beyond several hundred samples and fail to fully scale with the sample budget.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=0xUEBQV54B", "arxiv_id": "2407.21787", "paper_pdf": "papers/0xUEBQV54B.pdf", "paper_pdf_sha256": "9f3351a18bb10567ab64500bc630641732f0320233e171ae16579e80a5a1d577", "paper_pdf_bytes": 4521578, "paper_pdf_source": "openreview", "code_url": "https://github.com/ScalingIntelligence/large_language_monkeys", "code_repository": "ScalingIntelligence/large_language_monkeys", "code_commit": "ed2bdb06bcfcbda1ed7d9c2bff87c2db10c3ee78", "code_archive": "repos/0xUEBQV54B.zip", "code_archive_sha256": "40a582ea6b1fc2586a8b2742eeb07de2281d6c696281966226439aab494d88b9", "code_archive_bytes": 36856, "code_file_count": 21, "code_extensions": {".py": 21}, "github_disk_usage_kb": 30, "github_languages": {"Python": 72916, "Dockerfile": 768}, "github_archived": false, "github_pushed_at": "2024-09-25T21:28:38Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/large-language-monkeys-scaling-inference"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "XdSYtriYfI", "year": 2024, "status": "rejected", "title": "Federated Ensemble-Directed Offline Reinforcement Learning", "authors": ["Desik Rengarajan", "Nitin Ragothaman", "Dileep Kalathil", "Srinivas Shakkottai"], "authorids": ["~Desik_Rengarajan1", "~Nitin_Ragothaman1", "~Dileep_Kalathil1", "~Srinivas_Shakkottai1"], "authors_source": "OpenReview API", "abstract": "We consider the problem of federated offline reinforcement learning (RL), a scenario under which distributed learning agents must collaboratively learn a high-quality control policy only using small pre-collected datasets generated according to different unknown behavior policies. Naively combining a standard offline RL approach with a standard federated learning approach to solve this problem can lead to poorly performing policies.  In response, we develop the Federated Ensemble-Directed Offline Reinforcement Learning  Algorithm (FEDORA), which distills the collective wisdom of the clients using an ensemble learning approach.  We develop the FEDORA codebase to utilize distributed compute resources on a federated learning platform. We show that FEDORA significantly outperforms other approaches, including offline RL over the combined data pool, in various complex continuous control environments and real-world datasets. Finally, we demonstrate the performance of FEDORA in the real-world on a mobile robot.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=XdSYtriYfI", "arxiv_id": "2305.03097", "paper_pdf": "papers/XdSYtriYfI.pdf", "paper_pdf_sha256": "49176f17308865e93afc0d409c85f7f329bbf9ced58d91b309309618f56eaec1", "paper_pdf_bytes": 2173337, "paper_pdf_source": "openreview", "code_url": "https://github.com/DesikRengarajan/FEDORA", "code_repository": "DesikRengarajan/FEDORA", "code_commit": "3c355a867e0d38ab650fce18f78a742691afff0f", "code_archive": "repos/XdSYtriYfI.zip", "code_archive_sha256": "e4850d616fd553a2c0775040b3b94d6282022b6f4dd5208499ef82fef8fb77d4", "code_archive_bytes": 19874, "code_file_count": 10, "code_extensions": {".py": 9, ".sh": 1}, "github_disk_usage_kb": 29, "github_languages": {"Python": 44319, "Shell": 698}, "github_archived": false, "github_pushed_at": "2024-09-25T22:26:26Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/federated-ensemble-directed-offline"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ygYXtbb3og3", "year": 2023, "status": "rejected", "title": "Mutual Information Regularized Offline Reinforcement Learning", "authors": ["Xiao Ma", "Bingyi Kang", "Zhongwen Xu", "Min Lin", "Shuicheng YAN"], "authorids": ["~Xiao_Ma2", "~Bingyi_Kang1", "~Zhongwen_Xu1", "~Min_Lin1", "~Shuicheng_YAN3"], "authors_source": "OpenReview API", "abstract": "Offline reinforcement learning (RL) aims at learning an effective policy from offline datasets without active interactions with the environment. The major challenge of offline RL is the distribution shift that appears when out-of-distribution actions are queried, which makes the policy improvement direction biased by extrapolation errors. Most existing methods address this problem by penalizing the policy for deviating from the behavior policy during policy improvement or making conservative updates for value functions during policy evaluation. In this work, we propose a novel MISA framework to approach offline RL from the perspective of Mutual Information between States and Actions in the dataset by directly constraining the policy improvement direction. Intuitively, mutual information measures the mutual dependence of actions and states, which reflects how a behavior agent reacts to certain environment states during data collection. To effectively utilize this information to facilitate policy learning, MISA constructs lower bounds of mutual information parameterized by the policy and Q-values. We show that optimizing this lower bound is equivalent to maximizing the likelihood of a one-step improved policy on the offline dataset. In this way, we constrain the policy improvement direction to lie in the data manifold. The resulting algorithm simultaneously augments the policy evaluation and improvement by adding a mutual information regularization. MISA is a general offline RL framework that unifies conservative Q-learning (CQL) and behavior regularization methods (e.g., TD3+BC) as special cases. Our experiments show that MISA performs significantly better than existing methods and achieves new state-of-the-art on various tasks of the D4RL benchmark.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=ygYXtbb3og3", "arxiv_id": "2210.07484", "paper_pdf": "papers/ygYXtbb3og3.pdf", "paper_pdf_sha256": "1e9e9c98bcb97f24b45d4b6854cb4af21028bfd48f1afbd9a12f19a653915446", "paper_pdf_bytes": 742416, "paper_pdf_source": "openreview", "code_url": "https://github.com/sail-sg/MISA", "code_repository": "sail-sg/MISA", "code_commit": "bd350e5ee650ad0a212d80a85c02b6fadea9104a", "code_archive": "repos/ygYXtbb3og3.zip", "code_archive_sha256": "1d2b3d6ebb2fbeb23ff768c5d1bbd9b9662cf531aa9fa5a53fc8a50cbf4a1875", "code_archive_bytes": 38636, "code_file_count": 18, "code_extensions": {".py": 18}, "github_disk_usage_kb": 32, "github_languages": {"Python": 100673}, "github_archived": false, "github_pushed_at": "2022-10-19T03:38:53Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/mutual-information-regularized-offline-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "hW2kwAcXq5w", "year": 2022, "status": "rejected", "title": "Discriminator-Weighted Offline Imitation Learning from Suboptimal Demonstrations", "authors": ["Haoran Xu", "Xianyuan Zhan", "Honglei Yin", "Huiling Qin"], "authorids": ["~Haoran_Xu4", "~Xianyuan_Zhan1", "~Honglei_Yin1", "~Huiling_Qin1"], "authors_source": "OpenReview API", "abstract": "We study the problem of offline Imitation Learning (IL) where an agent aims to learn an optimal expert behavior policy without additional online environment interactions. Instead, the agent is provided with a static offline dataset of state-action-next state transition triples from both optimal and non-optimal expert behaviors. This strictly offline imitation learning problem arises in many real-world problems, where environment interactions and expert annotations are costly. Prior works that address the problem either require that expert data occupies the majority proportion of the offline dataset, or need to learn a reward function and perform offline reinforcement learning (RL) based on the learned reward function. In this paper, we propose an imitation learning algorithm to address the problem without additional steps of reward learning and offline RL training for the case when demonstrations containing large-proportion of suboptimal data. Built upon behavioral cloning (BC), we introduce an additional discriminator to distinguish expert and non-expert data, we propose a cooperation strategy to boost the performance of both tasks, this will result in a new policy learning objective and surprisingly, we find its equivalence to a generalized BC objective, where the outputs of discriminator serve as the weights of the BC loss function. Experimental results show that the proposed algorithm can learn behavior policies that are much closer to the optimal policies than policies learned by baseline algorithms.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=hW2kwAcXq5w", "arxiv_id": "2207.10050", "paper_pdf": "papers/hW2kwAcXq5w.pdf", "paper_pdf_sha256": "bb1da13f3b8789df5b81ba3c2b6b6c2bcf61740ff52ec2fb5e3ab39b8feff90c", "paper_pdf_bytes": 603665, "paper_pdf_source": "openreview", "code_url": "https://github.com/ryanxhr/DWBC", "code_repository": "ryanxhr/DWBC", "code_commit": "a1e9d8a068c478128dea0b07d55a375b064eab7e", "code_archive": "repos/hW2kwAcXq5w.zip", "code_archive_sha256": "9ab3af66ec9eaf72c2e131aa661558255e9c7af4c0e6b307f0303575e5978d39", "code_archive_bytes": 10563, "code_file_count": 6, "code_extensions": {".py": 5, ".sh": 1}, "github_disk_usage_kb": 30, "github_languages": {"Python": 29638, "Shell": 1208}, "github_archived": false, "github_pushed_at": "2023-01-05T05:02:20Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/discriminator-weighted-offline-imitation-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "thhdrl4IdMm", "year": 2021, "status": "rejected", "title": "A Chain Graph Interpretation of Real-World Neural Networks", "authors": ["Yuesong Shen", "Daniel Cremers"], "authorids": ["~Yuesong_Shen1", "~Daniel_Cremers1"], "authors_source": "OpenReview API", "abstract": "The last decade has witnessed a boom of deep learning research and applications achieving state-of-the-art results in various domains. However, most advances have been established empirically, and their theoretical analysis remains lacking. One major issue is that our current interpretation of neural networks (NNs) as function approximators is too generic to support in-depth analysis. In this paper, we remedy this by proposing an alternative interpretation that identifies NNs as chain graphs (CGs) and feed-forward as an approximate inference procedure. The CG interpretation specifies the nature of each NN component within the rich theoretical framework of probabilistic graphical models, while at the same time remains general enough to cover real-world NNs with arbitrary depth, multi-branching and varied activations, as well as common structures including convolution / recurrent layers, residual block and dropout. We demonstrate with concrete examples that the CG interpretation can provide novel theoretical support and insights for various NN techniques, as well as derive new deep learning approaches such as the concept of partially collapsed feed-forward inference. It is thus a promising framework that deepens our understanding of neural networks and provides a coherent theoretical formulation for future deep learning research.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=thhdrl4IdMm", "arxiv_id": "2006.16856", "paper_pdf": "papers/thhdrl4IdMm.pdf", "paper_pdf_sha256": "ba1f08c018193d5b9d64cdb2ec70bf73ac92238ca133ea054367efdea9bb44ae", "paper_pdf_bytes": 591560, "paper_pdf_source": "openreview", "code_url": "https://github.com/tum-vision/nnascg", "code_repository": "tum-vision/nnascg", "code_commit": "9eecd523d2a4d327144a33b3cadd5e3d5de4b734", "code_archive": "repos/thhdrl4IdMm.zip", "code_archive_sha256": "48fde91eb584b40b7e2a582ba3b3e42c0e2958c6fcff9e58d29b2f0172db5e96", "code_archive_bytes": 40424, "code_file_count": 11, "code_extensions": {".py": 9, ".sh": 2}, "github_disk_usage_kb": 34, "github_languages": {"Python": 79755, "Shell": 2616}, "github_archived": false, "github_pushed_at": "2020-07-10T15:41:03Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/deriving-neural-network-design-and-learning"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "rylxpA4YwH", "year": 2020, "status": "rejected", "title": "On the Evaluation of Conditional GANs", "authors": ["Terrance DeVries", "Adriana Romero", "Luis Pineda", "Graham W. Taylor", "Michal Drozdzal"], "authorids": ["terrance@uoguelph.ca", "adrianars@fb.com", "lep@fb.com", "gwtaylor@uoguelph.ca", "mdrozdzal@fb.com"], "authors_source": "OpenReview API", "abstract": "Conditional Generative Adversarial Networks (cGANs) are finding increasingly widespread use in many application domains. Despite outstanding progress, quantitative evaluation of such models often involves multiple distinct metrics to assess different desirable properties, such as image quality, conditional consistency, and intra-conditioning diversity. In this setting, model benchmarking becomes a challenge, as each metric may indicate a different \"best\" model. In this paper, we propose the Frechet Joint Distance (FJD), which is defined as the Frechet distance between joint distributions of images and conditioning, allowing it to implicitly capture the aforementioned properties in a single metric. We conduct proof-of-concept experiments on a controllable synthetic dataset, which consistently highlight the benefits of FJD when compared to currently established metrics. Moreover, we use the newly introduced metric to compare existing cGAN-based models for a variety of conditioning modalities (e.g. class labels, object masks, bounding boxes, images, and text captions). We show that FJD can be used as a promising single metric for model benchmarking.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=rylxpA4YwH", "arxiv_id": "1907.08175", "paper_pdf": "papers/rylxpA4YwH.pdf", "paper_pdf_sha256": "86cc696a9e3fdf67be1d6111386c0312321a2852cf24b1d8ddf2436087972998", "paper_pdf_bytes": 8542957, "paper_pdf_source": "openreview", "code_url": "https://github.com/facebookresearch/fjd", "code_repository": "facebookresearch/fjd", "code_commit": "6e48d77162b16a20ebaea832470cfa25eb2021ef", "code_archive": "repos/rylxpA4YwH.zip", "code_archive_sha256": "2d0a735da3cd716997d21f6ac6b4202f2939c9d270cbbb95fb05c003ea9d8fe6", "code_archive_bytes": 46694, "code_file_count": 11, "code_extensions": {".py": 10, ".ipynb": 1}, "github_disk_usage_kb": 42, "github_languages": {"Python": 78642, "Jupyter Notebook": 27554}, "github_archived": true, "github_pushed_at": "2019-11-27T17:57:57Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/on-the-evaluation-of-conditional-gans"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "BkMWx309FX", "year": 2019, "status": "rejected", "title": "Reinforcement Learning with Perturbed Rewards", "authors": ["Jingkang Wang", "Yang Liu", "Bo Li"], "authorids": ["wangjksjtu_01@sjtu.edu.cn", "yangliu@ucsc.edu", "lxbosky@gmail.com"], "authors_source": "OpenReview API", "abstract": "Recent studies have shown the vulnerability of reinforcement learning (RL) models in noisy settings. The sources of noises differ across scenarios. For instance, in practice, the observed reward channel is often subject to noise (e.g., when observed rewards are collected through sensors), and thus observed rewards may not be credible as a result. Also, in applications such as robotics, a deep reinforcement learning (DRL) algorithm can be manipulated to produce arbitrary errors. In this paper, we consider noisy RL problems where observed rewards by RL agents are generated with a reward confusion matrix. We call such observed rewards as perturbed rewards. We develop an unbiased reward estimator aided robust RL framework that enables RL agents to learn in noisy environments while observing only perturbed rewards. Our framework draws upon approaches for supervised learning with noisy data. The core ideas of our solution include estimating a reward confusion matrix and defining a set of unbiased surrogate rewards. We prove the convergence and sample complexity of our approach. Extensive experiments on different DRL platforms show that policies based on our estimated surrogate reward can achieve higher expected rewards, and converge faster than existing baselines. For instance, the state-of-the-art PPO algorithm is able to obtain 67.5% and 46.7% improvements in average on five Atari games, when the error rates are 10% and 30% respectively. ", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=BkMWx309FX", "arxiv_id": "1810.01032", "paper_pdf": "papers/BkMWx309FX.pdf", "paper_pdf_sha256": "f88345c5ee3b037231cf85c768753d4cdaf0d61507eb4108601d785a3e4b01c4", "paper_pdf_bytes": 7436475, "paper_pdf_source": "openreview", "code_url": "https://github.com/wangjksjtu/rl-perturbed-reward", "code_repository": "wangjksjtu/rl-perturbed-reward", "code_commit": "c894ca5dcdeadd0a0907770bb093b703092e1da1", "code_archive": "repos/BkMWx309FX.zip", "code_archive_sha256": "d839df87bf77f18b3f67e81d3a60f03804bd8cf1e9c1f93f6ac9f7ea4de36e74", "code_archive_bytes": 177778, "code_file_count": 99, "code_extensions": {".py": 84, ".sh": 15}, "github_disk_usage_kb": 164, "github_languages": {"Python": 477858, "Shell": 76269}, "github_archived": false, "github_pushed_at": "2024-08-02T16:11:32Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/reinforcement-learning-with-perturbed-rewards"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "r1YqWz-R-", "year": 2018, "status": "rejected", "title": "Improving Conditional Sequence Generative Adversarial Networks by Stepwise Evaluation", "authors": ["Yi-Lin Tuan", "Hung-yi Lee"], "authorids": ["pascaltuan@gmail.com", "hungyilee@ntu.edu.tw"], "authors_source": "OpenReview API", "abstract": "Conditional sequence generation is a widely researched topic. One of the most important tasks is dialogue generation, which is composed of input-output pairs with the one-to-many property. Given the recent success of generative adversarial networks (GANs), GANs have been used for sequence generation. However, there is still limited work of its application on conditional sequence generation. We investigate the influence of GAN on conditional sequence generation with three artificial grammars and  dialogue generation. Moreover, we propose stepwise GAN (StepGAN) for conditional sequence generation, which predicts the reward at each time-step. StepGAN can be seen as the general version of SeqGAN. It estimates the expected returns predicted by Monte-Carlo Search in SeqGAN, but it has a lower computational cost than Monte-Carlo Search. Experimental results show that stepwise GAN can outperform other state-of-the-art algorithms in most tasks.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=r1YqWz-R-", "arxiv_id": "1808.05599", "paper_pdf": "papers/r1YqWz-R-.pdf", "paper_pdf_sha256": "6be0b57413bf711934e2ab5bbbf1d3a34f115a4e533c490b798d6fe8d3f62ca0", "paper_pdf_bytes": 862427, "paper_pdf_source": "openreview", "code_url": "https://github.com/Pascalson/Conditional-Seq-GANs", "code_repository": "Pascalson/Conditional-Seq-GANs", "code_commit": "6e7594a7c21826b14d34972b085478e3b3010375", "code_archive": "repos/r1YqWz-R-.zip", "code_archive_sha256": "9e5334f373c9f234803cc14a0fdf4d189ec65bbd7697228af1aed29dc9154ad1", "code_archive_bytes": 222557, "code_file_count": 11, "code_extensions": {".py": 10, ".sh": 1}, "github_disk_usage_kb": 227, "github_languages": {"Python": 97581, "Shell": 2139}, "github_archived": false, "github_pushed_at": "2019-07-13T11:18:56Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/improving-conditional-sequence-generative"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "lsLb0WvlqA", "year": 2026, "status": "rejected", "title": "Efficient Attention via Pre-Scoring: Prioritizing Informative Keys in Transformers", "authors": ["Yutong Bao", "Haoyu Wang", "Zhexiang Li", "David Woodruff"], "authorids": ["~Yutong_Bao1", "~Haoyu_Wang57", "~Zhexiang_Li1", "~David_Woodruff1"], "authors_source": "OpenReview API", "abstract": "Recent advances in transformer architectures deeply enhanced long-context language modeling. Among them, HyperAttention achieves competitive efficiency by combining a single-level LSH-based clustering with uniform residual sampling. However, HyperAttention fails to find all significant keys, which in turn raises the overall perplexity. We propose a pre-scoring mechanism that prioritizes significant keys before applying HyperAttention. We introduce three scoring methods: $k$-means and kernel $k$-means clustering, $k$-median clustering, and leverage score-based ranking (inspired by LevAttention) to filter keys effectively. We further replace HyperAttention's original uniform residual sampling, relying exclusively on our pre-scoring mechanism. Experiments on ChatGLM2 (131k token context) reduce perplexity from 12 to 8.3, which outperforms standard HyperAttention. Moreover, when running on the Vision-Transformer (ViT), our method shows that it can guarantee similar accuracy compared with LevAttention, and will surpass LevAttention given specific parameters. Although this method introduces some computational overhead, its combination with HyperAttention remains 20 times faster than FlashAttention, providing a balanced trade-off between speed and modeling accuracy. Our results highlight the effectiveness of integrating pre-scoring into hierarchical attention mechanisms, significantly improving transformer efficiency.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=lsLb0WvlqA", "arxiv_id": "2505.11040", "paper_pdf": "papers/lsLb0WvlqA.pdf", "paper_pdf_sha256": "f12242204e300c8735d1d6bd574343ec71c3ca2f6ded69722013d4c9877d1767", "paper_pdf_bytes": 602507, "paper_pdf_source": "openreview", "code_url": "https://github.com/BruceLZX/prescored-transformer", "code_repository": "BruceLZX/prescored-transformer", "code_commit": "87e83d8a060d476b72e4bc2c81e3689e8c8d2645", "code_archive": "repos/lsLb0WvlqA.zip", "code_archive_sha256": "126265f0d03425eece58038d3da8a5a63b7560c6f741f828805e3dcd979c1da6", "code_archive_bytes": 133948, "code_file_count": 12, "code_extensions": {".py": 12}, "github_disk_usage_kb": 139, "github_languages": {"Python": 126956}, "github_archived": false, "github_pushed_at": "2026-04-25T06:06:33Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/2505-11040"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "NnwDdPDwUq", "year": 2025, "status": "rejected", "title": "Contextual Bandits with Entropy-based Human Feedback", "authors": ["Raihan Seraj", "Tristan Sylvain", "Lili Meng"], "authorids": ["~Raihan_Seraj1", "~Tristan_Sylvain2", "~Lili_Meng2"], "authors_source": "OpenReview API", "abstract": "In recent years, preference-based human feedback mechanisms have become integral to improving model performance across a range of applications, including conversational AI systems like ChatGPT. However, existing methodologies often overlook critical factors such as model uncertainty and variability in feedback quality. To address these limitations, we propose an innovative entropy-based human feedback framework designed for contextual bandits, which balances exploration and exploitation by soliciting expert feedback when model entropy surpasses a predefined threshold. Our method is model-agnostic and adaptable to any contextual bandit agent employing stochastic policies. Through rigorous experimentation, we demonstrate that our approach requires minimal human feedback to achieve significant performance gains, even with suboptimal feedback quality. Our work not only introduces a novel feedback solicitation strategy but also underscores the robustness of integrating human guidance into machine learning systems. Our code is publicly available: \\url{https://anonymous.4open.science/r/CBHF-33C5}", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=NnwDdPDwUq", "arxiv_id": "2502.08759", "paper_pdf": "papers/NnwDdPDwUq.pdf", "paper_pdf_sha256": "c9b7029fa311897cb7ca0dd4978772ad01f733d4b292dd4d24b0a9ef48a6a5e1", "paper_pdf_bytes": 1401169, "paper_pdf_source": "openreview", "code_url": "https://github.com/BorealisAI/CBHF", "code_repository": "BorealisAI/CBHF", "code_commit": "46fdc310b814fd6e0850804c1800fe386d8acca4", "code_archive": "repos/NnwDdPDwUq.zip", "code_archive_sha256": "f3b6ce069637d1176d7f7aff45bfb83cc97e6e37b63c7ec88983102525f4ca8d", "code_archive_bytes": 43640, "code_file_count": 30, "code_extensions": {".py": 21, ".sh": 9}, "github_disk_usage_kb": 32, "github_languages": {"Python": 106712, "Shell": 6517}, "github_archived": false, "github_pushed_at": "2024-05-31T20:35:32Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/contextual-bandits-with-entropy-based-human"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "PIl69UIAWL", "year": 2024, "status": "rejected", "title": "GraphLLM: Boosting Graph Reasoning Ability of Large Language Model", "authors": ["Ziwei Chai", "Tianjie Zhang", "Liang Wu", "Kaiqiao Han", "Xiaohai Hu", "Xuanwen Huang", "Yang Yang"], "authorids": ["~Ziwei_Chai1", "~Tianjie_Zhang1", "~Liang_Wu8", "~Kaiqiao_Han1", "~Xiaohai_Hu1", "~Xuanwen_Huang1", "~Yang_Yang35"], "authors_source": "OpenReview API", "abstract": "The advancement of Large Language Models (LLMs) has remarkably pushed the boundaries towards artificial general intelligence (AGI), with their exceptional ability on understanding diverse types of information,  including but not limited to images and audio.  Despite this progress, a critical gap remains in empowering LLMs to proficiently understand and reason on graph data. Recent studies underscore LLMs' underwhelming performance on fundamental graph reasoning tasks.  In this paper, we endeavor to unearth the obstacles that impede LLMs in graph reasoning, pinpointing the common practice of converting graphs into natural language descriptions (Graph2Text) as a fundamental bottleneck. To overcome this impediment, we introduce GraphLLM, a pioneering end-to-end approach that synergistically integrates graph learning models with LLMs. This synergy equips LLMs with the ability to proficiently interpret and reason on graph data, harnessing the superior expressive power of graph learning models. Our empirical evaluations across four fundamental graph reasoning tasks validate the effectiveness of GraphLLM. The results exhibit a substantial average accuracy enhancement of 54.44%, alongside a noteworthy context reduction of 96.45% across various graph reasoning tasks.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=PIl69UIAWL", "arxiv_id": "2310.05845", "paper_pdf": "papers/PIl69UIAWL.pdf", "paper_pdf_sha256": "c5928f6df18b1fb0790d7b1a2e0ccf8376d103fdcddcc4d21ea92469502f4e69", "paper_pdf_bytes": 1404711, "paper_pdf_source": "openreview", "code_url": "https://github.com/mistyreed63849/Graph-LLM", "code_repository": "mistyreed63849/Graph-LLM", "code_commit": "ceeba92039dd0066eb808861cc5e3c7e2232a167", "code_archive": "repos/PIl69UIAWL.zip", "code_archive_sha256": "3f7dd940eeca6185ac1c1855651d6d861b4849d523f61415865294c1221668e2", "code_archive_bytes": 24705, "code_file_count": 15, "code_extensions": {".py": 11, ".sh": 4}, "github_disk_usage_kb": 29, "github_languages": {"Python": 92814, "Shell": 458}, "github_archived": false, "github_pushed_at": "2026-01-29T00:17:59Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/graphllm-boosting-graph-reasoning-ability-of"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "usa87QW3_r9", "year": 2023, "status": "rejected", "title": "Everyone's Preference Changes Differently: Weighted Multi-Interest Retrieval Model", "authors": ["Hui Shi", "Yupeng Gu", "Yitong Zhou", "Bo Zhao", "Sicun Gao", "Jishen Zhao"], "authorids": ["~Hui_Shi3", "~Yupeng_Gu1", "~Yitong_Zhou1", "bozhao@pinterest.com", "~Sicun_Gao1", "~Jishen_Zhao1"], "authors_source": "OpenReview API", "abstract": "User embeddings (vectorized representations of a user) are essential in recommendation systems. Numerous approaches have been proposed to construct a representation for the user in order to find similar items for retrieval tasks, and they have been proven effective in industrial recommendation systems. Recently people have discovered the power of using multiple embeddings to represent a user, with the hope that each embedding represents the user's interest in a certain topic. With multi-interest representation, it's important to model the user's preference over the different topics and how the preference change with time. However, existing approaches either fail to estimate the user's affinity to each interest or unreasonably assume every interest of every user fades with an equal rate with time, thus hurting the performance of candidate retrieval. In this paper, we propose the Multi-Interest Preference (MIP) model, an approach that not only produces multi-interest for users by using the user's sequential engagement more effectively but also automatically learns a set of weights to represent the preference over each embedding so that the candidates can be retrieved from each interest proportionally. Extensive experiments have been done on various industrial-scale datasets to demonstrate the effectiveness of our approach. ", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=usa87QW3_r9", "arxiv_id": "2207.06652", "paper_pdf": "papers/usa87QW3_r9.pdf", "paper_pdf_sha256": "67dfc593e3fe9ef84908c0953a71dac45e1fc774012b3e40b8b9e4c990f20819", "paper_pdf_bytes": 835789, "paper_pdf_source": "openreview", "code_url": "https://github.com/shihui2010/MIP", "code_repository": "shihui2010/MIP", "code_commit": "891c071c298cc34faaabd4dc33f34e9cac9d8015", "code_archive": "repos/usa87QW3_r9.zip", "code_archive_sha256": "a75ec2c6afbf94aefbc65d20162669fb80f5a71ee2b0067946eba8466d45efc7", "code_archive_bytes": 30293, "code_file_count": 9, "code_extensions": {".py": 9}, "github_disk_usage_kb": 32, "github_languages": {"Python": 74802}, "github_archived": false, "github_pushed_at": "2023-03-19T04:53:17Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/every-preference-changes-differently-neural"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "kcrIligNnl", "year": 2022, "status": "rejected", "title": "Direct Molecular Conformation Generation", "authors": ["Jinhua Zhu", "Yingce Xia", "Chang Liu", "Lijun Wu", "Shufang Xie", "Wengang Zhou", "Tao Qin", "Houqiang Li", "Tie-Yan Liu"], "authorids": ["~Jinhua_Zhu1", "~Yingce_Xia1", "~Chang_Liu10", "~Lijun_Wu1", "~Shufang_Xie1", "~Wengang_Zhou1", "~Tao_Qin1", "~Houqiang_Li1", "~Tie-Yan_Liu1"], "authors_source": "OpenReview API", "abstract": "Molecular conformation generation, which is to generate 3 dimensional coordinates of all the atoms in a molecule, is an important task for bioinformatics and pharmacology. Most existing machine learning based methods first predict interatomic distances and then generate conformations based on them. This two-stage approach has a potential limitation that the predicted distances may conflict with each other, e.g., violating the triangle inequality. In this work, we propose a method that directly outputs the coordinates of atoms, so that there is no violation of constraints. The conformation generator of our method stacks multiple blocks, and each block outputs a conformation which is then refined by the following block.  We adopt the variational auto-encoder (VAE) framework and use a latent variable to generate diverse conformations. To handle the roto-translation equivariance, we adopt a loss that is invariant to rotation and translation of molecule coordinates, by computing the minimal achievable distance after any rotation and translation. Our method outperforms strong baselines on four public datasets, which shows the effectiveness of our method and the great potential of the direct approach. The code is released at \\url{https://github.com/DirectMolecularConfGen/DMCG}. ", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=kcrIligNnl", "arxiv_id": "2202.01356", "paper_pdf": "papers/kcrIligNnl.pdf", "paper_pdf_sha256": "72643a536ccc9d646d84a8fca47e0501f176064d4ae067db15590e91c66d11c2", "paper_pdf_bytes": 1071282, "paper_pdf_source": "openreview", "code_url": "https://github.com/DirectMolecularConfGen/DMCG", "code_repository": "DirectMolecularConfGen/DMCG", "code_commit": "23e90f7231eac9a065093ee44b378cbf65181a7c", "code_archive": "repos/kcrIligNnl.zip", "code_archive_sha256": "cb4c76c77be2da4eb8133a9d3649a05f697e2b5f839fbb804ec7fe5b19133d80", "code_archive_bytes": 33308, "code_file_count": 12, "code_extensions": {".py": 11, ".sh": 1}, "github_disk_usage_kb": 31, "github_languages": {"Python": 123858, "Dockerfile": 1829, "Shell": 1548}, "github_archived": false, "github_pushed_at": "2022-10-25T09:15:40Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/direct-molecular-conformation-generation-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "hcCao_UYd6O", "year": 2021, "status": "rejected", "title": "Adversarial Feature Desensitization", "authors": ["Pouya Bashivan", "Mojtaba Faramarzi", "Touraj Laleh", "Blake Aaron Richards", "Irina Rish"], "authorids": ["~Pouya_Bashivan1", "~Mojtaba_Faramarzi1", "~Touraj_Laleh1", "~Blake_Aaron_Richards1", "~Irina_Rish1"], "authors_source": "OpenReview API", "abstract": "Deep neural networks can now perform many tasks that were once thought to be only feasible for humans. While reaching impressive performance under standard settings, such networks are known to be  susceptible to  adversarial attacks -- slight but carefully constructed perturbations of the inputs which drastically decrease the network performance. Here we propose a new way to improve the network robustness against adversarial attacks by   focusing on robust representation learning based on  adversarial training procedure, called here Adversarial Feature Desensitization (AFD). AFD desensitizes the representation via an adversarial game between the embedding network and an adversarial discriminator introduced on top of  the standard predictive model, which is trained to distinguish between the clean and perturbed inputs from their high-level representations. Our method substantially improves the state-of-the-art in robust classification on MNIST, CIFAR10, and CIFAR100 datasets. More importantly, we demonstrate that AFD has better generalization ability than previous methods, as the learned features maintain their robustness across a wide range of perturbations, including perturbations not seen during training. These results indicate that reducing feature sensitivity is a promising approach for ameliorating the problem of adversarial attacks in deep neural networks. ", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=hcCao_UYd6O", "arxiv_id": "2006.04621", "paper_pdf": "papers/hcCao_UYd6O.pdf", "paper_pdf_sha256": "035e2a462e76b0f89f6d5e4f1c75bfbb4407c0161bd1d8da4356ddbe98637a4b", "paper_pdf_bytes": 11877574, "paper_pdf_source": "openreview", "code_url": "https://github.com/BashivanLab/afd", "code_repository": "BashivanLab/afd", "code_commit": "7631421528c9b7378863a25d06bb7b4e8f62419e", "code_archive": "repos/hcCao_UYd6O.zip", "code_archive_sha256": "f0ad78dd1159ee9217477a79b46f0590357340b078b990edf3891a8d9c84705f", "code_archive_bytes": 34568, "code_file_count": 10, "code_extensions": {".py": 9, ".ipynb": 1}, "github_disk_usage_kb": 36, "github_languages": {"Python": 72082, "Jupyter Notebook": 13929}, "github_archived": false, "github_pushed_at": "2022-01-20T01:08:17Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/adversarial-feature-desensitization"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "HJxWl0NKPB", "year": 2020, "status": "rejected", "title": "Combining MixMatch and Active Learning for Better Accuracy with Fewer Labels", "authors": ["Shuang Song", "David Berthelot", "Afshin Rostamizadeh"], "authorids": ["shuangsong@google.com", "dberth@google.com", "rostami@google.com"], "authors_source": "OpenReview API", "abstract": "We propose using active learning based techniques to further improve the state-of-the-art semi-supervised learning MixMatch algorithm. We provide a thorough empirical evaluation of several active-learning and baseline methods, which successfully demonstrate a significant improvement on the benchmark CIFAR-10, CIFAR-100, and SVHN datasets (as much as 1.5% in absolute accuracy). \nWe also provide an empirical analysis of the cost trade-off between incrementally gathering more labeled versus unlabeled data. This analysis can be used to measure the relative value of labeled/unlabeled data at different points of the learning curve, where we find that although the incremental value of labeled data can be as much as 20x that of unlabeled, it quickly diminishes to less than 3x once more than 2,000 labeled example are observed.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=HJxWl0NKPB", "arxiv_id": "1912.00594", "paper_pdf": "papers/HJxWl0NKPB.pdf", "paper_pdf_sha256": "ee5ed0a00519ee232b60fa2f28551565b5d8a4b81207e198945aa3b7d65ac3a8", "paper_pdf_bytes": 370563, "paper_pdf_source": "openreview", "code_url": "https://github.com/google-research/mma", "code_repository": "google-research/mma", "code_commit": "7d08be0837869db83048168be0a6a4cff0f0fcc6", "code_archive": "repos/HJxWl0NKPB.zip", "code_archive_sha256": "bb4ef8e628ea16b3d3d7e505bcfaf02bb10b913506c03ed5e7bddc669b9c49c3", "code_archive_bytes": 44956, "code_file_count": 9, "code_extensions": {".py": 9}, "github_disk_usage_kb": 42, "github_languages": {"Python": 80004}, "github_archived": true, "github_pushed_at": "2023-03-24T22:13:31Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/combining-mixmatch-and-active-learning-for-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "r1esnoAqt7", "year": 2019, "status": "rejected", "title": "Morpho-MNIST: Quantitative Assessment and Diagnostics for Representation Learning", "authors": ["Daniel C. Castro", "Jeremy Tan", "Bernhard Kainz", "Ender Konukoglu", "Ben Glocker"], "authorids": ["d.coelho-de-castro15@imperial.ac.uk", "j.tan17@imperial.ac.uk", "b.kainz@imperial.ac.uk", "kender@vision.ee.ethz.ch", "b.glocker@imperial.ac.uk"], "authors_source": "OpenReview API", "abstract": "Revealing latent structure in data is an active field of research, having introduced exciting technologies such as variational autoencoders and adversarial networks, and is essential to push machine learning towards unsupervised knowledge discovery. However, a major challenge is the lack of suitable benchmarks for an objective and quantitative evaluation of learned representations. To address this issue we introduce Morpho-MNIST, a framework that aims to answer: \"to what extent has my model learned to represent specific factors of variation in the data?\" We extend the popular MNIST dataset by adding a morphometric analysis enabling quantitative comparison of trained models, identification of the roles of latent variables, and characterisation of sample diversity. We further propose a set of quantifiable perturbations to assess the performance of unsupervised and supervised methods on challenging tasks such as outlier detection and domain adaptation.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=r1esnoAqt7", "arxiv_id": "1809.10780", "paper_pdf": "papers/r1esnoAqt7.pdf", "paper_pdf_sha256": "68a5af606e5b1da44f70f70a0bd5a25a291b3d6151bb3f8822e98d5284400166", "paper_pdf_bytes": 3645792, "paper_pdf_source": "openreview", "code_url": "https://github.com/dccastro/Morpho-MNIST", "code_repository": "dccastro/Morpho-MNIST", "code_commit": "3a20ed3b4a9b80ed0e01d6bf7d9f2b8b896a5ffe", "code_archive": "repos/r1esnoAqt7.zip", "code_archive_sha256": "d4d6936c394dc2a50681fc5bf50fd5ca9cae61b179c039adcb0d7d2bf9cc3480", "code_archive_bytes": 96149, "code_file_count": 40, "code_extensions": {".py": 40}, "github_disk_usage_kb": 168, "github_languages": {"Python": 139927}, "github_archived": false, "github_pushed_at": "2025-01-23T17:14:29Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/morpho-mnist-quantitative-assessment-and"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "BJ6anzb0Z", "year": 2018, "status": "rejected", "title": "Multimodal Sentiment Analysis To Explore the Structure of Emotions", "authors": ["Anthony Hu", "Seth Flaxman"], "authorids": ["anthony.hu@stats.ox.ac.uk", "s.flaxman@imperial.ac.uk"], "authors_source": "OpenReview API", "abstract": "We propose a novel approach to multimodal sentiment analysis using deep neural\nnetworks combining visual recognition and natural language processing. Our\ngoal is different than the standard sentiment analysis goal of predicting whether\na sentence expresses positive or negative sentiment; instead, we aim to infer the\nlatent emotional state of the user. Thus, we focus on predicting the emotion word\ntags attached by users to their Tumblr posts, treating these as “self-reported emotions.”\nWe demonstrate that our multimodal model combining both text and image\nfeatures outperforms separate models based solely on either images or text. Our\nmodel’s results are interpretable, automatically yielding sensible word lists associated\nwith emotions. We explore the structure of emotions implied by our model\nand compare it to what has been posited in the psychology literature, and validate\nour model on a set of images that have been used in psychology studies. Finally,\nour work also provides a useful tool for the growing academic study of images—\nboth photographs and memes—on social networks.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=BJ6anzb0Z", "arxiv_id": "1805.10205", "paper_pdf": "papers/BJ6anzb0Z.pdf", "paper_pdf_sha256": "3d370dc8f8cd9fd4d0cd445f41bd470b2c7febef07b3e297913bb31439fb30aa", "paper_pdf_bytes": 3934971, "paper_pdf_source": "openreview", "code_url": "https://github.com/deepsentiment/deepsentiment", "code_repository": "deepsentiment/deepsentiment", "code_commit": "620fc7175fc7cedfe13ab9c59d48fcfdb8c47ec8", "code_archive": "repos/BJ6anzb0Z.zip", "code_archive_sha256": "971ef646ceadd63003b95f6bc5cf66dd160caf602b781cd4a61e96c794cfb3ba", "code_archive_bytes": 334694, "code_file_count": 73, "code_extensions": {".py": 67, ".sh": 5, ".ipynb": 1}, "github_disk_usage_kb": 275, "github_languages": {"Python": 667877, "Jupyter Notebook": 46257, "Shell": 10688}, "github_archived": false, "github_pushed_at": "2017-10-27T23:02:48Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/multimodal-sentiment-analysis-to-explore-the"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "HSWE9aceZb", "year": 2026, "status": "rejected", "title": "CXMArena: Unified Dataset to benchmark performance in realistic CXM Scenarios", "authors": ["Raghav Garg", "Karan Gupta", "Kapil Sharma"], "authorids": ["~Raghav_Garg1", "~Karan_Gupta4", "~Kapil_Sharma1"], "authors_source": "OpenReview API", "abstract": "Large Language Models (LLMs) hold immense potential for revolutionizing Customer Experience Management (CXM), particularly in contact center operations. However, evaluating their practical utility in complex operational environments is hindered by data scarcity (due to privacy concerns) and the limitations of current benchmarks. Existing benchmarks often lack realism, failing to incorporate deep knowledge base (KB) integration, real-world noise, or critical operational tasks beyond conversational fluency.\n\nTo bridge this gap, we introduce CXMdataset, a novel, large-scale synthetic benchmark dataset specifically designed for evaluating AI in operational CXM contexts. Given the diversity in possible contact center features, we have developed a scalable LLM-powered pipeline that simulates the brand's CXM entities that form the foundation of our datasets—such as knowledge articles including product specifications, issue taxonomies, and contact center conversations. The entities closely represent real-world distribution because of controlled noise injection (informed by domain experts) and rigorous automated validation.\n\nBuilding on this, we release CXMdataset, which provides dedicated benchmarks targeting five important operational tasks: Knowledge Base Refinement, Intent Prediction, Agent Quality Adherence, Article Search, and Multi-turn RAG with Integrated Tools.\n\nOur baseline experiments underscore the benchmark's difficulty: even state-of-the-art embedding and generation models achieve only 68% accuracy on article search, while standard embedding methods yield a low F1 score of 0.3 for knowledge base refinement, highlighting significant challenges for current models necessitating complex pipelines and solutions over conventional techniques.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=HSWE9aceZb", "arxiv_id": "2505.09436", "paper_pdf": "papers/HSWE9aceZb.pdf", "paper_pdf_sha256": "241723888c50213be7e73a6e66eaa0c7e2abd673ddd88cc1708ada59dd56f46f", "paper_pdf_bytes": 2084781, "paper_pdf_source": "openreview", "code_url": "https://github.com/kapilsprinklr/CXMArena", "code_repository": "kapilsprinklr/CXMArena", "code_commit": "755a7c46976d0446aa30c0b210846771f4391bb4", "code_archive": "repos/HSWE9aceZb.zip", "code_archive_sha256": "ad625e4d2ecc7be309a964b480b69a2ba87fbe941c1a5240b258df46e349eb47", "code_archive_bytes": 62764, "code_file_count": 7, "code_extensions": {".py": 7}, "github_disk_usage_kb": 140, "github_languages": {"Python": 50274}, "github_archived": false, "github_pushed_at": "2026-09-01T11:09:42Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/cxmarena-unified-dataset-to-benchmark"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "F0Zd3knG9j", "year": 2025, "status": "rejected", "title": "How transformers learn structured data: insights from hierarchical filtering", "authors": ["Jerome Garnier-Brun", "Marc Mezard", "Emanuele Moscato", "Luca Saglietti"], "authorids": ["~Jerome_Garnier-Brun1", "~Marc_Mezard1", "~Emanuele_Moscato1", "~Luca_Saglietti1"], "authors_source": "OpenReview API", "abstract": "Understanding the learning process and the embedded computation in transformers is becoming a central goal for the development of interpretable AI. In the present study, we introduce a hierarchical filtering procedure for generative models of sequences on trees, allowing us to hand-tune the range of positional correlations in the data. Leveraging this controlled setting, we provide evidence that vanilla encoder-only transformers can approximate the exact inference algorithm when trained on root classification and masked language modeling tasks, and study *how* this computation is discovered and implemented. We find that correlations at larger distances, corresponding to increasing layers of the hierarchy, are sequentially included by the network during training. Moreover, by comparing attention maps from models trained with varying degrees of filtering and by probing the different encoder levels, we find clear evidence of a reconstruction of correlations on successive length scales corresponding to the various levels of the hierarchy, which we relate to a plausible implementation of the exact inference algorithm within the same architecture.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=F0Zd3knG9j", "arxiv_id": "2408.15138", "paper_pdf": "papers/F0Zd3knG9j.pdf", "paper_pdf_sha256": "3e9a2eb5c64b7be06b8522fa6fc5b4e7830b49a03565445e6c0a60dc034e56af", "paper_pdf_bytes": 794508, "paper_pdf_source": "openreview", "code_url": "https://github.com/emanuele-moscato/tree-language-paper-submission", "code_repository": "emanuele-moscato/tree-language-paper-submission", "code_commit": "3dd5853c96aa9cdba4c943dc49c9ad51e080201e", "code_archive": "repos/F0Zd3knG9j.zip", "code_archive_sha256": "9714d543d1100ad0ad2bef0cab7e2381d31bcd5ac371c8804c9f95797b1bf29c", "code_archive_bytes": 43241, "code_file_count": 13, "code_extensions": {".py": 13}, "github_disk_usage_kb": 33, "github_languages": {"Python": 134562}, "github_archived": false, "github_pushed_at": "2024-12-09T16:02:21Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/how-transformers-learn-structured-data"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "L9kwewFGQZ", "year": 2024, "status": "rejected", "title": "Mitigating Interference in the Knowledge Continuum through Attention-Guided Incremental Learning", "authors": ["Prashant Shivaram Bhat", "Bharath Chennamkulam Renjith", "Bahram Zonooz", "Elahe Arani"], "authorids": ["~Prashant_Shivaram_Bhat1", "~Bharath_Chennamkulam_Renjith1", "~Bahram_Zonooz1", "~Elahe_Arani1"], "authors_source": "OpenReview API", "abstract": "Continual learning (CL) remains a significant challenge for deep neural networks, as it is prone to forgetting previously acquired knowledge. Several approaches have been proposed in the literature, such as experience rehearsal, regularization, and parameter isolation, to address this problem. Although almost zero forgetting can be achieved in task-incremental learning, class-incremental learning remains highly challenging due to the problem of inter-task class separation. Limited access to previous task data makes it difficult to discriminate between classes of current and previous tasks. To address this issue, we propose `Attention-Guided Incremental Learning' (AGILE), a novel rehearsal-based CL approach that incorporates compact task-attention to effectively reduce interference between tasks. AGILE utilizes lightweight, learnable task projection vectors to transform the latent representations of a shared task-attention module toward task distribution. Through extensive empirical evaluation we show that AGILE significantly improves generalization performance by mitigating task interference and outperforms rehearsal-based approaches in several CL scenarios. Furthermore AGILE can scale well to a large number of tasks with minimal overhead while remaining well-calibrated with reduced task-recency bias.", "decision": "Reject", "meta_review": "", "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=L9kwewFGQZ", "arxiv_id": "2405.13978", "paper_pdf": "papers/L9kwewFGQZ.pdf", "paper_pdf_sha256": "b9b22608a98e83717b0e283c93282db4b1a4f1b6baa5f8ad90122246f5b570b1", "paper_pdf_bytes": 816257, "paper_pdf_source": "openreview", "code_url": "https://github.com/NeurAI-Lab/AGILE", "code_repository": "NeurAI-Lab/AGILE", "code_commit": "40220bf9ff2cf88fb785354b3a49e32c7e5eb328", "code_archive": "repos/L9kwewFGQZ.zip", "code_archive_sha256": "d10c038b2d94112ca5d9c62ddd0a876328547c2a17b34eb9524c96ab15fee590", "code_archive_bytes": 37059, "code_file_count": 29, "code_extensions": {".py": 29}, "github_disk_usage_kb": 29, "github_languages": {"Python": 107042}, "github_archived": false, "github_pushed_at": "2024-05-27T12:30:43Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/mitigating-interference-in-the-knowledge"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Om_QvnjjBL2", "year": 2023, "status": "rejected", "title": "D2Match: Leveraging Deep Learning and Degeneracy for  Subgraph Matching", "authors": ["Xuanzhou Liu", "Lin Zhang", "Jiaqi Sun", "Yujiu Yang", "Haiqin Yang"], "authorids": ["~Xuanzhou_Liu1", "~Lin_Zhang9", "~Jiaqi_Sun1", "~Yujiu_Yang2", "~Haiqin_Yang2"], "authors_source": "OpenReview API", "abstract": "Subgraph matching is a fundamental building block for many graph-based applications and is challenging due to its high-order combinatorial nature.  However, previous methods usually tackle it by combinatorial optimization or representation learning and suffer from exponential computational cost or matching without theoretical guarantees.  In this paper, we develop D2Match by leveraging the efficiency of Deep learning and Degeneracy for subgraph matching. More specifically, we prove that subgraph matching can degenerate to subtree matching, and subsequently is equivalent to finding a perfect matching on a bipartite graph.  This matching procedure can be implemented by the built-in tree-structured aggregation mechanism on graph neural networks, which yields linear time complexity.  Moreover, circle structures, abstracted as {\\em supernodes}, and node attributes can be easily incorporated in D2Match to boost the matching. Finally, we conduct extensive experiments to show the superior performance of our D2Match and confirm that our D2Match indeed tries to exploit the subtrees and differs from existing learning-based subgraph matching methods that depend on memorizing the data distribution divergence.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=Om_QvnjjBL2", "arxiv_id": "2306.06380", "paper_pdf": "papers/Om_QvnjjBL2.pdf", "paper_pdf_sha256": "0717cc33b43b26a720732cc720ccac889aca81b3ceba7448a711e93be0d092a2", "paper_pdf_bytes": 427383, "paper_pdf_source": "openreview", "code_url": "https://github.com/XuanzhouLiu/D2Match-ICML23", "code_repository": "XuanzhouLiu/D2Match-ICML23", "code_commit": "b444f90a05d6b06bed59e0b82007c374b2562a39", "code_archive": "repos/Om_QvnjjBL2.zip", "code_archive_sha256": "c39a1a4d1635becf4af91014a3f0bfba1bc12276c784ec75061c785c2e4de3e4", "code_archive_bytes": 27478, "code_file_count": 10, "code_extensions": {".py": 10}, "github_disk_usage_kb": 33, "github_languages": {"Python": 130463}, "github_archived": false, "github_pushed_at": "2023-12-30T06:36:38Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/d2match-leveraging-deep-learning-and"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "VQyHD2R3Aq", "year": 2022, "status": "rejected", "title": "SPIDE: A Purely Spike-based Method for Training Feedback Spiking Neural Networks", "authors": ["Mingqing Xiao", "Qingyan Meng", "Zongpeng Zhang", "Yisen Wang", "Zhouchen Lin"], "authorids": ["~Mingqing_Xiao1", "~Qingyan_Meng1", "~Zongpeng_Zhang1", "~Yisen_Wang1", "~Zhouchen_Lin1"], "authors_source": "OpenReview API", "abstract": "Spiking neural networks (SNNs) with event-based computation are promising brain-inspired models for energy-efficient applications on neuromorphic hardware. However, most supervised SNN training methods require complex computation or impractical neuron models, which hinders them from spike-based energy-efficient training. Among them, the recently proposed method, implicit differentiation on the equilibrium state (IDE), for training feedback SNNs is a promising way that is possible for generalization to locally spike-based learning with flexible network structures. In this paper, we study spike-based implicit differentiation on the equilibrium state (SPIDE) that extends the IDE method for supervised local learning with spikes, which could be possible for energy-efficient training on neuromorphic hardware. Specifically, we first introduce ternary spiking neuron couples that can realize ternary outputs with the common neuron model, and we prove that implicit differentiation can be solved by spikes based on this design. With this approach, the whole training procedure can be made as event-driven spike computation and weights are updated locally with two-stage average firing rates. Then to reduce the approximation error of spikes due to the finite simulation time steps, we propose to modify the resting membrane potential. Based on it, the average firing rate, when viewed as a stochastic estimator, achieves an unbiased estimation of iterative solution for implicit differentiation and the variance of this estimator is reduced. With these key components, we can train SNNs with either feedback or feedforward structures in a small number of time steps. Further, the firing sparsity during training demonstrates the great potential for energy efficiency. Meanwhile, even with these constraints, our trained models could still achieve competitive results on MNIST, CIFAR-10 and CIFAR-100. Our proposed method demonstrates the great potential for energy-efficient training of SNNs on neuromorphic hardware.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=VQyHD2R3Aq", "arxiv_id": "2302.00232", "paper_pdf": "papers/VQyHD2R3Aq.pdf", "paper_pdf_sha256": "132df82ffe8425d6421ac7f8e8699bfc4a10f94c4e7c6894ddca788deef1de6f", "paper_pdf_bytes": 484510, "paper_pdf_source": "openreview", "code_url": "https://github.com/pkuxmq/SPIDE-FSNN", "code_repository": "pkuxmq/SPIDE-FSNN", "code_commit": "54e4338d5d09a864909cdc284a02c8d73676e3c4", "code_archive": "repos/VQyHD2R3Aq.zip", "code_archive_sha256": "948fedfc0077c7ce3a0e5d4a48b3b83e941bdb9116874edc1c9bb40a344341e1", "code_archive_bytes": 50397, "code_file_count": 20, "code_extensions": {".py": 20}, "github_disk_usage_kb": 32, "github_languages": {"Python": 183260}, "github_archived": false, "github_pushed_at": "2023-02-02T03:29:30Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/spide-a-purely-spike-based-method-for-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "1UtnrqVUeNE", "year": 2021, "status": "rejected", "title": "Detecting Misclassification Errors in Neural Networks with a Gaussian Process Model", "authors": ["Xin Qiu", "Risto Miikkulainen"], "authorids": ["~Xin_Qiu1", "~Risto_Miikkulainen1"], "authors_source": "OpenReview API", "abstract": "As neural network classifiers are deployed in real-world applications, it is crucial that their predictions are not just accurate, but trustworthy as well. One practical solution is to assign confidence scores to each prediction, then filter out low-confidence predictions. However, existing confidence metrics are not yet sufficiently reliable for this role. This paper presents a new framework that produces more reliable confidence scores for detecting misclassification errors. This framework, RED, calibrates the classifier's inherent confidence indicators and estimates uncertainty of the calibrated confidence scores using Gaussian Processes. Empirical comparisons with other confidence estimation methods on 125 UCI datasets demonstrate that this approach is effective. An experiment on a vision task with a large deep learning architecture further confirms that the method can scale up, and a case study involving out-of-distribution and adversarial samples shows potential of the proposed method to improve robustness of neural network classifiers more broadly in the future.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=1UtnrqVUeNE", "arxiv_id": "2010.02065", "paper_pdf": "papers/1UtnrqVUeNE.pdf", "paper_pdf_sha256": "b40f23c343100d080e8b44ff9015e6d01f0ed0bc910a2dbc4d6a0d411308d017", "paper_pdf_bytes": 639420, "paper_pdf_source": "openreview", "code_url": "https://github.com/cognizant-ai-labs/red-paper", "code_repository": "cognizant-ai-labs/red-paper", "code_commit": "d88c77808954b13d8f59ca6727517892756548c5", "code_archive": "repos/1UtnrqVUeNE.zip", "code_archive_sha256": "2c46a671611cf1362d3f2d93e06c2140fb9a8ae27dc4bcf3aca0e01ef184075b", "code_archive_bytes": 58294, "code_file_count": 12, "code_extensions": {".py": 12}, "github_disk_usage_kb": 38, "github_languages": {"Python": 254942}, "github_archived": false, "github_pushed_at": "2025-09-28T09:57:32Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/detecting-misclassification-errors-in-neural-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Hye4WaVYwr", "year": 2020, "status": "rejected", "title": "Bootstrapping the Expressivity with Model-based Planning", "authors": ["Kefan Dong", "Yuping Luo", "Tengyu Ma"], "authorids": ["dkf16@mails.tsinghua.edu.cn", "yupingl@cs.princeton.edu", "tengyuma@cs.stanford.edu"], "authors_source": "OpenReview API", "abstract": "We compare the model-free reinforcement learning with the model-based approaches through the lens of the expressive power of neural networks for policies, $Q$-functions, and dynamics.  We show, theoretically and empirically, that even for one-dimensional continuous state space, there are many MDPs whose optimal $Q$-functions and policies are much more complex than the dynamics. We hypothesize many real-world MDPs also have a similar property. For these MDPs, model-based planning is a favorable algorithm, because the resulting policies can approximate the optimal policy significantly better than a neural network parameterization can, and model-free or model-based policy optimization rely on policy parameterization. Motivated by the theory, we apply a simple multi-step model-based bootstrapping planner (BOOTS) to bootstrap a weak $Q$-function into a stronger policy. Empirical results show that applying BOOTS on top of model-based or model-free policy optimization algorithms at the test time improves the performance on MuJoCo benchmark tasks. ", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=Hye4WaVYwr", "arxiv_id": null, "paper_pdf": "papers/Hye4WaVYwr.pdf", "paper_pdf_sha256": "6bab9bebe0e2277af06dbeff6bf2227da182ca4e00fbf0847430b51a9bf963ff", "paper_pdf_bytes": 1320829, "paper_pdf_source": "openreview", "code_url": "https://github.com/roosephu/boots", "code_repository": "roosephu/boots", "code_commit": "2f4f500f54feb95cf36abd863f3de4510d6f4950", "code_archive": "repos/Hye4WaVYwr.zip", "code_archive_sha256": "b03d052aaa850a46682c2ae00b46e7763fbc4dbc85c7696b46868a7e326c1529", "code_archive_bytes": 61173, "code_file_count": 69, "code_extensions": {".py": 69}, "github_disk_usage_kb": 45, "github_languages": {"Python": 112974}, "github_archived": false, "github_pushed_at": "2019-10-14T08:35:32Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/bootstrapping-the-expressivity-with-model-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "rylKB3A9Fm", "year": 2019, "status": "rejected", "title": "Assessing Generalization in Deep Reinforcement Learning", "authors": ["Charles Packer*", "Katelyn Gao*", "Jernej Kos", "Philipp Krahenbuhl", "Vladlen Koltun", "Dawn Song"], "authorids": ["cpacker@berkeley.edu", "katelyn.gao@intel.com", "jernej@kos.mx", "philkr@cs.utexas.edu", "vladlen.koltun@intel.com", "dawnsong@berkeley.edu"], "authors_source": "OpenReview API", "abstract": "Deep reinforcement learning (RL) has achieved breakthrough results on many tasks, but has been shown to be sensitive to system changes at test time. As a result, building deep RL agents that generalize has become an active research area. Our aim is to catalyze and streamline community-wide progress on this problem by providing the first benchmark and a common experimental protocol for investigating generalization in RL. Our benchmark contains a diverse set of environments and our evaluation methodology covers both in-distribution and out-of-distribution generalization. To provide a set of baselines for future research, we conduct a systematic evaluation of state-of-the-art algorithms, including those that specifically tackle the problem of generalization. The experimental results indicate that in-distribution generalization may be within the capacity of current algorithms, while out-of-distribution generalization is an exciting challenge for future work.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=rylKB3A9Fm", "arxiv_id": "1810.12282", "paper_pdf": "papers/rylKB3A9Fm.pdf", "paper_pdf_sha256": "293ca7ce737060d337bb86a1a211c1d8af8d1c969e26cb0e7f609e3003b6a623", "paper_pdf_bytes": 1864848, "paper_pdf_source": "openreview", "code_url": "https://github.com/sunblaze-ucb/rl-generalization", "code_repository": "sunblaze-ucb/rl-generalization", "code_commit": "5e05aa108aeadd4cf21f45d56128e9da3bddc14c", "code_archive": "repos/rylKB3A9Fm.zip", "code_archive_sha256": "287436828bfec9fd97753474c9ece96f66b69c3312c30757f5514f7f26a61d91", "code_archive_bytes": 212964, "code_file_count": 49, "code_extensions": {".py": 49}, "github_disk_usage_kb": 172, "github_languages": {"Python": 169497}, "github_archived": false, "github_pushed_at": "2019-01-22T05:01:26Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/assessing-generalization-in-deep"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "BJvWjcgAZ", "year": 2018, "status": "rejected", "title": "Sample-Efficient Deep Reinforcement Learning via Episodic Backward Update", "authors": ["Su Young Lee", "Sungik Choi", "Sae-Young Chung"], "authorids": ["sy9424@kaist.ac.kr", "si_choi@kaist.ac.kr", "schung@kaist.ac.kr"], "authors_source": "OpenReview API", "abstract": "We propose Episodic Backward Update - a new algorithm to boost the performance of a deep reinforcement learning agent by fast reward propagation. In contrast to the conventional use of the replay memory with uniform random sampling, our agent samples a whole episode and successively propagates the value of a state into its previous states. Our computationally efficient recursive algorithm allows sparse and delayed rewards to propagate effectively throughout the sampled episode. We evaluate our algorithm on 2D MNIST Maze Environment and 49 games of the Atari 2600 Environment and show that our agent improves sample efficiency with a competitive computational cost.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=BJvWjcgAZ", "arxiv_id": "1805.12375", "paper_pdf": "papers/BJvWjcgAZ.pdf", "paper_pdf_sha256": "6d180fcf3562e9ff8b0594b803974e7d9569bb3cc2b3c027d24cc1275a92999c", "paper_pdf_bytes": 1380708, "paper_pdf_source": "openreview", "code_url": "https://github.com/suyoung-lee/Episodic-Backward-Update", "code_repository": "suyoung-lee/Episodic-Backward-Update", "code_commit": "d4bbcfee39e8d4ee51422015115b55ab3cfa1829", "code_archive": "repos/BJvWjcgAZ.zip", "code_archive_sha256": "7a25ad561b6d5d615899f3ffabce9577863b3f3cdf5190672f265da6d244cf30", "code_archive_bytes": 286798, "code_file_count": 9, "code_extensions": {".py": 9}, "github_disk_usage_kb": 300, "github_languages": {"Python": 72839}, "github_archived": false, "github_pushed_at": "2019-09-24T08:06:59Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/sample-efficient-deep-reinforcement-learning-2"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "yyqbLqLhGl", "year": 2026, "status": "rejected", "title": "FLASH: Latent-Aware Semi-Autoregressive Speculative Decoding for Multimodal Tasks", "authors": ["Zihua Wang", "Ruibo Li", "Haozhe Du", "Joey Tianyi Zhou", "Yu Zhang", "Xu Yang"], "authorids": ["~Zihua_Wang1", "~Ruibo_Li2", "~Haozhe_Du2", "~Joey_Tianyi_Zhou1", "~Yu_Zhang21", "~Xu_Yang5"], "authors_source": "OpenReview API", "abstract": "Large language and multimodal models (LLMs and LMMs) exhibit strong inference capabilities but are often limited by slow decoding speeds. This challenge is especially acute in LMMs, where visual inputs typically comprise more tokens with lower information density than text --- an issue exacerbated by recent trends toward finer-grained visual tokenizations to boost performance. Speculative decoding has been effective in accelerating LLM inference by using a smaller draft model to generate candidate tokens, which are then selectively verified by the target model, improving speed without sacrificing output quality. While this strategy has been extended to LMMs, existing methods largely overlook the unique properties of visual inputs and depend solely on text-based draft models.\nIn this work, we propose \\textbf{FLASH} (Fast Latent-Aware Semi-Autoregressive Heuristics), a speculative decoding framework designed specifically for LMMs, which leverages two key properties of multimodal data to design the draft model. First, to address redundancy in visual tokens, we propose a lightweight latent-aware token compression mechanism. Second, recognizing that visual objects often co-occur within a scene, we employ a semi-autoregressive decoding strategy to generate multiple tokens per forward pass. \nExperiments show that FLASH consistently outperforms prior speculative decoding approaches in both unimodal and multimodal settings, achieving up to \\textbf{2.68$\\times$} speed-up on video captioning and \\textbf{2.55$\\times$} on visual instruction tuning tasks compared to the original LMM.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=yyqbLqLhGl", "arxiv_id": "2505.12728", "paper_pdf": "papers/yyqbLqLhGl.pdf", "paper_pdf_sha256": "aa690d43d850c4e7dd735b72fa9ebd4f668c59caf114fb6f517dd4aaf94bab0f", "paper_pdf_bytes": 550004, "paper_pdf_source": "openreview", "code_url": "https://github.com/ZihuaEvan/FlashSD", "code_repository": "ZihuaEvan/FlashSD", "code_commit": "2a93afd06cda1450d07e489a34d4fc11c12dbe3c", "code_archive": "repos/yyqbLqLhGl.zip", "code_archive_sha256": "cb92f5a6f2a41ac06967c6cb01891847983c489689fe0393bc510d2521c271df", "code_archive_bytes": 178575, "code_file_count": 34, "code_extensions": {".py": 34}, "github_disk_usage_kb": 142, "github_languages": {"Python": 647839}, "github_archived": false, "github_pushed_at": "2026-08-01T09:22:31Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/flash-latent-aware-semi-autoregressive"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "6BjEqGn1OO", "year": 2025, "status": "rejected", "title": "Modeling Real-Time Interactive Conversations as Timed Diarized Transcripts", "authors": ["Garrett Tanzer", "Gustaf Ahdritz", "Luke Melas-Kyriazi"], "authorids": ["~Garrett_Tanzer1", "~Gustaf_Ahdritz2", "~Luke_Melas-Kyriazi1"], "authors_source": "OpenReview API", "abstract": "Chatbots built upon language models have exploded in popularity, but they have largely been limited to synchronous, turn-by-turn dialogues. In this paper we present a simple yet general method to simulate real-time interactive conversations using pretrained text-only language models, by modeling timed diarized transcripts and decoding them with causal rejection sampling. We demonstrate the promise of this method with two case studies: instant messenger dialogues and spoken conversations, which require generation at about 30 tok/s and 20 tok/s respectively to maintain real-time interactivity. These capabilities can be added into language models using relatively little data and run on commodity hardware.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=6BjEqGn1OO", "arxiv_id": "2405.13203", "paper_pdf": "papers/6BjEqGn1OO.pdf", "paper_pdf_sha256": "f3ed0676a59c4129cff3f198e3f125410510c10d8dd67eea02afaa8b04b710e6", "paper_pdf_bytes": 417819, "paper_pdf_source": "openreview", "code_url": "https://github.com/gahdritz/rtic", "code_repository": "gahdritz/rtic", "code_commit": "0ed62af2cda9c174dfb1309b66ba453ce724a0d8", "code_archive": "repos/6BjEqGn1OO.zip", "code_archive_sha256": "b6ac20f5f13f3d7f59b4c07bb32768692686df170624ff69479e68966f8b97c1", "code_archive_bytes": 28570, "code_file_count": 14, "code_extensions": {".py": 14}, "github_disk_usage_kb": 34, "github_languages": {"Python": 61707}, "github_archived": false, "github_pushed_at": "2024-06-03T07:10:51Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/modeling-real-time-interactive-conversations"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "1gkePTsAWf", "year": 2024, "status": "rejected", "title": "Self-Taught Optimizer (STOP): Recursively Self-Improving Code Generation", "authors": ["Eric Zelikman", "Eliana Lorch", "Lester Mackey", "Adam Tauman Kalai"], "authorids": ["~Eric_Zelikman1", "eliana@lorien.ai", "~Lester_Mackey1", "~Adam_Tauman_Kalai1"], "authors_source": "OpenReview API", "abstract": "Several recent advances in AI systems (e.g., Tree-of-Thoughts and Program-Aided Language Models) solve problems by providing a \"scaffolding\" program that structures multiple calls to language models to generate better outputs. A scaffolding program is written in a programming language such as Python. In this work, we use a language-model-infused scaffolding program to improve itself. We start with a seed \"improver\" that improves an input program according to a given utility function by querying a language model several times and returning the best solution. We then run this seed improver to improve itself. Across a small set of downstream tasks, the resulting improved improver generates programs with significantly better performance than its seed improver. Afterward, we analyze the variety of self-improvement strategies proposed by the language model, including beam search, genetic algorithms, and simulated annealing. Since the language models themselves are not altered, this is not full recursive self-improvement. Nonetheless, it demonstrates that a modern language model, GPT-4 in our proof-of-concept experiments, is capable of writing code that can call itself to improve itself. We critically consider concerns around the development of self-improving technologies and evaluate the frequency with which the generated code bypasses a sandbox.", "decision": "Reject", "meta_review": "", "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=1gkePTsAWf", "arxiv_id": "2310.02304", "paper_pdf": "papers/1gkePTsAWf.pdf", "paper_pdf_sha256": "77a23375914992ce6e6238087e278753f7ae79949cc2fc46b28e6d5b2aaaa27d", "paper_pdf_bytes": 600577, "paper_pdf_source": "openreview", "code_url": "https://github.com/microsoft/stop", "code_repository": "microsoft/stop", "code_commit": "0d6780c54306b2486dd36e9c4ae9b49aceb27ea4", "code_archive": "repos/1gkePTsAWf.zip", "code_archive_sha256": "5452c2032b207e9076866faa26ce4333bc38e59b43a5f11fce571d110c7222c4", "code_archive_bytes": 42182, "code_file_count": 31, "code_extensions": {".py": 31}, "github_disk_usage_kb": 31, "github_languages": {"Python": 89274}, "github_archived": false, "github_pushed_at": "2024-01-01T20:28:46Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/self-taught-optimizer-stop-recursively-self"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Z4lOwCEJQ8Z", "year": 2023, "status": "rejected", "title": "Training Normalizing Flows from Dependent Data", "authors": ["Matthias Kirchler", "Christoph Lippert", "Marius Kloft"], "authorids": ["~Matthias_Kirchler1", "~Christoph_Lippert1", "~Marius_Kloft1"], "authors_source": "OpenReview API", "abstract": "Normalizing flows are powerful non-parametric statistical models that function as a hybrid between density estimators and generative models. Current learning algorithms for normalizing flows assume that data points are sampled independently, an assumption that is frequently violated in practice, which may lead to erroneous density estimation and data generation. We propose a likelihood objective of normalizing flows incorporating dependencies between the data points, for which we derive a flexible and efficient learning algorithm suitable for different dependency structures. We show that respecting dependencies between observations can improve empirical results on both synthetic and real-world data.\n\n", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=Z4lOwCEJQ8Z", "arxiv_id": "2209.14933", "paper_pdf": "papers/Z4lOwCEJQ8Z.pdf", "paper_pdf_sha256": "04d4516c4c72a8ff42c615f1af893c4b701f64673893d2d794c5940bfac0470a", "paper_pdf_bytes": 688970, "paper_pdf_source": "openreview", "code_url": "https://github.com/mkirchler/dependent_data_flows", "code_repository": "mkirchler/dependent_data_flows", "code_commit": "78f40b72215ac57bf3cfb916641ffc73362b6712", "code_archive": "repos/Z4lOwCEJQ8Z.zip", "code_archive_sha256": "49f1b4319a5b96c2b05b36552d07a90bf99ed0f30ac781ad245cf30e279a5869", "code_archive_bytes": 65461, "code_file_count": 13, "code_extensions": {".py": 13}, "github_disk_usage_kb": 35, "github_languages": {"Python": 99433}, "github_archived": false, "github_pushed_at": "2023-10-31T10:56:06Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/training-normalizing-flows-from-dependent"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "I_RLPhVUfw8", "year": 2022, "status": "rejected", "title": "Dense Gaussian Processes for Few-Shot Segmentation", "authors": ["Joakim Johnander", "Johan Edstedt", "Michael Felsberg", "Fahad Khan", "Martin Danelljan"], "authorids": ["~Joakim_Johnander1", "~Johan_Edstedt1", "~Michael_Felsberg2", "~Fahad_Khan1", "~Martin_Danelljan4"], "authors_source": "OpenReview API", "abstract": "Few-shot segmentation is a challenging dense prediction task, which entails segmenting a novel query image given only a small annotated support set. The key problem is thus to design a method that aggregates detailed information from the support set, while being robust to large variations in appearance and context. To this end, we propose a few-shot segmentation method based on dense Gaussian process (GP) regression. Given the support set, our dense GP learns the mapping from local deep image features to mask values, capable of capturing complex appearance distributions. Furthermore, it provides a principled means of capturing uncertainty, which serves as another powerful cue for the final segmentation, obtained by a CNN decoder. Instead of a one-dimensional mask output, we further exploit the end-to-end learning capabilities of our approach to learn a high-dimensional output space for the GP. Our approach sets a new state-of-the-art for both 1-shot and 5-shot FSS on the PASCAL-5$^i$ and COCO-20$^i$ benchmarks, achieving an absolute gain of $+14.9$ mIoU in the COCO-20$^i$ 5-shot setting. Furthermore, the segmentation quality of our approach scales gracefully when increasing the support set size, while achieving robust cross-dataset transfer. ", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=I_RLPhVUfw8", "arxiv_id": "2110.03674", "paper_pdf": "papers/I_RLPhVUfw8.pdf", "paper_pdf_sha256": "e6fbc6d933d4d8f6ade1707021dbabc78b401b62f673b567a8e3e8af938741bf", "paper_pdf_bytes": 3132078, "paper_pdf_source": "openreview", "code_url": "https://github.com/joakimjohnander/DGPNet", "code_repository": "joakimjohnander/DGPNet", "code_commit": "c610c14554ae73abf8f968ea4f73eb8211b532c2", "code_archive": "repos/I_RLPhVUfw8.zip", "code_archive_sha256": "3a8deaf08c38be593a4611f00dfb0f86245350f11edb58385b140df3086af19e", "code_archive_bytes": 41080, "code_file_count": 16, "code_extensions": {".py": 15, ".sh": 1}, "github_disk_usage_kb": 33, "github_languages": {"Python": 122861, "Shell": 551}, "github_archived": false, "github_pushed_at": "2022-09-28T07:11:19Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/dense-gaussian-processes-for-few-shot"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "YjXnezbeCwG", "year": 2021, "status": "rejected", "title": "Learning to Use Future Information in Simultaneous Translation", "authors": ["Xueqing Wu", "Yingce Xia", "Lijun Wu", "Shufang Xie", "Weiqing Liu", "Tao Qin", "Tie-Yan Liu"], "authorids": ["~Xueqing_Wu1", "~Yingce_Xia1", "~Lijun_Wu1", "~Shufang_Xie1", "weiqing.liu@microsoft.com", "~Tao_Qin1", "~Tie-Yan_Liu1"], "authors_source": "OpenReview API", "abstract": "Simultaneous neural machine translation (briefly, NMT) has attracted much attention recently. In contrast to standard NMT, where the NMT system can access the full input sentence, simultaneous NMT is a prefix-to-prefix problem, where the system can only utilize the prefix of the input sentence and thus more uncertainty and difficulty are introduced to decoding. Wait-k inference is a simple yet effective strategy for simultaneous NMT, where the decoder generates the output sequence $k$ words behind the input words. For wait-k inference, we observe that wait-m training with $m>k$ in simultaneous NMT (i.e., using more future information for training than inference) generally outperforms wait-k training. Based on this observation, we propose a method that automatically learns how much future information to use in training for simultaneous NMT. Specifically, we introduce a controller to adaptively select wait-m training strategies according to the network status of the translation model and current training sentence pairs, and the controller is jointly trained with the translation model through bi-level optimization. Experiments on four datasets show that our method brings 1 to 3 BLEU point improvement over baselines under the same latency. Our code is available at https://github.com/P2F-research/simulNMT .", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=YjXnezbeCwG", "arxiv_id": null, "paper_pdf": "papers/YjXnezbeCwG.pdf", "paper_pdf_sha256": "d5ca022554f0e1a7303c0e24f6f76d052b68b765635eb3b94227fbe4cf6b8df3", "paper_pdf_bytes": 579358, "paper_pdf_source": "openreview", "code_url": "https://github.com/P2F-research/simulNMT", "code_repository": "P2F-research/simulNMT", "code_commit": "4721707f7cf0beea39d7ee8accb058eb2c7394fc", "code_archive": "repos/YjXnezbeCwG.zip", "code_archive_sha256": "bb69a76dd4f44b4bffefd0c120ad26ccf78a6ffa04424eb9b52211ffb5fc345d", "code_archive_bytes": 58090, "code_file_count": 28, "code_extensions": {".py": 22, ".sh": 6}, "github_disk_usage_kb": 40, "github_languages": {"Python": 163307, "Shell": 6267}, "github_archived": false, "github_pushed_at": "2020-10-05T11:42:02Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/learning-to-use-future-information-in"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "rygePJHYPH", "year": 2020, "status": "rejected", "title": "Towards trustworthy predictions from deep neural networks with fast adversarial calibration", "authors": ["Christian Tomani", "Florian Buettner"], "authorids": ["christian.tomani@gmail.com", "fbuettner.phys@gmail.com"], "authors_source": "OpenReview API", "abstract": "To facilitate a wide-spread acceptance of AI systems guiding decision making in real-world applications, trustworthiness of deployed models is key. That is, it is crucial for predictive models to be uncertainty-aware and yield well-calibrated (and thus trustworthy) predictions for both in-domain samples as well as under domain shift. Recent efforts to account for predictive uncertainty include post-processing steps for trained neural networks, Bayesian neural networks as well as alternative non-Bayesian approaches such as ensemble approaches and evidential deep learning. Here, we propose an efficient yet general modelling approach for obtaining well-calibrated, trustworthy probabilities for samples obtained after a domain shift. We introduce a new training strategy combining an entropy-encouraging loss term with an adversarial calibration loss term and demonstrate that this results in well-calibrated and technically trustworthy predictions for a wide range of perturbations. We comprehensively evaluate previously proposed approaches on different data modalities, a large range of data sets, network architectures and perturbation strategies and observe that our modelling approach substantially outperforms existing state-of-the-art approaches, yielding well-calibrated predictions for both in-domain and out-of domain samples. ", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=rygePJHYPH", "arxiv_id": "2012.10923", "paper_pdf": "papers/rygePJHYPH.pdf", "paper_pdf_sha256": "2017b808a8159561c24a00ecd67ea98cf2b20a0f856210d5f272d2809179ae9d", "paper_pdf_bytes": 1644403, "paper_pdf_source": "openreview", "code_url": "https://github.com/tochris/falcon", "code_repository": "tochris/falcon", "code_commit": "6e0ff0f4359b6b8e8b4c843d7e6eb097f6e1a39f", "code_archive": "repos/rygePJHYPH.zip", "code_archive_sha256": "afc03e3eb65bf2a720e57180928ce5f9b0c6f8ae11f22e52a63c48bb42ce3ca6", "code_archive_bytes": 50864, "code_file_count": 29, "code_extensions": {".py": 29}, "github_disk_usage_kb": 45, "github_languages": {"Python": 198841}, "github_archived": false, "github_pushed_at": "2021-02-24T10:40:55Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/towards-trustworthy-predictions-from-deep"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "rkgwuiA9F7", "year": 2019, "status": "rejected", "title": "Cramer-Wold AutoEncoder", "authors": ["Jacek Tabor", "Szymon Knop", "Przemysław Spurek", "Igor Podolak", "Marcin Mazur", "Stanisław Jastrzębski"], "authorids": ["jacek.tabor@uj.edu.pl", "szymon.knop@doctoral.uj.edu.pl", "przemyslaw.spurek@uj.edu.pl", "igor.podolak@uj.edu.pl", "marcin.mazur@uj.edu.pl", "staszek.jastrzebski@gmail.com"], "authors_source": "OpenReview API", "abstract": "Assessing distance betweeen the true and the sample distribution is a key component of many state of the art generative models, such as Wasserstein Autoencoder (WAE). Inspired by prior work on Sliced-Wasserstein Autoencoders (SWAE) and\nkernel smoothing we construct a new generative model – Cramer-Wold AutoEncoder (CWAE). CWAE cost function, based on introduced Cramer-Wold distance between samples, has a simple closed-form in the case of normal prior. As a consequence, while simplifying the optimization procedure (no need of sampling necessary to evaluate the distance function in the training loop), CWAE performance matches quantitatively and qualitatively that of WAE-MMD (WAE using maximum mean discrepancy based distance function) and often improves upon SWAE.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=rkgwuiA9F7", "arxiv_id": "1805.09235", "paper_pdf": "papers/rkgwuiA9F7.pdf", "paper_pdf_sha256": "36fda4f9111166f7bbbd9d4da07d25478bd74e2a4acafdab9c995fde268789a1", "paper_pdf_bytes": 7236018, "paper_pdf_source": "openreview", "code_url": "https://github.com/gmum/cwae", "code_repository": "gmum/cwae", "code_commit": "50592903c321de25f339f3b00cbd2143741e5037", "code_archive": "repos/rkgwuiA9F7.zip", "code_archive_sha256": "4528136d0c84d4d22b8a19b7b3b7f623834c56e6e8d1095ca8147eee6840b821", "code_archive_bytes": 185385, "code_file_count": 20, "code_extensions": {".py": 19, ".ipynb": 1}, "github_disk_usage_kb": 180, "github_languages": {"Jupyter Notebook": 232896, "Python": 37925}, "github_archived": false, "github_pushed_at": "2020-09-17T16:31:56Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/cramer-wold-autoencoder"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ryCM8zWRb", "year": 2018, "status": "rejected", "title": "Recurrent Neural Networks with Top-k Gains for Session-based Recommendations", "authors": ["Balázs Hidasi", "Alexandros Karatzoglou"], "authorids": ["hidasib@gmail.com", "alexk@tid.es"], "authors_source": "OpenReview API", "abstract": "RNNs have been shown to be excellent models for sequential data and in particular for session-based user behavior. The use of RNNs provides impressive performance benefits over classical methods in session-based recommendations. In this work we introduce a novel ranking loss function tailored for RNNs in recommendation settings. The better performance of such loss over alternatives, along with further tricks and improvements described in this work, allow to achieve an overall improvement of up to 35% in terms of MRR and Recall@20 over previous session-based RNN solutions and up to 51% over classical collaborative filtering approaches. Unlike data augmentation-based improvements, our method does not increase training times significantly.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=ryCM8zWRb", "arxiv_id": "1706.03847", "paper_pdf": "papers/ryCM8zWRb.pdf", "paper_pdf_sha256": "1978893be01f8b5332a1b09b0cea54253d33ff20199d669afe4d44c7f7a6440d", "paper_pdf_bytes": 499988, "paper_pdf_source": "openreview", "code_url": "https://github.com/hidasib/GRU4Rec", "code_repository": "hidasib/GRU4Rec", "code_commit": "a4ed5fbdba35bcc18bf1d4a9b76692ef462bb284", "code_archive": "repos/ryCM8zWRb.zip", "code_archive_sha256": "b2ad2479ffb6d320d3f2f3b929b6bd084d8943a36be9dee83dd06635a8c4a230", "code_archive_bytes": 185985, "code_file_count": 22, "code_extensions": {".py": 22}, "github_disk_usage_kb": 312, "github_languages": {"Python": 132123}, "github_archived": false, "github_pushed_at": "2023-08-24T15:57:50Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/recurrent-neural-networks-with-top-k-gains"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "5y8R8XSHkG", "year": 2026, "status": "rejected", "title": "Smoothie: Smoothing Diffusion on Token Embeddings for Text Generation", "authors": ["Alexander Shabalin", "Viacheslav Meshchaninov", "Dmitry Vetrov"], "authorids": ["~Alexander_Shabalin1", "~Viacheslav_Meshchaninov1", "~Dmitry_Vetrov1"], "authors_source": "OpenReview API", "abstract": "Diffusion models have achieved state-of-the-art performance in generating images, audio, and video, but their adaptation to text remains challenging due to its discrete nature. Prior approaches either apply Gaussian diffusion in continuous latent spaces, which inherits semantic structure but struggles with token decoding, or operate in categorical simplex space, which respect discreteness but disregard semantic relation between tokens. In this paper, we propose Smoothing Diffusion on Token Embeddings (Smoothie), a novel diffusion method that combines the strengths of both approaches by progressively smoothing token embeddings based on semantic similarity. This technique enables gradual information removal while maintaining a natural decoding process. Experimental results on several sequence-to-sequence generation tasks demonstrate that Smoothie outperforms existing diffusion-based models in generation quality. Furthermore, ablation studies show that our proposed diffusion space yields better performance than both the standard embedding space and the categorical simplex.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=5y8R8XSHkG", "arxiv_id": "2505.18853", "paper_pdf": "papers/5y8R8XSHkG.pdf", "paper_pdf_sha256": "fd878313ad5515bf0d512bc16e0cefcaf144c95e037528870dad4eb96303a295", "paper_pdf_bytes": 1212515, "paper_pdf_source": "openreview", "code_url": "https://github.com/ashaba1in/smoothie", "code_repository": "ashaba1in/smoothie", "code_commit": "79d3a409fb5044c2108be749634e2fc1ea7fb163", "code_archive": "repos/5y8R8XSHkG.zip", "code_archive_sha256": "c105df33895b0145356c140cd79a78cbae7d29765d327fc1fc0ab4fbd2e02f25", "code_archive_bytes": 45316, "code_file_count": 25, "code_extensions": {".py": 25}, "github_disk_usage_kb": 160, "github_languages": {"Python": 148052}, "github_archived": false, "github_pushed_at": "2026-02-02T15:58:15Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/smoothie-smoothing-diffusion-on-token"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "MM197t8WlM", "year": 2025, "status": "rejected", "title": "Local Flow Matching Generative Models", "authors": ["Chen Xu", "Xiuyuan Cheng", "Yao Xie"], "authorids": ["~Chen_Xu12", "~Xiuyuan_Cheng1", "~Yao_Xie2"], "authors_source": "OpenReview API", "abstract": "Flow Matching (FM) is a simulation-free method for learning a continuous and invertible flow to interpolate between two distributions, and in particular to generate data from noise in generative modeling. In this paper, we introduce Local Flow Matching ($\\texttt{LFM}$), which consecutively learns a sequence of FM sub-models and each matches a diffusion process up to the time of the step size in the data-to-noise direction. In each step, the two distributions to be interpolated by the sub-model are closer to each other than data vs. noise, and this enables the use of smaller models with faster training. The stepwise structure of $\\texttt{LFM}$ is natural to be distilled and different distillation techniques can be adopted to speed up generation. Theoretically, we prove a generation guarantee of the proposed flow model in terms of the $\\chi^2$-divergence between the generated and true data distributions. In experiments, we demonstrate the improved training efficiency and competitive generative performance of $\\texttt{LFM}$ compared to FM on the unconditional generation of tabular data and image datasets, and also on the conditional generation of robotic manipulation policies.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=MM197t8WlM", "arxiv_id": "2410.02548", "paper_pdf": "papers/MM197t8WlM.pdf", "paper_pdf_sha256": "682ac3ac54133fc803f7b9fe17e2ff94cd6460a5e6b421d6049c844d5259252f", "paper_pdf_bytes": 2658300, "paper_pdf_source": "openreview", "code_url": "https://github.com/hamrel-cxu/LocalFlowMatching", "code_repository": "hamrel-cxu/LocalFlowMatching", "code_commit": "b249c1f28b45693b2624b0486df4eef6339b71d3", "code_archive": "repos/MM197t8WlM.zip", "code_archive_sha256": "01e68c5f0f02ed18b81f0fb50659f935ff7ab0642178deb21b00602976eb04b6", "code_archive_bytes": 41362, "code_file_count": 8, "code_extensions": {".py": 8}, "github_disk_usage_kb": 34, "github_languages": {"Python": 43438}, "github_archived": false, "github_pushed_at": "2025-11-10T21:25:11Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/local-flow-matching-generative-models"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "0IaTFNJner", "year": 2024, "status": "rejected", "title": "On the Embedding Collapse When Scaling up Recommendation Models", "authors": ["Xingzhuo Guo", "Junwei Pan", "Ximei Wang", "Baixu Chen", "Jie Jiang", "Mingsheng Long"], "authorids": ["~Xingzhuo_Guo1", "~Junwei_Pan1", "~Ximei_Wang1", "~Baixu_Chen2", "~Jie_Jiang3", "~Mingsheng_Long5"], "authors_source": "OpenReview API", "abstract": "Recent advances in deep foundation models have led to a promising trend of developing large recommendation models to leverage vast amounts of available data. However, we experiment to scale up existing recommendation models and observe that the enlarged models do not improve satisfactorily. In this context, we investigate the embedding layers of enlarged models and identify a phenomenon of *embedding collapse*, which ultimately hinders scalability, wherein the embedding matrix tends to reside in a low-dimensional subspace. Through empirical and theoretical analysis, we demonstrate that the feature interaction module specific to recommendation models has a *two-sided effect*. On the one hand, the interaction restricts embedding learning when interacting with collapsed embeddings, exacerbating the collapse issue. On the other hand, feature interaction is crucial in mitigating the fitting of spurious features, thereby improving scalability. Based on this analysis, we propose a simple yet effective *multi-embedding* design incorporating embedding-set-specific interaction modules to capture diverse patterns and reduce collapse. Extensive experiments demonstrate that this proposed design provides consistent scalability for various recommendation models.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=0IaTFNJner", "arxiv_id": "2310.04400", "paper_pdf": "papers/0IaTFNJner.pdf", "paper_pdf_sha256": "61bed8f96538f35463a03eb5a446a3f24d35786118d26899a86f27c029d22cf4", "paper_pdf_bytes": 2016960, "paper_pdf_source": "openreview", "code_url": "https://github.com/thuml/Multi-Embedding", "code_repository": "thuml/Multi-Embedding", "code_commit": "55078571bf1922dced112505514958f85ca4804c", "code_archive": "repos/0IaTFNJner.zip", "code_archive_sha256": "f3a9d4c906e13a9e12a476736cd2b9a64b142b9592833dc598a54d5c8c99996e", "code_archive_bytes": 40339, "code_file_count": 32, "code_extensions": {".py": 32}, "github_disk_usage_kb": 31, "github_languages": {"Python": 114761}, "github_archived": false, "github_pushed_at": "2024-08-04T00:56:50Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/on-the-embedding-collapse-when-scaling-up"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "GKpwIa9wgwR", "year": 2023, "status": "rejected", "title": "Efficient Data Subset Selection to Generalize Training Across Models: Transductive and Inductive Networks", "authors": ["Eeshaan Jain", "Tushar Nandy", "Gaurav Aggarwal", "Ashish V. Tendulkar", "Rishabh K Iyer", "Abir De"], "authorids": ["~Eeshaan_Jain1", "~Tushar_Nandy1", "~Gaurav_Aggarwal4", "~Ashish_V._Tendulkar1", "~Rishabh_K_Iyer2", "~Abir_De1"], "authors_source": "OpenReview API", "abstract": "Subset selection, in recent times, has emerged as a successful approach toward efficient training of models by significantly reducing the amount of data and computational resources required. However, existing methods employ discrete combinatorial and model-specific approaches which lack generalizability--- for each new model, the algorithm has to be executed from the beginning. Therefore, for data subset selection for an unseen architecture, one cannot use the subset chosen for a different model. In this work, we propose SubSelNet, a non-adaptive  subset selection framework, which tackles these problems with two main components. First, we introduce an attention-based neural gadget that leverages the graph structure of architectures and acts as a surrogate to trained deep neural networks for quick model prediction. Then, we use these predictions to build subset samplers. This leads us to develop two variants of  SubSelNet. The first variant is transductive (called as Transductive-SubSelNet) which computes the subset separately for each model by solving a small optimization problem. Such an optimization is still super fast, thanks to the replacement of explicit model training by the model approximator. The second variant is inductive (called as Inductive-SubSelNet) which computes the subset using a trained subset selector, without any optimization.  Most state-of-the-art data subset selection approaches are adaptive, in that the subset selection adapts as the training progresses, and as a result, they require access to the entire data at training time.  Our approach, in contrast, is non-adaptive and does the subset selection only once in the beginning, thereby achieving resource and memory efficiency along with compute-efficiency at training time. Our experiments show that both transductive and inductive variants of our models outperform several methods on the quality of the subset chosen and further demonstrate that our method can be used for choosing the best architecture from a set of architectures.\n", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=GKpwIa9wgwR", "arxiv_id": "2409.12255", "paper_pdf": "papers/GKpwIa9wgwR.pdf", "paper_pdf_sha256": "581117c8d34f86e758a7c520ab2a822181addc574bd3bcdbde56bc663cb35310", "paper_pdf_bytes": 1498817, "paper_pdf_source": "openreview", "code_url": "https://github.com/structlearning/subselnet", "code_repository": "structlearning/subselnet", "code_commit": "8e75a6c18224940a59de0721e9a09bf562d4f30b", "code_archive": "repos/GKpwIa9wgwR.zip", "code_archive_sha256": "9df6b1abbb98d39157c2eeefe8aa396e69aab37421407a84a52bba598d11f19b", "code_archive_bytes": 36720, "code_file_count": 23, "code_extensions": {".py": 23}, "github_disk_usage_kb": 39, "github_languages": {"Python": 83913}, "github_archived": false, "github_pushed_at": "2024-01-28T10:31:01Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/efficient-data-subset-selection-to-generalize-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "L2a_bcarHcF", "year": 2022, "status": "rejected", "title": "Linear algebra with transformers", "authors": ["Francois Charton"], "authorids": ["~Francois_Charton1"], "authors_source": "OpenReview API", "abstract": "Most applications of transformers to mathematics, from integration to theorem proving, focus on symbolic computation. In this paper, we show that transformers can be trained to perform numerical calculations with high accuracy. We consider problems of linear algebra:  matrix transposition, addition, multiplication, eigenvalues and vectors, singular value decomposition, and inversion. Training small transformers (up to six layers) over datasets of random matrices, we achieve high accuracies (over 90%) on all problems. We also show that trained models can generalize out of their training distribution, and that out-of-domain accuracy can be greatly improved by working from more diverse datasets (in particular, by training from matrices with non-independent and identically distributed coefficients). Finally, we show that few-shot learning can be leveraged to retrain models to solve larger problems.\n\n", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=L2a_bcarHcF", "arxiv_id": "2112.01898", "paper_pdf": "papers/L2a_bcarHcF.pdf", "paper_pdf_sha256": "bc131c4a0592ef313bb305cb8dd0244d46446b6a1d5b740f7d6e058a905939af", "paper_pdf_bytes": 838859, "paper_pdf_source": "openreview", "code_url": "https://github.com/facebookresearch/LAWT", "code_repository": "facebookresearch/LAWT", "code_commit": "50bba0451c605b7112f914e827961300ad10b170", "code_archive": "repos/L2a_bcarHcF.zip", "code_archive_sha256": "bb45209ffe3b027cb67152262b4578d362bf8d38fe67a209dd94e7cff7152b3e", "code_archive_bytes": 63847, "code_file_count": 15, "code_extensions": {".py": 14, ".ipynb": 1}, "github_disk_usage_kb": 34, "github_languages": {"Python": 163819, "Jupyter Notebook": 12325}, "github_archived": true, "github_pushed_at": "2024-08-14T03:33:17Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/linear-algebra-with-transformers-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "muppfCkU9H1", "year": 2021, "status": "rejected", "title": "Multi-hop Attention Graph Neural Network", "authors": ["Guangtao Wang", "Zhitao Ying", "Jing Huang", "Jure Leskovec"], "authorids": ["~Guangtao_Wang1", "~Zhitao_Ying1", "~Jing_Huang3", "~Jure_Leskovec1"], "authors_source": "OpenReview API", "abstract": "Self-attention mechanism in graph neural networks (GNNs) led to state-of-the-art performance on many graph representation learning task. Currently, at every layer, attention is computed between connected pairs of nodes and depends solely on the representation of the two nodes. However, such attention mechanism does not account for nodes that are not directly connected but provide important network context, which could lead to improved predictive performance. Here we propose Multi-hop Attention Graph Neural Network (MAGNA), a principled way to incorporate multi-hop context information into attention computation, enabling long-range interactions at every layer of the GNN.  To compute attention between nodes that are not directly connected,\nMAGNA diffuses the attention scores across the network, which increases the ''receptive field'' for every layer of the GNN.\nUnlike previous approaches, MAGNA uses a diffusion prior on attention values, to efficiently account for all paths between the pair of disconnected nodes. This helps MAGNA capture large-scale structural information in every layer, and learn more informative attention. Experimental results on node classification as well as the knowledge graph completion benchmarks show that MAGNA achieves state-of-the-art results: MAGNA achieves up to 5.7% relative error reduction over the previous state-of-the-art on Cora, Citeseer, and Pubmed. MAGNA also obtains the best performance on a large-scale Open Graph Benchmark dataset. On knowledge graph completion  MAGNA advances state-of-the-art on WN18RR and FB15k-237 across four different performance metrics.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=muppfCkU9H1", "arxiv_id": "2009.14332", "paper_pdf": "papers/muppfCkU9H1.pdf", "paper_pdf_sha256": "8750d1d4a7c780f108c6c741cb7084932cfb180cc5192f9f7bb656042f90cc8c", "paper_pdf_bytes": 424519, "paper_pdf_source": "openreview", "code_url": "https://github.com/xjtuwgt/GNN-MAGNA", "code_repository": "xjtuwgt/GNN-MAGNA", "code_commit": "b37162830b5fe2582316390ca7853b4f30bbd129", "code_archive": "repos/muppfCkU9H1.zip", "code_archive_sha256": "065b7da814b19e8e98715c85be45b6b5a30d5a49fdab89d2f7bae2c6c5c594b5", "code_archive_bytes": 57644, "code_file_count": 28, "code_extensions": {".py": 28}, "github_disk_usage_kb": 40, "github_languages": {"Python": 175959}, "github_archived": false, "github_pushed_at": "2021-08-25T19:14:08Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/direct-multi-hop-attention-based-graph-neural"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "rkltE0VKwH", "year": 2020, "status": "rejected", "title": "Coordinated Exploration via Intrinsic Rewards for Multi-Agent Reinforcement Learning", "authors": ["Shariq Iqbal", "Fei Sha"], "authorids": ["shariqiqbal2810@gmail.com", "fsha@google.com"], "authors_source": "OpenReview API", "abstract": "Solving tasks with sparse rewards is one of the most important challenges in reinforcement learning. In the single-agent setting, this challenge has been addressed by introducing intrinsic rewards that motivate agents to explore unseen regions of their state spaces. Applying these techniques naively to the multi-agent setting results in agents exploring independently, without any coordination among themselves. We argue that learning in cooperative multi-agent settings can be accelerated and improved if agents coordinate with respect to what they have explored. In this paper we propose an approach for learning how to dynamically select between different types of intrinsic rewards which consider not just what an individual agent has explored, but all agents, such that the agents can coordinate their exploration and maximize extrinsic returns. Concretely, we formulate the approach as a hierarchical policy where a high-level controller selects among sets of policies trained on different types of intrinsic rewards and the low-level controllers learn the action policies of all agents under these specific rewards. We demonstrate the effectiveness of the proposed approach in a multi-agent gridworld domain with sparse rewards, and then show that our method scales up to more complex settings by evaluating on the VizDoom platform.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=rkltE0VKwH", "arxiv_id": "1905.12127", "paper_pdf": "papers/rkltE0VKwH.pdf", "paper_pdf_sha256": "239cb76ff4f87fa4d4ff2c71ac36a2612331d8a5fc4c5becf1e51678cc45c4b9", "paper_pdf_bytes": 1209301, "paper_pdf_source": "openreview", "code_url": "https://github.com/shariqiqbal2810/Multi-Explore", "code_repository": "shariqiqbal2810/Multi-Explore", "code_commit": "116cfbec87f97c83dd16c271616287d0b1dcb0dd", "code_archive": "repos/rkltE0VKwH.zip", "code_archive_sha256": "9a03d94a0eab37a6fb2b293f529bc9486ec7c339cc299ad1e276770255758752", "code_archive_bytes": 61094, "code_file_count": 12, "code_extensions": {".py": 12}, "github_disk_usage_kb": 47, "github_languages": {"Python": 142794}, "github_archived": false, "github_pushed_at": "2021-05-22T18:45:30Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/coordinated-exploration-via-intrinsic-rewards"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "m3FOf6nKnU", "year": 2026, "status": "rejected", "title": "Forget Vectors at Play: Universal Input Perturbations Driving Machine Unlearning in Image Classification", "authors": ["Changchang Sun", "Ren Wang", "Yihua Zhang", "Jinghan Jia", "Jiancheng Liu", "Gaowen Liu", "Yan Yan", "Sijia Liu"], "authorids": ["~Changchang_Sun1", "~Ren_Wang1", "~Yihua_Zhang1", "~Jinghan_Jia1", "~Jiancheng_Liu2", "~Gaowen_Liu4", "~Yan_Yan6", "~Sijia_Liu1"], "authors_source": "OpenReview API", "abstract": "Machine unlearning (MU), which seeks to erase the influence of specific unwanted data from already-trained models, is becoming increasingly vital in model editing, particularly to comply with evolving data regulations like the \"right to be forgotten\". Conventional approaches are predominantly model-based, typically requiring retraining or fine-tuning the model's weights to meet unlearning requirements. In this work, we approach the MU problem from an input perturbation-based perspective, where the model weights remain intact throughout the unlearning process. We demonstrate the existence of a proactive input-based unlearning strategy, referred to forget vector, which can be generated as an input-agnostic data perturbation and remains as effective as model-based approximate unlearning approaches. We also explore forget vector arithmetic, whereby multiple class-specific forget vectors can be combined through simple operations (e.g., linear combinations) to generate new forget vectors for unseen unlearning tasks, such as forgetting arbitrary subsets across classes. Extensive experiments validate the effectiveness and adaptability of the forget vector, showcasing its competitive performance relative to state-of-the-art model-based methods while achieving superior parameter efficiency.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=m3FOf6nKnU", "arxiv_id": "2412.16780", "paper_pdf": "papers/m3FOf6nKnU.pdf", "paper_pdf_sha256": "8f9b0ebb8cf819d23b4d76ca4f466a460c455820f7494cea8f2291e953246aea", "paper_pdf_bytes": 3305568, "paper_pdf_source": "openreview", "code_url": "https://github.com/Changchangsun/Forget-Vector", "code_repository": "Changchangsun/Forget-Vector", "code_commit": "6abd0eb5c3d22d6dfd530bdfdd4c65abc944121f", "code_archive": "repos/m3FOf6nKnU.zip", "code_archive_sha256": "547c90f7714d7f86aa518a608ff43f151c659eb4ab6cfdc5e75085138524d3f3", "code_archive_bytes": 170253, "code_file_count": 47, "code_extensions": {".py": 47}, "github_disk_usage_kb": 203, "github_languages": {"Python": 453103}, "github_archived": false, "github_pushed_at": "2026-01-06T00:46:36Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/forget-vectors-at-play-universal-input"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "vJmpg0exYA", "year": 2025, "status": "rejected", "title": "DiscQuant: A Quantization Method for Neural Networks Inspired by Discrepancy Theory", "authors": ["Jerry Chee", "Arturs Backurs", "Rainie Heck", "Li Zhang", "Janardhan Kulkarni", "Thomas Rothvoss", "Sivakanth Gopi"], "authorids": ["~Jerry_Chee1", "~Arturs_Backurs1", "~Rainie_Heck1", "~Li_Zhang11", "~Janardhan_Kulkarni2", "~Thomas_Rothvoss1", "~Sivakanth_Gopi1"], "authors_source": "OpenReview API", "abstract": "Quantizing the weights of a neural network has two steps: (1) Finding a good low bit-complexity representation for weights (which we call the quantization grid) and (2) Rounding the original weights to values in the quantization grid. In this paper, we study the problem of rounding optimally given any quantization grid. The simplest and most commonly used way to round is Round-to-Nearest (RTN). By rounding in a data-dependent way instead, one can improve the quality of the quantized model significantly.\n\nWe study the rounding problem from the lens of \\emph{discrepancy theory}, which studies how well we can round a continuous solution to a discrete solution without affecting solution quality too much. We prove that given $m=poly(1/\\epsilon)$ samples from the data distribution, we can round all but $O(m)$ model weights such that the expected approximation error of the quantized model on the true data distribution is $\\le \\epsilon$ as long as the space of gradients of the original model is approximately low rank (which we empirically validate).\n\nOur proof, which is algorithmic, inspired a  simple and practical rounding algorithm called \\emph{DiscQuant}. In our experiments, we demonstrate that DiscQuant significantly improves over the prior state-of-the-art rounding method called GPTQ and the baseline RTN over a range of benchmarks on Phi3mini-3.8B and Llama3.1-8B. For example, rounding Phi3mini-3.8B to a fixed quantization grid with 3.25 bits per parameter using DiscQuant gets 64\\% accuracy on the GSM8k dataset, whereas GPTQ achieves 54\\% and RTN achieves 31\\% (the original model achieves 84\\%).", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=vJmpg0exYA", "arxiv_id": "2501.06417", "paper_pdf": "papers/vJmpg0exYA.pdf", "paper_pdf_sha256": "80f4912851402231ebd62268d7269a6c4ea54ac1d6e8befe107e4f9af33e771b", "paper_pdf_bytes": 926711, "paper_pdf_source": "openreview", "code_url": "https://github.com/jerry-chee/DiscQuant", "code_repository": "jerry-chee/DiscQuant", "code_commit": "b1db18b079803982d70d2e74e3679e6a96a683e5", "code_archive": "repos/vJmpg0exYA.zip", "code_archive_sha256": "c44e0ee2ae20182a892f5101629dc1704215f42a4fcaebc795890306f4f4a1c3", "code_archive_bytes": 38876, "code_file_count": 15, "code_extensions": {".py": 12, ".sh": 3}, "github_disk_usage_kb": 34, "github_languages": {"Python": 233732, "Shell": 4702}, "github_archived": false, "github_pushed_at": "2025-01-10T04:16:30Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/discquant-a-quantization-method-for-neural"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "IhWtRwIbos", "year": 2024, "status": "rejected", "title": "Discovering Environments with XRM", "authors": ["Mohammad Pezeshki", "Diane Bouchacourt", "Mark Ibrahim", "Nicolas Ballas", "Pascal Vincent", "David Lopez-Paz"], "authorids": ["~Mohammad_Pezeshki1", "~Diane_Bouchacourt3", "~Mark_Ibrahim1", "~Nicolas_Ballas1", "~Pascal_Vincent1", "~David_Lopez-Paz2"], "authors_source": "OpenReview API", "abstract": "Successful out-of-distribution generalization requires environment annotations. Unfortunately, these are resource-intensive to obtain, and their relevance to model performance is limited by the expectations and perceptual biases of human annotators. Therefore, to enable robust AI systems across applications, we must develop algorithms to automatically discover environments inducing broad generalization. Current proposals, which divide examples based on their training error, suffer from one fundamental problem. These methods add hyper-parameters and early-stopping criteria that are impossible to tune without a validation set with human-annotated environments, the very information subject to discovery. In this paper, we propose Cross-Risk-Minimization (XRM) to address this issue. XRM trains two twin networks, each learning from one random half of the training data, while imitating confident held-out mistakes made by its sibling. XRM provides a recipe for hyper-parameter tuning, does not require early-stopping, and can discover environments for all training and validation data. Domain generalization algorithms built on top of XRM environments achieve oracle worst-group-accuracy, solving a long-standing problem in out-of-distribution generalization.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=IhWtRwIbos", "arxiv_id": "2309.16748", "paper_pdf": "papers/IhWtRwIbos.pdf", "paper_pdf_sha256": "530817f989fd5a1b94db56b513bd3e3595723eb4254ca1e29fc33e274f43431d", "paper_pdf_bytes": 1253272, "paper_pdf_source": "openreview", "code_url": "https://github.com/facebookresearch/XRM", "code_repository": "facebookresearch/XRM", "code_commit": "5e90815250aedea546b9e17ffd5d7e0a7c34c20f", "code_archive": "repos/IhWtRwIbos.zip", "code_archive_sha256": "9f5ff3e8126cebd856716093df2e8f37adc07f5ed500e631b6a25694f33c0660", "code_archive_bytes": 32953, "code_file_count": 9, "code_extensions": {".py": 9}, "github_disk_usage_kb": 32, "github_languages": {"Python": 78103}, "github_archived": true, "github_pushed_at": "2024-12-06T22:23:08Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/discovering-environments-with-xrm"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "WcTLZrpzfe", "year": 2023, "status": "rejected", "title": "Orientation-Aware Graph Neural Networks for Protein Structure Representation Learning", "authors": ["Jiahan Li", "Shitong Luo", "Congyue Deng", "Chaoran Cheng", "Jiaqi Guan", "Leonidas Guibas", "Jianzhu Ma", "Jian Peng"], "authorids": ["~Jiahan_Li2", "~Shitong_Luo1", "~Congyue_Deng1", "~Chaoran_Cheng2", "~Jiaqi_Guan1", "~Leonidas_Guibas1", "~Jianzhu_Ma2", "~Jian_Peng1"], "authors_source": "OpenReview API", "abstract": "By folding to particular 3D structures, proteins play a key role in living beings. To learn meaningful representation from a protein structure for downstream tasks, not only the global backbone topology but the local fine-grained orientational relations between amino acids should also be considered. In this work, we propose the Orientation-Aware Graph Neural Networks (OAGNNs) to better sense the geometric characteristics in protein structure (e.g. inner-residue torsion angles, inter-residue orientations). Extending a single weight from a scalar to a 3D vector, we construct a rich set of geometric-meaningful operations to process both the classical and SO(3) representations of a given structure. To plug our designed perceptron unit into existing Graph Neural Networks, we further introduce an equivariant message passing paradigm, showing superior versatility in maintaining SO(3)-equivariance at the global scale. Experiments have shown that our OAGNNs have a remarkable ability to sense geometric orientational features compared to classical networks. OAGNNs have also achieved state-of-the-art performance on various computational biology applications related to protein 3D structures.\n\n", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=WcTLZrpzfe", "arxiv_id": "2201.13299", "paper_pdf": "papers/WcTLZrpzfe.pdf", "paper_pdf_sha256": "bbbfb7b1c456222f86bebe6a7d19126a6fa3e902167dd10ada454684a4b822f5", "paper_pdf_bytes": 1797323, "paper_pdf_source": "openreview", "code_url": "https://github.com/Ced3-han/OAGNN", "code_repository": "Ced3-han/OAGNN", "code_commit": "63307a57c9c4d02ec4f1bee780c117e72c2b93c3", "code_archive": "repos/WcTLZrpzfe.zip", "code_archive_sha256": "bbd1ba1b98ea2ee611018ce54b190df26c727b0f6a0352aceb06a95a1026b63f", "code_archive_bytes": 34628, "code_file_count": 18, "code_extensions": {".py": 18}, "github_disk_usage_kb": 40, "github_languages": {"Python": 88581}, "github_archived": false, "github_pushed_at": "2025-05-27T14:14:15Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/directed-weight-neural-networks-for-protein"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Py8WbvKH_wv", "year": 2022, "status": "rejected", "title": "DRIBO: Robust Deep Reinforcement Learning via Multi-View Information Bottleneck", "authors": ["Jiameng Fan", "Wenchao Li"], "authorids": ["~Jiameng_Fan1", "~Wenchao_Li1"], "authors_source": "OpenReview API", "abstract": "Deep reinforcement learning (DRL) agents are often sensitive to visual changes that were unseen in their training environments. To address this problem, we leverage the sequential nature of RL to learn robust representations that encode only task-relevant information from observations based on the unsupervised multi-view setting. Specifically, we introduce a novel contrastive version of Multi-View Information Bottleneck (MIB) objective for temporal data. We train RL agents from pixels with this auxiliary objective to learn robust representations that can compress away task-irrelevant information and are predictive of task-relevant dynamics. This approach enables us to train high-performance policies that are robust to visual distractions and can generalize well to unseen environments. We demonstrate that our approach can achieve SOTA performance on diverse visual control tasks on the DeepMind Control Suite when the background is replaced with natural videos. In addition, we show that our approach outperforms well-established baselines for generalization to unseen environments on the Procgen benchmark.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=Py8WbvKH_wv", "arxiv_id": "2102.13268", "paper_pdf": "papers/Py8WbvKH_wv.pdf", "paper_pdf_sha256": "d6b14c9f4426cb524a4a2e5fc0b41a735fe1446b6921365c8626b3e1bad3dba0", "paper_pdf_bytes": 5377900, "paper_pdf_source": "openreview", "code_url": "https://github.com/BU-DEPEND-Lab/DRIBO", "code_repository": "BU-DEPEND-Lab/DRIBO", "code_commit": "bb3374e2ae861ed9aaf7606b43a027247f3fabc6", "code_archive": "repos/Py8WbvKH_wv.zip", "code_archive_sha256": "b7bfaa71ddba14541a5c0c724c44277a0752d054c1aae9d33e9afd9b23c523f9", "code_archive_bytes": 42557, "code_file_count": 21, "code_extensions": {".py": 15, ".sh": 6}, "github_disk_usage_kb": 36, "github_languages": {"Python": 136324, "Shell": 13326}, "github_archived": false, "github_pushed_at": "2022-06-14T13:51:08Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/robust-deep-reinforcement-learning-via-multi"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "w5bNwUzj33", "year": 2021, "status": "rejected", "title": "Cross-Domain Few-Shot Learning by Representation Fusion", "authors": ["Thomas Adler", "Johannes Brandstetter", "Michael Widrich", "Andreas Mayr", "David Kreil", "Michael K Kopp", "Günter Klambauer", "Sepp Hochreiter"], "authorids": ["~Thomas_Adler1", "~Johannes_Brandstetter1", "~Michael_Widrich2", "~Andreas_Mayr2", "openreview20@kreil.org", "~Michael_K_Kopp1", "~Günter_Klambauer1", "~Sepp_Hochreiter1"], "authors_source": "OpenReview API", "abstract": "In order to quickly adapt to new data, few-shot learning aims at learning from few examples, often by using already acquired knowledge. The new data often differs from the previously seen data due to a domain shift, that is, a change of the input-target distribution. While several methods perform well on small domain shifts like new target classes with similar inputs, larger domain shifts are still challenging. Large domain shifts may result in abstract concepts that are not shared between the original and the new domain. However, low-level concepts like edges in images might still be shared and useful. For cross-domain few-shot learning, we suggest representation fusion to unify different abstraction levels of a deep neural network into one representation. We propose Cross-domain Hebbian Ensemble Few-shot learning (CHEF), which consists of representation fusion by an ensemble of Hebbian learners acting on different layers of a deep neural network that was trained on the original domain. On the few-shot datasets miniImagenet and tieredImagenet, where the domain shift is small, CHEF is competitive with state-of-the-art methods. On cross-domain few-shot benchmark challenges with larger domain shifts, CHEF obtains state-of-the-art results in all categories. We further apply CHEF on a real-world cross-domain application in drug discovery. We consider a domain shift from bioactive molecules to environmental chemicals and drugs with twelve associated toxicity prediction tasks. On these tasks that are highly relevant for computational drug discovery, CHEF significantly outperforms all its competitors. ", "decision": "Reject", "meta_review": null, "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=w5bNwUzj33", "arxiv_id": "2010.06498", "paper_pdf": "papers/w5bNwUzj33.pdf", "paper_pdf_sha256": "0a842fb5522b231c3a02cf0c4be8d17993e87922e692fd9f4841916b9b039408", "paper_pdf_bytes": 364312, "paper_pdf_source": "openreview", "code_url": "https://github.com/tomte812/chef", "code_repository": "tomte812/chef", "code_commit": "ace4fa3621669c4454044bf1932b46ef0a59af44", "code_archive": "repos/w5bNwUzj33.zip", "code_archive_sha256": "d2930c076c14dc2f30236c099a12f62a6cf6737224cabb7c18c601e9664cdc59", "code_archive_bytes": 55919, "code_file_count": 24, "code_extensions": {".py": 24}, "github_disk_usage_kb": 41, "github_languages": {"Python": 160628}, "github_archived": false, "github_pushed_at": "2020-10-05T07:05:07Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/cross-domain-few-shot-learning-by-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "BJevJCVYvB", "year": 2020, "status": "rejected", "title": "Training Neural Networks for and by Interpolation", "authors": ["Leonard Berrada", "Andrew Zisserman", "Pawan M. Kumar"], "authorids": ["lberrada@robots.ox.ac.uk", "az@robots.ox.ac.uk", "pawan@robots.ox.ac.uk"], "authors_source": "OpenReview API", "abstract": "In modern supervised learning, many deep neural networks are able to interpolate the data: the empirical loss can be driven to near zero on all samples simultaneously. In this work, we explicitly exploit this interpolation property for the design of a new optimization algorithm for deep learning. Specifically, we use it to compute an adaptive learning-rate in closed form at each iteration. This results in the Adaptive Learning-rates for Interpolation with Gradients (ALI-G) algorithm. ALI-G retains the main advantage of SGD which is a low computational cost per iteration. But unlike SGD, the learning-rate of ALI-G uses a single constant hyper-parameter and does not require a decay schedule, which makes it considerably easier to tune. We provide convergence guarantees of ALI-G in the stochastic convex setting. Notably, all our convergence results tackle the realistic case where the interpolation property is satisfied up to some tolerance. We provide experiments on a variety of architectures and tasks: (i) learning a differentiable neural computer; (ii) training a wide residual network on the SVHN data set; (iii) training a Bi-LSTM on the SNLI data set; and (iv) training wide residual networks and densely connected networks on the CIFAR data sets. ALI-G produces state-of-the-art results among adaptive methods, and even yields comparable performance with SGD, which requires manually tuned learning-rate schedules. Furthermore, ALI-G is simple to implement in any standard deep learning framework and can be used as a drop-in replacement in existing code.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=BJevJCVYvB", "arxiv_id": "1906.05661", "paper_pdf": "papers/BJevJCVYvB.pdf", "paper_pdf_sha256": "ceb527e116770d06f06aea1025ac9cd14923fabe563394be0a379d5173852316", "paper_pdf_bytes": 605903, "paper_pdf_source": "openreview", "code_url": "https://github.com/oval-group/ali-g", "code_repository": "oval-group/ali-g", "code_commit": "568947a2bbb651be3401fbae60229e2fcbf8a505", "code_archive": "repos/BJevJCVYvB.zip", "code_archive_sha256": "c9fb31b7b9dc9fcfb3e119b323f6c87e474c98342869771893ef60fd3e8fb4e7", "code_archive_bytes": 37466, "code_file_count": 30, "code_extensions": {".py": 30}, "github_disk_usage_kb": 48, "github_languages": {"Python": 76994}, "github_archived": false, "github_pushed_at": "2021-03-07T15:57:21Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/training-neural-networks-for-and-by"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "HygUOoC5KX", "year": 2019, "status": "rejected", "title": "Are Generative Classifiers More Robust to Adversarial Attacks?", "authors": ["Yingzhen Li", "John Bradshaw", "Yash Sharma"], "authorids": ["yl494@cam.ac.uk", "jab255@cam.ac.uk", "ysharma1126@gmail.com"], "authors_source": "OpenReview API", "abstract": "There is a rising interest in studying the robustness of deep neural network classifiers against adversaries, with both advanced attack and defence techniques being actively developed. However, most recent work focuses on discriminative classifiers, which only model the conditional distribution of the labels given the inputs. In this paper, we propose and investigate the deep Bayes classifier, which improves classical naive Bayes with conditional deep generative models. We further develop detection methods for adversarial examples, which reject inputs with low likelihood under the generative model. Experimental results suggest that deep Bayes classifiers are more robust than deep discriminative classifiers, and that the proposed detection methods are effective against many recently proposed attacks.", "decision": null, "meta_review": "Reject", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=HygUOoC5KX", "arxiv_id": "1802.06552", "paper_pdf": "papers/HygUOoC5KX.pdf", "paper_pdf_sha256": "c4119443dc3e9b71f4f1b3fea836e27cee60a0f22268cd445847096f30517167", "paper_pdf_bytes": 2991445, "paper_pdf_source": "openreview", "code_url": "https://github.com/deepgenerativeclassifier/DeepBayes", "code_repository": "deepgenerativeclassifier/DeepBayes", "code_commit": "b7d78330638b8dfc8b64b1c19f460c74944bebda", "code_archive": "repos/HygUOoC5KX.zip", "code_archive_sha256": "c2489fc08e6b2e69b5d2edc231a5ef9f945fb10f9a0e30d0d010a35b757e49ee", "code_archive_bytes": 265888, "code_file_count": 73, "code_extensions": {".py": 73}, "github_disk_usage_kb": 228, "github_languages": {"Python": 472032}, "github_archived": false, "github_pushed_at": "2019-05-10T10:29:49Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/are-generative-classifiers-more-robust-to"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "HklpCzC6-", "year": 2018, "status": "rejected", "title": "Image Segmentation by Iterative Inference from Conditional Score Estimation", "authors": ["Adriana Romero", "Michal Drozdzal", "Akram Erraqabi", "Simon Jégou", "Yoshua Bengio"], "authorids": ["adriana.romsor@gmail.com", "michal.drozdzal@gmail.com", "akram.er-raqabi@umontreal.ca", "simon.jegou@gmail.com", "yoshua.umontreal@gmail.com"], "authors_source": "OpenReview API", "abstract": "Inspired by the combination of feedforward and iterative computations in the visual cortex, and taking advantage of the ability of denoising autoencoders to estimate the score of a joint distribution, we propose a novel approach to iterative inference for capturing and exploiting the complex joint distribution of output variables conditioned on some input variables. This approach is applied to image pixel-wise segmentation, with the estimated conditional score used to perform gradient ascent towards a mode of the estimated conditional distribution. This extends previous work on score estimation by denoising autoencoders to the case of a conditional distribution, with a novel use of a corrupted feedforward predictor replacing Gaussian corruption. An advantage of this approach over more classical ways to perform iterative inference for structured outputs, like conditional random fields (CRFs), is that it is not any more necessary to define an explicit energy function linking the output variables. To keep computations tractable, such energy function parametrizations are typically fairly constrained, involving only a few neighbors of each of the output variables in each clique. We experimentally find that the proposed iterative inference from conditional score estimation by conditional denoising autoencoders performs better than comparable models based on CRFs or those not using any explicit modeling of the conditional joint distribution of outputs.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=HklpCzC6-", "arxiv_id": "1705.07450", "paper_pdf": "papers/HklpCzC6-.pdf", "paper_pdf_sha256": "bb797bccb459d007147e87e7736b81eb9d93ba0635d1bb475675e706e25a3b14", "paper_pdf_bytes": 3639524, "paper_pdf_source": "openreview", "code_url": "https://github.com/adri-romsor/iterative_inference_segm", "code_repository": "adri-romsor/iterative_inference_segm", "code_commit": "33e3a988034b9c807ad15c9d0c2ab8bb32d79a39", "code_archive": "repos/HklpCzC6-.zip", "code_archive_sha256": "ad43313e1c86ef87627b75a49c52fc03dccf205db50e6188d8551439283ed79d", "code_archive_bytes": 101174, "code_file_count": 25, "code_extensions": {".py": 23, ".ipynb": 1, ".sh": 1}, "github_disk_usage_kb": 453, "github_languages": {"Python": 196495, "Jupyter Notebook": 60338, "Shell": 193}, "github_archived": false, "github_pushed_at": "2017-08-07T21:42:44Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/image-segmentation-by-iterative-inference"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "rpblsD3eXG", "year": 2026, "status": "rejected", "title": "Communication-Efficient Multi-Device Inference Acceleration for Transformer Models", "authors": ["Xiao Liu", "Lijun Zhang", "Deepak Ganesan", "Hui Guan"], "authorids": ["~Xiao_Liu9", "~Lijun_Zhang4", "~Deepak_Ganesan1", "~Hui_Guan1"], "authors_source": "OpenReview API", "abstract": "Transformer models power many AI applications but suffer from high inference latency, limiting their use in real-time settings. Multi-device inference can reduce latency by parallelizing computation. Yet, existing methods require high inter-device bandwidth, making them impractical for bandwidth-constrained environments. We propose ASTRA, a communication-efficient framework that accelerates Transformer inference through a novel integration of sequence parallelism and a Mixed-Precision Attention mechanism designed to minimize inter-device communication. ASTRA compresses non-local token embeddings via vector quantization and preserves task accuracy through two optimizations, Noise-Augmented Quantization and Distributed Class Tokens. Experiments on ViT and GPT2 across vision and NLP tasks show that Astra achieves up to 2.64$\\times$ speedups over single-device inference and up to 15.25$\\times$ speedups over state-of-the-art multi-device inferences, while operating under bandwidths as low as 10 Mbps.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=rpblsD3eXG", "arxiv_id": "2505.19342", "paper_pdf": "papers/rpblsD3eXG.pdf", "paper_pdf_sha256": "b8aa60f7841b593dec32649b2cb76c11783fcbb6cd2ae6db3262dde2112d6299", "paper_pdf_bytes": 942944, "paper_pdf_source": "openreview", "code_url": "https://github.com/xl1990/Astra", "code_repository": "xl1990/Astra", "code_commit": "c959791453fb5627027825b019ab20ca9384e6ad", "code_archive": "repos/rpblsD3eXG.zip", "code_archive_sha256": "1ee487edbae1c968c037536f1d429a3fb0716102b983ab3e51d1c1d12658a99a", "code_archive_bytes": 519372, "code_file_count": 14, "code_extensions": {".py": 9, ".sh": 5}, "github_disk_usage_kb": 224, "github_languages": {"Python": 90349, "Shell": 6165}, "github_archived": false, "github_pushed_at": "2025-05-25T22:05:47Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/communication-efficient-multi-device"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "rnRBGMNYa2", "year": 2025, "status": "rejected", "title": "Response Tuning: Aligning Large Language Models without Instruction", "authors": ["Seokhyun An", "Hyounghun Kim"], "authorids": ["~Seokhyun_An1", "~Hyounghun_Kim1"], "authors_source": "OpenReview API", "abstract": "Instruction tuning—supervised fine-tuning using instruction-response pairs—is a foundational step in transitioning pre-trained Large Language Models (LLMs) into helpful and safe chat assistants. Our hypothesis is that establishing an adequate output distribution can enable such a transition given the capabilities inherent in pre-trained LLMs. To verify this, we propose Response Tuning (RT), which eliminates the instruction-conditioning step in instruction tuning and solely focuses on response distribution supervision. Our experiments demonstrate that RT models, trained only using responses, can effectively respond to a wide range of instructions and exhibit helpfulness comparable to that of their instruction-tuned counterparts. Furthermore, we observe that controlling the training response distribution can substantially improve their user preference or elicit target behaviors such as refusing assistance for unsafe queries. Our findings illuminate the role of establishing an adequate output distribution in alignment, highlighting the potential of the extensive inherent capabilities of pre-trained LLMs.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=rnRBGMNYa2", "arxiv_id": "2410.02465", "paper_pdf": "papers/rnRBGMNYa2.pdf", "paper_pdf_sha256": "2c546f7a0663f82b345270f848d742a2fb25ab5b7f735db5088ec4c85feedfcf", "paper_pdf_bytes": 1123010, "paper_pdf_source": "openreview", "code_url": "https://github.com/seokhyunan/response-tuning", "code_repository": "seokhyunan/response-tuning", "code_commit": "8db38cbbf14740ff2ddf8cf16cd033f09d2af449", "code_archive": "repos/rnRBGMNYa2.zip", "code_archive_sha256": "d08bea568563f111beb5bd88dd6bb85cfb8296d89daff07ad4f7f25f57993c06", "code_archive_bytes": 47126, "code_file_count": 26, "code_extensions": {".py": 15, ".sh": 11}, "github_disk_usage_kb": 35, "github_languages": {"Python": 84369, "Shell": 9165}, "github_archived": false, "github_pushed_at": "2025-09-12T06:49:37Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/response-tuning-aligning-large-language"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "EE75tyB5Ay", "year": 2024, "status": "rejected", "title": "On the Generalization of Training-based ChatGPT Detection Methods", "authors": ["Han Xu", "Jie Ren", "Pengfei He", "Shenglai Zeng", "Yingqian Cui", "Amy Liu", "Hui Liu", "Jiliang Tang"], "authorids": ["~Han_Xu1", "~Jie_Ren6", "~Pengfei_He2", "~Shenglai_Zeng2", "~Yingqian_Cui1", "amyziyuliu@gmail.com", "~Hui_Liu8", "~Jiliang_Tang1"], "authors_source": "OpenReview API", "abstract": "ChatGPT is one of the most popular language models which achieve amazing performance on various natural language tasks. Consequently, there is also an urgent need to detect the texts generated ChatGPT from human written. One of the extensively studied methods trains classification models to distinguish both. However, existing studies also demonstrate that the trained models may suffer from distribution shifts (during test), i.e., they are ineffective to predict the generated texts from unseen  language tasks or topics. In this work, we aim to have a comprehensive investigation on these methods' generalization behaviors under distribution shift caused by a wide range of factors, including prompts, text lengths, topics, and language tasks. To achieve this goal, we first collect a new dataset with human and ChatGPT texts, and then we conduct extensive studies on the collected dataset. Our studies unveil insightful findings which provide guidance for developing future methodologies or data collection strategies for ChatGPT detection.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=EE75tyB5Ay", "arxiv_id": "2310.01307", "paper_pdf": "papers/EE75tyB5Ay.pdf", "paper_pdf_sha256": "3abe01c531e2b31c13d7c77ee27d3d4154e73af831b6966d10527fb55d4b9c9a", "paper_pdf_bytes": 3363618, "paper_pdf_source": "openreview", "code_url": "https://github.com/hannxu123/hcvar", "code_repository": "hannxu123/hcvar", "code_commit": "c5c6aed6c013610bf91606e0cdef0d052b4da8e4", "code_archive": "repos/EE75tyB5Ay.zip", "code_archive_sha256": "8f77bd09813a4f4a5c4351265c8b8ec45064c9297467f83aacc145b6adb4cddc", "code_archive_bytes": 7866, "code_file_count": 3, "code_extensions": {".py": 3}, "github_disk_usage_kb": 33, "github_languages": {"Python": 17523}, "github_archived": false, "github_pushed_at": "2023-10-03T16:26:35Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/on-the-generalization-of-training-based"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "vKXd1m74DkN", "year": 2023, "status": "rejected", "title": "Artificial Replay: A Meta-Algorithm for Harnessing Historical Data in Bandits", "authors": ["Sid Banerjee", "Sean R. Sinclair", "Milind Tambe", "Lily Xu", "Christina Yu"], "authorids": ["~Sid_Banerjee1", "~Sean_R._Sinclair1", "~Milind_Tambe1", "~Lily_Xu1", "~Christina_Yu1"], "authors_source": "OpenReview API", "abstract": "While standard bandit algorithms sometimes incur high regret, their performance can be greatly improved by \"warm starting\" with historical data. Unfortunately, how best to incorporate historical data is unclear: naively initializing reward estimates using all historical samples can suffer from spurious data and imbalanced data coverage, leading to computational and storage issues - particularly in continuous action spaces. We address these two challenges by proposing Artificial Replay, a meta-algorithm for incorporating historical data into any arbitrary base bandit algorithm. Artificial Replayuses only a subset of the historical data as needed to reduce computation and storage. We show that for a broad class of base algorithms that satisfy independence of irrelevant data (IIData), a novel property that we introduce, our method achieves equal regret as a full warm-start approach while potentially using only a fraction of historical data. We complement these theoretical results with a case study of $K$-armed and continuous combinatorial bandit algorithms, including on a green security domain using real poaching data, to show the practical benefits of Artificial Replayin achieving optimal regret alongside low computational and storage costs.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=vKXd1m74DkN", "arxiv_id": "2210.00025", "paper_pdf": "papers/vKXd1m74DkN.pdf", "paper_pdf_sha256": "f867ff4aad5a9c20617f0cfbc636afe98cbc552aa9a13841e5faefc7df7dccc1", "paper_pdf_bytes": 706286, "paper_pdf_source": "openreview", "code_url": "https://github.com/lily-x/artificial-replay", "code_repository": "lily-x/artificial-replay", "code_commit": "7602ffcdc5a9766eefe806e4ad152a31a2d303b0", "code_archive": "repos/vKXd1m74DkN.zip", "code_archive_sha256": "b53aa425ab10e14e0758d820d94a46ba6d92e10f0403d25584e2711b9a6c24d4", "code_archive_bytes": 47228, "code_file_count": 26, "code_extensions": {".py": 26}, "github_disk_usage_kb": 40, "github_languages": {"Python": 122475}, "github_archived": false, "github_pushed_at": "2023-01-26T22:41:17Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/artificial-replay-a-meta-algorithm-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "a1m8Jba-N6l", "year": 2022, "status": "rejected", "title": "$k$-Mixup Regularization for Deep Learning via Optimal Transport", "authors": ["Kristjan Greenewald", "Anming Gu", "Mikhail Yurochkin", "Justin Solomon", "Edward Chien"], "authorids": ["~Kristjan_Greenewald1", "~Anming_Gu1", "~Mikhail_Yurochkin1", "~Justin_Solomon1", "~Edward_Chien1"], "authors_source": "OpenReview API", "abstract": "Mixup is a popular regularization technique for training deep neural networks that can improve generalization  and increase adversarial robustness. It perturbs input training data in the direction of other randomly-chosen instances in the training set. To better leverage the structure of the data, we extend mixup to $k$-mixup by perturbing $k$-batches of training points in the direction of other $k$-batches using displacement interpolation, i.e. interpolation under the Wasserstein metric. We demonstrate theoretically and in simulations that $k$-mixup preserves cluster and manifold structures, and we extend theory studying the efficacy of standard mixup to the $k$-mixup case. Our empirical results show that training with $k$-mixup further improves generalization and robustness across several network architectures and benchmark datasets of differing modalities.", "decision": "Reject", "meta_review": null, "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=a1m8Jba-N6l", "arxiv_id": "2106.02933", "paper_pdf": "papers/a1m8Jba-N6l.pdf", "paper_pdf_sha256": "f963cac5a1157591e5423c4dbc99fcde7c1cf76eba645a292c9f2493488ad84b", "paper_pdf_bytes": 7687462, "paper_pdf_source": "openreview", "code_url": "https://github.com/AnmingGu/kmixup-cifar10", "code_repository": "AnmingGu/kmixup-cifar10", "code_commit": "ed2ecfd47226baf375f90e12cad5b4a97899a149", "code_archive": "repos/a1m8Jba-N6l.zip", "code_archive_sha256": "bdc208c551f27d744334db5dcc959c74034c94814117959f9aad4ccc1a78e6cc", "code_archive_bytes": 21750, "code_file_count": 13, "code_extensions": {".py": 13}, "github_disk_usage_kb": 37, "github_languages": {"Python": 66782}, "github_archived": false, "github_pushed_at": "2023-08-28T20:12:31Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/k-mixup-regularization-for-deep-learning-via"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "tqc8n6oHCtZ", "year": 2021, "status": "rejected", "title": "Length-Adaptive Transformer: Train Once with Length Drop, Use Anytime with Search", "authors": ["Gyuwan Kim", "Kyunghyun Cho"], "authorids": ["~Gyuwan_Kim1", "~Kyunghyun_Cho1"], "authors_source": "OpenReview API", "abstract": "Although transformers have achieved impressive accuracies in various tasks in natural language processing, they often come with a prohibitive computational cost, that prevents their use in scenarios with limited computational resources for inference. This need for computational efficiency in inference has been addressed by for instance PoWER-BERT (Goyal et al., 2020) which gradually decreases the length of a sequence as it is passed through layers. These approaches however often assume that the target computational complexity is known in advance at the time of training. This implies that a separate model must be trained for each inference scenario with its distinct computational budget. In this paper, we extend PoWER-BERT to address this issue of inefficiency and redundancy. The proposed extension enables us to train a large-scale transformer, called Length-Adaptive Transformer, once and uses it for various inference scenarios without re-training it. To do so, we train a transformer with LengthDrop, a structural variant of dropout, which stochastically determines the length of a sequence at each layer. We then use a multi-objective evolutionary search to find a length configuration that maximizes the accuracy and minimizes the computational complexity under any given computational budget. Additionally, we significantly extend the applicability of PoWER-BERT beyond sequence-level classification into token-level classification such as span-based question-answering, by introducing the idea of Drop-and-Restore. With Drop-and-Restore, word-vectors are dropped temporarily in intermediate layers and restored at the last layer if necessary. We empirically verify the utility of the proposed approach by demonstrating the superior accuracy-efficiency trade-off under various setups, including SQuAD 1.1, MNLI-m, and SST-2. Upon publication, the code to reproduce our work will be open-sourced.\n", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=tqc8n6oHCtZ", "arxiv_id": "2010.07003", "paper_pdf": "papers/tqc8n6oHCtZ.pdf", "paper_pdf_sha256": "1498257af2d552ddc7840717dc3eb672b62b2d55e4cb2ba7efe22686eba77c28", "paper_pdf_bytes": 308856, "paper_pdf_source": "openreview", "code_url": "https://github.com/clovaai/length-adaptive-transformer", "code_repository": "clovaai/length-adaptive-transformer", "code_commit": "8c35cb3c2ea76112d973e8e5f867330af9f52113", "code_archive": "repos/tqc8n6oHCtZ.zip", "code_archive_sha256": "290896c6a9ac7a1f9080700f6b3a021b9916f682f1e1cb3509462f936ba40225", "code_archive_bytes": 47571, "code_file_count": 10, "code_extensions": {".py": 10}, "github_disk_usage_kb": 43, "github_languages": {"Python": 151470}, "github_archived": false, "github_pushed_at": "2020-11-02T12:22:07Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/length-adaptive-transformer-train-once-with-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "BJxt2aVFPr", "year": 2020, "status": "rejected", "title": "Optimizing Data Usage via Differentiable Rewards", "authors": ["Xinyi Wang", "Hieu Pham", "Paul Michel", "Antonios Anastasopoulos", "Graham Neubig", "Jaime Carbonell"], "authorids": ["xinyiw1@cs.cmu.edu", "hyhieu@cmu.edu", "pmichel1@cs.cmu.edu", "aanastas@andrew.cmu.edu", "gneubig@cs.cmu.edu", "jgc@cs.cmu.edu"], "authors_source": "OpenReview API", "abstract": "To acquire a new skill, humans learn better and faster if a tutor, based on their current knowledge level, informs them of how much attention they should pay to particular content or practice problems. Similarly, a machine learning model could potentially be trained better with a scorer that “adapts” to its current learning state and estimates the importance of each training data instance. Training such an adaptive scorer efficiently is a challenging problem; in order to precisely quantify the effect of a data instance at a given time during the training, it is typically necessary to first complete the entire training process. To efficiently optimize data usage, we propose a reinforcement learning approach called Differentiable Data Selection (DDS). In DDS, we formulate a scorer network as a learnable function of the training data, which can be efficiently updated along with the main model being trained. Specifically, DDS updates the scorer with an intuitive reward signal: it should up-weigh the data that has a similar gradient with a dev set upon which we would finally like to perform well. Without significant computing overhead, DDS delivers strong and consistent improvements over several strong baselines on two very different tasks of machine translation and image classification.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=BJxt2aVFPr", "arxiv_id": "1911.10088", "paper_pdf": "papers/BJxt2aVFPr.pdf", "paper_pdf_sha256": "e2e75774eaa658db7b1b16c602f01d4f0a3c0d0d4f7b4d6d0113b34ebc57667f", "paper_pdf_bytes": 932144, "paper_pdf_source": "openreview", "code_url": "https://github.com/cindyxinyiwang/DataSelection", "code_repository": "cindyxinyiwang/DataSelection", "code_commit": "b7cbb2cd819640aa9ed0e1090c18fce2095082ae", "code_archive": "repos/BJxt2aVFPr.zip", "code_archive_sha256": "9d583ea78e867281933b50c814a52c88f2f6c93d4de6ac99aa5bdd1eba06d87d", "code_archive_bytes": 61741, "code_file_count": 24, "code_extensions": {".py": 17, ".sh": 7}, "github_disk_usage_kb": 52, "github_languages": {"Python": 262569, "Shell": 12989, "Perl": 4412}, "github_archived": false, "github_pushed_at": "2021-06-20T00:19:55Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/optimizing-data-usage-via-differentiable-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "rJiaRbk0-", "year": 2018, "status": "rejected", "title": "Towards Binary-Valued Gates for Robust LSTM Training ", "authors": ["Zhuohan Li", "Di He", "Fei Tian", "Wei Chen", "Tao Qin", "Liwei Wang", "Tie-Yan Liu"], "authorids": ["lizhuohan@pku.edu.cn", "di_he@pku.edu.cn", "fetia@microsoft.com", "wche@microsoft.com", "taoqin@microsoft.com", "wanglw@cis.pku.edu.cn", "tyliu@microsoft.com"], "authors_source": "OpenReview API", "abstract": "Long Short-Term Memory (LSTM) is one of the most widely used recurrent structures in sequence modeling. Its goal is to use gates to control the information flow (e.g., whether to skip some information/transformation or not) in the recurrent computations, although its practical implementation based on soft gates only partially achieves this goal and is easy to overfit. In this paper, we propose a new way for LSTM training, which pushes the values of the gates towards 0 or 1. By doing so, we can (1) better control the information flow: the gates are mostly open or closed, instead of in a middle state; and (2) avoid overfitting to certain extent: the gates operate at their flat regions, which is shown to correspond to better generalization ability. However, learning towards discrete values of the gates is generally difficult. To tackle this challenge, we leverage the recently developed Gumbel-Softmax trick from the field of variational methods, and make the model trainable with standard backpropagation. Experimental results on language modeling and machine translation show that (1) the values of the gates generated by our method are more reasonable and intuitively interpretable, and (2) our proposed method generalizes better and achieves better accuracy on test sets in all tasks. Moreover, the learnt models are not sensitive to low-precision approximation and low-rank approximation of the gate parameters due to the flat loss surface.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=rJiaRbk0-", "arxiv_id": "1806.02988", "paper_pdf": "papers/rJiaRbk0-.pdf", "paper_pdf_sha256": "30f2ae910e4e16a7636eb28772281d71599260ec5674b38373054756f369fef1", "paper_pdf_bytes": 763886, "paper_pdf_source": "openreview", "code_url": "https://github.com/zhuohan123/g2-lstm", "code_repository": "zhuohan123/g2-lstm", "code_commit": "542aee7c014f3d1c24da23c5363994c3d7b3999e", "code_archive": "repos/rJiaRbk0-.zip", "code_archive_sha256": "51c6e7d06c403f4783d97ef9c50686407d75681247b799b8f5a7c438cf1e374e", "code_archive_bytes": 676052, "code_file_count": 53, "code_extensions": {".py": 52, ".sh": 1}, "github_disk_usage_kb": 642, "github_languages": {"Python": 345728, "Perl": 15002, "Shell": 783}, "github_archived": false, "github_pushed_at": "2018-07-22T18:45:38Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/towards-binary-valued-gates-for-robust-lstm"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "qPwSoOQPrI", "year": 2026, "status": "rejected", "title": "Leave it to the Specialist: Repair Sparse LLMs with Sparse Fine-Tuning via Sparsity Evolution", "authors": ["Qiao Xiao", "Alan Ansell", "Boqian Wu", "Lu Yin", "Mykola Pechenizkiy", "Shiwei Liu", "Decebal Constantin Mocanu"], "authorids": ["~Qiao_Xiao1", "~Alan_Ansell1", "~Boqian_Wu1", "~Lu_Yin1", "~Mykola_Pechenizkiy1", "~Shiwei_Liu2", "~Decebal_Constantin_Mocanu1"], "authors_source": "OpenReview API", "abstract": "Large language models (LLMs) have achieved remarkable success across various tasks but face deployment challenges due to their massive computational demands. While post-training pruning methods like SparseGPT and Wanda can effectively reduce the model size, but struggle to maintain model performance at high sparsity levels, limiting their utility for downstream tasks. Existing fine-tuning methods, such as full fine-tuning and LoRA, fail to preserve sparsity as they require updating the whole dense matrices, not well-suited for sparse LLMs. In this paper, we propose \\textbf{Sparsity Evolution Fine-Tuning (SEFT)}, a novel method designed specifically for sparse LLMs. SEFT dynamically evolves the sparse topology of pruned models during fine-tuning, while preserving the overall sparsity throughout the process. The strengths of SEFT lie in its ability to perform task-specific adaptation through a weight drop-and-grow strategy, enabling the pruned model to self-adapt its sparse connectivity pattern based on the target dataset. Furthermore, a sensitivity-driven pruning criterion is employed to ensure that the desired sparsity level is consistently maintained throughout fine-tuning. Our experiments on various LLMs, including LLaMA families, DeepSeek, and Mistral, across a diverse set of benchmarks demonstrate that SEFT achieves stronger performance while offering superior memory and computation efficiency compared to existing baselines. The code is provided in the supplementary material and will be released publicly.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=qPwSoOQPrI", "arxiv_id": "2505.24037", "paper_pdf": "papers/qPwSoOQPrI.pdf", "paper_pdf_sha256": "f4bbfe734ca000370d6ce2688093827945d9112641c022ae27d4a56fe8f0f8af", "paper_pdf_bytes": 1611300, "paper_pdf_source": "openreview", "code_url": "https://github.com/QiaoXiao7282/SEFT", "code_repository": "QiaoXiao7282/SEFT", "code_commit": "3fea1d28ee6d1ac0711acb295ec9d3af2ef8e55f", "code_archive": "repos/qPwSoOQPrI.zip", "code_archive_sha256": "94a361b1e6228d54a5cd8ab835163e8d23c1643555b4ec2059c71194e95c7a98", "code_archive_bytes": 303543, "code_file_count": 111, "code_extensions": {".py": 109, ".cpp": 1, ".cu": 1}, "github_disk_usage_kb": 253, "github_languages": {"Python": 1033131, "C++": 4498, "Dockerfile": 3924, "Cuda": 3062, "Makefile": 1432}, "github_archived": false, "github_pushed_at": "2025-06-06T15:22:50Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/leave-it-to-the-specialist-repair-sparse-llms"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "sBpYRQOrMn", "year": 2025, "status": "rejected", "title": "New Paradigm of Adversarial Training: Breaking Inherent Trade-Off between Accuracy and Robustness via Dummy Classes", "authors": ["Yanyun Wang", "Li Liu", "Zi Liang", "Qingqing Ye", "Haibo Hu"], "authorids": ["~Yanyun_Wang1", "~Li_Liu8", "~Zi_Liang1", "~Qingqing_Ye1", "~Haibo_Hu2"], "authors_source": "OpenReview API", "abstract": "Adversarial Training (AT) is recognized as one of the most effective methods to enhance the robustness of Deep Neural Networks (DNNs). However, existing AT methods suffer from an inherent trade-off between adversarial robustness and clean accuracy, which seriously hinders their real-world deployment. Previous works have studied this trade-off within the current AT paradigm, exploring various factors such as perturbation intensity, label noise and class margin. Despite these efforts, current AT methods still typically experience a reduction in clean accuracy by over 10% to date, without significant improvements in robustness compared with simple baselines like PGD-AT. This inherent trade-off raises a question: whether the current AT paradigm, which assumes to learn the corresponding benign and adversarial samples as the same class, inappropriately combines clean and robust objectives that may be essentially inconsistent. In this work, we surprisingly reveal that up to 40% of CIFAR-10 adversarial samples always fail to satisfy such an assumption across various AT methods and robust models, explicitly indicating the improvement room for the current AT paradigm. Accordingly, to relax the tension between clean and robust learning derived from this overstrict assumption, we propose a new AT paradigm by introducing an additional dummy class for each original class, aiming to accommodate the hard adversarial samples with shifted distribution after perturbation. The robustness w.r.t. these adversarial samples can be achieved by runtime recovery from the predicted dummy classes to their corresponding original ones, eliminating the compromise with clean learning. Building on this new paradigm, we propose a novel plug-and-play AT technology named DUmmy Classes-based Adversarial Training (DUCAT). Extensive experiments on CIFAR-10, CIFAR-100, and Tiny-ImageNet demonstrate that the DUCAT concurrently improves clean accuracy and adversarial robustness compared with state-of-the-art benchmarks, effectively releasing the existing inherent trade-off. The code is available at https://anonymous.4open.science/r/DUCAT.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=sBpYRQOrMn", "arxiv_id": "2410.12671", "paper_pdf": "papers/sBpYRQOrMn.pdf", "paper_pdf_sha256": "bca40401ca5de6221f3286d08b21a6307fbbd5e19b5c0a79aa141369b7842eeb", "paper_pdf_bytes": 4853463, "paper_pdf_source": "openreview", "code_url": "https://github.com/FlaAI/DUCAT", "code_repository": "FlaAI/DUCAT", "code_commit": "6bb27b31cbecfa6ce04c7d559203b54732eca3c9", "code_archive": "repos/sBpYRQOrMn.zip", "code_archive_sha256": "579780961626ef8adbc81a367e253950d3e67fc692c9b6f19f114d3f5f5d9447", "code_archive_bytes": 34170, "code_file_count": 32, "code_extensions": {".py": 32}, "github_disk_usage_kb": 36, "github_languages": {"Python": 86427}, "github_archived": false, "github_pushed_at": "2024-10-03T13:57:55Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/new-paradigm-of-adversarial-training-breaking"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "7Scc7Nl7lg", "year": 2024, "status": "rejected", "title": "Revealing Vision-Language Integration in the Brain with Multimodal Networks", "authors": ["Vighnesh Subramaniam", "Colin Conwell", "Christopher Wang", "Gabriel Kreiman", "Boris Katz", "Ignacio Cases", "Andrei Barbu"], "authorids": ["~Vighnesh_Subramaniam1", "~Colin_Conwell1", "~Christopher_Wang1", "~Gabriel_Kreiman1", "~Boris_Katz1", "~Ignacio_Cases2", "~Andrei_Barbu3"], "authors_source": "OpenReview API", "abstract": "We use multimodal deep neural networks to identify sites of multimodal integration in the human brain. These are regions where a multimodal language-vision model is better at predicting neural recordings (stereoelectroencephalography, SEEG) than either a unimodal language, unimodal vision model, or a linearly-integrated language-vision model. We use a wide range of state-of-the-art models spanning different architectures including Transformers and CNNs (ALBEF, BLIP, Flava, ConvNeXt, BEIT, SIMCLR, CLIP, SLIP) with different multimodal integration approaches to model the SEEG signal while subjects watched movies. As a key enabling step, we first demonstrate that the approach has the resolution to distinguish trained from randomly-initialized models for both language and vision; the inability to do so would fundamentally hinder further analysis. We show that trained models systematically outperform randomly initialized models in their ability to predict the SEEG signal. We then compare unimodal and multimodal models against one another. A key contribution is standardizing the methodology for doing so while carefully avoiding statistical artifacts. Since models all have different architectures, number of parameters, and training sets which can obscure the results, we then carry out a test between two controlled models: SLIP-Combo and SLIP-SimCLR which keep all of these attributes the same aside from multimodal input. Using this method, we identify neural sites (on average 141 out of 1090 total sites or 12.94\\%) and brain regions where multimodal integration is occurring. We find numerous new sites of multimodal integration, many of which lie around the temporoparietal junction, long theorized to be a hub of multimodal integration.", "decision": "Reject", "meta_review": "", "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=7Scc7Nl7lg", "arxiv_id": "2406.14481", "paper_pdf": "papers/7Scc7Nl7lg.pdf", "paper_pdf_sha256": "d8188eeba8ed6256030fa6fb09958db5ecd0cc41e0636a257da8d41811e8e8b2", "paper_pdf_bytes": 30079476, "paper_pdf_source": "openreview", "code_url": "https://github.com/vsubramaniam851/brain-multimodal", "code_repository": "vsubramaniam851/brain-multimodal", "code_commit": "87175f724d908cb5e2db33d65ce7e94d777dc430", "code_archive": "repos/7Scc7Nl7lg.zip", "code_archive_sha256": "e13a58b235f8dd49dcc3411768251a6fca2e9fd9e56d048cf3d4568b20527d85", "code_archive_bytes": 27408, "code_file_count": 9, "code_extensions": {".py": 8, ".ipynb": 1}, "github_disk_usage_kb": 34, "github_languages": {"Python": 90948, "Jupyter Notebook": 13584}, "github_archived": false, "github_pushed_at": "2024-07-25T22:27:34Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/revealing-vision-language-integration-in-the"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "_d2f3hRn0hT", "year": 2023, "status": "rejected", "title": "Single-level Adversarial Data Synthesis based on Neural Tangent Kernels", "authors": ["Yu-Rong Zhang", "Reddy Su", "Sheng-Yen Chou", "Shan-Hung Wu"], "authorids": ["~Yu-Rong_Zhang1", "~Reddy_Su1", "~Sheng-Yen_Chou1", "~Shan-Hung_Wu1"], "authors_source": "OpenReview API", "abstract": "Generative adversarial networks (GANs) have achieved impressive performance in data synthesis and have driven the development of many applications. How- ever, GANs are known to be hard to train due to their bilevel objective, which leads to the problems of convergence, mode collapse, and gradient vanishing. In this paper, we propose a new generative model called the generative adversarial NTK (GA-NTK) that has a single-level objective. The GA-NTK keeps the spirit of adversarial learning (which helps generate plausible data) while avoiding the training difficulties of GANs. This is done by modeling the discriminator as a Gaussian process with a neural tangent kernel (NTK-GP) whose training dynam- ics can be completely described by a closed-form formula. We analyze the conver- gence behavior of GA-NTK trained by gradient descent and give some sufficient conditions for convergence. We also conduct extensive experiments to study the advantages and limitations of GA-NTK and propose some techniques that make GA-NTK more practical.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=_d2f3hRn0hT", "arxiv_id": "2204.04090", "paper_pdf": "papers/_d2f3hRn0hT.pdf", "paper_pdf_sha256": "689b9da743b479c97d98057268704bde84e85b5cbb82f40347c555a84b990951", "paper_pdf_bytes": 33549887, "paper_pdf_source": "openreview", "code_url": "https://github.com/ga-ntk/ga-ntk", "code_repository": "ga-ntk/ga-ntk", "code_commit": "8151ac9a39c05689f8070e1b57645361d49f14f9", "code_archive": "repos/_d2f3hRn0hT.zip", "code_archive_sha256": "dd702a7af463567e6b470be3b97d8d7e3782924cad60fcfbb1120def9269ae83", "code_archive_bytes": 44767, "code_file_count": 13, "code_extensions": {".py": 13}, "github_disk_usage_kb": 42, "github_languages": {"Python": 181458}, "github_archived": false, "github_pushed_at": "2022-09-29T10:42:26Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/generative-adversarial-method-based-on-neural"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "zLb9oSWy933", "year": 2022, "status": "rejected", "title": "Fast Finite Width Neural Tangent Kernel", "authors": ["Roman Novak", "Jascha Sohl-Dickstein", "Samuel Stern Schoenholz"], "authorids": ["~Roman_Novak2", "~Jascha_Sohl-Dickstein2", "~Samuel_Stern_Schoenholz1"], "authors_source": "OpenReview API", "abstract": "The Neural Tangent Kernel (NTK), defined as the outer product of the neural network (NN) Jacobians, $\\Theta_\\theta(x_1, x_2) = \\left[\\partial f(\\theta, x_1)\\big/\\partial \\theta\\right] \\left[\\partial f(\\theta, x_2)\\big/\\partial \\theta\\right]^T$, has emerged as a central object of study in deep learning. In the infinite width limit, the NTK can sometimes be computed analytically and is useful for understanding training and generalization of NN architectures. At finite widths, the NTK is also used to better initialize NNs, compare the conditioning across models, perform architecture search, and do meta-learning. Unfortunately, the finite-width NTK is notoriously expensive to compute, which severely limits its practical utility. \n\nWe perform the first in-depth analysis of the compute and memory requirements for NTK computation in finite width networks. \nLeveraging the structure of neural networks, we further propose two novel algorithms that change the exponent of the compute and memory requirements of the finite width NTK, dramatically improving efficiency.\n\nWe open-source (https://github.com/iclr2022anon/fast_finite_width_ntk) our two algorithms as general-purpose JAX function transformations that apply to any differentiable computation (convolutions, attention, recurrence, etc.) and introduce no new hyper-parameters.\n", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=zLb9oSWy933", "arxiv_id": "2206.08720", "paper_pdf": "papers/zLb9oSWy933.pdf", "paper_pdf_sha256": "59d5dff4295d4aebd90c0472af833f3302c6ac4e2c5acf20e5e6575b480b1cf7", "paper_pdf_bytes": 1028775, "paper_pdf_source": "openreview", "code_url": "https://github.com/iclr2022anon/fast_finite_width_ntk", "code_repository": "iclr2022anon/fast_finite_width_ntk", "code_commit": "09a6c2e805d61f8acebf38c590020b79b491f733", "code_archive": "repos/zLb9oSWy933.zip", "code_archive_sha256": "67c517b5d642af8d4d737c08c83c9a23ea4f37ce30d2879a080ba974aca15573", "code_archive_bytes": 39704, "code_file_count": 9, "code_extensions": {".py": 8, ".ipynb": 1}, "github_disk_usage_kb": 37, "github_languages": {"Python": 104975, "Jupyter Notebook": 54565}, "github_archived": false, "github_pushed_at": "2022-06-20T18:37:38Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/fast-finite-width-neural-tangent-kernel-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "5L8XMh667qz", "year": 2021, "status": "rejected", "title": "Encoded Prior Sliced Wasserstein AutoEncoder for learning latent manifold representations", "authors": ["Sanjukta Krishnagopal", "Jacob Bedrossian"], "authorids": ["~Sanjukta_Krishnagopal1", "jacob@math.umd.edu"], "authors_source": "OpenReview API", "abstract": "While variational autoencoders have been successful in a variety of tasks, the use of conventional Gaussian or Gaussian mixture priors are limited in their ability to encode underlying structure of data in the latent representation.\nIn this work, we introduce an Encoded Prior Sliced Wasserstein AutoEncoder (EPSWAE) wherein an additional prior-encoder network facilitates learns an embedding of the data manifold which preserves topological and geometric properties of the data, thus improving the structure of latent space.\nThe autoencoder and prior-encoder networks are iteratively trained using the Sliced Wasserstein (SW) distance, which efficiently measures the distance between two \\textit{arbitrary} sampleable distributions without being constrained to a specific form as in the KL divergence, and without requiring expensive adversarial training.\nTo improve the representation, we use (1) a structural consistency term in the loss that encourages isometry between feature space and latent space and (2) a nonlinear variant of the SW distance which averages over random nonlinear shearing.\nThe effectiveness of the learned manifold encoding is best explored by traversing the latent space through interpolations along \\textit{geodesics} which generate samples that lie on the manifold and hence are advantageous compared to standard Euclidean interpolation.\nTo this end, we introduce a graph-based algorithm for interpolating along network-geodesics in latent space by maximizing the density of samples along the path while minimizing total energy. We use the 3D-spiral data to show that the prior does indeed encode the geometry underlying the data and to demonstrate the advantages of the network-algorithm for interpolation.\nAdditionally, we apply our framework to MNIST, and CelebA datasets, and show that outlier generations, latent representations, and geodesic interpolations are comparable to the state of the art.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=5L8XMh667qz", "arxiv_id": "2010.01037", "paper_pdf": "papers/5L8XMh667qz.pdf", "paper_pdf_sha256": "e5fe10b8d71575c08ccc3563ada6149268d7207f66386d947d8894d66c3bbfed", "paper_pdf_bytes": 12219545, "paper_pdf_source": "openreview", "code_url": "https://github.com/chimeraki/EPSWAE", "code_repository": "chimeraki/EPSWAE", "code_commit": "64c1d7c7d66a481cc449bffc28c05c5616744c62", "code_archive": "repos/5L8XMh667qz.zip", "code_archive_sha256": "0b9834d346718e1e8c53ecff1048efa16107686dfd5ba73b1a9183c435aaa85c", "code_archive_bytes": 32340, "code_file_count": 6, "code_extensions": {".py": 6}, "github_disk_usage_kb": 44, "github_languages": {"Python": 55281}, "github_archived": false, "github_pushed_at": "2020-10-05T12:01:21Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/encoded-prior-sliced-wasserstein-autoencoder"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "SJlJSaEFwS", "year": 2020, "status": "rejected", "title": "Robust Cross-lingual Embeddings from Parallel Sentences ", "authors": ["Ali Sabet", "Prakhar Gupta", "Jean-Baptiste Cordonnier", "Robert West", "Martin Jaggi"], "authorids": ["asabet@uwaterloo.ca", "prakhar.gupta@epfl.ch", "jean-baptiste.cordonnier@epfl.ch", "robert.west@epfl.ch", "martin.jaggi@epfl.ch"], "authors_source": "OpenReview API", "abstract": "Recent advances in cross-lingual word embeddings have primarily relied on mapping-based methods, which project pretrained word embeddings from different languages into a shared space through a linear transformation. However, these approaches assume word embedding spaces are isomorphic between different languages, which has been shown not to  hold in practice (Søgaard et al., 2018), and fundamentally limits their performance. This motivates investigating joint learning methods which can overcome this impediment, by simultaneously learning embeddings across languages via a cross-lingual term in the training objective. Given the abundance of parallel data available (Tiedemann, 2012), we propose a bilingual extension of the CBOW method which leverages sentence-aligned corpora to obtain robust cross-lingual word and sentence representations. Our approach significantly improves cross-lingual sentence retrieval performance over all other approaches, as well as convincingly outscores mapping methods while maintaining parity with jointly trained methods on word-translation. It also achieves parity with a deep RNN method on a zero-shot cross-lingual document classification task, requiring far fewer computational resources for training and inference. As an additional advantage, our bilingual method also improves the quality of monolingual word vectors despite training on much smaller datasets.  We make our code and models publicly available.\n", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=SJlJSaEFwS", "arxiv_id": "1912.12481", "paper_pdf": "papers/SJlJSaEFwS.pdf", "paper_pdf_sha256": "31b33466673e3ebfb027e3dbfc334e1d8ee56e3611209bdf7cc5e04ac2d2e9ac", "paper_pdf_bytes": 300711, "paper_pdf_source": "openreview", "code_url": "https://github.com/epfml/Bi-Sent2Vec", "code_repository": "epfml/Bi-Sent2Vec", "code_commit": "b3bf758f86a7006112a7a65bf9955d88ef3465d7", "code_archive": "repos/SJlJSaEFwS.zip", "code_archive_sha256": "62849656fb4d22a4428fcfb2e67fcf3349ece2b3ff3321f33f8581c21832eb0f", "code_archive_bytes": 40961, "code_file_count": 24, "code_extensions": {".h": 12, ".cc": 11, ".py": 1}, "github_disk_usage_kb": 52, "github_languages": {"C++": 119853, "Python": 5657, "Makefile": 1647}, "github_archived": false, "github_pushed_at": "2020-06-27T17:28:41Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/robust-cross-lingual-embeddings-from-parallel-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ByN7Yo05YX", "year": 2019, "status": "rejected", "title": "Adaptive Neural Trees", "authors": ["Ryutaro Tanno", "Kai Arulkumaran", "Daniel C. Alexander", "Antonio Criminisi", "Aditya Nori"], "authorids": ["ryutaro.tanno.15@ucl.ac.uk", "kailash.arulkumaran13@imperial.ac.uk", "d.alexander@ucl.ac.uk", "antcrim@microsoft.com", "adityan@microsoft.com"], "authors_source": "OpenReview API", "abstract": "Deep neural networks and decision trees operate on largely separate paradigms; typically, the former performs representation learning with pre-specified architectures, while the latter is characterised by learning hierarchies over pre-specified features with data-driven architectures. We unite the two via adaptive neural trees (ANTs), a model that incorporates representation learning into edges, routing functions and leaf nodes of a decision tree, along with a backpropagation-based training algorithm that adaptively grows the architecture from primitive modules (e.g., convolutional layers). ANTs allow increased interpretability via hierarchical clustering, e.g., learning meaningful class associations, such as separating natural vs. man-made objects. We demonstrate this on classification and regression tasks, achieving over 99% and 90% accuracy on the MNIST and CIFAR-10 datasets, and outperforming standard neural networks, random forests and gradient boosted trees on the SARCOS dataset. Furthermore, ANT optimisation naturally adapts the architecture to the size and complexity of the training data.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=ByN7Yo05YX", "arxiv_id": "1807.06699", "paper_pdf": "papers/ByN7Yo05YX.pdf", "paper_pdf_sha256": "9be575e73fe2dda1a1b86a32502d469dbaab3109e56e95337d028c47271b3ce6", "paper_pdf_bytes": 1975257, "paper_pdf_source": "openreview", "code_url": "https://github.com/rtanno21609/AdaptiveNeuralTrees", "code_repository": "rtanno21609/AdaptiveNeuralTrees", "code_commit": "8d7835fd21b16e43801ef20ad530380057f53a0c", "code_archive": "repos/ByN7Yo05YX.zip", "code_archive_sha256": "ef03d66c454a8cbd7a46a4a93f7c57e70a4c4faaf6d876e38f96ec7802c8a422", "code_archive_bytes": 482040, "code_file_count": 11, "code_extensions": {".py": 6, ".ipynb": 4, ".sh": 1}, "github_disk_usage_kb": 279, "github_languages": {"Jupyter Notebook": 550686, "Python": 137273, "Shell": 777}, "github_archived": false, "github_pushed_at": "2019-07-03T10:52:28Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/adaptive-neural-trees"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "SJQO7UJCW", "year": 2018, "status": "rejected", "title": "Adversarial Learning for Semi-Supervised Semantic Segmentation", "authors": ["Wei-Chih Hung", "Yi-Hsuan Tsai", "Yan-Ting Liou", "Yen-Yu Lin", "Ming-Hsuan Yang"], "authorids": ["whung8@ucmerced.edu", "ytsai@nec-labs.com", "lyt@csie.ntu.edu.tw", "yylin@citi.sinica.edu.tw", "mhyang@ucmerced.edu"], "authors_source": "OpenReview API", "abstract": "We propose a method for semi-supervised semantic segmentation using the adversarial network. While most existing discriminators are trained to classify input images as real or fake on the image level, we design a discriminator in a fully convolutional manner to differentiate the predicted probability maps from the ground truth segmentation distribution with the consideration of the spatial resolution. We show that the proposed discriminator can be used to improve the performance on semantic segmentation by coupling the adversarial loss with the standard cross entropy loss on the segmentation network. In addition, the fully convolutional discriminator enables the semi-supervised learning through discovering the trustworthy regions in prediction results of unlabeled images, providing additional supervisory signals. In contrast to existing methods that utilize weakly-labeled images, our method leverages unlabeled images without any annotation to enhance the segmentation model. Experimental results on both the PASCAL VOC 2012 dataset and the Cityscapes dataset demonstrate the effectiveness of our algorithm.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=SJQO7UJCW", "arxiv_id": "1802.07934", "paper_pdf": "papers/SJQO7UJCW.pdf", "paper_pdf_sha256": "504cc9935023e0374bf77bd040099a15014757a8b7938c56cf6728ace804c08c", "paper_pdf_bytes": 4570902, "paper_pdf_source": "openreview", "code_url": "https://github.com/hfslyc/AdvSemiSeg", "code_repository": "hfslyc/AdvSemiSeg", "code_commit": "841d546c927605747594be726622363bf781cefb", "code_archive": "repos/SJQO7UJCW.zip", "code_archive_sha256": "e72f40de01137b360bf1f1ec7e7601d5aee87d6ab4d1abcbb595aabfa66532e8", "code_archive_bytes": 849235, "code_file_count": 7, "code_extensions": {".py": 7}, "github_disk_usage_kb": 834, "github_languages": {"Python": 49841}, "github_archived": false, "github_pushed_at": "2021-04-21T10:48:11Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/adversarial-learning-for-semi-supervised"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "eChWBrh9mc", "year": 2026, "status": "rejected", "title": "EntryPrune: Neural Network Feature Selection using First Impressions", "authors": ["Felix Zimmer", "Patrik Okanovic", "Torsten Hoefler"], "authorids": ["~Felix_Zimmer1", "~Patrik_Okanovic1", "~Torsten_Hoefler1"], "authors_source": "OpenReview API", "abstract": "There is an ongoing effort to develop feature selection algorithms to improve interpretability, reduce computational resources, and minimize overfitting in predictive models. Neural networks stand out as architectures on which to build feature selection methods, and recently, neuron pruning and regrowth have emerged from the sparse neural network literature as promising new tools. We introduce EntryPrune, a novel supervised feature selection algorithm using a dense neural network with a dynamic sparse input layer. It employs entry-based pruning, a novel approach that compares neurons based on their relative change induced when they have entered the network. Extensive experiments on 13 different datasets show that our approach generally outperforms the current state-of-the-art methods, and in particular improves the average accuracy on low-dimensional datasets. Furthermore, we show that EntryPruning surpasses traditional techniques such as magnitude pruning within the EntryPrune framework and that EntryPrune achieves lower runtime than competing approaches. Our code is available in the supplementary material.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=eChWBrh9mc", "arxiv_id": "2410.02344", "paper_pdf": "papers/eChWBrh9mc.pdf", "paper_pdf_sha256": "ef4c85abfb5c75fca4a06f79c550257b0836932a9c5e08f37154c63cb10f2c29", "paper_pdf_bytes": 4484287, "paper_pdf_source": "openreview", "code_url": "https://github.com/flxzimmer/entryprune", "code_repository": "flxzimmer/entryprune", "code_commit": "81cd448d5a1037902aaa86fc649d38cdf5f780eb", "code_archive": "repos/eChWBrh9mc.zip", "code_archive_sha256": "c8f482782c3b32c9342851c7f8a082e5703a9058e856bf69fd8493ad657d6a03", "code_archive_bytes": 166864, "code_file_count": 8, "code_extensions": {".py": 8}, "github_disk_usage_kb": 261, "github_languages": {"Python": 35650}, "github_archived": false, "github_pushed_at": "2025-05-20T14:28:28Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/relchanet-neural-network-feature-selection"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "r6aX67YhD9", "year": 2025, "status": "rejected", "title": "Learning to Watermark LLM-generated Text via Reinforcement Learning", "authors": ["Xiaojun Xu", "Yuanshun Yao", "Yang Liu"], "authorids": ["~Xiaojun_Xu1", "~Yuanshun_Yao2", "~Yang_Liu3"], "authors_source": "OpenReview API", "abstract": "We study how to watermark LLM outputs, i.e. embedding algorithmically detectable signals into LLM-generated text to track misuse. Unlike the current mainstream methods that work with a fixed LLM, we expand the watermark design space by including the LLM tuning stage in the watermark pipeline. While prior works focus on token-level watermark that embeds signals into the output, we design a model-level watermark that embeds signals into the LLM weights, and such signals can be detected by a paired detector. We propose a co-training framework based on reinforcement learning that iteratively (1) trains a detector to detect the generated watermarked text and (2) tunes the LLM to generate text easily detectable by the detector while keeping its normal utility. We empirically show that our watermarks are more accurate, robust, and adaptable (to new attacks) with no generation overhead. It also allows watermarked model open-sourcing. In addition, if used together with alignment, the extra overhead introduced is low -- only training an extra reward model (i.e. our detector). We hope our work can bring more effort into studying a broader watermark design that is not limited to working with a fixed LLM.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=r6aX67YhD9", "arxiv_id": "2403.10553", "paper_pdf": "papers/r6aX67YhD9.pdf", "paper_pdf_sha256": "f3c93c28a2c15aade4c9c9939bafa6ec3526949f4a78c21634b7f111a4306555", "paper_pdf_bytes": 476124, "paper_pdf_source": "openreview", "code_url": "https://github.com/xiaojunxu/learning-to-watermark-llm", "code_repository": "xiaojunxu/learning-to-watermark-llm", "code_commit": "09e6bf65a9c1ef867ee5356bbf67000caec651f6", "code_archive": "repos/r6aX67YhD9.zip", "code_archive_sha256": "48e090ca7194c97ad7cf49bd5cfbdb1f6a40a14999465d0e51cbf3c789c29641", "code_archive_bytes": 42777, "code_file_count": 13, "code_extensions": {".py": 13}, "github_disk_usage_kb": 37, "github_languages": {"Python": 151486}, "github_archived": false, "github_pushed_at": "2024-03-19T17:02:33Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/learning-to-watermark-llm-generated-text-via"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "C6zFUEvgiU", "year": 2024, "status": "rejected", "title": "Feedback-guided Data Synthesis for Imbalanced Classification", "authors": ["Reyhane Askari Hemmat", "Mohammad Pezeshki", "Florian Bordes", "Michal Drozdzal", "Adriana Romero-Soriano"], "authorids": ["~Reyhane_Askari_Hemmat1", "~Mohammad_Pezeshki1", "~Florian_Bordes1", "~Michal_Drozdzal1", "~Adriana_Romero-Soriano1"], "authors_source": "OpenReview API", "abstract": "Current status quo in machine learning is to use static datasets of real images for training, which often come from long-tailed distributions. With the recent advances in generative models, researchers have started augmenting these static datasets with synthetic data, reporting moderate performance improvements on classification tasks. We hypothesize that these performance gains are limited by the lack of feedback from the classifier to the generative model, which would promote the usefulness of the generated samples to improve the classifier’s performance. In this work, we introduce a framework for augmenting static datasets with useful synthetic samples, which leverages one-shot feedback from the classifier to drive the sampling of the generative model. In order for the framework to be effective, we find that the samples must be close to the support of the real data of the task at hand, and be sufficiently diverse. We validate three feedback criteria on a long-tailed dataset (ImageNet-LT) as well as a group-imbalanced dataset (NICO++). On ImageNet-LT, we achieve state-of-the-art results, with over 4% improvement on underrepresented classes while being twice efficient in terms of the number of generated synthetic samples. NICO++ also enjoys marked boosts of over 5% in worst group accuracy. With these results, our framework paves the path towards effectively leveraging state-of-the-art text-to-image models as data sources that can be queried to improve downstream applications.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=C6zFUEvgiU", "arxiv_id": "2310.00158", "paper_pdf": "papers/C6zFUEvgiU.pdf", "paper_pdf_sha256": "4f620a92d5c286c2f9ee06b5cc2931782f68ef71b082d09972843b41ada10ee3", "paper_pdf_bytes": 15629518, "paper_pdf_source": "openreview", "code_url": "https://github.com/facebookresearch/Feedback-guided-Data-Synthesis", "code_repository": "facebookresearch/Feedback-guided-Data-Synthesis", "code_commit": "60dd027b7f5a4c133fcf6b43e98bc5dbdd5cacb6", "code_archive": "repos/C6zFUEvgiU.zip", "code_archive_sha256": "31126bad0da3a98c3d5b16ab4b163e33127d7f587a65653b1c6a6f97f8b2ac59", "code_archive_bytes": 38679, "code_file_count": 4, "code_extensions": {".py": 4}, "github_disk_usage_kb": 35, "github_languages": {"Python": 17277}, "github_archived": true, "github_pushed_at": "2024-09-09T21:22:42Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/feedback-guided-data-synthesis-for-imbalanced"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "K2spEiswXVf", "year": 2023, "status": "rejected", "title": "MALIBO: Meta-Learning for Likelihood-free Bayesian Optimization", "authors": ["Jiarong Pan", "Stefan Falkner", "Felix Berkenkamp", "Joaquin Vanschoren"], "authorids": ["~Jiarong_Pan2", "~Stefan_Falkner1", "~Felix_Berkenkamp1", "~Joaquin_Vanschoren1"], "authors_source": "OpenReview API", "abstract": "Bayesian Optimization (BO) is a popular method to optimize expensive black-box functions. Typically, BO only uses observations from the current task. Recently proposed methods try to warm-start BO by exploiting knowledge from related tasks, yet suffer from scalability issues and sensitivity to heterogeneous scale across multiple datasets. We propose a novel approach to solve these problems by combining a meta-learning technique and a likelihood-free acquisition function. The meta-learning model simultaneously learns the underlying (task-agnostic) data distribution and a latent feature representation for individual tasks. The likelihood-free BO technique has less stringent assumptions about the problems and works with any classification algorithm, making it computation efficient and robust to different scales across tasks. Finally, gradient boosting is used as a residual model on top to adapt to distribution drifts between new and prior tasks, which might otherwise weaken the usefulness of the meta-learned features. Experiments show that the meta-model learns an effective prior for warm-starting optimization algorithms, while being cheap to evaluate and invariant to changes of scale across different datasets.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=K2spEiswXVf", "arxiv_id": "2307.03565", "paper_pdf": "papers/K2spEiswXVf.pdf", "paper_pdf_sha256": "d3e672a633b22012cb3e2d6f94a7eed612c4e145530d1df7bd8b773843e90884", "paper_pdf_bytes": 7692730, "paper_pdf_source": "openreview", "code_url": "https://github.com/boschresearch/meta-learning-likelihood-free-bayesian-optimization", "code_repository": "boschresearch/meta-learning-likelihood-free-bayesian-optimization", "code_commit": "87f1b3d2ed59441f8197b38ffdf68116bb90c2d8", "code_archive": "repos/K2spEiswXVf.zip", "code_archive_sha256": "bc582087f97d95af0ef8f1a4e0577f898871e73b28fb1b32a6d99bafbeae54bc", "code_archive_bytes": 45516, "code_file_count": 14, "code_extensions": {".py": 14}, "github_disk_usage_kb": 43, "github_languages": {"Python": 77655}, "github_archived": true, "github_pushed_at": "2024-09-11T14:24:19Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/malibo-meta-learning-for-likelihood-free"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ZAA0Ol4z2i4", "year": 2022, "status": "rejected", "title": "Explaining Off-Policy Actor-Critic From A Bias-Variance Perspective", "authors": ["Ting-Han Fan", "Peter Ramadge"], "authorids": ["~Ting-Han_Fan1", "~Peter_Ramadge1"], "authors_source": "OpenReview API", "abstract": "Off-policy Actor-Critic algorithms have demonstrated phenomenal experimental performance but still require better explanations. To this end, we show its policy evaluation error on the distribution of transitions decomposes into: a Bellman error, a bias from policy mismatch, and a variance term from sampling. By comparing the magnitude of bias and variance, we explain the success of the Emphasizing Recent Experience sampling and 1/age weighted sampling. Both sampling strategies yield smaller bias and variance and are hence preferable to uniform sampling.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=ZAA0Ol4z2i4", "arxiv_id": "2110.02421", "paper_pdf": "papers/ZAA0Ol4z2i4.pdf", "paper_pdf_sha256": "ddcf988c2390ec48386e434e20ab08079f8319b5496670d2c91e7eac2e743c8e", "paper_pdf_bytes": 676711, "paper_pdf_source": "openreview", "code_url": "https://github.com/sunfex/weighted-sac", "code_repository": "sunfex/weighted-sac", "code_commit": "2cc75a420f09b5a0a17bd25021e70db1425985a8", "code_archive": "repos/ZAA0Ol4z2i4.zip", "code_archive_sha256": "803a2a08daf1e1886bea4b7c1870c89cf06bc2824b3024b49b22655e588ece2d", "code_archive_bytes": 35145, "code_file_count": 7, "code_extensions": {".py": 6, ".sh": 1}, "github_disk_usage_kb": 39, "github_languages": {"Python": 28818, "Shell": 54}, "github_archived": false, "github_pushed_at": "2021-10-02T22:05:03Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/explaining-off-policy-actor-critic-from-a"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "heFdS9_tkzc", "year": 2021, "status": "rejected", "title": "Distantly Supervised Relation Extraction in Federated Settings", "authors": ["Dianbo Sui", "Yubo Chen", "Kang Liu", "Jun Zhao"], "authorids": ["~Dianbo_Sui1", "~Yubo_Chen1", "~Kang_Liu1", "~Jun_Zhao4"], "authors_source": "OpenReview API", "abstract": "Distant supervision is widely used in relation extraction in order to create a large-scale training dataset by aligning a knowledge base with unstructured text. Most existing studies in this field have assumed there is a great deal of centralized unstructured text. However, in practice, text may be distributed on different platforms and cannot be centralized due to privacy restrictions. Therefore, it is worthwhile to investigate distant supervision in the federated learning paradigm, which decouples the training of the model from the need for direct access to the raw text. However, overcoming label noise of distant supervision becomes more difficult in federated settings, because the sentences containing the same entity pair scatter around different platforms. In this paper, we propose a federated denoising framework to suppress label noise in federated settings. The core of this framework is a multiple instance learning based denoising method that is able to select reliable sentences via cross-platform collaboration. Various experimental results on New York Times dataset and miRNA gene regulation relation dataset demonstrate the effectiveness of the proposed method.", "decision": "Reject", "meta_review": null, "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=heFdS9_tkzc", "arxiv_id": "2008.05049", "paper_pdf": "papers/heFdS9_tkzc.pdf", "paper_pdf_sha256": "003b498d2b85ce84b06bbc885441fe95b9dc6c0a7c5853543a6bdb944a5cbbab", "paper_pdf_bytes": 704607, "paper_pdf_source": "openreview", "code_url": "https://github.com/DianboWork/FedDS", "code_repository": "DianboWork/FedDS", "code_commit": "fd5eec9ce2386d83da910854ea5dd7934f39e138", "code_archive": "repos/heFdS9_tkzc.zip", "code_archive_sha256": "d8ef2308082edba1608e6dc87be9f2bb851dd494e65521be26948892783c10b8", "code_archive_bytes": 44395, "code_file_count": 24, "code_extensions": {".py": 24}, "github_disk_usage_kb": 46, "github_languages": {"Python": 148604}, "github_archived": false, "github_pushed_at": "2021-09-18T06:41:53Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/distantly-supervised-relation-extraction-in"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "SJl3CANKvB", "year": 2020, "status": "rejected", "title": "A SIMPLE AND EFFECTIVE FRAMEWORK FOR PAIRWISE DEEP METRIC LEARNING", "authors": ["Qi Qi", "Yan Yan", "Zixuan Wu", "Xiaoyu Wang", "Tianbao Yang"], "authorids": ["qi-qi@uiowa.edu", "yanyan.tju@gmail.com", "wuzu@bc.edu", "fanghuaxue@gmail.com", "tianbao-yang@uiowa.edu"], "authors_source": "OpenReview API", "abstract": "Deep metric learning (DML) has received much attention in deep learning due to its wide applications in computer vision. Previous studies have focused on designing complicated losses and hard example mining methods, which are mostly heuristic and lack of theoretical understanding. In this paper, we cast DML as a simple pairwise binary classification problem that classifies a pair of examples as similar or dissimilar. It identifies the most critical issue in this problem---imbalanced data pairs. To tackle this issue, we propose a simple and effective framework to sample pairs in a batch of data for updating the model. The key to this framework is to define a robust loss for all pairs over a mini-batch of data, which is formulated by distributionally robust optimization. The flexibility in constructing the  {\\it uncertainty decision set} of the dual variable allows us to recover state-of-the-art complicated losses and also to induce novel variants.  Empirical studies on several benchmark data sets demonstrate that our simple and effective method outperforms the state-of-the-art results.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=SJl3CANKvB", "arxiv_id": "1912.11194", "paper_pdf": "papers/SJl3CANKvB.pdf", "paper_pdf_sha256": "c0dc6b29592b7834ec9ef2e83d1433cbbc2fb67f873c1e06b879efaa372dcd27", "paper_pdf_bytes": 427281, "paper_pdf_source": "openreview", "code_url": "https://github.com/qiqi-helloworld/A-Simple-and-Effective-Framework-for-Pairewise-Distance-Metric-Learning", "code_repository": "qiqi-helloworld/A-Simple-and-Effective-Framework-for-Pairewise-Distance-Metric-Learning", "code_commit": "e12a417513a6cb4098bd955c0063c24a4c8e3543", "code_archive": "repos/SJl3CANKvB.zip", "code_archive_sha256": "a19f516da2cf3560984bf93098b182f69a61145003c459eb46fc3ced5fcbf60b", "code_archive_bytes": 81756, "code_file_count": 49, "code_extensions": {".py": 46, ".sh": 3}, "github_disk_usage_kb": 64, "github_languages": {"Python": 138404, "Shell": 5653}, "github_archived": false, "github_pushed_at": "2020-05-03T15:54:58Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/a-simple-and-effective-framework-for-pairwise-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "BklKFo09YX", "year": 2019, "status": "rejected", "title": "Mol-CycleGAN - a generative model for molecular optimization", "authors": ["Łukasz Maziarka", "Agnieszka Pocha", "Jan Kaczmarczyk", "Michał Warchoł"], "authorids": ["l.maziarka@gmail.com", "lamiane.chan@gmail.com", "jan.kaczmarczyk@ardigen.com", "michal.warchol@ardigen.com"], "authors_source": "OpenReview API", "abstract": "Designing a molecule with desired properties is one of the biggest challenges in drug development, as it requires optimization of chemical compound structures with respect to many complex properties. To augment the compound design process we introduce Mol-CycleGAN -- a CycleGAN-based model that generates optimized compounds with a chemical scaffold of interest. Namely, given a molecule our model generates a structurally similar one with an optimized value of the considered property. We evaluate the performance of the model on selected optimization objectives related to structural properties (presence of halogen groups, number of aromatic rings) and to a physicochemical property (penalized logP). In the task of optimization of penalized logP of drug-like molecules our model significantly outperforms previous results. ", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=BklKFo09YX", "arxiv_id": "1902.02119", "paper_pdf": "papers/BklKFo09YX.pdf", "paper_pdf_sha256": "1d54c705989d1df4520b8c8721288cce8f109206f13ba39a1af403c7db1e741b", "paper_pdf_bytes": 1392295, "paper_pdf_source": "openreview", "code_url": "https://github.com/ardigen/mol-cycle-gan", "code_repository": "ardigen/mol-cycle-gan", "code_commit": "feb8d7504d0078798ee70d6d5cda3f37b4f7a903", "code_archive": "repos/BklKFo09YX.zip", "code_archive_sha256": "8b90b216b33b7d950f4cb5a179c50b6e04d4bc41283f27e99c2837abc57d8256", "code_archive_bytes": 359906, "code_file_count": 17, "code_extensions": {".py": 10, ".sh": 5, ".ipynb": 2}, "github_disk_usage_kb": 346, "github_languages": {"Jupyter Notebook": 493617, "Python": 38177, "Shell": 1006}, "github_archived": false, "github_pushed_at": "2019-02-06T14:30:24Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/mol-cyclegan-a-generative-model-for-molecular"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "MWtm2f1tyG", "year": 2026, "status": "rejected", "title": "IMPACT: Industrial Machine Perception via Acoustic Cognitive Transformer", "authors": ["Changheon Han", "Yuseop Sim", "Hoin Jung", "Jiho Lee", "Hojun Lee", "Yun Seok Kang", "Sucheol Woo", "Garam Kim", "Hyung Wook Park", "Martin Byung-Guk Jun"], "authorids": ["~Changheon_Han2", "~Yuseop_Sim1", "~Hoin_Jung1", "~Jiho_Lee8", "~Hojun_Lee4", "~Yun_Seok_Kang1", "~Sucheol_Woo1", "~Garam_Kim3", "~Hyung_Wook_Park1", "~Martin_Byung-Guk_Jun1"], "authors_source": "OpenReview API", "abstract": "Industrial acoustic signals encode machine state, yet prevailing data-driven approaches are task-specific supervised pipelines that generalize poorly beyond their design conditions. Progress is further limited by the scarcity of large-scale datasets and pretrained models tailored to active shop floor audio. To address this, we introduce DINOS (Diverse INdustrial Operation Sounds), a dataset of 74,149 recordings totaling over 1,093 hours collected from active manufacturing lines across diverse processes and operating regimes. We also provide IMPACT(Industrial Machine Perception via Acoustic Cognitive Transformer), a reference model pretrained on DINOS to standardize evaluation. Our benchmark is structured in four machine-specific steps: (1) baseline discrimination, (2) moderate operational complexity, (3) scalability to unseen equipment, and (4) domain shift and sensor modality adaptation. Across tasks, models pretrained or fine-tuned on DINOS consistently outperform general-purpose audio models, demonstrating the value of domain-specific pretraining for industrial acoustic perception.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=MWtm2f1tyG", "arxiv_id": "2507.06481", "paper_pdf": "papers/MWtm2f1tyG.pdf", "paper_pdf_sha256": "5485256ffdedfbaa6a9e5ebc82992d8ffc1dc7cb199d064c6c4acb31a67aab5b", "paper_pdf_bytes": 1451593, "paper_pdf_source": "openreview", "code_url": "https://github.com/hanprd/DINOS", "code_repository": "hanprd/DINOS", "code_commit": "2cfbc7786551cd2a734763aeb821243560453721", "code_archive": "repos/MWtm2f1tyG.zip", "code_archive_sha256": "c7125ed0b06d0edd167f736edc5a96e7855b4191d9b659449228c6f647625f40", "code_archive_bytes": 326926, "code_file_count": 15, "code_extensions": {".py": 14, ".sh": 1}, "github_disk_usage_kb": 278, "github_languages": {"Python": 282806, "Shell": 446}, "github_archived": false, "github_pushed_at": "2026-05-17T22:35:18Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/impact-industrial-machine-perception-via"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ejVuTFFkl6", "year": 2025, "status": "rejected", "title": "EEG-ImageNet: An Electroencephalogram Dataset and Benchmarks with Image Visual Stimuli of Multi-Granularity Labels", "authors": ["Shuqi Zhu", "Ziyi Ye", "Qingyao Ai", "Yiqun LIU"], "authorids": ["~Shuqi_Zhu1", "~Ziyi_Ye2", "~Qingyao_Ai1", "~Yiqun_LIU1"], "authors_source": "OpenReview API", "abstract": "Exploring how brain activity translates into visual perception offers valuable insights into the biological visual system's representation of the world. Recent advancements have enabled effective image classification and high-quality reconstruction using brain signals obtained through Functional Magnetic Resonance Imaging (fMRI) or magnetoencephalography (MEG). However, the cost and bulkiness of these technologies hinder their practical application. In contrast, Electroencephalography (EEG) presents advantages such as ease of use, affordability, high temporal resolution, and non-invasive operation, yet it remains underutilized in related research due to a shortage of comprehensive datasets. To fill this gap, we introduce EEG-ImageNet, a novel EEG dataset featuring recordings from 16 participants exposed to 4000 images sourced from the ImageNet dataset. This dataset offers five times the number of EEG-image pairs compared to existing benchmarks. EEG-ImageNet includes image stimuli labeled with varying levels of granularity, comprising 40 images with coarse labels and 40 with fine labels. We establish benchmarks for both object classification and image reconstruction based on this dataset. Experiments with several commonly used models show that the best-performing models can achieve object classification with an accuracy around 60% and image reconstruction with two-way identification around 64%. These findings highlight the dataset's potential to enhance EEG-based visual brain-computer interfaces, deepen our understanding of visual perception in biological systems, and suggest promising applications for improving machine vision models.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=ejVuTFFkl6", "arxiv_id": "2406.07151", "paper_pdf": "papers/ejVuTFFkl6.pdf", "paper_pdf_sha256": "11fcebeeb116af283350768ee4210d2b0e9702089b32ff3a6b27c6c8c0f719ad", "paper_pdf_bytes": 9831475, "paper_pdf_source": "openreview", "code_url": "https://github.com/Promise-Z5Q2SQ/EEG-ImageNet-Dataset", "code_repository": "Promise-Z5Q2SQ/EEG-ImageNet-Dataset", "code_commit": "7c78157100d3737690d73f206f4385e72613290d", "code_archive": "repos/ejVuTFFkl6.zip", "code_archive_sha256": "ad670acc5e055a42b19287596ee17ded6c7c3cfa80a712bdbbb104a9cfbf8a26", "code_archive_bytes": 26518, "code_file_count": 16, "code_extensions": {".py": 13, ".sh": 3}, "github_disk_usage_kb": 44, "github_languages": {"Python": 40544, "Shell": 2053}, "github_archived": false, "github_pushed_at": "2025-05-15T13:26:29Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/eeg-imagenet-an-electroencephalogram-dataset"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "mBzsKsrXf9", "year": 2024, "status": "rejected", "title": "ArtWhisperer: A Dataset for Characterizing Human-AI Interactions in Artistic Creations", "authors": ["Kailas Vodrahalli", "James Zou"], "authorids": ["~Kailas_Vodrahalli1", "~James_Zou1"], "authors_source": "OpenReview API", "abstract": "As generative AI becomes more prevalent, it is important to study how human users interact with such models. In this work, we investigate how people use text-to-image models to generate desired target images. To study this interaction, we created ArtWhisperer, an online game where users are given a target image and are tasked with iteratively finding a prompt that creates a similar-looking image as the target. Through this game, we recorded over 50,000 human-AI interactions; each interaction corresponds to one text prompt created by a user and the corresponding generated image. The majority of these are repeated interactions where a user iterates to find the best prompt for their target image, making this a unique sequential dataset for studying human-AI collaborations. In an initial analysis of this dataset, we identify several characteristics of prompt interactions and user strategies. People submit diverse prompts and are able to discover a variety of text descriptions that generate similar images. Interestingly, prompt diversity does not decrease as users find better prompts. We further propose a new metric to quantify the steerability of AI using our dataset. We define steerability as the expected number of interactions required to adequately complete a task. We estimate this value by fitting a Markov chain for each target task and calculating the expected time to reach an adequate score in the Markov chain. We  quantify and compare AI steerability across different types of target images and two different models, finding that images of cities and natural world images are more steerable than artistic and fantasy images. These findings provide insights into human-AI interaction behavior, present a concrete method of assessing AI steerability, and demonstrate the general utility of the ArtWhisperer dataset.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=mBzsKsrXf9", "arxiv_id": "2306.08141", "paper_pdf": "papers/mBzsKsrXf9.pdf", "paper_pdf_sha256": "d7217c23549ff7f50da0f9f3d07bda846e233fbdd65dd4987491fb6c5fdcdf82", "paper_pdf_bytes": 3923407, "paper_pdf_source": "openreview", "code_url": "https://github.com/kailas-v/ArtWhisperer", "code_repository": "kailas-v/ArtWhisperer", "code_commit": "246b8361cd217c7b679e95b139dabdada7cdda16", "code_archive": "repos/mBzsKsrXf9.zip", "code_archive_sha256": "419847fd432eda559548f21cfbd0caee6ac9769d8091c1b3c077df5bd7dd98d8", "code_archive_bytes": 32075, "code_file_count": 11, "code_extensions": {".py": 11}, "github_disk_usage_kb": 35, "github_languages": {"Python": 105880}, "github_archived": false, "github_pushed_at": "2023-08-29T12:48:18Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/artwhisperer-a-dataset-for-characterizing"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "iMy1hOrqiVE", "year": 2023, "status": "rejected", "title": "Exclusive Supermask Subnetwork Training for Continual Learning", "authors": ["Prateek Yadav", "Mohit Bansal"], "authorids": ["~Prateek_Yadav1", "~Mohit_Bansal2"], "authors_source": "OpenReview API", "abstract": "Continual Learning (CL) methods mainly focus on avoiding catastrophic forgetting and learning representations that are transferable to new tasks. Recently, Wortsman et al. (2020) proposed a CL method, SupSup, which uses a randomly initialized, fixed base network (model) and finds a supermask for each new task that selectively keeps or removes each weight to produce a subnetwork. They prevent forgetting as the network weights are not being updated. Although there is no forgetting, the performance of supermask is sub-optimal because fixed weights restrict its representational power. Furthermore, there is no accumulation or transfer of knowledge inside the model when new tasks are learned. Hence, we propose ExSSNeT (Exclusive Supermask SubNEtwork Training), that performs exclusive and non-overlapping subnetwork weight training. This avoids conflicting updates to the shared weights by subsequent tasks to improve performance while still preventing forgetting. Furthermore, we propose a novel KNN-based Knowledge Transfer (KKT) module that dynamically initializes a new task's mask based on previous tasks for improving knowledge transfer. We demonstrate that ExSSNeT outperforms SupSup and other strong previous methods on both text classification and vision tasks while preventing forgetting. Moreover, ExSSNeT is particularly advantageous for sparse masks that activate 2-10% of the model parameters, resulting in an average improvement of 8.3% over SupSup. Additionally, ExSSNeT scales to a large number of tasks (100), and our KKT module helps to learn new tasks faster while improving the overall performance.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=iMy1hOrqiVE", "arxiv_id": "2210.10209", "paper_pdf": "papers/iMy1hOrqiVE.pdf", "paper_pdf_sha256": "121713a735c66b2116797bab35e8a34d49f8afa42aca8e5db3e458c973dcdafb", "paper_pdf_bytes": 596011, "paper_pdf_source": "openreview", "code_url": "https://github.com/prateeky2806/exessnet", "code_repository": "prateeky2806/exessnet", "code_commit": "1e768f855eeb708b7a5e9189acc725f89bd7c98e", "code_archive": "repos/iMy1hOrqiVE.zip", "code_archive_sha256": "f12391621a870c289f94b8b7a81fe8e92834b76fdf2aea54edc51f19a3b0a0c9", "code_archive_bytes": 53725, "code_file_count": 29, "code_extensions": {".py": 29}, "github_disk_usage_kb": 44, "github_languages": {"Python": 171218}, "github_archived": false, "github_pushed_at": "2022-10-20T14:59:31Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/exclusive-supermask-subnetwork-training-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "v-f7ifhKYps", "year": 2022, "status": "rejected", "title": "Maximum Entropy Population Based Training for Zero-Shot Human-AI Coordination", "authors": ["Rui Zhao", "Jinming Song", "Hu Haifeng", "Yang Gao", "Yi Wu", "Zhongqian Sun", "Yang Wei"], "authorids": ["~Rui_Zhao1", "joshuasong@tencent.com", "~Hu_Haifeng1", "~Yang_Gao1", "~Yi_Wu1", "sallensun@tencent.com", "~Yang_Wei2"], "authors_source": "OpenReview API", "abstract": "An AI agent should be able to coordinate with humans to solve tasks. We consider the problem of training a Reinforcement Learning (RL) agent without using any human data, i.e., in a zero-shot setting, to make it capable of collaborating with humans. Standard RL agents learn through self-play. Unfortunately, these agents only know how to collaborate with themselves and normally do not perform well with unseen partners, such as humans. The methodology of how to train a robust agent in a zero-shot fashion is still subject to research. Motivated from the maximum entropy RL, we derive a centralized population entropy objective to facilitate learning of a diverse population of agents, which is later used to train a robust AI agent to collaborate with unseen partners. The proposed method shows its effectiveness compared to baseline methods, including self-play PPO, the standard Population-Based Training (PBT), and trajectory diversity-based PBT, in the popular Overcooked game environment. We also conduct online experiments with real humans and further demonstrate the efficacy of the method in the real world.", "decision": "Reject", "meta_review": null, "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=v-f7ifhKYps", "arxiv_id": "2112.11701", "paper_pdf": "papers/v-f7ifhKYps.pdf", "paper_pdf_sha256": "adb015a81ba67976c35b71e206708da9acba6d57cd04b0211f160f851f363e6c", "paper_pdf_bytes": 2582698, "paper_pdf_source": "openreview", "code_url": "https://github.com/ruizhaogit/maximum_entropy_population_based_training", "code_repository": "ruizhaogit/maximum_entropy_population_based_training", "code_commit": "518ea6a2f9ccc75c1c252bafaf946de62dc96e05", "code_archive": "repos/v-f7ifhKYps.zip", "code_archive_sha256": "73887f6251f025da74b83bcc1b0de66c6f8ba823085a030b5fbd07f16f7a411b", "code_archive_bytes": 48551, "code_file_count": 10, "code_extensions": {".py": 10}, "github_disk_usage_kb": 42, "github_languages": {"Python": 144349}, "github_archived": false, "github_pushed_at": "2022-11-29T12:00:07Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/maximum-entropy-population-based-training-for-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "L5b6jUonKFB", "year": 2021, "status": "rejected", "title": "Deep Continuous Networks", "authors": ["Nergis Tomen", "Silvia Laura Pintea", "Jan van Gemert"], "authorids": ["~Nergis_Tomen1", "~Silvia_Laura_Pintea1", "~Jan_van_Gemert1"], "authors_source": "OpenReview API", "abstract": "CNNs and computational models of biological vision share some fundamental principles, which, combined with recent developments in deep learning, have opened up new avenues of research in neuroscience. However, in contrast to biological models, conventional CNN architectures are based on spatio-temporally discrete representations, and thus cannot accommodate certain aspects of biological complexity such as continuously varying receptive field sizes and temporal dynamics of neuronal responses. Here we propose deep continuous networks (DCNs), which combine spatially continuous convolutional filter representations, with the continuous time framework of neural ODEs. This allows us to learn the spatial support of the filters during training, as well as model the temporal evolution of feature maps, linking DCNs closely to biological models. We show that DCNs are versatile. Experimentally, we demonstrate their applicability to a standard classification problem, where they allow for parameter reductions and meta-parametrization. We illustrate the biological plausibility of the scale distributions learned by DCNs and explore their performance in a pattern completion task, which is inspired by models from computational neuroscience. Finally, we suggest that the continuous representations learned by DCNs may enable computationally efficient implementations.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=L5b6jUonKFB", "arxiv_id": "2402.01557", "paper_pdf": "papers/L5b6jUonKFB.pdf", "paper_pdf_sha256": "835884eaa3a8bbdef28749a8e3328707813d164ce29c262e8118fc165338cf18", "paper_pdf_bytes": 2603410, "paper_pdf_source": "openreview", "code_url": "https://github.com/ntomen/Deep-Continuous-Networks", "code_repository": "ntomen/Deep-Continuous-Networks", "code_commit": "81cd9b651179aa70a495feccf883526cc28f78bb", "code_archive": "repos/L5b6jUonKFB.zip", "code_archive_sha256": "3023d0264a9830b1d8ad20a64d74ba9b5d28fc1c5eb002fce44c3f57439d6db2", "code_archive_bytes": 123038, "code_file_count": 42, "code_extensions": {".py": 41, ".sh": 1}, "github_disk_usage_kb": 48, "github_languages": {"Python": 364470, "Shell": 135}, "github_archived": false, "github_pushed_at": "2021-09-27T16:51:19Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/deep-continuous-networks"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "rylkma4twr", "year": 2020, "status": "rejected", "title": "Min-Max Optimization without Gradients: Convergence and Applications to Adversarial ML", "authors": ["Sijia Liu", "Songtao Lu", "Xiangyi Chen", "Yao Feng", "Kaidi Xu", "Abdullah Al-Dujaili", "Minyi Hong", "Una-May Obelilly"], "authorids": ["sijia.liu@ibm.com", "songtao@ibm.com", "chen5719@umn.edu", "feng-y16@mails.tsinghua.edu.cn", "xu.kaid@husky.neu.edu", "aldujail@mit.edu", "mhong@umn.edu", "unamay@csail.mit.edu"], "authors_source": "OpenReview API", "abstract": "In this paper, we study the problem of constrained robust (min-max) optimization ina black-box setting, where the desired optimizer cannot access the gradients of the objective function but may query its values. We present a principled optimization framework, integrating a zeroth-order (ZO) gradient estimator with an alternating projected stochastic gradient descent-ascent method, where the former only requires a small number of function queries and the later needs just one-step descent/ascent update. We show that the proposed framework, referred to as ZO-Min-Max, has a sub-linear convergence rate under mild conditions and scales gracefully with problem size. From an application side, we explore a promising connection between black-box min-max optimization and black-box evasion and poisoning attacks in adversarial machine learning (ML). Our empirical evaluations on these use cases demonstrate the effectiveness of our approach and its scalability to dimensions that prohibit using recent black-box solvers.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=rylkma4twr", "arxiv_id": "1909.13806", "paper_pdf": "papers/rylkma4twr.pdf", "paper_pdf_sha256": "3a6ac0973684ffe30c17b4d89023ea6ab561b2eaa07386771e203ee6a2e8dd27", "paper_pdf_bytes": 838837, "paper_pdf_source": "openreview", "code_url": "https://github.com/KaidiXu/ZO-minmax", "code_repository": "KaidiXu/ZO-minmax", "code_commit": "17d1b1183a5e9d07d70875efa33e6f0134f85958", "code_archive": "repos/rylkma4twr.zip", "code_archive_sha256": "6a8ee763ebbfefedaacf239cd28db3d6ab92319c1034407525da4d5a5350c70e", "code_archive_bytes": 64966, "code_file_count": 13, "code_extensions": {".py": 13}, "github_disk_usage_kb": 69, "github_languages": {"Python": 322618}, "github_archived": false, "github_pushed_at": "2020-06-28T19:49:36Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/min-max-optimization-without-gradients"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "rkxt8oC9FQ", "year": 2019, "status": "rejected", "title": "Perfect Match: A Simple Method for Learning Representations For Counterfactual Inference With Neural Networks", "authors": ["Patrick Schwab", "Lorenz Linhardt", "Walter Karlen"], "authorids": ["patrick.schwab@hest.ethz.ch", "llorenz@student.ethz.ch", "walter.karlen@hest.ethz.ch"], "authors_source": "OpenReview API", "abstract": "Learning representations for counterfactual inference from observational data is of high practical relevance for many domains, such as healthcare, public policy and economics. Counterfactual inference enables one to answer \"What if...?\" questions, such as \"What would be the outcome if we gave this patient treatment $t_1$?\". However, current methods for training neural networks for counterfactual inference on observational data are either overly complex, limited to settings with only two available treatment options, or both. Here, we present Perfect Match (PM), a method for training neural networks for counterfactual inference that is easy to implement, compatible with any architecture, does not add computational complexity or hyperparameters, and extends to any number of treatments. PM is based on the idea of augmenting samples within a minibatch with their propensity-matched nearest neighbours. Our experiments demonstrate that PM outperforms a number of more complex state-of-the-art methods in inferring counterfactual outcomes across several real-world and semi-synthetic datasets.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=rkxt8oC9FQ", "arxiv_id": "1810.00656", "paper_pdf": "papers/rkxt8oC9FQ.pdf", "paper_pdf_sha256": "2765311a94ecf1808988ffb56d2dab6c973efdca46bfeea6225cae0114350475", "paper_pdf_bytes": 1009524, "paper_pdf_source": "openreview", "code_url": "https://github.com/d909b/perfect_match", "code_repository": "d909b/perfect_match", "code_commit": "37673e0d8b4029e734a9bf2a4a164e06a8105b84", "code_archive": "repos/rkxt8oC9FQ.zip", "code_archive_sha256": "7ebadd0d62c1e35a1f8909209756c4edd81c3175920e874fdb676dd0d45c4c23", "code_archive_bytes": 398094, "code_file_count": 51, "code_extensions": {".py": 45, ".r": 5, ".sh": 1}, "github_disk_usage_kb": 369, "github_languages": {"Python": 324313, "R": 18317, "Shell": 3828}, "github_archived": false, "github_pushed_at": "2023-03-24T23:48:55Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/perfect-match-a-simple-method-for-learning"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "BkVf1AeAZ", "year": 2018, "status": "rejected", "title": "Label Embedding Network: Learning Label Representation for Soft Training of Deep Networks", "authors": ["Xu Sun", "Bingzhen Wei", "Xuancheng Ren", "Shuming Ma"], "authorids": ["xusun@pku.edu.cn", "weibz@pku.edu.cn", "renxc@pku.edu.cn", "shumingma@pku.edu.cn"], "authors_source": "OpenReview API", "abstract": "We propose a method, called Label Embedding Network, which can learn label representation (label embedding) during the training process of deep networks. With the proposed method, the label embedding is adaptively and automatically learned through back propagation. The original one-hot represented loss function is converted into a new loss function with soft distributions, such that the originally unrelated labels have continuous interactions with each other during the training process. As a result, the trained model can achieve substantially higher accuracy and with faster convergence speed. Experimental results based on competitive tasks demonstrate the effectiveness of the proposed method, and the learned label embedding is reasonable and interpretable. The proposed method achieves comparable or even better results than the state-of-the-art systems.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=BkVf1AeAZ", "arxiv_id": "1710.10393", "paper_pdf": "papers/BkVf1AeAZ.pdf", "paper_pdf_sha256": "18e98059f62a50a1e7f0c41b7c5adf5a8b3bca2b3448368b0aa23807b5ef4627", "paper_pdf_bytes": 499393, "paper_pdf_source": "openreview", "code_url": "https://github.com/lancopku/label-embedding-network", "code_repository": "lancopku/label-embedding-network", "code_commit": "ecd9d9b0610ab50b5f0c2630078be72bd4e4591a", "code_archive": "repos/BkVf1AeAZ.zip", "code_archive_sha256": "a8378e4a5539d50605168fde403054f3d22ccb3511b52855cfd62a28da1ac0e6", "code_archive_bytes": 1552845, "code_file_count": 25, "code_extensions": {".py": 22, ".pl": 3}, "github_disk_usage_kb": 1352, "github_languages": {"Python": 209409, "Perl": 21387}, "github_archived": false, "github_pushed_at": "2018-01-29T07:15:51Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/label-embedding-network-learning-label"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "1CoJWNcpmm", "year": 2026, "status": "rejected", "title": "You Do Not Fully Utilize Transformer's Representation Capacity", "authors": ["Gleb Gerasimov", "Yaroslav Aksenov", "Nikita Balagansky", "Viacheslav Sinii", "Daniil Gavrilov"], "authorids": ["~Gleb_Gerasimov1", "~Yaroslav_Aksenov1", "~Nikita_Balagansky3", "~Viacheslav_Sinii1", "~Daniil_Gavrilov1"], "authors_source": "OpenReview API", "abstract": "In contrast to RNNs, which compress their history into a single hidden state, Transformers can attend to all past tokens directly. However, standard Transformers rely solely on the hidden state from the previous layer to represent the entire context. We show that this design choice induces representation collapse and degrades performance. To address this issue, we introduce Layer-Integrated Memory (LIMe), a lightweight extension that leverages existing key–value buffers and learns per-head, per-layer routing weights to integrate representations from all previous layers with negligible overhead. Through extensive experiments—including language modeling, synthetic reasoning benchmarks, and very deep architectures—LIMe consistently achieves faster convergence, lower perplexity per FLOP, and substantial accuracy improvements on synthetic tasks while preserving higher value–vector entropy and improved token separability. Finally, our analysis of the learned routing weights reveals systematic reuse of both local and long-distance features, demonstrating how LIMe mitigates collapse, unlocks richer representations without increasing hidden-state size, and points to promising directions for future research.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=1CoJWNcpmm", "arxiv_id": "2502.09245", "paper_pdf": "papers/1CoJWNcpmm.pdf", "paper_pdf_sha256": "35e926d45626558187a1a65f5b193e1642de1915247c336e34d974d284427231", "paper_pdf_bytes": 1462274, "paper_pdf_source": "openreview", "code_url": "https://github.com/corl-team/lime", "code_repository": "corl-team/lime", "code_commit": "424281209236fc4e461d3e32160df26245f6ba4c", "code_archive": "repos/1CoJWNcpmm.zip", "code_archive_sha256": "dae1aec6fe0c6e72bb38cd8edefb0c1e4b824375e3c14f01524a191707e20e40", "code_archive_bytes": 291665, "code_file_count": 14, "code_extensions": {".py": 14}, "github_disk_usage_kb": 280, "github_languages": {"Python": 76088}, "github_archived": false, "github_pushed_at": "2025-05-28T12:07:57Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/you-do-not-fully-utilize-transformer-s"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "VvHVLVUD6m", "year": 2025, "status": "rejected", "title": "On the Mode-Seeking Properties of Langevin Dynamics", "authors": ["Xiwei Cheng", "Kexin Fu", "Farzan Farnia"], "authorids": ["~Xiwei_Cheng2", "~Kexin_Fu2", "~Farzan_Farnia1"], "authors_source": "OpenReview API", "abstract": "The Langevin Dynamics framework, which aims to generate samples from the score function of a probability distribution, is widely used for analyzing and interpreting score-based generative modeling. While the convergence behavior of Langevin Dynamics under unimodal distributions has been extensively studied in the literature, in practice the data distribution could consist of multiple distinct modes. In this work, we investigate Langevin Dynamics in producing samples from multimodal distributions and theoretically study its mode-seeking properties. We prove that under a variety of sub-Gaussian mixtures, Langevin Dynamics is unlikely to find all mixture components within a sub-exponential number of steps in the data dimension. To reduce the mode-seeking tendencies of Langevin Dynamics, we propose Chained Langevin Dynamics, which divides the data vector into patches of constant size and generates every patch sequentially conditioned on the previous patches. We perform a theoretical analysis of Chained Langevin Dynamics by reducing it to sampling from a constant-dimensional distribution. We present the results of several numerical experiments on synthetic and real image datasets, supporting our theoretical results on the iteration complexities of sample generation from mixture distributions using the chained and vanilla Langevin Dynamics.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=VvHVLVUD6m", "arxiv_id": "2406.02017", "paper_pdf": "papers/VvHVLVUD6m.pdf", "paper_pdf_sha256": "2834ee1d036272eff2c65d236c9a4f1741b4ba4933f50300578dd38f196ab192", "paper_pdf_bytes": 25124044, "paper_pdf_source": "openreview", "code_url": "https://github.com/Xiwei-Cheng/Chained_LD", "code_repository": "Xiwei-Cheng/Chained_LD", "code_commit": "8f8da77696037112eb4f29c93674913dc5d45115", "code_archive": "repos/VvHVLVUD6m.zip", "code_archive_sha256": "9b8ed26fbbf94e0b1653456bbf2a1295a90606937ade403017c976d03630976e", "code_archive_bytes": 43217, "code_file_count": 17, "code_extensions": {".py": 17}, "github_disk_usage_kb": 44, "github_languages": {"Python": 180335}, "github_archived": false, "github_pushed_at": "2024-05-28T11:28:03Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/on-the-mode-seeking-properties-of-langevin"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "hkQOYyUChL", "year": 2024, "status": "rejected", "title": "Learning and Forgetting Unsafe Examples in Large Language Models", "authors": ["Jiachen ZHAO", "Zhun Deng", "David Madras", "James Zou", "Mengye Ren"], "authorids": ["~Jiachen_ZHAO1", "~Zhun_Deng1", "~David_Madras1", "~James_Zou1", "~Mengye_Ren1"], "authors_source": "OpenReview API", "abstract": "As the number of large language models (LLMs) available to the public grows, there is a pressing need to understand the safety implications associated with these models learning from third-party custom finetuning data.  We explore the behavior of LLMs finetuned on unsafe content, represented by datasets that contain biases, toxicity, and harmfulness, finding that while LLMs can readily learn this unsafe content, they also tend to forget it when subsequently finetuned on safer content.  Drawing inspiration from this forgetting behavior, we introduce the ``\\ff{}'' algorithm, which filters unsafe data based on how strong the model's forgetting signal is for that data. We find that the \\ff{} algorithm outperforms alternative strategies like replay and moral self-correction in curbing LLMs' ability to assimilate unsafe content during custom finetuning, e.g. 75\\% lower than not applying any safety measures and 62\\% lower than using self-correction in toxicity score.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=hkQOYyUChL", "arxiv_id": "2312.12736", "paper_pdf": "papers/hkQOYyUChL.pdf", "paper_pdf_sha256": "88253a52b7eef8b698331fd1dd51417a16b23d29daf9160ec901e46f683c7cc9", "paper_pdf_bytes": 588135, "paper_pdf_source": "openreview", "code_url": "https://github.com/andotalao24/learn-forget-unsafe-llm", "code_repository": "andotalao24/learn-forget-unsafe-llm", "code_commit": "9c2faf7d1a404455e5661c8610dfa550e1fad8dc", "code_archive": "repos/hkQOYyUChL.zip", "code_archive_sha256": "df4c1ccd44cf70b4194e40721129e3e89c8a15f52dd516c83bd8c5fb6bd61e39", "code_archive_bytes": 22964, "code_file_count": 12, "code_extensions": {".py": 9, ".sh": 3}, "github_disk_usage_kb": 38, "github_languages": {"Python": 60039, "Shell": 2298}, "github_archived": false, "github_pushed_at": "2024-07-15T03:32:56Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/learning-and-forgetting-unsafe-examples-in"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "wtr-9AKxCI5", "year": 2023, "status": "rejected", "title": "MobileViTv3: Mobile-Friendly Vision Transformer with Simple and Effective Fusion of Local, Global and Input Features", "authors": ["Shakti Nagnath Wadekar", "Abhishek Chaurasia"], "authorids": ["~Shakti_Nagnath_Wadekar1", "~Abhishek_Chaurasia2"], "authors_source": "OpenReview API", "abstract": "MobileViT (MobileViTv1) combines convolutional neural networks (CNNs) and vision transformers (ViTs) to create light-weight models for mobile vision tasks. Though the main MobileViTv1-block helps to achieve competitive state-of-the-art results, the fusion block inside MobileViTv1-block, creates scaling challenges and has a complex learning task. We propose changes to the fusion block that are simple and effective to create MobileViTv3-block, which addresses the scaling and simplifies the learning task. Our proposed MobileViTv3-block used to create MobileViTv3-XXS, XS and S models outperform MobileViTv1 on ImageNet-1k, ADE20K, COCO and PascalVOC2012 datasets. On ImageNet-1K, MobileViTv3-XXS and MobileViTv3-XS surpasses MobileViTv1-XXS and MobileViTv1-XS by 2% and 1.9% respectively. Recently published MobileViTv2 architecture removes fusion block and uses linear complexity transformers to perform better than MobileViTv1. We add our proposed fusion block to MobileViTv2 to create MobileViTv3-0.5, 0.75 and 1.0 models. MobileViTv3-0.5 and MobileViTv3-0.75 outperforms MobileViTv2-0.5 and MobileViTv2-0.75 by 2.1% and 1.0% respectively on ImageNet-1K dataset. For segmentation task, MobileViTv3-1.0 achieves 2.07% and 1.1% better mIOU compared to MobileViTv2-1.0 on ADE20K dataset and PascalVOC2012 dataset respectively.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=wtr-9AKxCI5", "arxiv_id": "2209.15159", "paper_pdf": "papers/wtr-9AKxCI5.pdf", "paper_pdf_sha256": "f0c407ee71df02bfed77b3fd7d63af69e1defcc35d36caf352d02dc460e280a5", "paper_pdf_bytes": 5990095, "paper_pdf_source": "openreview", "code_url": "https://github.com/micronDLA/MobileViTv3", "code_repository": "micronDLA/MobileViTv3", "code_commit": "d381beb017ae3c244686afaa48064f95865df7b9", "code_archive": "repos/wtr-9AKxCI5.zip", "code_archive_sha256": "5525ed62c8215cc988be5684e90865092eb1fb335c4aea0d56b4d771152bd8e6", "code_archive_bytes": 56595, "code_file_count": 12, "code_extensions": {".py": 12}, "github_disk_usage_kb": 45, "github_languages": {"Python": 100763}, "github_archived": true, "github_pushed_at": "2022-10-06T17:19:32Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/mobilevitv3-mobile-friendly-vision"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "SGOma2sAF7Q", "year": 2022, "status": "rejected", "title": "LCS: Learning Compressible Subspaces for Adaptive Network Compression at Inference Time", "authors": ["Maxwell Horton", "Elvis Nunez", "Anish Prabhu", "Anurag Ranjan", "Ali Farhadi", "Mohammad Rastegari"], "authorids": ["~Maxwell_Horton1", "~Elvis_Nunez1", "~Anish_Prabhu1", "~Anurag_Ranjan1", "~Ali_Farhadi4", "~Mohammad_Rastegari2"], "authors_source": "OpenReview API", "abstract": "When deploying deep learning models to a device, it is traditionally assumed that available computational resources (compute, memory, and power) remain static. However, real-world computing systems do not always provide stable resource guarantees. Computational resources need to be conserved when load from other processes is high or battery power is low. Inspired by recent works on neural network subspaces, we propose a method for training a \"compressible subspace\" of neural networks that contains a fine-grained spectrum of models that range from highly efficient to highly accurate. Our models require no retraining, thus our subspace of models can be deployed entirely on-device to allow adaptive network compression at inference time. We present results for achieving arbitrarily fine-grained accuracy-efficiency trade-offs at inference time for structured and unstructured sparsity. We achieve accuracies on-par with standard models when testing our uncompressed models, and maintain high accuracy for sparsity rates above 90% when testing our compressed models. We also demonstrate that our algorithm extends to quantization at variable bit widths, achieving accuracy on par with individually trained networks.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=SGOma2sAF7Q", "arxiv_id": "2110.04252", "paper_pdf": "papers/SGOma2sAF7Q.pdf", "paper_pdf_sha256": "b351cec7d73cd0bc6456716a829564bd79be90f442bcac316187b46c2cb4ccf1", "paper_pdf_bytes": 1591332, "paper_pdf_source": "openreview", "code_url": "https://github.com/apple/learning-compressible-subspaces", "code_repository": "apple/learning-compressible-subspaces", "code_commit": "308f5216c7b0c15a0ebc88c7cec3e7d28c82b6c1", "code_archive": "repos/SGOma2sAF7Q.zip", "code_archive_sha256": "f484dc27f7964049847fd1025144fb77cb1318fddff1009813b75bf0e37eaf02", "code_archive_bytes": 79518, "code_file_count": 31, "code_extensions": {".py": 29, ".sh": 2}, "github_disk_usage_kb": 59, "github_languages": {"Python": 168374, "Shell": 225}, "github_archived": false, "github_pushed_at": "2021-10-27T16:22:32Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/lcs-learning-compressible-subspaces-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "bK-rJMKrOsm", "year": 2021, "status": "rejected", "title": "Multi-Head Attention: Collaborate Instead of Concatenate", "authors": ["Jean-Baptiste Cordonnier", "Andreas Loukas", "Martin Jaggi"], "authorids": ["~Jean-Baptiste_Cordonnier2", "~Andreas_Loukas1", "~Martin_Jaggi1"], "authors_source": "OpenReview API", "abstract": "Attention layers are widely used in natural language processing (NLP) and are beginning to influence computer vision architectures. However, they suffer from over-parameterization. For instance, it was shown that the majority of attention heads could be pruned without impacting accuracy. This work aims to enhance current understanding on how multiple heads interact. Motivated by the observation that trained attention heads share common key/query projections, we propose a collaborative multi-head attention layer that enables heads to learn shared projections. Our scheme decreases the number of parameters in an attention layer and can be used as a drop-in replacement in any transformer architecture.For instance, by allowing heads to collaborate on a neural machine translation task, we can reduce the key dimension by 4× without any loss in performance. We also show that it is possible to re-parametrize a pre-trained multi-head attention layer into our collaborative attention layer. Even without retraining, collaborative multi-head attention manages to reduce the size of the key and query projections by half without sacrificing accuracy. Our code is public.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=bK-rJMKrOsm", "arxiv_id": "2006.16362", "paper_pdf": "papers/bK-rJMKrOsm.pdf", "paper_pdf_sha256": "2c090efed555025bf1ec70aaaf02df4a259d6752ce8c48947808eacf1d1dc5da", "paper_pdf_bytes": 1342898, "paper_pdf_source": "openreview", "code_url": "https://github.com/epfml/collaborative-attention", "code_repository": "epfml/collaborative-attention", "code_commit": "5b05408640e339970f1009e06cac678e2a1ab534", "code_archive": "repos/bK-rJMKrOsm.zip", "code_archive_sha256": "2e38005b5627bae9e850a705ea04bfdbe629d2b0a890af1ebb5c512581daf089", "code_archive_bytes": 23532, "code_file_count": 11, "code_extensions": {".py": 11}, "github_disk_usage_kb": 49, "github_languages": {"Python": 35139}, "github_archived": false, "github_pushed_at": "2023-06-12T21:27:03Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/multi-head-attention-collaborate-instead-of"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "H1xKBCEYDr", "year": 2020, "status": "rejected", "title": "Black-box Adversarial Attacks with Bayesian Optimization", "authors": ["Satya Narayan Shukla", "Anit Kumar Sahu", "Devin Willmott", "J. Zico Kolter"], "authorids": ["snshukla@cs.umass.edu", "anit.sahu@gmail.com", "devin.willmott@uky.edu", "zkolter@cs.cmu.edu"], "authors_source": "OpenReview API", "abstract": "We focus on the problem of black-box adversarial attacks, where the aim is to generate adversarial examples using information limited to loss function evaluations of input-output pairs. We use Bayesian optimization (BO) to specifically\ncater to scenarios involving low query budgets to develop query efficient adversarial attacks. We alleviate the issues surrounding BO in regards to optimizing high dimensional deep learning models by effective dimension upsampling techniques. Our proposed approach achieves performance comparable to the state of the art black-box adversarial attacks albeit with a much lower average query count. In particular, in low query budget regimes, our proposed method reduces the query count up to 80% with respect to the state of the art methods.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=H1xKBCEYDr", "arxiv_id": "1909.13857", "paper_pdf": "papers/H1xKBCEYDr.pdf", "paper_pdf_sha256": "a5807cf9e806a8e545d93e0bb80c00f7b6b99809ed294526c61975f6c573948e", "paper_pdf_bytes": 689993, "paper_pdf_source": "openreview", "code_url": "https://github.com/snu-mllab/parsimonious-blackbox-attack", "code_repository": "snu-mllab/parsimonious-blackbox-attack", "code_commit": "9061d845b161fc100933c38b74c98b8b9375b80a", "code_archive": "repos/H1xKBCEYDr.zip", "code_archive_sha256": "99e7f1364ac6977b8893b249a41250653f95a102ab46fd4936eab91407f3f9d1", "code_archive_bytes": 72077, "code_file_count": 13, "code_extensions": {".py": 13}, "github_disk_usage_kb": 69, "github_languages": {"Python": 94094}, "github_archived": false, "github_pushed_at": "2020-12-07T04:45:10Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/black-box-adversarial-attacks-with-bayesian"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "HJeKCi0qYX", "year": 2019, "status": "rejected", "title": "MILE: A Multi-Level Framework for Scalable Graph Embedding", "authors": ["Jiongqian Liang", "Saket Gurukar", "Srinivasan Parthasarathy"], "authorids": ["liang.albert@outlook.com", "gurukar.1@osu.edu", "srini@cse.ohio-state.edu"], "authors_source": "OpenReview API", "abstract": "Recently there has been a surge of interest in designing graph embedding methods. Few, if any, can scale to a large-sized graph with millions of nodes due to both computational complexity and memory requirements. In this paper, we relax this limitation by introducing the MultI-Level Embedding (MILE) framework – a generic methodology allowing contemporary graph embedding methods to scale to large graphs. MILE repeatedly coarsens the graph into smaller ones using a hybrid matching technique to maintain the backbone structure of the graph. It then applies existing embedding methods on the coarsest graph and refines the embeddings to the original graph through a novel graph convolution neural network that it learns. The proposed MILE framework is agnostic to the underlying graph embedding techniques and can be applied to many existing graph embedding methods without modifying them. We employ our framework on several popular graph embedding techniques and conduct embedding for real-world graphs. Experimental results on five large-scale datasets demonstrate that MILE significantly boosts the speed (order of magnitude) of graph embedding while also often generating embeddings of better quality for the task of node classification. MILE can comfortably scale to a graph with 9 million nodes and 40 million edges, on which existing methods run out of memory or take too long to compute on a modern workstation.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=HJeKCi0qYX", "arxiv_id": "1802.09612", "paper_pdf": "papers/HJeKCi0qYX.pdf", "paper_pdf_sha256": "a635b28a127ea382124e28578ab2e7375d43838be9614294e9b425f414a4fe1c", "paper_pdf_bytes": 960423, "paper_pdf_source": "openreview", "code_url": "https://github.com/jiongqian/MILE", "code_repository": "jiongqian/MILE", "code_commit": "47940a33ab9bbb452af88fbc863020bdbe924fa2", "code_archive": "repos/HJeKCi0qYX.zip", "code_archive_sha256": "1ec4d3bc1c09e396071700a7af620d152daf891c95ee7bb30d7c4efc034109f1", "code_archive_bytes": 350324, "code_file_count": 12, "code_extensions": {".py": 12}, "github_disk_usage_kb": 432, "github_languages": {"Python": 55560}, "github_archived": false, "github_pushed_at": "2020-07-19T19:41:53Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/mile-a-multi-level-framework-for-scalable"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "BJlrSmbAZ", "year": 2018, "status": "rejected", "title": "Bayesian Uncertainty Estimation for Batch Normalized Deep Networks", "authors": ["Mattias Teye", "Hossein Azizpour", "Kevin Smith"], "authorids": ["teye@kth.se", "azizpour@kth.se", "ksmith@kth.se"], "authors_source": "OpenReview API", "abstract": "Deep neural networks have led to a series of breakthroughs, dramatically improving the state-of-the-art in many domains. The techniques driving these advances, however, lack a formal method to account for model uncertainty. While the Bayesian approach to learning provides a solid theoretical framework to handle uncertainty, inference in Bayesian-inspired deep neural networks is difficult. In this paper, we provide a practical approach to Bayesian learning that relies on a regularization technique found in nearly every modern network, batch normalization. We show that training a deep network using batch normalization is equivalent to approximate inference in Bayesian models, and we demonstrate how this finding allows us to make useful estimates of the model uncertainty. Using our approach, it is possible to make meaningful uncertainty estimates using conventional architectures without modifying the network or the training procedure. Our approach is thoroughly validated in a series of empirical experiments on different tasks and using various measures, showing it to outperform baselines on a majority of datasets with strong statistical significance.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=BJlrSmbAZ", "arxiv_id": "1802.06455", "paper_pdf": "papers/BJlrSmbAZ.pdf", "paper_pdf_sha256": "343389233ec47193f90410316e2fd1ab0aac5e05fcd71f80949826d0f79f808c", "paper_pdf_bytes": 1928150, "paper_pdf_source": "openreview", "code_url": "https://github.com/icml-mcbn/mcbn", "code_repository": "icml-mcbn/mcbn", "code_commit": "7c3338272eff0096c27dd139278037ea57c90cf7", "code_archive": "repos/BJlrSmbAZ.zip", "code_archive_sha256": "0f965632f9598d34d5f9ed3365ca129bd562a379035b20fe9806e916ff07b81b", "code_archive_bytes": 2319851, "code_file_count": 34, "code_extensions": {".py": 22, ".ipynb": 7, ".sh": 5}, "github_disk_usage_kb": 2251, "github_languages": {"Python": 106881, "Jupyter Notebook": 64125, "Shell": 4139, "Makefile": 387}, "github_archived": false, "github_pushed_at": "2018-05-04T22:44:30Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/bayesian-uncertainty-estimation-for-batch"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "bkU1bQUSQD", "year": 2026, "status": "rejected", "title": "VeriReason: Reinforcement Learning with Testbench Feedback for Reasoning-Enhanced Verilog Generation", "authors": ["Yiting Wang", "Guoheng Sun", "Wanghao Ye", "Gang Qu", "Ang Li"], "authorids": ["~Yiting_Wang8", "~Guoheng_Sun1", "~Wanghao_Ye2", "~Gang_Qu2", "~Ang_Li6"], "authors_source": "OpenReview API", "abstract": "Automating Register Transfer Level (RTL) code generation using Large Language Models (LLMs) offers substantial promise for streamlining digital circuit design and reducing human effort. However, current LLM-based approaches for RTL code generation face significant challenges. Methods such as supervised fine-tuning (SFT), in-context learning, and chain-of-thought (CoT) struggle with several critical limitations in the RTL domain: the scarcity of high-quality training data, poor alignment between natural language specifications and generated code, lack of built-in verification mechanisms, and difficulty balancing between model generalization and domain specialization. Inspired by groundbreaking research such as DeepSeek-R1, which combines reinforcement learning with reasoning capabilities, we introduce VeriReason, a comprehensive framework that integrates supervised fine-tuning with Guided Reward Proximal Optimization (GRPO) reinforcement learning specifically tailored for RTL code generation. Using our curated high-quality training examples alongside a feedback-driven reward model, VeriReason combines testbench evaluations with structural heuristics to improve specification-code alignment and eliminate hallucinations. Iterative GRPO embeds intrinsic self-checking and reasoning capabilities, enabling the model to autonomously detect and correct functional errors. On the VerilogEval Benchmark, VeriReason delivers significant improvements: achieving 83.1% functional correctness on the VerilogEval Machine benchmark, substantially outperforming both comparable-sized models and much larger commercial systems like GPT-4 Turbo. Additionally, our approach demonstrates up to a 2.8× increase in first-attempt functional correctness compared to baseline methods and exhibits robust generalization to unseen designs. To our knowledge, VeriReason represents the first system to successfully integrate explicit reasoning capabilities with reinforcement learning for Verilog generation, establishing a new state-of-the-art for automated RTL synthesis. The code is available at: https://anonymous.4open.science/r/VeriReason-E625.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=bkU1bQUSQD", "arxiv_id": "2505.11849", "paper_pdf": "papers/bkU1bQUSQD.pdf", "paper_pdf_sha256": "48334b758dea98d967ffedc9eb2034e444c37d1bc3b1eb1c1f3845dcc353645a", "paper_pdf_bytes": 397020, "paper_pdf_source": "openreview", "code_url": "https://github.com/NellyW8/VeriReason", "code_repository": "NellyW8/VeriReason", "code_commit": "d215b7fe1b3db6dd4ca725f7d9399c49414c7531", "code_archive": "repos/bkU1bQUSQD.zip", "code_archive_sha256": "a8f766e82a9735827429b94b64d7783379615a8dba7a962322f6b0d25da5c49a", "code_archive_bytes": 199100, "code_file_count": 5, "code_extensions": {".py": 4, ".sh": 1}, "github_disk_usage_kb": 284, "github_languages": {"Python": 103440, "HTML": 71729, "Shell": 1229}, "github_archived": false, "github_pushed_at": "2025-09-25T11:45:37Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/verireason-reinforcement-learning-with-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "UW0zetsx8X", "year": 2025, "status": "rejected", "title": "Prompt Optimization with Human Feedback", "authors": ["Xiaoqiang Lin", "Zhongxiang Dai", "Arun Verma", "See-Kiong Ng", "Patrick Jaillet", "Bryan Kian Hsiang Low"], "authorids": ["~Xiaoqiang_Lin1", "~Zhongxiang_Dai1", "~Arun_Verma1", "~See-Kiong_Ng1", "~Patrick_Jaillet1", "~Bryan_Kian_Hsiang_Low1"], "authors_source": "OpenReview API", "abstract": "Large language models (LLMs) have demonstrated remarkable performances in various tasks. However, the performances of LLMs heavily depend on the input prompt. This has given rise to a number of recent works on prompt optimization. However, the previous works often require the availability of a numeric score to assess the quality of every prompt. Unfortunately, when a human user interacts with a black-box LLM, it is often infeasible and unreliable to attain such a score. Instead, it is usually significantly easier and more reliable to obtain preference feedback from a human user, i.e., showing the user the responses generated from a pair of prompts and asking the user which one is preferred. Therefore, in this paper, we study the problem of prompt optimization with human feedback (POHF), in which we aim to optimize the prompt for a black-box LLM using only human preference feedback. By drawing inspirations from dueling bandits, we design a theoretically principled strategy to select a pair of prompts to query for preference feedback in every iteration, and hence introduce our algorithm named automated POHF (APOHF). We apply our APOHF algorithm to a variety of tasks, including optimizing user instructions, prompt optimization for text-to-image generative models, and response optimization with human feedback (i.e., further refining the response using a variant of our APOHF). The results demonstrate that our APOHF can efficiently find a good prompt using a small number of preference feedback instances.", "decision": "Reject", "meta_review": "", "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=UW0zetsx8X", "arxiv_id": "2405.17346", "paper_pdf": "papers/UW0zetsx8X.pdf", "paper_pdf_sha256": "3b7d626c1285a47cc6c52cbfdfa107259ac1c4f614f8c0f42431a0083800638a", "paper_pdf_bytes": 1520400, "paper_pdf_source": "openreview", "code_url": "https://github.com/xqlin98/APOHF", "code_repository": "xqlin98/APOHF", "code_commit": "26eea14b1f89efc1b6d61bed8d154f44fa236b79", "code_archive": "repos/UW0zetsx8X.zip", "code_archive_sha256": "795ace631dad3634d807850a2374f1c9fc7b8ff945b7a89ba07440b29d7565f1", "code_archive_bytes": 50363, "code_file_count": 17, "code_extensions": {".py": 14, ".sh": 3}, "github_disk_usage_kb": 44, "github_languages": {"Python": 164262, "Shell": 1311}, "github_archived": false, "github_pushed_at": "2024-08-07T16:59:20Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/prompt-optimization-with-human-feedback"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "5BoXZXTJvL", "year": 2024, "status": "rejected", "title": "Beyond Size: How Gradients Shape Pruning Decisions in Large Language Models", "authors": ["Rocktim Jyoti Das", "Liqun Ma", "Zhiqiang Shen"], "authorids": ["~Rocktim_Jyoti_Das2", "~Liqun_Ma1", "~Zhiqiang_Shen1"], "authors_source": "OpenReview API", "abstract": "Large Language Models (LLMs) with a billion or more parameters are prime targets for network pruning, which aims to reduce a portion of the network weights without compromising performance. Prior approaches such as Weights Magnitude, SparseGPT, and Wanda, either concentrated solely on weights or integrated weights with activations for sparsity. However, they overlooked the informative gradients derived from pretrained large language models. In this paper, we present a novel sparsity-centric pruning method for pretrained LLMs, termed **G**radient-**b**ased **L**anguage **M**odel **P**runer (**GBLM-Pruner**). Distinctively, GBLM-Pruner operates in a training-free manner by harnessing normalized gradients, and substantially outperforms competitive counterparts like SparseGPT and Wanda in multiple benchmarks. Intriguing, after incorporating gradients, the unstructured pruning method tends to reveal some structural patterns post-pruning, which mirrors the geometric interdependence inherent in the LLMs' parameter structure. Additionally, GBLM-Pruner functions without any subsequent retraining or weight updates to maintain its simplicity as other counterparts. Extensive evaluations on LLaMA-1 and LLaMA-2 across various language benchmarks and perplexity show that GBLM-Pruner surpasses magnitude pruning, Wanda (*weights+activations*), and SparseGPT (*weights+activations+weight update*) by significant margins. Our code and models will be publicly available.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=5BoXZXTJvL", "arxiv_id": "2311.04902", "paper_pdf": "papers/5BoXZXTJvL.pdf", "paper_pdf_sha256": "105bf01bb58e356452a1975b942bb4fd4d510bfb0fc75908858f828bec26cb60", "paper_pdf_bytes": 2882889, "paper_pdf_source": "openreview", "code_url": "https://github.com/VILA-Lab/GBLM-Pruner", "code_repository": "VILA-Lab/GBLM-Pruner", "code_commit": "b59221570017d2e1dac45cb1c16f1c54ba4d8ea1", "code_archive": "repos/5BoXZXTJvL.zip", "code_archive_sha256": "1aee469d557f5cbd5f5bd74e35c8dbf4b9ff2e68fd0d13a086e15bd717bb9e24", "code_archive_bytes": 31903, "code_file_count": 12, "code_extensions": {".py": 7, ".sh": 5}, "github_disk_usage_kb": 38, "github_languages": {"Python": 45290, "Shell": 1001}, "github_archived": false, "github_pushed_at": "2025-08-23T11:51:17Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/beyond-size-how-gradients-shape-pruning"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "9rRhMKNOkeT", "year": 2023, "status": "rejected", "title": "Concept-based Explanations for Out-of-Distribution Detectors", "authors": ["Jihye Choi", "Jayaram Raghuram", "Ryan Feng", "Jiefeng Chen", "Somesh Jha", "Atul Prakash"], "authorids": ["~Jihye_Choi1", "~Jayaram_Raghuram1", "~Ryan_Feng1", "~Jiefeng_Chen2", "~Somesh_Jha1", "~Atul_Prakash1"], "authors_source": "OpenReview API", "abstract": "Out-of-distribution (OOD) detection plays a crucial role in ensuring the safe deployment of deep neural network (DNN) classifiers.\nWhile a myriad of methods have focused on improving the performance of OOD detectors, a critical gap remains in interpreting their decisions.\nWe help bridge this gap by providing explanations for OOD detectors based on learned high-level concepts.\nWe first propose two new metrics for assessing the effectiveness of a particular set of concepts for explaining OOD detectors: 1) $\\textit{detection completeness}$, which quantifies the sufficiency of concepts for explaining an OOD-detector's decisions, and 2) $\\textit{concept separability}$, which captures the distributional separation between in-distribution and OOD data in the concept space.\nBased on these metrics, we propose a framework for learning a set of concepts that satisfy the desired properties of detection completeness and concept separability, and demonstrate the framework's effectiveness in providing concept-based explanations for diverse OOD detection techniques.\nWe also show how to identify prominent concepts that contribute to the detection results via a modified Shapley value-based importance score.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=9rRhMKNOkeT", "arxiv_id": "2203.02586", "paper_pdf": "papers/9rRhMKNOkeT.pdf", "paper_pdf_sha256": "2f77f67b4b7fde208d63bee04a537bff7da4191ff63320b69188d052264d8881", "paper_pdf_bytes": 2695565, "paper_pdf_source": "openreview", "code_url": "https://github.com/jihyechoi77/concepts-for-ood", "code_repository": "jihyechoi77/concepts-for-ood", "code_commit": "28be657c6cb8fb01b3760db29b866eb6dfd4448d", "code_archive": "repos/9rRhMKNOkeT.zip", "code_archive_sha256": "d2ab9e0b33d10a9ea6d057ea6fec36bd8e71e04e08530a8e063b307751604f91", "code_archive_bytes": 48741, "code_file_count": 12, "code_extensions": {".py": 12}, "github_disk_usage_kb": 46, "github_languages": {"Python": 145394}, "github_archived": false, "github_pushed_at": "2023-06-08T23:48:45Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/concept-based-explanations-for-out-of"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "TNxKD3z_tPZ", "year": 2022, "status": "rejected", "title": "Persistent Homology Captures the Generalization of Neural Networks Without A Validation Set", "authors": ["Asier Gutiérrez-Fandiño", "David Pérez Fernández", "Jordi Armengol-Estapé", "Marta Villegas"], "authorids": ["~Asier_Gutiérrez-Fandiño1", "~David_Pérez_Fernández1", "~Jordi_Armengol-Estapé1", "~Marta_Villegas2"], "authors_source": "OpenReview API", "abstract": "The training of neural networks is usually monitored with a validation (holdout) set to estimate the generalization of the model. This is done instead of measuring intrinsic properties of the model to determine whether it is learning appropriately. In this work, we suggest studying the training of neural networks with Algebraic Topology, specifically Persistent Homology (PH). Using simplicial complex representations of neural networks, we study the PH diagram distance evolution on the neural network learning process with different architectures and several datasets. Results show that the PH diagram distance between consecutive neural network states correlates with the validation accuracy, implying that the generalization error of a neural network could be intrinsically estimated without any holdout set.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=TNxKD3z_tPZ", "arxiv_id": "2106.00012", "paper_pdf": "papers/TNxKD3z_tPZ.pdf", "paper_pdf_sha256": "6b0dd30151bd111d9d1e0521ea91732bf34d60f06b81274516ee66db104119e3", "paper_pdf_bytes": 866979, "paper_pdf_source": "openreview", "code_url": "https://github.com/asier-gutierrez/nn-evolution", "code_repository": "asier-gutierrez/nn-evolution", "code_commit": "311211821b0ebffd2458061e9636bd6dd68e7609", "code_archive": "repos/TNxKD3z_tPZ.zip", "code_archive_sha256": "f19c3432fe4dafa338be2973e2b322a5fddd35e9d763096c873b99544440ae5a", "code_archive_bytes": 62312, "code_file_count": 32, "code_extensions": {".py": 32}, "github_disk_usage_kb": 60, "github_languages": {"Python": 156507}, "github_archived": false, "github_pushed_at": "2021-07-17T08:21:26Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/persistent-homology-captures-the"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "_TGlfdZOHY3", "year": 2021, "status": "rejected", "title": "On Episodes, Prototypical Networks, and Few-Shot Learning", "authors": ["Steinar Laenen", "Luca Bertinetto"], "authorids": ["~Steinar_Laenen1", "~Luca_Bertinetto1"], "authors_source": "OpenReview API", "abstract": "Episodic learning is a popular practice among researchers and practitioners interested in few-shot learning. It consists of organising training in a series of learning problems, each relying on small “support” and “query” sets to mimic the few-shot circumstances encountered during evaluation.\nIn this paper, we investigate the usefulness of episodic learning in Prototypical Networks, one of the most popular algorithms making use of this practice.\nSurprisingly, in our experiments we found that, for Prototypical Networks, it is detrimental to use the episodic learning strategy of separating training samples between support and query set, as it is\na data-inefficient way to exploit training batches. This “non-episodic” version of Prototypical Networks, which corresponds to the classic Neighbourhood Component Analysis, reliably improves over its episodic counterpart in multiple datasets, achieving an accuracy that is competitive with the state-of-the-art, despite being extremely simple.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=_TGlfdZOHY3", "arxiv_id": "2012.09831", "paper_pdf": "papers/_TGlfdZOHY3.pdf", "paper_pdf_sha256": "98f819786ee8688b553d7c79f4ec4d1fb048460dbb4b4a2ee65e06d1ee844dae", "paper_pdf_bytes": 504019, "paper_pdf_source": "openreview", "code_url": "https://github.com/fiveai/on-episodes-fsl", "code_repository": "fiveai/on-episodes-fsl", "code_commit": "90f2fe1244d5add4a93eea0768b49472ce42e504", "code_archive": "repos/_TGlfdZOHY3.zip", "code_archive_sha256": "6d0f00124ef08cc6b65cd40c6a68a1c56b8c1dd7692fd82099dcc7426bbee4ed", "code_archive_bytes": 72290, "code_file_count": 68, "code_extensions": {".sh": 46, ".py": 22}, "github_disk_usage_kb": 52, "github_languages": {"Python": 86516, "Shell": 43569}, "github_archived": false, "github_pushed_at": "2021-11-30T09:03:18Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/on-episodes-prototypical-networks-and-few-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ryxmrpNtvH", "year": 2020, "status": "rejected", "title": "Deeper Insights into Weight Sharing in Neural Architecture Search", "authors": ["Yuge Zhang", "Quanlu Zhang", "Junyang Jiang", "Zejun Lin", "Yujing Wang"], "authorids": ["scottyugochang@gmail.com", "quanlu.zhang@microsoft.com", "jyjiang97@gmail.com", "gdzejlin@gmail.com", "yujing.wang@microsoft.com"], "authors_source": "OpenReview API", "abstract": "With the success of deep neural networks, Neural Architecture Search (NAS) as a way of automatic model design has attracted wide attention. As training every child model from scratch is very time-consuming, recent works leverage weight-sharing to speed up the model evaluation procedure. These approaches greatly reduce computation by maintaining a single copy of weights on the super-net and share the weights among every child model. However, weight-sharing has no theoretical guarantee and its impact has not been well studied before. In this paper, we conduct comprehensive experiments to reveal the impact of weight-sharing: (1) The best-performing models from different runs or even from consecutive epochs within the same run have significant variance; (2) Even with high variance, we can extract valuable information from training the super-net with shared weights; (3) The interference between child models is a main factor that induces high variance; (4) Properly reducing the degree of weight sharing could effectively reduce variance and improve performance.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=ryxmrpNtvH", "arxiv_id": "2001.01431", "paper_pdf": "papers/ryxmrpNtvH.pdf", "paper_pdf_sha256": "c4559211c46f944c66f11a60b3b8c97880c11f3107cce8d85c0253e7c16eb119", "paper_pdf_bytes": 828821, "paper_pdf_source": "openreview", "code_url": "https://github.com/ultmaster/deeper-insights-weight-sharing", "code_repository": "ultmaster/deeper-insights-weight-sharing", "code_commit": "2a7f852cd22190b0ccd2bad3e7a1c208f01986b5", "code_archive": "repos/ryxmrpNtvH.zip", "code_archive_sha256": "6e6d34f04b91ffcb3bea90576ea93040991e4a57234d5373c73236c9b95836bb", "code_archive_bytes": 97620, "code_file_count": 27, "code_extensions": {".py": 24, ".sh": 3}, "github_disk_usage_kb": 69, "github_languages": {"Python": 124967, "Batchfile": 8031, "Shell": 1174}, "github_archived": false, "github_pushed_at": "2022-12-08T06:17:40Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/deeper-insights-into-weight-sharing-in-neural-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "r1xYr3C5t7", "year": 2019, "status": "rejected", "title": "Neural Message Passing for Multi-Label Classification", "authors": ["Jack Lanchantin", "Arshdeep Sekhon", "Yanjun Qi"], "authorids": ["jjl5sw@virginia.edu", "as5cu@virginia.edu", "yq2h@virginia.edu"], "authors_source": "OpenReview API", "abstract": "Multi-label classification (MLC) is the task of assigning a set of target labels for a given sample. Modeling the combinatorial label interactions in MLC has been a long-haul challenge. Recurrent neural network (RNN) based encoder-decoder models have shown state-of-the-art performance for solving MLC. However, the sequential nature of modeling label dependencies through an RNN limits its ability in parallel computation, predicting dense labels, and providing interpretable results. In this paper, we propose Message Passing Encoder-Decoder (MPED) Networks,  aiming to provide fast, accurate, and interpretable MLC. MPED networks model the joint prediction of labels by replacing all RNNs in the encoder-decoder architecture with message passing mechanisms and dispense with autoregressive inference entirely.  The proposed models are simple, fast, accurate, interpretable, and structure-agnostic (can be used on known or unknown structured data). Experiments on seven real-world MLC datasets show the proposed models outperform autoregressive RNN models across five different metrics with a significant speedup during training and testing time.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=r1xYr3C5t7", "arxiv_id": "1904.08049", "paper_pdf": "papers/r1xYr3C5t7.pdf", "paper_pdf_sha256": "216d0696d2fe98a0440e0f74dce95c7cc3ebdc6c06dc6746d2ee27b22f2b8f53", "paper_pdf_bytes": 1381301, "paper_pdf_source": "openreview", "code_url": "https://github.com/QData/LaMP", "code_repository": "QData/LaMP", "code_commit": "c4152b9370c59dec825b98f857756d80679a5230", "code_archive": "repos/r1xYr3C5t7.zip", "code_archive_sha256": "2befdd3fb4b532a5fef28f506f536d02bba26e40ec471ec9831f734ca6600cda", "code_archive_bytes": 125706, "code_file_count": 20, "code_extensions": {".py": 20}, "github_disk_usage_kb": 452, "github_languages": {"Python": 118220}, "github_archived": false, "github_pushed_at": "2024-07-21T20:28:01Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/neural-message-passing-for-multi-label-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "H1U_af-0-", "year": 2018, "status": "rejected", "title": "Quadrature-based features for kernel approximation", "authors": ["Marina Munkhoeva", "Yermek Kapushev", "Evgeny Burnaev", "Ivan Oseledets"], "authorids": ["marina.munkhoeva@skolkovotech.ru", "kapushev@gmail.com", "e.burnaev@skoltech.ru", "i.oseledets@skoltech.ru"], "authors_source": "OpenReview API", "abstract": "We consider the problem of improving kernel approximation via feature maps. These maps arise as Monte Carlo approximation to integral representations of kernel functions and scale up kernel methods for larger datasets. We propose to use more efficient numerical integration technique to obtain better estimates of the integrals compared to the state-of-the-art methods. Our approach allows to use information about the integrand to enhance approximation and facilitates fast computations. We derive the convergence behavior and conduct an extensive empirical study that supports our hypothesis.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=H1U_af-0-", "arxiv_id": "1802.03832", "paper_pdf": "papers/H1U_af-0-.pdf", "paper_pdf_sha256": "823493ecc0fc69703dd2500ae9b70fc7d104465a2d623a27395091c0630ee5cc", "paper_pdf_bytes": 1686033, "paper_pdf_source": "openreview", "code_url": "https://github.com/quffka/quffka", "code_repository": "quffka/quffka", "code_commit": "1599e6ce580e7482a4525d4df5fc654b10a72e96", "code_archive": "repos/H1U_af-0-.zip", "code_archive_sha256": "927db02153ce7dbd1b266ef3f991413da803517357d494c7614e9e9e6f912ed4", "code_archive_bytes": 4323707, "code_file_count": 29, "code_extensions": {".py": 19, ".ipynb": 9, ".sh": 1}, "github_disk_usage_kb": 3783, "github_languages": {"Jupyter Notebook": 2551738, "Python": 63776, "Shell": 1198}, "github_archived": false, "github_pushed_at": "2018-04-17T23:06:30Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/quadrature-based-features-for-kernel"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "PbiQPxFLtQ", "year": 2026, "status": "rejected", "title": "How Explanations Leak the Decision Logic: Stealing Graph Neural Networks via Explanation Alignment", "authors": ["Bin Ma", "Yuyuan Feng", "Minhua Lin", "Enyan Dai"], "authorids": ["~Bin_Ma9", "~Yuyuan_Feng2", "~Minhua_Lin1", "~Enyan_Dai1"], "authors_source": "OpenReview API", "abstract": "Graph Neural Networks (GNNs) have become essential tools for analyzing graph-structured data in domains such as drug discovery and financial analysis, leading to growing demands for model transparency. Recent advances in explainable GNNs have addressed this need by revealing important subgraphs that influence predictions, but these explanation mechanisms may inadvertently expose models to security risks. This paper investigates how such explanations potentially leak critical decision logic that can be exploited for model stealing. We propose {\\method}, a novel stealing framework that integrates explanation alignment for capturing decision logic with guided data augmentation for efficient training under limited queries, enabling effective replication of both the predictive behavior and underlying reasoning patterns of target models. Experiments on molecular graph datasets demonstrate that our approach shows advantages over conventional methods in model stealing. This work highlights important security considerations for the deployment of explainable GNNs in sensitive domains and suggests the need for protective measures against explanation-based attacks. Our code is available at https://anonymous.4open.science/r/EGSteal-2BF7.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=PbiQPxFLtQ", "arxiv_id": "2506.03087", "paper_pdf": "papers/PbiQPxFLtQ.pdf", "paper_pdf_sha256": "dc3a2a0092de33ca0ea243fa4728f4dd91d4b852d61067c22f1afc8f13530290", "paper_pdf_bytes": 951582, "paper_pdf_source": "openreview", "code_url": "https://github.com/beanmah/EGSteal", "code_repository": "beanmah/EGSteal", "code_commit": "47a65a5dbec40d4156bcffac6eebb00322ae74a6", "code_archive": "repos/PbiQPxFLtQ.zip", "code_archive_sha256": "38ed2416e3cce64dd46681bd928ffad381b851cfad3af1f679f2e9a772a8caa1", "code_archive_bytes": 306624, "code_file_count": 15, "code_extensions": {".py": 15}, "github_disk_usage_kb": 294, "github_languages": {"Python": 217907}, "github_archived": false, "github_pushed_at": "2025-06-03T08:27:48Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/how-explanations-leak-the-decision-logic"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "NV5p50EkT6", "year": 2025, "status": "rejected", "title": "Channel-aware Contrastive Conditional Diffusion for Multivariate Probabilistic Time Series Forecasting", "authors": ["Siyang Li", "Yize Chen", "Hui Xiong"], "authorids": ["~Siyang_Li9", "~Yize_Chen1", "~Hui_Xiong1"], "authors_source": "OpenReview API", "abstract": "Forecasting faithful trajectories of multivariate time series from practical scopes is essential for reasonable decision-making. Recent methods majorly tailor generative conditional diffusion models to estimate the target temporal predictive distribution. However, it remains an obstacle to enhance the exploitation efficiency of given implicit temporal predictive information to bolster conditional diffusion learning. To this end, we propose a generic channel-aware contrastive conditional diffusion model termed CCDM to achieve desirable multivariate probabilistic forecasting, obviating the need for curated temporal conditioning inductive biases. In detail, we first design a channel-centric conditional denoising network to manage intra-variate variations and cross-variate correlations, which can lead to scalability on diverse prediction horizons and channel numbers. Then, we devise an ad-hoc denoising-based temporal contrastive learning to explicitly amplify the predictive mutual information between past observations and future forecasts. It can coherently complement naive step-wise denoising diffusion training and improve the forecasting accuracy and generality on unknown test time series. Besides, we offer theoretic insights on the benefits of such auxiliary contrastive training refinement from both neural mutual information and temporal distribution generalization aspects. The proposed CCDM can exhibit superior forecasting capability compared to current state-of-the-art diffusion forecasters over a comprehensive benchmark, with best MSE and CRPS outcomes on 79.17% and 87.5% cases.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=NV5p50EkT6", "arxiv_id": "2410.02168", "paper_pdf": "papers/NV5p50EkT6.pdf", "paper_pdf_sha256": "225900e3e1ba2eb27fb1f149a72c79ebd0db158d0bb8363850843bae74ccb9e9", "paper_pdf_bytes": 14416893, "paper_pdf_source": "openreview", "code_url": "https://github.com/LSY-Cython/CCDM", "code_repository": "LSY-Cython/CCDM", "code_commit": "a9a4998d397861b5a3f081ed7f05baeedc88cd2b", "code_archive": "repos/NV5p50EkT6.zip", "code_archive_sha256": "154d248927c9f002bea3a149b186135d4683dc0bddebac000013d4b4f43426a5", "code_archive_bytes": 23141, "code_file_count": 14, "code_extensions": {".py": 14}, "github_disk_usage_kb": 47, "github_languages": {"Python": 75397}, "github_archived": false, "github_pushed_at": "2025-10-09T10:47:48Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/channel-aware-contrastive-conditional"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "QzQSR56JZr", "year": 2024, "status": "rejected", "title": "Harnessing the Power of Large Language Models for Natural Language to First-Order Logic Translation", "authors": ["Yuan Yang", "Siheng Xiong", "Ali Payani", "Ehsan Shareghi", "Faramarz Fekri"], "authorids": ["~Yuan_Yang1", "~Siheng_Xiong1", "~Ali_Payani1", "~Ehsan_Shareghi1", "~Faramarz_Fekri1"], "authors_source": "OpenReview API", "abstract": "Translating natural language sentences to first-order logic (NL-FOL translation) remains a critical task in many logic-based NLP systems, as it enables ML models to reason logically over text.\nHowever, existing translation methods still struggle to scale to real-world tasks due to the lack of a large and high-quality dataset and a model family with high precision and coverage.\n\nIn this work, we approach this longstanding challenge by harnessing the power of pre-trained large language models (LLMs).\nTo do so, we present MALLS (large language **M**odel gener**A**ted N**L**-FO**L** pair**S**), a dataset of 28K diverse and verified sentence-level NL-FOL pairs collected from GPT-4.\nWe create MALLS by implementing an adaptive pipeline that prompts GPT-4 for pairs with rich and diverse contexts. \nTo ensure the validity of FOL rules and their alignment with the NL sentences, we utilized a combined strategy of FOL rule parsing, human annotation, and automatic filtering.\n\nWe also present LogicLLaMA, an LLaMA2-7B/13B model family fine-tuned on MALLS for NL-FOL translation. \nLogicLLaMA is capable of directly translating natural language into FOL rules, which outperforms GPT-3.5. \nLogicLLaMA is also equipped to correct FOL rules predicted by GPT-3.5, and can achieve similar performance as GPT-4 with a fraction of the cost.\nThis correction ability was achieved by a novel reinforcement learning with human feedback (RLHF) framework, which initially trains on synthetically perturbed NL-FOL pairs to encourage chain-of-thought reasoning and then fine-tunes with RLHF on GPT-3.5 outputs using a FOL verifier as the reward model. \nCodes and data are available [here](https://www.dropbox.com/sh/t0f69776773e9er/AABKaWuvepUvhSp-0u2w-b2Pa?dl=0).", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=QzQSR56JZr", "arxiv_id": "2305.15541", "paper_pdf": "papers/QzQSR56JZr.pdf", "paper_pdf_sha256": "c7fd4b01fbe5f423eef75311e3bd78fd126cb38dbb392e7757b85983cbce7ebf", "paper_pdf_bytes": 749189, "paper_pdf_source": "openreview", "code_url": "https://github.com/gblackout/LogicLLaMA", "code_repository": "gblackout/LogicLLaMA", "code_commit": "785a2c08e8fe964c8b2a10bb183ce5a08867aa3b", "code_archive": "repos/QzQSR56JZr.zip", "code_archive_sha256": "57d3792140e8239e651c6f1e7ec70eae49ff4bbb65d41b8aa7b7e976faba3555", "code_archive_bytes": 42110, "code_file_count": 16, "code_extensions": {".py": 9, ".sh": 6, ".ipynb": 1}, "github_disk_usage_kb": 40, "github_languages": {"Jupyter Notebook": 288177, "Python": 60800, "Shell": 2479}, "github_archived": false, "github_pushed_at": "2023-10-25T20:44:38Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/harnessing-the-power-of-large-language-models"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "CcXTudu9bvu", "year": 2023, "status": "rejected", "title": "DELTA: Diverse Client Sampling for Fasting Federated Learning", "authors": ["Lin Wang", "Yongxin Guo", "Tao Lin", "Xiaoying Tang"], "authorids": ["~Lin_Wang14", "~Yongxin_Guo1", "~Tao_Lin1", "~Xiaoying_Tang2"], "authors_source": "OpenReview API", "abstract": "Partial client participation has been widely adopted in Federated Learning (FL) to efficiently reduce the communication burden. However, an improper client sampling scheme will select unrepresentative subsets, which will cause a large variance in the model update and slows down the convergence. Existing sampling methods are either biased or can be further improved to accelerate the convergence. In this paper, we propose an unbiased sampling scheme, termed DELTA, to alleviate this problem. In particular, DELTA characterizes the impact of client diversity and local variance and samples the representative clients who carry valuable information for global model updates. Moreover, DELTA is a provably optimal unbiased sampling scheme that minimizes the variance caused by partial client participation and achieves better convergence than other unbiased sampling schemes. We corroborate our results with experiments on both synthetic and real data sets.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=CcXTudu9bvu", "arxiv_id": null, "paper_pdf": "papers/CcXTudu9bvu.pdf", "paper_pdf_sha256": "34866af809800423b45fa26f7caf56dd6d8c0a609047519e44b9cd4f1125e0f7", "paper_pdf_bytes": 2641510, "paper_pdf_source": "openreview", "code_url": "https://github.com/T-Lab-CUHKSZ/DELTA_FL", "code_repository": "T-Lab-CUHKSZ/DELTA_FL", "code_commit": "bc8e3a7aed6691ea6a954837656437721edf81e7", "code_archive": "repos/CcXTudu9bvu.zip", "code_archive_sha256": "c546d3d1e18990a27aaed08d18faa60cdc94d42187f2c802c8813801240b91c3", "code_archive_bytes": 111062, "code_file_count": 32, "code_extensions": {".py": 32}, "github_disk_usage_kb": 47, "github_languages": {"Python": 325235}, "github_archived": false, "github_pushed_at": "2024-04-02T05:14:35Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/delta-diverse-client-sampling-for-fasting"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "3GHHpYrYils", "year": 2022, "status": "rejected", "title": "On Anytime Learning at Macroscale", "authors": ["Lucas Caccia", "Jing Xu", "Myle Ott", "MarcAurelio Ranzato", "Ludovic Denoyer"], "authorids": ["~Lucas_Caccia1", "~Jing_Xu5", "~Myle_Ott1", "~MarcAurelio_Ranzato1", "~Ludovic_Denoyer1"], "authors_source": "OpenReview API", "abstract": " Classical machine learning frameworks assume access to a possibly large dataset in order to train a predictive model. In many practical applications however, data does not arrive all at once, but in batches over time. This creates a natural trade-off between accuracy of a model and time to obtain such a model. A greedy predictor could produce non-trivial predictions by immediately training on batches as soon as these become available but, it may also make sub-optimal use of future data. On the other hand, a tardy predictor could wait for a long time to aggregate several batches into a larger dataset, but ultimately deliver a much better performance.  In this work, we consider such a streaming learning setting,  which we dub {\\em anytime learning at macroscale} (ALMA). It is an instance of anytime learning applied not at the level of a single chunk of data, but at the level of the entire  sequence of large batches. We first formalize this learning setting, we then introduce metrics to assess how well learners perform on the given task for a given memory and compute budget, and finally we test about thirty baseline approaches on three standard benchmarks repurposed for anytime learning at macroscale. Our findings indicate that no model strikes the best trade-off across the board. While replay-based methods attain the lowest error rate, they also incur in a 5 to 10 times increase of compute. Approaches that grow capacity over time do offer better scaling in terms of training flops, but they also underperform simpler ensembling methods in terms of error rate. Overall, ALMA offers both a good abstraction of the typical learning setting faced everyday by practitioners, and a set of unsolved modeling problems for those interested in efficient learning of dynamic models.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=3GHHpYrYils", "arxiv_id": "2106.09563", "paper_pdf": "papers/3GHHpYrYils.pdf", "paper_pdf_sha256": "582db19e47086a146cc981abd12dcfafe2453b697d086d500789939581ed1e15", "paper_pdf_bytes": 454568, "paper_pdf_source": "openreview", "code_url": "https://github.com/facebookresearch/alma", "code_repository": "facebookresearch/alma", "code_commit": "02a6d0e76fedc6af315dd906698153edabbaea5a", "code_archive": "repos/3GHHpYrYils.zip", "code_archive_sha256": "d94e19c3f6d30199b213f3a04d1de10ede9f501117a7817f064cbb9114e96069", "code_archive_bytes": 84934, "code_file_count": 31, "code_extensions": {".py": 31}, "github_disk_usage_kb": 60, "github_languages": {"Python": 188093}, "github_archived": true, "github_pushed_at": "2021-11-24T20:57:21Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/on-anytime-learning-at-macroscale"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "-J9xYzP2HD", "year": 2021, "status": "rejected", "title": "Chameleon: Learning Model Initializations Across Tasks With Different Schemas", "authors": ["Lukas Brinkmeyer", "Rafael Rego Drumond", "Randolf Scholz", "Josif Grabocka", "Lars Schmidt-Thieme"], "authorids": ["~Lukas_Brinkmeyer1", "radrumond@ismll.uni-hildesheim.de", "scholz@ismll.uni-hildesheim.de", "~Josif_Grabocka1", "~Lars_Schmidt-Thieme1"], "authors_source": "OpenReview API", "abstract": "Parametric models, and particularly neural networks, require weight initialization as a starting point for gradient-based optimization. Recent work shows that an initial parameter set can be learned from a population of supervised learning tasks that enables a fast convergence for unseen tasks even when only a handful of instances is available (model-agnostic meta-learning). \nCurrently, methods for learning model initializations are limited to a population of tasks sharing the same schema, i.e., the same number, order, type, and semantics of predictor and target variables.\nIn this paper, we address the problem of meta-learning weight initialization across tasks with different schemas, for example, if the number of predictors varies across tasks, while they still share some variables. We propose Chameleon, a model that learns to align different predictor schemas to a common representation. \nIn experiments on 23 datasets of the OpenML-CC18 benchmark, we show that Chameleon can successfully learn parameter initializations across tasks with different schemas, presenting, to the best of our knowledge, the first cross-dataset few-shot classification approach for unstructured data.", "decision": "Reject", "meta_review": null, "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=-J9xYzP2HD", "arxiv_id": "1909.13576", "paper_pdf": "papers/-J9xYzP2HD.pdf", "paper_pdf_sha256": "5cca826662fa60afd6abbc0be4bba477d3399efff780dcb8380841e022d18272", "paper_pdf_bytes": 632014, "paper_pdf_source": "openreview", "code_url": "https://github.com/radrumond/Chameleon", "code_repository": "radrumond/Chameleon", "code_commit": "e7c2e86c81b7292a7ffb5bea0dc3b0a71f5bb1b8", "code_archive": "repos/-J9xYzP2HD.zip", "code_archive_sha256": "1fd83b1a661cb326c80a8153b3e956fd4a55b7b624624fa146e47fc0a66401a4", "code_archive_bytes": 44459, "code_file_count": 24, "code_extensions": {".py": 24}, "github_disk_usage_kb": 61, "github_languages": {"Python": 117074}, "github_archived": false, "github_pushed_at": "2020-03-20T15:48:06Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/chameleon-learning-model-initializations"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "BkgF4kSFPB", "year": 2020, "status": "rejected", "title": "Hallucinative Topological Memory for Zero-Shot Visual Planning", "authors": ["Kara Liu", "Thanard Kurutach", "Pieter Abbeel", "Aviv Tamar"], "authorids": ["karamarieliu@berkeley.edu", "thanard.kurutach@berkeley.edu", "pabbeel@cs.berkeley.edu", "aviv.tamar.mail@gmail.com"], "authors_source": "OpenReview API", "abstract": "In visual planning (VP), an agent learns to plan goal-directed behavior from observations of a dynamical system obtained offline, e.g., images obtained from self-supervised robot interaction. VP algorithms essentially combine data-driven perception and planning, and are important for robotic manipulation and navigation domains, among others. A recent and promising approach to VP is the semi-parametric topological memory (SPTM) method, where image samples are treated as nodes in a graph, and the connectivity in the graph is learned using deep image classification. Thus, the learned graph represents the topological connectivity of the data, and planning can be performed using conventional graph search methods. However, training SPTM necessitates a suitable loss function for the connectivity classifier, which requires non-trivial manual tuning. More importantly, SPTM is constricted in its ability to generalize to changes in the domain, as its graph is constructed from direct observations and thus requires collecting new samples for planning. In this paper, we propose Hallucinative Topological Memory (HTM), which overcomes these shortcomings. In HTM, instead of training a discriminative classifier we train an energy function using contrastive predictive coding. In addition, we learn a conditional VAE model that generates samples given a context image of the domain, and use these hallucinated samples for building the connectivity graph, allowing for zero-shot generalization to domain changes. In simulated domains, HTM outperforms conventional SPTM and visual foresight methods in terms of both plan quality and success in long-horizon planning. ", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=BkgF4kSFPB", "arxiv_id": "2002.12336", "paper_pdf": "papers/BkgF4kSFPB.pdf", "paper_pdf_sha256": "83688d472624b80f280220069766928741253d656665d2e51fad7cf2e4403ae0", "paper_pdf_bytes": 3414096, "paper_pdf_source": "openreview", "code_url": "https://github.com/thanard/hallucinative-topological-memory", "code_repository": "thanard/hallucinative-topological-memory", "code_commit": "5e75a0fb926dbee2323e292a82846b8ea81f99fa", "code_archive": "repos/BkgF4kSFPB.zip", "code_archive_sha256": "17474f39bf4c8971fdfec403717338366a44166844364c4555f4de52eb45ef5c", "code_archive_bytes": 34465, "code_file_count": 19, "code_extensions": {".py": 14, ".sh": 5}, "github_disk_usage_kb": 71, "github_languages": {"Python": 54931, "Shell": 1428}, "github_archived": false, "github_pushed_at": "2020-08-04T04:03:59Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/hallucinative-topological-memory-for-zero-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "HyGySsAct7", "year": 2019, "status": "rejected", "title": "Targeted Adversarial Examples for Black Box Audio Systems", "authors": ["Rohan Taori", "Amog Kamsetty", "Brenton Chu", "Nikita Vemuri"], "authorids": ["rohantaori@berkeley.edu", "amogkamsetty@berkeley.edu", "brentonlongchu@berkeley.edu", "nikitavemuri@berkeley.edu"], "authors_source": "OpenReview API", "abstract": "The application of deep recurrent networks to audio transcription has led to impressive gains in automatic speech recognition (ASR) systems. Many have demonstrated that small adversarial perturbations can fool deep neural networks into incorrectly predicting a specified target with high confidence. Current work on fooling ASR systems have focused on white-box attacks, in which the model architecture and parameters are known. In this paper, we adopt a black-box approach to adversarial generation, combining the approaches of both genetic algorithms and gradient estimation to solve the task. We achieve a 89.25% targeted attack similarity after 3000 generations while maintaining 94.6% audio file similarity.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=HyGySsAct7", "arxiv_id": "1805.07820", "paper_pdf": "papers/HyGySsAct7.pdf", "paper_pdf_sha256": "815b739b4090b9da68f044e7975924cb0c2cfbd6be6d245f097e9fa1418642e6", "paper_pdf_bytes": 1122854, "paper_pdf_source": "openreview", "code_url": "https://github.com/rtaori/Black-Box-Audio", "code_repository": "rtaori/Black-Box-Audio", "code_commit": "3d6c9001faed2988b39ee142200d11e7a54846a3", "code_archive": "repos/HyGySsAct7.zip", "code_archive_sha256": "96ce550898c6c061c6279b0fc145726d4498fd6bb74672ff42ec17952b7be21b", "code_archive_bytes": 469586, "code_file_count": 3, "code_extensions": {".py": 3}, "github_disk_usage_kb": 466, "github_languages": {"Python": 17625}, "github_archived": false, "github_pushed_at": "2020-08-27T19:40:32Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/targeted-adversarial-examples-for-black-box"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "B1spAqUp-", "year": 2018, "status": "rejected", "title": "Pixel Deconvolutional Networks", "authors": ["Hongyang Gao", "Hao Yuan", "Zhengyang Wang", "Shuiwang Ji"], "authorids": ["hongyang.gao@wsu.edu", "hao.yuan@wsu.edu", "zwang6@eecs.wsu.edu", "sji@eecs.wsu.edu"], "authors_source": "OpenReview API", "abstract": "Deconvolutional layers have been widely used in a variety of deep\nmodels for up-sampling, including encoder-decoder networks for\nsemantic segmentation and deep generative models for unsupervised\nlearning. One of the key limitations of deconvolutional operations\nis that they result in the so-called checkerboard problem. This is\ncaused by the fact that no direct relationship exists among adjacent\npixels on the output feature map. To address this problem, we\npropose the pixel deconvolutional layer (PixelDCL) to establish\ndirect relationships among adjacent pixels on the up-sampled feature\nmap. Our method is based on a fresh interpretation of the regular\ndeconvolution operation. The resulting PixelDCL can be used to\nreplace any deconvolutional layer in a plug-and-play manner without\ncompromising the fully trainable capabilities of original models.\nThe proposed PixelDCL may result in slight decrease in efficiency,\nbut this can be overcome by an implementation trick. Experimental\nresults on semantic segmentation demonstrate that PixelDCL can\nconsider spatial features such as edges and shapes and yields more\naccurate segmentation outputs than deconvolutional layers. When used\nin image generation tasks, our PixelDCL can largely overcome the\ncheckerboard problem suffered by regular deconvolution operations.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=B1spAqUp-", "arxiv_id": "1705.06820", "paper_pdf": "papers/B1spAqUp-.pdf", "paper_pdf_sha256": "225d2d063947da936db5e382ba17ddbb63eda29732d07ce110a773bf56c688e3", "paper_pdf_bytes": 3512620, "paper_pdf_source": "openreview", "code_url": "https://github.com/divelab/PixelTCN", "code_repository": "divelab/PixelTCN", "code_commit": "40806e88da8b174ed38f7fae1fcc5f345e3e67c2", "code_archive": "repos/B1spAqUp-.zip", "code_archive_sha256": "0e00f408b9ff95de60708ae79935ed862ff9c5e0452341e361738319e6c6ee93", "code_archive_bytes": 7724808, "code_file_count": 7, "code_extensions": {".py": 7}, "github_disk_usage_kb": 5072, "github_languages": {"Python": 33457}, "github_archived": false, "github_pushed_at": "2018-12-13T17:15:23Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/pixel-deconvolutional-networks"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "G9bpAoc47a", "year": 2026, "status": "rejected", "title": "It Takes a Graph to Know a Graph: Rewiring for Homophily with a Reference Graph", "authors": ["Harel Mendelman", "Haggai Maron", "Ronen Talmon"], "authorids": ["~Harel_Mendelman1", "~Haggai_Maron1", "~Ronen_Talmon2"], "authors_source": "OpenReview API", "abstract": "Graph Neural Networks (GNNs) excel at analyzing graph-structured data but struggle on heterophilic graphs, where connected nodes often belong to different classes. While this challenge is commonly addressed with specialized GNN architectures, graph rewiring remains an underexplored strategy in this context. We provide theoretical foundations linking edge homophily, GNN embedding smoothness, and node classification performance, motivating the need to enhance homophily. Building on this insight, we introduce a rewiring framework that increases graph homophily using a reference graph, with theoretical guarantees on the homophily of the rewired graph. To broaden applicability, we propose a label-driven diffusion approach for constructing a homophilic reference graph from node features and training labels. Through extensive simulations, we analyze how the homophily of both the original and reference graphs influences the rewired graph homophily and downstream GNN performance. We evaluate our method on 13 real-world heterophilic datasets and show that it outperforms existing rewiring techniques and specialized GNNs for heterophilic graphs, achieving improved node classification accuracy while remaining efficient and scalable to large graphs.", "decision": "Reject", "meta_review": "", "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=G9bpAoc47a", "arxiv_id": "2505.12411", "paper_pdf": "papers/G9bpAoc47a.pdf", "paper_pdf_sha256": "18cd23519a8058ee7a8f078e7f58c68036750efd9b3d82bf7d1047cdd7b05a64", "paper_pdf_bytes": 9253210, "paper_pdf_source": "openreview", "code_url": "https://github.com/harel147/REFine", "code_repository": "harel147/REFine", "code_commit": "288eef57e78a7c0030cf3ba4be905b5278eff39f", "code_archive": "repos/G9bpAoc47a.zip", "code_archive_sha256": "a48ff9c7842f79d2cbe6e5172af4b7b87380f42f0e229f020809c6e6eb4a55a4", "code_archive_bytes": 335922, "code_file_count": 35, "code_extensions": {".py": 34, ".sh": 1}, "github_disk_usage_kb": 311, "github_languages": {"Python": 186327, "Makefile": 1154, "Shell": 394}, "github_archived": false, "github_pushed_at": "2026-01-26T14:58:13Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/it-takes-a-graph-to-know-a-graph-rewiring-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "fEEbTDoecM", "year": 2025, "status": "rejected", "title": "RL$^3$: Boosting Meta Reinforcement Learning via RL inside RL$^2$", "authors": ["Abhinav Bhatia", "Samer B. Nashed", "Shlomo Zilberstein"], "authorids": ["~Abhinav_Bhatia1", "~Samer_B._Nashed1", "~Shlomo_Zilberstein1"], "authors_source": "OpenReview API", "abstract": "Meta reinforcement learning (meta-RL) methods such as \\rlsquare have emerged as promising approaches for learning data-efficient RL algorithms tailored to a given task distribution. However, they show poor asymptotic performance and struggle with out-of-distribution tasks because they rely on sequence models, such as recurrent neural networks or transformers, to process experiences rather than summarize them using general-purpose RL components such as value functions. In contrast, traditional RL algorithms are data-inefficient as they do not use domain knowledge, but do converge to an optimal policy in the limit. We propose RL$^3$, a principled hybrid approach that incorporates action-values, learned per task via traditional RL, in the inputs to meta-RL. We show that RL$^3$ earns greater cumulative reward in the long term compared to RL$^2$ while drastically reducing meta-training time and generalizes better to out-of-distribution tasks. Experiments are conducted on both custom and benchmark discrete domains from the meta-RL literature that exhibit a range of short-term, long-term, and complex dependencies.", "decision": "Reject", "meta_review": "", "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=fEEbTDoecM", "arxiv_id": "2306.15909", "paper_pdf": "papers/fEEbTDoecM.pdf", "paper_pdf_sha256": "ef7644aba83f86aa707328f336b05b3a741b0482f5b5b696eb57dde99f1ed242", "paper_pdf_bytes": 5781210, "paper_pdf_source": "openreview", "code_url": "https://github.com/bhatiaabhinav/RL3", "code_repository": "bhatiaabhinav/RL3", "code_commit": "0ae3447be0d25a268df17391e6e279d11402c127", "code_archive": "repos/fEEbTDoecM.zip", "code_archive_sha256": "da0856caf38f058dfd30efda6c7bb21239ccc4015460536eb8f1fd4443d2313c", "code_archive_bytes": 50491, "code_file_count": 6, "code_extensions": {".jl": 6}, "github_disk_usage_kb": 48, "github_languages": {"Julia": 72661}, "github_archived": false, "github_pushed_at": "2024-12-08T16:11:10Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/rl-3-boosting-meta-reinforcement-learning-via"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "WnEnU2K3Rb", "year": 2024, "status": "rejected", "title": "Beyond the Benchmark: Detecting Diverse Anomalies in Videos", "authors": ["Yoav Arad", "Michael Werman"], "authorids": ["~Yoav_Arad1", "~Michael_Werman1"], "authors_source": "OpenReview API", "abstract": "Video Anomaly Detection (VAD) plays a crucial role in modern surveillance systems, aiming to  identify various anomalies in real-world situations. However, current benchmark datasets predominantly emphasize simple, single-frame anomalies such as novel object detection. This narrow focus restricts the advancement of VAD models. In this research, we advocate for an expansion of VAD investigations to encompass intricate anomalies that extend beyond conventional benchmark boundaries.\nTo facilitate this, we introduce two  datasets, HMDB-AD and HMDB-Violence, to challenge models with diverse action-based anomalies. These datasets are derived from the HMDB51 action recognition dataset.\nWe further present Multi-Frame Anomaly Detection (MFAD), a novel method built upon the AI-VAD framework. \nAI-VAD utilizes single-frame features such as pose estimation and deep image encoding, and two-frame features such as object velocity. They then apply a density estimation algorithm to compute anomaly scores. To address complex multi-frame anomalies, we add a deep video encoding features capturing long-range temporal dependencies, and logistic regression  to enhance final score calculation.\nExperimental results confirm our assumptions, highlighting existing models limitations with new anomaly types. MFAD excels in both simple and complex anomaly detection scenarios.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=WnEnU2K3Rb", "arxiv_id": "2310.01904", "paper_pdf": "papers/WnEnU2K3Rb.pdf", "paper_pdf_sha256": "c8b07d25c6bfb8fb55a14e652d41c699034747cb903ff672c50f6ae744227b01", "paper_pdf_bytes": 10393905, "paper_pdf_source": "openreview", "code_url": "https://github.com/yoavarad/MFAD", "code_repository": "yoavarad/MFAD", "code_commit": "a2a3d3fbb3c609081af537af34ac4288c4f39954", "code_archive": "repos/WnEnU2K3Rb.zip", "code_archive_sha256": "4d597a422be7606cf94ac232bf5e9732cc7895bb36e93323e6602ca1a25c4442", "code_archive_bytes": 17922, "code_file_count": 10, "code_extensions": {".py": 9, ".sh": 1}, "github_disk_usage_kb": 40, "github_languages": {"Python": 59588, "Shell": 1282}, "github_archived": false, "github_pushed_at": "2023-09-28T10:57:56Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/beyond-the-benchmark-detecting-diverse"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "j3AyKG-H3uM", "year": 2023, "status": "rejected", "title": "Improved Group Robustness via Classifier Retraining on Independent Splits", "authors": ["Thien Hang Nguyen", "Hongyang Ryan Zhang", "Huy Nguyen"], "authorids": ["~Thien_Hang_Nguyen1", "~Hongyang_Ryan_Zhang1", "~Huy_Nguyen1"], "authors_source": "OpenReview API", "abstract": "Deep neural networks learned by minimizing the average risk can achieve strong average performance, but their performance for a subgroup may degrade, if the subgroup is underrepresented in the overall data population. Group distributionally robust optimization (Sagawa et al., 2020a, GDRO) is a standard baseline for learning models with strong worst-group performance. However, GDRO requires group labels for every example during training and can be prone to overfitting, often requiring careful model capacity control via regularization or early stopping. When only a limited amount of group labels is available, Just Train Twice (Liu et al., 2021, JTT) is a popular approach which infers a pseudo-group-label for every unlabeled example. The process of inferring pseudo labels can be highly sensitive during model selection. To alleviate overfitting for GDRO and the pseudo labeling process for JTT, we propose a new method via classifier retraining on independent splits (of the training data). We find that using a novel sample splitting procedure achieves robust worst-group performance in the fine-tuning step. When evaluated on benchmark image and text classification tasks, our approach consistently reduces the requirement of group labels and hyperparameter search during training. Experimental results confirm that our approach performs favorably compared with existing methods (including GDRO and JTT) when either group labels are available during training or are only available during validation.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=j3AyKG-H3uM", "arxiv_id": "2204.09583", "paper_pdf": "papers/j3AyKG-H3uM.pdf", "paper_pdf_sha256": "281a45e1840587f8b41835a912b54684bbf7aa2f6d525608e7ea1117f2dd7e64", "paper_pdf_bytes": 1218406, "paper_pdf_source": "openreview", "code_url": "https://github.com/timmytonga/crois", "code_repository": "timmytonga/crois", "code_commit": "33930be328adeb87f38792630b924e077bcc9d3a", "code_archive": "repos/j3AyKG-H3uM.zip", "code_archive_sha256": "1a49355239eec7d4bb7f77d8cfcd12f90bbf037d660a090230ecec3e7d701050", "code_archive_bytes": 57308, "code_file_count": 26, "code_extensions": {".py": 26}, "github_disk_usage_kb": 49, "github_languages": {"Python": 179333}, "github_archived": false, "github_pushed_at": "2023-07-27T15:26:55Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/improved-worst-group-robustness-via"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "46lmrnVBHBL", "year": 2022, "status": "rejected", "title": "Explanatory Learning: Beyond Empiricism in Neural Networks", "authors": ["Antonio Norelli", "Giorgio Mariani", "Luca Moschella", "Andrea Santilli", "Giambattista Parascandolo", "Simone Melzi", "Emanuele Rodolà"], "authorids": ["~Antonio_Norelli2", "~Giorgio_Mariani1", "~Luca_Moschella1", "~Andrea_Santilli1", "~Giambattista_Parascandolo1", "~Simone_Melzi2", "~Emanuele_Rodolà1"], "authors_source": "OpenReview API", "abstract": "We introduce Explanatory Learning (EL), an explanation-driven machine learning framework to use existing knowledge buried in symbolic sequences expressed in an unknown language. In EL, the burden of interpreting explanations is not left to humans or human-coded compilers, as done in Program Synthesis. Rather, EL calls for a learned interpreter, built upon existing explanations paired with observations of several phenomena. This interpreter can then be used to make predictions on novel phenomena, and even find an explanation for them. We formulate the EL problem as a simple binary classification task, so that common end-to-end approaches aligned with the dominant empiricist view of machine learning could, in principle, solve it. To these models, we oppose Critical Rationalist Networks (CRNs), which instead embrace a rationalist view on the acquisition of knowledge. CRNs express several desired properties by construction, they are truly explainable, can adjust their processing at test-time for harder inferences, and can offer strong confidence guarantees on their predictions.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=46lmrnVBHBL", "arxiv_id": "2201.10222", "paper_pdf": "papers/46lmrnVBHBL.pdf", "paper_pdf_sha256": "926fd69698fdd89e21e85bc08fb3481222090d8aa755bcb519635c5b18e4f734", "paper_pdf_bytes": 8725249, "paper_pdf_source": "openreview", "code_url": "https://github.com/gladia-research-group/explanatory-learning", "code_repository": "gladia-research-group/explanatory-learning", "code_commit": "40d32f3ea8a2b7d875452de288cc8b611369423c", "code_archive": "repos/46lmrnVBHBL.zip", "code_archive_sha256": "c6883c7e9e619a07d5d9fcbbab0d7f2b7c3e4993dcee4d9e96556013535c9bda", "code_archive_bytes": 59820, "code_file_count": 29, "code_extensions": {".py": 29}, "github_disk_usage_kb": 61, "github_languages": {"Python": 168574}, "github_archived": false, "github_pushed_at": "2022-05-17T11:18:37Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/explanatory-learning-beyond-empiricism-in-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "tckGH8K9y6o", "year": 2021, "status": "rejected", "title": "Symmetric Wasserstein Autoencoders", "authors": ["Sun Sun", "Hongyu Guo"], "authorids": ["~Sun_Sun1", "~Hongyu_Guo1"], "authors_source": "OpenReview API", "abstract": "Leveraging the framework of Optimal Transport, we introduce a new family of generative autoencoders with a learnable prior, called Symmetric Wasserstein Autoencoders (SWAEs). We propose to symmetrically match the joint distributions of the observed data and the latent representation induced by the encoder and the decoder. The resultant algorithm jointly optimizes the modelling losses in both the data and the latent spaces with  the loss in the data space leading to the denoising effect. With the symmetric treatment of the data and the latent representation, the algorithm implicitly preserves the local structure of the data in the latent space. To further improve the latent representation, we incorporate a reconstruction loss into the objective, which significantly benefits both the generation and reconstruction. We empirically show the superior performance of SWAEs over several state-of-the-art generative autoencoders in terms of classification, reconstruction, and generation.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=tckGH8K9y6o", "arxiv_id": "2106.13024", "paper_pdf": "papers/tckGH8K9y6o.pdf", "paper_pdf_sha256": "e1987ec91ea20e18b648ac668f21c02ccae1f2474361caaa6584ea3a3f1badfc", "paper_pdf_bytes": 38922030, "paper_pdf_source": "openreview", "code_url": "https://github.com/sunsunyyl/SWAE", "code_repository": "sunsunyyl/SWAE", "code_commit": "0554c3b67bddda189e4eeda40aab147963997281", "code_archive": "repos/tckGH8K9y6o.zip", "code_archive_sha256": "f8f68048b4e27d8222e8801e25b95baa42aa0716b28654fa1deeb839ddecf99f", "code_archive_bytes": 48086, "code_file_count": 23, "code_extensions": {".py": 22, ".sh": 1}, "github_disk_usage_kb": 62, "github_languages": {"Python": 195735, "Shell": 1463}, "github_archived": false, "github_pushed_at": "2021-06-29T19:32:36Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/symmetric-wasserstein-autoencoders-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "BkgStySKPB", "year": 2020, "status": "rejected", "title": "Contrastive Multiview Coding", "authors": ["Yonglong Tian", "Dilip Krishnan", "Phillip Isola"], "authorids": ["yonglong@mit.edu", "dilipkay@google.com", "phillipi@mit.edu"], "authors_source": "OpenReview API", "abstract": "Humans view the world through many sensory channels, e.g., the long-wavelength light channel, viewed by the left eye, or the high-frequency vibrations channel, heard by the right ear. Each view is noisy and incomplete, but important factors, such as physics, geometry, and semantics, tend to be shared between all views (e.g., a \"dog\" can be seen, heard, and felt). We hypothesize that a powerful representation is one that models view-invariant factors. Based on this hypothesis, we investigate a contrastive coding scheme, in which a representation is learned that aims to maximize mutual information between different views but is otherwise compact. Our approach scales to any number of views, and is view-agnostic. The resulting learned representations perform above the state of the art for downstream tasks such as object classification, compared to formulations based on predictive learning or single view reconstruction, and improve as more views are added. On the Imagenet linear readoff benchmark, we achieve 68.4% top-1 accuracy. ", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=BkgStySKPB", "arxiv_id": "1906.05849", "paper_pdf": "papers/BkgStySKPB.pdf", "paper_pdf_sha256": "7eaa203b54d639db823de64b9560e8d0698cbc15c4b6beda0e8ea7a43cd23cdf", "paper_pdf_bytes": 1495630, "paper_pdf_source": "openreview", "code_url": "https://github.com/HobbitLong/CMC", "code_repository": "HobbitLong/CMC", "code_commit": "7b227be0b10ef4e526c72af07664f5079ed9ee09", "code_archive": "repos/BkgStySKPB.zip", "code_archive_sha256": "79418356244c8040ee8d63d8e7b042c1ab45d8ae7d73cefe779c3c7761865fb9", "code_archive_bytes": 32632, "code_file_count": 12, "code_extensions": {".py": 12}, "github_disk_usage_kb": 76, "github_languages": {"Python": 98729}, "github_archived": false, "github_pushed_at": "2020-11-10T07:33:17Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/contrastive-multiview-coding"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "BkedwoC5t7", "year": 2019, "status": "rejected", "title": "Formal Limitations on the Measurement of Mutual Information", "authors": ["David McAllester", "Karl Stratos"], "authorids": ["mcallester@ttic.edu", "stratos@ttic.edu"], "authors_source": "OpenReview API", "abstract": "Motivated by applications to unsupervised learning, we consider the problem of measuring mutual information. Recent analysis has shown that naive kNN estimators of mutual information have serious statistical limitations motivating more refined methods. In this paper we prove that serious statistical limitations are inherent to any measurement method. More specifically, we show that any distribution-free high-confidence lower bound on mutual information cannot be larger than $O(\\ln N)$ where $N$ is the size of the data sample. We also analyze the Donsker-Varadhan lower bound on KL divergence in particular and show that, when simple statistical considerations are taken into account, this bound can never produce a high-confidence value larger than $\\ln N$. While large high-confidence lower bounds are impossible, in practice one can use estimators without formal guarantees. We suggest expressing mutual information as a difference of entropies and using cross entropy as an entropy estimator.  We observe that, although cross entropy is only an upper bound on entropy, cross-entropy estimates converge to the true cross entropy at the rate of $1/\\sqrt{N}$.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=BkedwoC5t7", "arxiv_id": "1811.04251", "paper_pdf": "papers/BkedwoC5t7.pdf", "paper_pdf_sha256": "09808aa345082d851bd65a9b259ed82c81d8ec3670116fa78ba58ae363c9f68b", "paper_pdf_bytes": 186582, "paper_pdf_source": "openreview", "code_url": "https://github.com/karlstratos/doe", "code_repository": "karlstratos/doe", "code_commit": "84b5d5c394d5e6be3455a48ced742c231e2fef5c", "code_archive": "repos/BkedwoC5t7.zip", "code_archive_sha256": "a99bb40e031255a74b2caae7b89a33d6fc7d422be17a9561f89f1b1700c65231", "code_archive_bytes": 250905, "code_file_count": 3, "code_extensions": {".py": 3}, "github_disk_usage_kb": 476, "github_languages": {"Python": 21116}, "github_archived": false, "github_pushed_at": "2022-04-13T19:15:25Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/formal-limitations-on-the-measurement-of"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "HytSvlWRZ", "year": 2018, "status": "rejected", "title": "Subspace Network: Deep Multi-Task Censored Regression for Modeling Neurodegenerative Diseases", "authors": ["Mengying Sun", "Inci M. Baytas", "Zhangyang Wang", "Jiayu Zhou"], "authorids": ["sunmeng2@msu.edu", "baytasin@msu.edu", "atlaswang@tamu.edu", "jiayuz@msu.edu"], "authors_source": "OpenReview API", "abstract": "Over the past decade a wide spectrum of machine learning models have been developed to model the neurodegenerative diseases, associating biomarkers, especially non-intrusive neuroimaging markers, with key clinical scores measuring the cognitive status of patients. Multi-task learning (MTL) has been extensively explored in these studies to address challenges associated to high dimensionality and small cohort size. However, most existing MTL approaches are based on linear models and suffer from two major limitations: 1) they cannot explicitly consider upper/lower bounds in these clinical scores; 2) they lack the capability to capture complicated non-linear effects among the variables. In this paper, we propose the Subspace Network, an efficient deep modeling approach for non-linear multi-task censored regression. Each layer of the subspace network performs a multi-task censored regression to improve upon the predictions from the last layer via sketching a low-dimensional subspace to perform knowledge transfer among learning tasks. We show that under mild assumptions, for each layer the parametric subspace can be recovered using only one pass of training data. In addition, empirical results demonstrate that the proposed subspace network quickly picks up correct parameter subspaces, and outperforms state-of-the-arts in predicting neurodegenerative clinical scores using information in brain imaging. ", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=HytSvlWRZ", "arxiv_id": "1802.06516", "paper_pdf": "papers/HytSvlWRZ.pdf", "paper_pdf_sha256": "85c20e92e1ff241feb59b97c98effc1118cda2c3a2c830d079c37acb223a404d", "paper_pdf_bytes": 410074, "paper_pdf_source": "openreview", "code_url": "https://github.com/illidanlab/subspace-net", "code_repository": "illidanlab/subspace-net", "code_commit": "39c811836550c4762cd40ff9c3b32382a7f5a807", "code_archive": "repos/HytSvlWRZ.zip", "code_archive_sha256": "bd3a27ec9909349e32e15caf1b70776e9002c79c4fdad9a7dc6f3220405dcf46", "code_archive_bytes": 9319542, "code_file_count": 19, "code_extensions": {".py": 19}, "github_disk_usage_kb": 9102, "github_languages": {"Python": 33399}, "github_archived": false, "github_pushed_at": "2018-02-25T04:45:13Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/subspace-network-deep-multi-task-censored"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "7lzq9mMVxq", "year": 2026, "status": "rejected", "title": "Feature Hedging: Correlated Features Break Narrow Sparse Autoencoders", "authors": ["David Chanin", "Tomáš Dulka", "Adrià Garriga-Alonso"], "authorids": ["~David_Chanin1", "~Tomáš_Dulka1", "~Adrià_Garriga-Alonso1"], "authors_source": "OpenReview API", "abstract": "It is assumed that sparse autoencoders (SAEs) decompose polysemantic activations into interpretable linear directions, as long as the activations are composed of sparse linear combinations of underlying features. However, we find that if an SAE is more narrow than the number of underlying \"true features\" on which it is trained, and there is correlation between features, the SAE will merge components of correlated features together, thus destroying monosemanticity. In LLM SAEs, these two conditions are almost certainly true. This phenomenon, which we call feature hedging, is caused by SAE reconstruction loss, and is more severe the narrower the SAE. In this work, we introduce the problem of feature hedging and study it both theoretically in toy models and empirically in SAEs trained on LLMs. We suspect that feature hedging may be one of the core reasons that SAEs consistently underperform supervised baselines. Finally, we use our understanding of feature hedging to propose an improved variant of matryoshka SAEs. Importantly, our work shows that SAE width is not a neutral hyperparameter: narrower SAEs suffer more from hedging than wider SAEs.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=7lzq9mMVxq", "arxiv_id": "2505.11756", "paper_pdf": "papers/7lzq9mMVxq.pdf", "paper_pdf_sha256": "8646991b763679c38c987cc748cd9ff22363e1e3b868d43039bf583054066e1b", "paper_pdf_bytes": 1400967, "paper_pdf_source": "openreview", "code_url": "https://github.com/chanind/feature-hedging-paper", "code_repository": "chanind/feature-hedging-paper", "code_commit": "6b6d7d64ce3f8871b8fec2e67557f6a5c7d00827", "code_archive": "repos/7lzq9mMVxq.zip", "code_archive_sha256": "b7ce0542f3dbab0051323fd259d8cd258cf9052a7356006b7eaa9ef6c1b676f1", "code_archive_bytes": 314993, "code_file_count": 69, "code_extensions": {".py": 65, ".ipynb": 4}, "github_disk_usage_kb": 332, "github_languages": {"Python": 340819, "Jupyter Notebook": 105284}, "github_archived": false, "github_pushed_at": "2025-11-26T15:15:58Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/feature-hedging-correlated-features-break"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "hkjcdmz8Ro", "year": 2024, "status": "rejected", "title": "Jailbreaking Black Box Large Language Models in Twenty Queries", "authors": ["Patrick Chao", "Alexander Robey", "Edgar Dobriban", "Hamed Hassani", "George J. Pappas", "Eric Wong"], "authorids": ["~Patrick_Chao1", "~Alexander_Robey1", "~Edgar_Dobriban1", "~Hamed_Hassani2", "~George_J._Pappas1", "~Eric_Wong1"], "authors_source": "OpenReview API", "abstract": "There is growing research interest in ensuring that large language models align with human safety and ethical guidelines. Adversarial attacks known as 'jailbreaks' pose a significant threat as they coax models into overriding alignment safeguards. Identifying these vulnerabilities through attacking a language model (red teaming) is instrumental in understanding inherent weaknesses and preventing misuse. We present Prompt Automatic Iterative Refinement (PAIR), which generates semantic jailbreaks with only black-box access \nto a language model.\nEmpirically, PAIR often requires fewer than 20 queries, orders of magnitude fewer than prior jailbreak attacks. PAIR draws inspiration from the human process of social engineering, and employs an attacker language model to automatically generate adversarial prompts in place of a human. The attacker model uses the target model's response as additional context to iteratively refine the adversarial prompt. PAIR achieves competitive jailbreaking success rates and transferability on open and closed-source language models, including GPT-3.5/4, Vicuna, and PaLM.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=hkjcdmz8Ro", "arxiv_id": "2310.08419", "paper_pdf": "papers/hkjcdmz8Ro.pdf", "paper_pdf_sha256": "8b9f505b06727faf677f4fbcb47886d2ec0d3b1dd614d2879b12cca15fdd5912", "paper_pdf_bytes": 2863510, "paper_pdf_source": "openreview", "code_url": "https://github.com/patrickrchao/JailbreakingLLMs", "code_repository": "patrickrchao/JailbreakingLLMs", "code_commit": "6379ef705a0fc745530f7d895963510c021b496a", "code_archive": "repos/hkjcdmz8Ro.zip", "code_archive_sha256": "75e7f0351db06f8b142a9e2315030330873571ffffd7d6a40d3603c0dd9c5e4c", "code_archive_bytes": 29111, "code_file_count": 12, "code_extensions": {".py": 12}, "github_disk_usage_kb": 41, "github_languages": {"Python": 73751, "Dockerfile": 362}, "github_archived": false, "github_pushed_at": "2025-07-02T00:03:40Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/jailbreaking-black-box-large-language-models"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ueEMZjY9WiM", "year": 2023, "status": "rejected", "title": "Compression-aware Training of Neural Networks using Frank-Wolfe", "authors": ["Max Zimmer", "Christoph Spiegel", "Sebastian Pokutta"], "authorids": ["~Max_Zimmer1", "~Christoph_Spiegel1", "~Sebastian_Pokutta1"], "authors_source": "OpenReview API", "abstract": "Many existing Neural Network pruning approaches either rely on retraining to compensate for pruning-caused performance degradation or they induce strong biases to converge to a specific sparse solution throughout training. A third paradigm, ’compression-aware’ training, obtains state-of-the-art dense models which are robust to a wide range of compression ratios using a single dense training run while also avoiding retraining. In that vein, we propose a constrained optimization framework centered around a versatile family of norm constraints and the Stochastic Frank-Wolfe (SFW) algorithm which together encourage convergence to well-performing solutions while inducing robustness towards convolutional filter pruning and low-rank matrix decomposition. Comparing our novel approaches to compression methods in these domains on benchmark image-classification architectures and datasets, we find that our proposed scheme is able to yield competitive results, often outperforming existing compression-aware approaches. In the case of low-rank matrix decomposition, our approach can require much less computational resources than nuclear-norm regularization based approaches by requiring only a fraction of the singular values in each iteration. As a special case, our proposed constraints can be extended to include the unstructured sparsity-inducing constraint proposed constraint by Pokutta et al. (2020) and Miao et al. (2022), which we improve upon. Our findings also indicate that the robustness of SFW-trained models largely depends on the gradient rescaling of the learning rate and we establish a theoretical foundation for that practice.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=ueEMZjY9WiM", "arxiv_id": "2205.11921", "paper_pdf": "papers/ueEMZjY9WiM.pdf", "paper_pdf_sha256": "71945e16dd49a9b53db08ca758b1bee9cb33bb83ebc3d460dac27bc13c4fb8f8", "paper_pdf_bytes": 564052, "paper_pdf_source": "openreview", "code_url": "https://github.com/ZIB-IOL/compression-aware-SFW", "code_repository": "ZIB-IOL/compression-aware-SFW", "code_commit": "6e9e0febc09f4af43596c1aceaf8c29e6214fdb6", "code_archive": "repos/ueEMZjY9WiM.zip", "code_archive_sha256": "86c89f0e738ea1361b4e957ad813fc6d9393d7dafca9ff5a3d68e943d4a0aa0c", "code_archive_bytes": 55203, "code_file_count": 19, "code_extensions": {".py": 19}, "github_disk_usage_kb": 49, "github_languages": {"Python": 198043, "TeX": 296}, "github_archived": false, "github_pushed_at": "2023-02-22T09:22:36Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/compression-aware-training-of-neural-networks"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "aJ_GcB4vcT0", "year": 2022, "status": "rejected", "title": "Unsupervised Learning of Neurosymbolic Encoders", "authors": ["Eric Zhan", "Jennifer J. Sun", "Ann Kennedy", "Yisong Yue", "Swarat Chaudhuri"], "authorids": ["~Eric_Zhan1", "~Jennifer_J._Sun1", "~Ann_Kennedy1", "~Yisong_Yue1", "~Swarat_Chaudhuri1"], "authors_source": "OpenReview API", "abstract": "We present a framework for the unsupervised learning of neurosymbolic encoders, i.e., encoders obtained by composing neural networks with symbolic programs from a domain-specific language. Such a framework can naturally incorporate symbolic expert knowledge into the learning process and lead to more interpretable and factorized latent representations than fully neural encoders. Also, models learned this way can have downstream impact, as many analysis workflows can benefit from having clean programmatic descriptions. We ground our learning algorithm in the variational autoencoding (VAE) framework, where we aim to learn a neurosymbolic encoder in conjunction with a standard decoder. Our algorithm integrates standard VAE-style training with modern program synthesis techniques. We evaluate our method on learning latent representations for real-world trajectory data from animal biology and sports analytics. We show that our approach offers significantly better separation than standard VAEs and leads to practical gains on downstream tasks.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=aJ_GcB4vcT0", "arxiv_id": "2107.13132", "paper_pdf": "papers/aJ_GcB4vcT0.pdf", "paper_pdf_sha256": "a5f1a50c761a88d7969b46cbee49f7b38798e99801fb54b1ae6e5d21c96f7b09", "paper_pdf_bytes": 3419080, "paper_pdf_source": "openreview", "code_url": "https://github.com/ezhan94/neurosymbolic-encoders", "code_repository": "ezhan94/neurosymbolic-encoders", "code_commit": "c52a83a33c99e02c52be7b176edeb2c6bdedec78", "code_archive": "repos/aJ_GcB4vcT0.zip", "code_archive_sha256": "feda2969087efd7f01a5a984eb29bf87928e74a4e7f594980dee37d587956e5d", "code_archive_bytes": 85769, "code_file_count": 52, "code_extensions": {".py": 50, ".sh": 2}, "github_disk_usage_kb": 65, "github_languages": {"Python": 227332, "Shell": 8642}, "github_archived": false, "github_pushed_at": "2023-06-05T02:39:57Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/unsupervised-learning-of-neurosymbolic"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "xFYXLlpIyPQ", "year": 2021, "status": "rejected", "title": "Guarantees for Tuning the Step Size using a Learning-to-Learn Approach", "authors": ["Xiang Wang", "Shuai Yuan", "Chenwei Wu", "Rong Ge"], "authorids": ["~Xiang_Wang1", "shuai@cs.duke.edu", "~Chenwei_Wu1", "~Rong_Ge1"], "authors_source": "OpenReview API", "abstract": "Learning-to-learn---using optimization algorithms to learn a new optimizer---has successfully trained efficient optimizers in practice. This approach relies on meta-gradient descent on a meta-objective based on the trajectory that the optimizer generates. However, there were few theoretical guarantees on how to avoid meta-gradient explosion/vanishing problem, or how to train an optimizer with good generalization performance. In this paper, we study the learning-to-learn approach on a simple problem of tuning the step size for quadratic loss. Our results show that although there is a way to design the meta-objective so that the meta-gradient remains polynomially bounded, computing the meta-gradient directly using backpropagation leads to numerical issues that look similar to gradient explosion/vanishing problems. We also characterize when it is necessary to compute the meta-objective on a separate validation set instead of the original training set. Finally, we verify our results empirically and show that a similar phenomenon appears even for more complicated learned optimizers parametrized by neural networks.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=xFYXLlpIyPQ", "arxiv_id": "2006.16495", "paper_pdf": "papers/xFYXLlpIyPQ.pdf", "paper_pdf_sha256": "b72b7f9b9c1e02c71497713fdd8639343487e8464d113bcd0a61994855dba06b", "paper_pdf_bytes": 749683, "paper_pdf_source": "openreview", "code_url": "https://github.com/shuaiyuan1996/learning-to-learn", "code_repository": "shuaiyuan1996/learning-to-learn", "code_commit": "5c937d6b9e4a893fd3ac4955c44aa46afb81012d", "code_archive": "repos/xFYXLlpIyPQ.zip", "code_archive_sha256": "ade990114b1baea1f9b2eb755dca54d42db606f1219d74c2a3aba42d6c9def5c", "code_archive_bytes": 119455, "code_file_count": 34, "code_extensions": {".py": 30, ".sh": 4}, "github_disk_usage_kb": 63, "github_languages": {"Python": 388958, "Shell": 1469}, "github_archived": false, "github_pushed_at": "2021-06-09T21:30:48Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/guarantees-for-tuning-the-step-size-using-a"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "SJxNzgSKvH", "year": 2020, "status": "rejected", "title": "Selective sampling for accelerating  training of deep neural networks", "authors": ["Berry Weinstein", "Shai Fine", "Yacov Hel-Or"], "authorids": ["berry.weinstein@post.idc.ac.il", "shai.fine@idc.ac.il", "toky@idc.ac.il"], "authors_source": "OpenReview API", "abstract": "We present a selective sampling method designed to accelerate the training of deep neural networks. To this end,  we introduce a novel measurement,  the {\\it minimal margin score} (MMS), which measures the minimal amount of displacement an input should take until its predicted classification is switched.   For multi-class linear classification,  the MMS measure is a natural generalization of the margin-based selection criterion, which was thoroughly studied in the binary classification setting.  In addition, the MMS measure provides an interesting insight into the progress of the training process and can be useful for designing and monitoring new training regimes. Empirically we demonstrate a substantial acceleration when training commonly used deep neural network architectures for popular image classification tasks.  The efficiency of our method is compared against the standard training procedures, and against commonly used selective sampling alternatives: Hard negative mining selection, and Entropy-based selection.\nFinally, we demonstrate an additional speedup when we adopt a more aggressive learning-drop regime while using the MMS selective sampling method.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=SJxNzgSKvH", "arxiv_id": "1911.06996", "paper_pdf": "papers/SJxNzgSKvH.pdf", "paper_pdf_sha256": "15cbc67e612a4d0e8b3d0a379a83536db80b510327334510460cd7133e9d88cb", "paper_pdf_bytes": 904527, "paper_pdf_source": "openreview", "code_url": "https://github.com/berryweinst/mms-select", "code_repository": "berryweinst/mms-select", "code_commit": "e1a9dda89c69ad5bec3edd5a0cd6bb37c22753d8", "code_archive": "repos/SJxNzgSKvH.zip", "code_archive_sha256": "94795d0f099ba1aa2b726b2b565aba6f4aa7a95729d6ec903989d2d9f5b22ffb", "code_archive_bytes": 89824, "code_file_count": 47, "code_extensions": {".py": 47}, "github_disk_usage_kb": 81, "github_languages": {"Python": 317969}, "github_archived": false, "github_pushed_at": "2019-10-02T07:26:46Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/selective-sampling-for-accelerating-training-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "r1GkMhAqYm", "year": 2019, "status": "rejected", "title": "CoDraw: Collaborative Drawing as a Testbed for Grounded Goal-driven Communication", "authors": ["Nikita Kitaev", "Jin-Hwa Kim", "Xinlei Chen", "Marcus Rohrbach", "Yuandong Tian", "Dhruv Batra", "Devi Parikh"], "authorids": ["kitaev@cs.berkeley.edu", "jnhwkim@gmail.com", "xinleic@fb.com", "maroffm@gmail.com", "yuandong@fb.com", "dbatra@gatech.edu", "parikh@gatech.edu"], "authors_source": "OpenReview API", "abstract": "In this work, we propose a goal-driven collaborative task that contains language, vision, and action in a virtual environment as its core components. Specifically, we develop a Collaborative image-Drawing game between two agents, called CoDraw. Our game is grounded in a virtual world that contains movable clip art objects. The game involves two players: a Teller and a Drawer. The Teller sees an abstract scene containing multiple clip art pieces in a semantically meaningful configuration, while the Drawer tries to reconstruct the scene on an empty canvas using available clip art pieces. The two players communicate via two-way communication using natural language. We collect the CoDraw dataset of ~10K dialogs consisting of ~138K messages exchanged between human agents. We define protocols and metrics to evaluate the effectiveness of learned agents on this testbed, highlighting the need for a novel \"crosstalk\" condition which pairs agents trained independently on disjoint subsets of the training data for evaluation. We present models for our task, including simple but effective baselines and neural network approaches trained using a combination of imitation learning and goal-driven training. All models are benchmarked using both fully automated evaluation and by playing the game with live human agents.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=r1GkMhAqYm", "arxiv_id": "1712.05558", "paper_pdf": "papers/r1GkMhAqYm.pdf", "paper_pdf_sha256": "3f47fca372a160081c23ec1c4c7b2e87024f34a3d54ddd173d195a1736bd75af", "paper_pdf_bytes": 4087502, "paper_pdf_source": "openreview", "code_url": "https://github.com/facebookresearch/CoDraw", "code_repository": "facebookresearch/CoDraw", "code_commit": "b209770a327f48fdd768724bbcf2783897b0c7fb", "code_archive": "repos/r1GkMhAqYm.zip", "code_archive_sha256": "5a589ea0cbdb9d28895d6d14d76b3641439317cac96fe7c853e834a0ee7b4d55", "code_archive_bytes": 925079, "code_file_count": 9, "code_extensions": {".js": 8, ".py": 1}, "github_disk_usage_kb": 552, "github_languages": {"JavaScript": 50428, "HTML": 6385, "CSS": 1853, "Python": 713}, "github_archived": true, "github_pushed_at": "2019-01-15T21:03:26Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/codraw-collaborative-drawing-as-a-testbed-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "79uLTCNMpJ", "year": 2026, "status": "rejected", "title": "From Emergence to Control: Probing and Modulating Self-Reflection in Language Models", "authors": ["Xudong Zhu", "Jiachen Jiang", "Mohammad Mahdi Khalili", "Zhihui Zhu"], "authorids": ["~Xudong_Zhu1", "~Jiachen_Jiang1", "~Mohammad_Mahdi_Khalili3", "~Zhihui_Zhu1"], "authors_source": "OpenReview API", "abstract": "Self-reflection---the ability of a large language model (LLM) to revisit, evaluate, and revise its own reasoning---has recently emerged as a powerful behavior enabled by reinforcement learning with verifiable rewards (RLVR). While self-reflection correlates with improved reasoning accuracy, its origin and underlying mechanisms remain poorly understood. In this work, we first show that self-reflection is not exclusive to RLVR fine-tuned models: it already emerges, albeit rarely, in pretrained models. To probe this latent ability, we introduce Reflection-Inducing Probing, a method that injects reflection-triggering reasoning traces from fine-tuned models into pretrained models. This intervention raises self-reflection frequency of Qwen2.5 from 0.6% to 18.6%, revealing a hidden capacity for reflection. Moreover, our analysis of internal representations shows that both pretrained and fine-tuned models maintain hidden states that distinctly separate self-reflective from non-reflective contexts. Leveraging this observation, we then construct a self-reflection vector, a direction in activation space associated with self-reflective reasoning. By manipulating this vector, we enable bidirectional control over the self-reflective behavior for both pretrained and fine-tuned models. Experiments across multiple reasoning benchmarks show that enhancing these vectors improves reasoning performance by up to 12%, while suppressing them reduces computational cost, providing a flexible mechanism to navigate the trade-off between reasoning quality and efficiency without requiring additional training. Our findings further our understanding of self-reflection and support a growing body of work showing that understanding model internals can enable precise behavioral control.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=79uLTCNMpJ", "arxiv_id": "2506.12217", "paper_pdf": "papers/79uLTCNMpJ.pdf", "paper_pdf_sha256": "5b7659603bea5a77158d40fa02b73aac414fff609befeddc301e1cbc19132658", "paper_pdf_bytes": 3683241, "paper_pdf_source": "openreview", "code_url": "https://github.com/xzAscC/ProbingReflection", "code_repository": "xzAscC/ProbingReflection", "code_commit": "33991f446ee887f959f7f22be330b03d07c2a5ca", "code_archive": "repos/79uLTCNMpJ.zip", "code_archive_sha256": "115836012fe713a2d0a534b4749d3e47068b118721a93e47bc71d53f34740f44", "code_archive_bytes": 293907, "code_file_count": 37, "code_extensions": {".py": 29, ".sh": 8}, "github_disk_usage_kb": 358, "github_languages": {"Python": 265817, "Shell": 11872}, "github_archived": false, "github_pushed_at": "2026-07-16T22:15:17Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/from-emergence-to-control-probing-and"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "E9GakjQype", "year": 2025, "status": "rejected", "title": "AdvPrompter: Fast Adaptive Adversarial Prompting for LLMs", "authors": ["Anselm Paulus", "Arman Zharmagambetov", "Chuan Guo", "Brandon Amos", "Yuandong Tian"], "authorids": ["~Anselm_Paulus1", "~Arman_Zharmagambetov1", "~Chuan_Guo1", "~Brandon_Amos1", "~Yuandong_Tian1"], "authors_source": "OpenReview API", "abstract": "While recently Large Language Models (LLMs) have achieved remarkable successes, they are vulnerable to certain `jailbreaking attacks` that lead to generation of inappropriate or harmful content. Manual red-teaming requires finding adversarial prompts that cause such jailbreaking, e.g. by appending a suffix to a given instruction, which is inefficient and time-consuming. \nOn the other hand, automatic adversarial prompt generation often leads to semantically meaningless attacks that can easily be detected by perplexity-based filters, may require gradient information from the TargetLLM, or do not scale well due to time-consuming discrete optimization processes over the token space. In this paper, we present a novel method that uses another LLM, called the `AdvPrompter`, to generate human-readable adversarial prompts in seconds, $\\sim800\\times$ faster than existing optimization-based approaches.\nWe train the AdvPrompter using a novel algorithm that `does not require gradients` of the TargetLLM. This process alternates between two steps: (1) generating high-quality target adversarial suffixes by optimizing the AdvPrompter predictions, and (2) fine-tuning of the AdvPrompter with the generated adversarial suffixes. The trained AdvPrompter generates suffixes that veil the input instruction without changing its meaning, such that the TargetLLM is lured to give a harmful response. Experimental results on popular open source TargetLLMs show state-of-the-art results on the AdvBench dataset, that also transfer to closed-source black-box LLM APIs. Further, we demonstrate that by fine-tuning on a synthetic dataset generated by AdvPrompter, LLMs can be made more robust against jailbreaking attacks while maintaining performance, i.e. high MMLU scores.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=E9GakjQype", "arxiv_id": "2404.16873", "paper_pdf": "papers/E9GakjQype.pdf", "paper_pdf_sha256": "d2beb5335bae944c14f9ed923fb1cfe4eda5f43f2d00b0b92fd9f11389d02c1f", "paper_pdf_bytes": 2188146, "paper_pdf_source": "openreview", "code_url": "https://github.com/facebookresearch/advprompter", "code_repository": "facebookresearch/advprompter", "code_commit": "802a500c91f1dcd7c8b76869d3e39bf8e40ed7d7", "code_archive": "repos/E9GakjQype.zip", "code_archive_sha256": "0cdb6d552ae22cb7d8d04cec9a49bb0b4d1015946f2e715ab772704a8b4a9079", "code_archive_bytes": 58385, "code_file_count": 5, "code_extensions": {".py": 5}, "github_disk_usage_kb": 49, "github_languages": {"Python": 93783}, "github_archived": true, "github_pushed_at": "2024-05-06T20:09:40Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/advprompter-fast-adaptive-adversarial"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "qRbkTbe8JT", "year": 2024, "status": "rejected", "title": "IMEX-Reg: Implicit-Explicit Regularization in the Function Space for Continual Learning", "authors": ["Prashant Shivaram Bhat", "Bharath Chennamkulam Renjith", "Bahram Zonooz", "Elahe Arani"], "authorids": ["~Prashant_Shivaram_Bhat1", "~Bharath_Chennamkulam_Renjith1", "~Bahram_Zonooz1", "~Elahe_Arani1"], "authors_source": "OpenReview API", "abstract": "Continual learning (CL) remains one of the long-standing challenges for deep\nneural networks due to catastrophic forgetting of previously acquired knowledge.\nAlthough rehearsal-based approaches have been fairly successful in mitigating\ncatastrophic forgetting, they suffer from overfitting on buffered samples and prior\ninformation loss, hindering generalization under low-buffer regimes. Inspired\nby how humans learn using strong inductive biases, we propose IMEX-Reg to\nimprove the generalization performance of experience rehearsal in CL under low\nbuffer regimes. Specifically, we employ a two-pronged implicit-explicit regular-\nization approach using contrastive representation learning (CRL) and consistency\nregularization. To further leverage the global relationship between representations\nlearned using CRL, we propose a novel regularization strategy to guide the clas-\nsifier toward the activation correlations in the unit hypersphere of the CRL. Our\nresults show that IMEX-Reg significantly improves generalization performance and\noutperforms rehearsal-based approaches in several CL scenarios. It is also robust\nto natural and adversarial corruptions with less task-recency bias. Additionally, we\nprovide theoretical insights to support our design decisions further.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=qRbkTbe8JT", "arxiv_id": "2404.18161", "paper_pdf": "papers/qRbkTbe8JT.pdf", "paper_pdf_sha256": "c60973ec84e37afe0d9317a98c09963f29841b574c735e885533505dc78ccba2", "paper_pdf_bytes": 489335, "paper_pdf_source": "openreview", "code_url": "https://github.com/NeurAI-Lab/IMEX-Reg", "code_repository": "NeurAI-Lab/IMEX-Reg", "code_commit": "74169d4617c1a86db2d3f1b11ef7182689cb9293", "code_archive": "repos/qRbkTbe8JT.zip", "code_archive_sha256": "8b28b80fa188503cae961903cb76834b8b0234c6c4e44046b9e883c91e947cee", "code_archive_bytes": 52460, "code_file_count": 41, "code_extensions": {".py": 41}, "github_disk_usage_kb": 42, "github_languages": {"Python": 158980}, "github_archived": false, "github_pushed_at": "2024-05-24T20:11:44Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/imex-reg-implicit-explicit-regularization-in"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "6OxI4WqGr6", "year": 2023, "status": "rejected", "title": "Semi-Supervised Offline Reinforcement Learning with Action-Free Trajectories", "authors": ["Qinqing Zheng", "Mikael Henaff", "Brandon Amos", "Aditya Grover"], "authorids": ["~Qinqing_Zheng1", "~Mikael_Henaff1", "~Brandon_Amos1", "~Aditya_Grover1"], "authors_source": "OpenReview API", "abstract": "Natural agents can effectively learn from multiple data sources that differ in size, quality, and types of measurements. We study this heterogeneity in the context of offline reinforcement learning (RL) by introducing a new, practically motivated semi-supervised setting. Here, an agent has access to two sets of trajectories: labelled trajectories containing state, action, reward triplets at every timestep, along with unlabelled trajectories that contain only state and reward information.  For this setting, we develop a simple meta-algorithmic pipeline that learns an inverse-dynamics model on the labelled data to obtain proxy-labels for the unlabelled data, followed by the use of any offline RL algorithm on the true and proxy-labelled trajectories. Empirically, we find this simple pipeline to be highly successful --- on several D4RL benchmarks~\\cite{fu2020d4rl}, certain offline RL algorithms can match the performance of variants trained on a fully labeled dataset even when we label only 10\\% trajectories from the low return regime. Finally, we perform a large-scale controlled empirical study investigating the interplay of data-centric properties of the labelled and unlabelled datasets, with algorithmic design choices (e.g., inverse dynamics, offline RL algorithm) to identify general trends and best practices for training RL agents on semi-supervised offline datasets.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=6OxI4WqGr6", "arxiv_id": "2210.06518", "paper_pdf": "papers/6OxI4WqGr6.pdf", "paper_pdf_sha256": "a2f78117fae4e7d2c63b2d5af3b61fdd8abf613d1b1aee5ecb27237ccb422172", "paper_pdf_bytes": 1146698, "paper_pdf_source": "openreview", "code_url": "https://github.com/facebookresearch/ssorl", "code_repository": "facebookresearch/ssorl", "code_commit": "cfefabed915e76aefb453f29344055014906d4ed", "code_archive": "repos/6OxI4WqGr6.zip", "code_archive_sha256": "efe378d8359c08fb7cfbd609168b18f0401b778ad6d216234bbde8a4bfd629cb", "code_archive_bytes": 59279, "code_file_count": 18, "code_extensions": {".py": 18}, "github_disk_usage_kb": 53, "github_languages": {"Python": 152547}, "github_archived": true, "github_pushed_at": "2023-07-16T20:08:02Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/semi-supervised-offline-reinforcement-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "s51gCxF70pq", "year": 2022, "status": "rejected", "title": "Learning Temporally-Consistent Representations for Data-Efficient Reinforcement Learning", "authors": ["Trevor McInroe", "Lukas Schäfer", "Stefano V Albrecht"], "authorids": ["~Trevor_McInroe1", "~Lukas_Schäfer1", "~Stefano_V_Albrecht1"], "authors_source": "OpenReview API", "abstract": "Deep reinforcement learning (RL) agents that exist in high-dimensional state spaces, such as those composed of images, have interconnected learning burdens. Agents must learn an action-selection policy that completes their given task, which requires them to learn a representation of the state space that discerns between useful and useless information. The reward function is the only supervised feedback that RL agents receive, which causes a representation learning bottleneck that can manifest in poor sample efficiency. We present $k$-Step Latent (KSL), a new representation learning method that enforces temporal consistency of representations via a self-supervised auxiliary task wherein agents learn to recurrently predict action-conditioned representations of the state space. The state encoder learned by KSL produces low-dimensional representations that make optimization of the RL task more sample efficient. Altogether, KSL produces state-of-the-art results in both data efficiency and asymptotic performance in the popular PlaNet benchmark suite. Our analyses show that KSL produces encoders that generalize better to new tasks unseen during training, and its representations are more strongly tied to reward, are more invariant to perturbations in the state space, and move more smoothly through the temporal axis of the RL problem than other methods such as DrQ, RAD, CURL, and SAC-AE.", "decision": "Reject", "meta_review": null, "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=s51gCxF70pq", "arxiv_id": "2110.04935", "paper_pdf": "papers/s51gCxF70pq.pdf", "paper_pdf_sha256": "77427493e7d80a45671bf60be24684de67c5ee544cc420d85428ae4edcc0ff18", "paper_pdf_bytes": 3816891, "paper_pdf_source": "openreview", "code_url": "https://github.com/anon-researcher-repo/ksl", "code_repository": "anon-researcher-repo/ksl", "code_commit": "72f0665322e8e4de33838a8e3a57356c192c5743", "code_archive": "repos/s51gCxF70pq.zip", "code_archive_sha256": "9c2e5dde0894d72f4527b68515946252de7caf329d800dfadbb7072428cda9cb", "code_archive_bytes": 55851, "code_file_count": 7, "code_extensions": {".py": 6, ".ipynb": 1}, "github_disk_usage_kb": 66, "github_languages": {"Jupyter Notebook": 56108, "Python": 48398}, "github_archived": false, "github_pushed_at": "2021-10-06T14:48:07Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/learning-temporally-consistent-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "lJuOUWlAC8i", "year": 2021, "status": "rejected", "title": "Learning Contextualized Knowledge Structures for Commonsense Reasoning", "authors": ["Jun Yan", "Mrigank Raman", "Tianyu Zhang", "Ryan Rossi", "Handong Zhao", "Sungchul Kim", "Nedim Lipka", "Xiang Ren"], "authorids": ["~Jun_Yan5", "mt1170736@iitd.ac.in", "zhang-ty17@mails.tsinghua.edu.cn", "~Ryan_Rossi1", "~Handong_Zhao3", "sukim@adobe.com", "lipka@adobe.com", "~Xiang_Ren1"], "authors_source": "OpenReview API", "abstract": "Recently, neural-symbolic architectures have achieved success on commonsense reasoning through effectively encoding relational structures retrieved from external knowledge graphs (KGs) and obtained state-of-the-art results in tasks such as (commonsense) question answering and natural language inference. However, current neural-symbolic reasoning methods rely on quality and contextualized knowledge structures (i.e., fact triples) that can be retrieved at the pre-processing stage and overlook challenges such as dealing with incompleteness of a KG (low coverage), limited expressiveness of its relations, and irrelevant retrieved facts in the reasoning context. \nIn this paper, we present a novel neural-symbolic approach, named Hybrid Graph Network (HGN), which jointly generates feature representations for new triples (as complement to the existing edges in the KG), determines relevance of the triples to the reasoning context, and learns graph model parameters for encoding the relational information. Our method learns a compact graph structure (comprising both retrieved and generated edges) through filtering edges that are unhelpful to the reasoning process. We show marked improvements on three commonsense reasoning benchmarks and demonstrate the superiority of the learned graph structures with user studies.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=lJuOUWlAC8i", "arxiv_id": "2010.12873", "paper_pdf": "papers/lJuOUWlAC8i.pdf", "paper_pdf_sha256": "8c713e67d2faba556ea45b1c39a904c3161730a0c00c720bec8c91c8143a7539", "paper_pdf_bytes": 628874, "paper_pdf_source": "openreview", "code_url": "https://github.com/INK-USC/HGN", "code_repository": "INK-USC/HGN", "code_commit": "42c391dc9753156c5dfb3fb686b615110d8c36ce", "code_archive": "repos/lJuOUWlAC8i.zip", "code_archive_sha256": "10f6cd9a9c3b1e13b9a49c4e5e9dade80ddac6b72d18b5694890478c45cbca38", "code_archive_bytes": 71478, "code_file_count": 21, "code_extensions": {".py": 20, ".sh": 1}, "github_disk_usage_kb": 68, "github_languages": {"Python": 259119, "Shell": 1569}, "github_archived": false, "github_pushed_at": "2022-11-24T09:41:17Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/learning-contextualized-knowledge-structures"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ryxtCpNtDS", "year": 2020, "status": "rejected", "title": "Autoencoders and Generative Adversarial Networks for Imbalanced Sequence Classification", "authors": ["Stephanie Ger", "Diego Klabjan"], "authorids": ["stephanieger@u.northwestern.edu", "d-klabjan@northwestern.edu"], "authors_source": "OpenReview API", "abstract": "We introduce a novel synthetic oversampling method for variable length, multi- feature sequence datasets based on autoencoders and generative adversarial net- works. We show that this method improves classification accuracy for highly imbalanced sequence classification tasks. We show that this method outperforms standard oversampling techniques that use techniques such as SMOTE and autoencoders. We also use generative adversarial networks on the majority class as an outlier detection method for novelty detection, with limited classification improvement. We show that the use of generative adversarial network based synthetic data improves classification model performance on a variety of sequence data sets.\n", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=ryxtCpNtDS", "arxiv_id": "1901.02514", "paper_pdf": "papers/ryxtCpNtDS.pdf", "paper_pdf_sha256": "86ce9fba8d962d3b3126fb94ed8b21a1ce29bf2ea78a37a0c3953bd4b3f89a3a", "paper_pdf_bytes": 286288, "paper_pdf_source": "openreview", "code_url": "https://github.com/stephanieger/imbalanced-sequence-classification", "code_repository": "stephanieger/imbalanced-sequence-classification", "code_commit": "3534ff8930aab1a1c98e18443fc89ce2c4a84aac", "code_archive": "repos/ryxtCpNtDS.zip", "code_archive_sha256": "43e2934862da677362062e517ebb8d7cfd418654f9579688be03853784a4f3d2", "code_archive_bytes": 46025, "code_file_count": 20, "code_extensions": {".py": 20}, "github_disk_usage_kb": 96, "github_languages": {"Python": 261239}, "github_archived": false, "github_pushed_at": "2022-12-08T05:31:33Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/autoencoders-and-generative-adversarial"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "S1eFtj0cKQ", "year": 2019, "status": "rejected", "title": "Generative Models from the perspective of Continual Learning", "authors": ["Timothée Lesort", "Hugo Caselles-Dupré", "Michael Garcia-Ortiz", "Jean-François Goudou", "David Filliat"], "authorids": ["timothee.lesort@thalesgroup.com", "caselles@ensta.fr", "mgarciaortiz@softbankrobotics.com", "jean-francois.goudou@thalesgroup.com", "david.filliat@ensta.fr"], "authors_source": "OpenReview API", "abstract": "Which generative model is the most suitable for Continual Learning? This paper aims at evaluating and comparing generative models on disjoint sequential image generation tasks. We investigate how several models learn and forget, considering various strategies: rehearsal, regularization, generative replay and fine-tuning. We used two quantitative metrics to estimate the generation quality and memory ability. We experiment with sequential tasks on three commonly used benchmarks for Continual Learning (MNIST, Fashion MNIST and CIFAR10). We found that among all models, the original GAN performs best and among Continual Learning strategies, generative replay outperforms all other methods. Even if we found satisfactory combinations on MNIST and Fashion MNIST, training generative models sequentially on CIFAR10 is particularly instable, and remains a challenge.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=S1eFtj0cKQ", "arxiv_id": "1812.09111", "paper_pdf": "papers/S1eFtj0cKQ.pdf", "paper_pdf_sha256": "4af51d5ff91c155bcae8ac010586b6b76a9ae40befcd9c5289ef184213b69573", "paper_pdf_bytes": 12967663, "paper_pdf_source": "openreview", "code_url": "https://github.com/TLESORT/Generative_Continual_Learning", "code_repository": "TLESORT/Generative_Continual_Learning", "code_commit": "66b121437c248993b41f154b5a2d6b7197278578", "code_archive": "repos/S1eFtj0cKQ.zip", "code_archive_sha256": "8eb5fcccad983692cb5e1089fe7478a79fcb7ff9d881575240d5f991abe9e85e", "code_archive_bytes": 565858, "code_file_count": 39, "code_extensions": {".py": 38, ".sh": 1}, "github_disk_usage_kb": 560, "github_languages": {"Python": 237477, "Shell": 4556}, "github_archived": false, "github_pushed_at": "2019-11-22T13:51:52Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/generative-models-from-the-perspective-of"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "dDvH0UA0CZ", "year": 2026, "status": "rejected", "title": "Sampling from Your Language Model One Byte at a Time", "authors": ["Jonathan Hayase", "Alisa Liu", "Noah A. Smith", "Sewoong Oh"], "authorids": ["~Jonathan_Hayase2", "~Alisa_Liu1", "~Noah_A._Smith2", "~Sewoong_Oh3"], "authors_source": "OpenReview API", "abstract": "Tokenization is used almost universally by modern language models, enabling efficient text representation using multi-byte or multi-character tokens. These models are typically invoked to autoregressively complete a text prompt by tokenizing the prompt, sampling more tokens to continue the tokenized prompt, and detokenizing the result. However, prior work has shown that this process can introduce distortion into the model's sampling distribution, leading to unexpected or undesirable generations. For example, users are often advised not to end their prompts with a space because it prevents the model from including the space as part of the next token. While this heuristic is effective in English, the underlying problem continues to affect languages such as Chinese as well as code generation, settings where word and syntactic boundaries may not line up with token boundaries. We present an optimal method to solve this \"Prompt Boundary Problem,\" which is based on an efficient online algorithm for Byte-Pair Encoding (BPE). This allows one to compute the next byte distribution conditioned on an arbitrary byte prefix, given only logit access to the original tokenizer-based model. This procedure can be applied iteratively to convert any autoregressive LM with a BPE tokenizer into a character-level or byte-level LM, without changing the generative distribution at the text level. We show that this significantly improves next-character prediction accuracy when computed on arbitrary prefixes. Moreover, our method is able to unify the vocabularies of language models with different tokenizers, allowing one to ensemble LMs with different tokenizers at inference time as well as transfer the post-training from one model to another using proxy-tuning. We demonstrate in experiments that the ensemble and proxy-tuned models outperform their constituents on downstream evals.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=dDvH0UA0CZ", "arxiv_id": "2506.14123", "paper_pdf": "papers/dDvH0UA0CZ.pdf", "paper_pdf_sha256": "6e2c4d13fcc52e1d60ee1148593c7210646f1c7ebaeaa0347fca218df95e973d", "paper_pdf_bytes": 437853, "paper_pdf_source": "openreview", "code_url": "https://github.com/SewoongLab/byte-sampler", "code_repository": "SewoongLab/byte-sampler", "code_commit": "f27de98fa3ad9d5df7704a31c8d30972a41d26ce", "code_archive": "repos/dDvH0UA0CZ.zip", "code_archive_sha256": "b4291266e14add6ea400764adb885c6eb11684760c47017dfbe460c9b22e3a56", "code_archive_bytes": 265698, "code_file_count": 13, "code_extensions": {".py": 13}, "github_disk_usage_kb": 361, "github_languages": {"Python": 147990}, "github_archived": false, "github_pushed_at": "2026-02-26T21:28:47Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/sampling-from-your-language-model-one-byte-at"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Nk1MegaPuG", "year": 2025, "status": "rejected", "title": "Evading Data Contamination Detection for Language Models is (too) Easy", "authors": ["Jasper Dekoninck", "Mark Niklas Mueller", "Maximilian Baader", "Marc Fischer", "Martin Vechev"], "authorids": ["~Jasper_Dekoninck1", "~Mark_Niklas_Mueller2", "~Maximilian_Baader1", "~Marc_Fischer1", "~Martin_Vechev1"], "authors_source": "OpenReview API", "abstract": "The benchmark performance of large language models (LLMs) has a high impact on their popularity and is thus of great importance to many model providers. However, the reliability of such benchmark scores as a measure of model quality gets compromised if the model is contaminated with benchmark data. While recent contamination detection methods try to address this issue, they overlook the possibility of deliberate contamination by malicious model providers aiming to evade detection. We propose a categorization of model providers based on their (de)contamination practices and argue that malicious contamination is of crucial importance as it casts doubt on the reliability of public benchmarks. To study this issue more rigorously, we analyze current contamination detection methods based on their assumptions. This analysis reveals a significant vulnerability in existing approaches: they do not account for rephrased benchmark data used during training by malicious actors. We demonstrate how exploiting this gap can result in significantly inflated benchmark scores while completely evading current detection methods.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=Nk1MegaPuG", "arxiv_id": "2402.02823", "paper_pdf": "papers/Nk1MegaPuG.pdf", "paper_pdf_sha256": "17493804626c3bf4e46d9bf8d0730495f1082805e02ca55ceae5d8fe767213c1", "paper_pdf_bytes": 1640940, "paper_pdf_source": "openreview", "code_url": "https://github.com/eth-sri/malicious-contamination", "code_repository": "eth-sri/malicious-contamination", "code_commit": "59d02ef6bd42ff1047a0520bbe58807f0324b22d", "code_archive": "repos/Nk1MegaPuG.zip", "code_archive_sha256": "69b85fd40a5aa0497cce2ae90f631693908288f012702721183d7d8030bd1416", "code_archive_bytes": 60655, "code_file_count": 30, "code_extensions": {".py": 22, ".sh": 7, ".ipynb": 1}, "github_disk_usage_kb": 51, "github_languages": {"Python": 129906, "Jupyter Notebook": 26259, "Shell": 6523}, "github_archived": false, "github_pushed_at": "2024-02-12T15:25:29Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/evading-data-contamination-detection-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "mt5NPvTp5a", "year": 2024, "status": "rejected", "title": "Improved Operator Learning by Orthogonal Attention", "authors": ["Zipeng Xiao", "Zhongkai Hao", "Bokai Lin", "Zhijie Deng", "Hang Su"], "authorids": ["~Zipeng_Xiao1", "~Zhongkai_Hao1", "~Bokai_Lin1", "~Zhijie_Deng1", "~Hang_Su3"], "authors_source": "OpenReview API", "abstract": "Neural operators, as an efficient surrogate model for learning the solutions of PDEs, have received extensive attention in the field of scientific machine learning.\nAmong them, attention-based neural operators have become one of the mainstreams in related research.\nHowever, existing approaches overfit the limited training data due to the considerable number of parameters in the attention mechanism.\nTo address this, we develop an orthogonal attention based on the eigendecomposition of the kernel integral operator and the neural approximation of eigenfunctions. \nThe orthogonalization naturally poses a proper regularization effect on the resulting neural operator, which aids in resisting overfitting and boosting generalization. \nExperiments on six standard neural operator benchmark datasets comprising both regular and irregular geometries show that our method can outperform competing baselines with decent margins.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=mt5NPvTp5a", "arxiv_id": "2310.12487", "paper_pdf": "papers/mt5NPvTp5a.pdf", "paper_pdf_sha256": "3c6067466dc44a9ea885a0eacc49070ffb35a709d441c7ba435932600ebcf5c6", "paper_pdf_bytes": 889287, "paper_pdf_source": "openreview", "code_url": "https://github.com/SJTU-DENG-Lab/Orthogonal-Neural-operator", "code_repository": "SJTU-DENG-Lab/Orthogonal-Neural-operator", "code_commit": "c9bc85dcd3630ffcb9822acc455a551cae76d662", "code_archive": "repos/mt5NPvTp5a.zip", "code_archive_sha256": "046b9e7bf4c13b010b2eb1608e30c65e3c3e16c32a9d78cb2e60bc3b24206473", "code_archive_bytes": 34180, "code_file_count": 10, "code_extensions": {".py": 10}, "github_disk_usage_kb": 42, "github_languages": {"Python": 114700}, "github_archived": false, "github_pushed_at": "2023-10-15T03:10:08Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/improved-operator-learning-by-orthogonal"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "nIGza1_wxk", "year": 2023, "status": "rejected", "title": "Model Transferability with Responsive Decision Subjects ", "authors": ["Yang Liu", "Yatong Chen", "Zeyu Tang", "Kun Zhang"], "authorids": ["~Yang_Liu3", "~Yatong_Chen1", "~Zeyu_Tang1", "~Kun_Zhang1"], "authors_source": "OpenReview API", "abstract": "This paper studies model transferability when human decision subjects respond to a deployed machine learning model. In our setting, an agent or a user corresponds to a sample $(X,Y)$ drawn from a distribution $\\mathcal{D}$ and will face a model $h$ and its classification result $h(X)$. Agents can modify $X$ to adapt to $h$, which will incur a distribution shift on $(X,Y)$. Therefore, when training $h$, the learner will need to consider the subsequently ``induced\" distribution when the output model is deployed. Our formulation is motivated by applications where the deployed machine learning models interact with human agents, and will ultimately face \\emph{responsive} and interactive data distributions. We formalize the discussions of the transferability of a model by studying how the model trained on the available source distribution (data) would translate to the performance on the induced domain. We provide both upper bounds for the performance gap due to the induced domain shift, as well as lower bound for the trade-offs that a classifier has to suffer on either the source training distribution or the induced target distribution. We provide further instantiated analysis for two popular domain adaptation settings with covariate shift and target shift.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=nIGza1_wxk", "arxiv_id": "2107.05911", "paper_pdf": "papers/nIGza1_wxk.pdf", "paper_pdf_sha256": "c2b6bca7159c19e9e0f659b7920d3c53865250ed8b1057b3c6b1fd8d6868e45e", "paper_pdf_bytes": 1596478, "paper_pdf_source": "openreview", "code_url": "https://github.com/UCSC-REAL/Model_Transferability", "code_repository": "UCSC-REAL/Model_Transferability", "code_commit": "25eefb48e784dc4f808b74ec902fdeba7fb7ab0c", "code_archive": "repos/nIGza1_wxk.zip", "code_archive_sha256": "b769c8b92240ce3bfbfc3666ea3b35c86d327b43ae39dddb41e4853f56b9a6fd", "code_archive_bytes": 50657, "code_file_count": 8, "code_extensions": {".py": 7, ".ipynb": 1}, "github_disk_usage_kb": 56, "github_languages": {"Jupyter Notebook": 124753, "Python": 61416}, "github_archived": false, "github_pushed_at": "2023-05-31T11:36:22Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/induced-domain-adaptation"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "jNB6vfl_680", "year": 2022, "status": "rejected", "title": "Global Magnitude Pruning With Minimum Threshold Is All We Need", "authors": ["Manas Gupta", "Vishandi Rudy Keneta", "Abhishek Vaidyanathan", "Ritwik Kanodia", "Efe Camci", "Chuan-Sheng Foo", "Jie Lin"], "authorids": ["~Manas_Gupta1", "~Vishandi_Rudy_Keneta1", "~Abhishek_Vaidyanathan1", "~Ritwik_Kanodia1", "~Efe_Camci1", "~Chuan-Sheng_Foo1", "~Jie_Lin1"], "authors_source": "OpenReview API", "abstract": "Neural network pruning remains a very important yet challenging problem to solve. Many pruning solutions have been proposed over the years with high degrees of algorithmic complexity. In this work, we shed light on a very simple pruning technique that achieves state-of-the-art (SOTA) performance. We showcase that magnitude based pruning, specifically, global magnitude pruning (GP) is sufficient to achieve SOTA performance on a range of neural network architectures. In certain architectures, the last few layers of a network may get over-pruned. For these cases, we introduce a straightforward method to mitigate this. We preserve a certain fixed number of weights in each layer of the network to ensure no layer is over-pruned. We call this the Minimum Threshold (MT). We find that GP combined with MT when needed, achieves SOTA performance on all datasets and architectures tested including ResNet-50 and MobileNet-V1 on ImageNet. Code available on github.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=jNB6vfl_680", "arxiv_id": null, "paper_pdf": "papers/jNB6vfl_680.pdf", "paper_pdf_sha256": "69d9bdeda6b27a9cb31ea8f8834509a402e7393b7bfc356b31a13adf3d44ce67", "paper_pdf_bytes": 1288947, "paper_pdf_source": "openreview", "code_url": "https://github.com/GPMT-Authors/Global-Pruning-With-Minimum-Threshold", "code_repository": "GPMT-Authors/Global-Pruning-With-Minimum-Threshold", "code_commit": "6e753e4bf39c3343d8514fffd4269c092908a9f8", "code_archive": "repos/jNB6vfl_680.zip", "code_archive_sha256": "e7ca2e42da25f81467874f3426fe01587c1a213909699b3270295dfeff894944", "code_archive_bytes": 28026, "code_file_count": 6, "code_extensions": {".py": 6}, "github_disk_usage_kb": 66, "github_languages": {"Python": 94785}, "github_archived": false, "github_pushed_at": "2021-10-06T12:48:04Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/global-magnitude-pruning-with-minimum"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "4Nt1F3qf9Gn", "year": 2021, "status": "rejected", "title": "CLOCS: Contrastive Learning of Cardiac Signals Across Space, Time, and Patients", "authors": ["Dani Kiyasseh", "Tingting Zhu", "David A. Clifton"], "authorids": ["~Dani_Kiyasseh1", "tingting.zhu@eng.ox.ac.uk", "~David_A._Clifton1"], "authors_source": "OpenReview API", "abstract": "The healthcare industry generates troves of unlabelled physiological data. This data can be exploited via contrastive learning, a self-supervised pre-training method that encourages representations of instances to be similar to one another. We propose a family of contrastive learning methods, CLOCS, that encourages representations across space, time, \\textit{and} patients to be similar to one another. We show that CLOCS consistently outperforms the state-of-the-art methods, BYOL and SimCLR, when performing a linear evaluation of, and fine-tuning on, downstream tasks. We also show that CLOCS achieves strong generalization performance with only 25\\% of labelled training data. Furthermore, our training procedure naturally generates patient-specific representations that can be used to quantify patient-similarity. ", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=4Nt1F3qf9Gn", "arxiv_id": "2005.13249", "paper_pdf": "papers/4Nt1F3qf9Gn.pdf", "paper_pdf_sha256": "adabb38f1737e5ffc4fb3ff16d86f833bce954d0895c9a8614a937f4d72eaee7", "paper_pdf_bytes": 1579474, "paper_pdf_source": "openreview", "code_url": "https://github.com/danikiyasseh/CLOCS", "code_repository": "danikiyasseh/CLOCS", "code_commit": "e58fb4f0369b086735169d781a9513730fa157d0", "code_archive": "repos/4Nt1F3qf9Gn.zip", "code_archive_sha256": "d4ebb1e655b44ff2a441510432f121d08daa448c2c14cd3ce5c40ccef28b6d09", "code_archive_bytes": 27970, "code_file_count": 8, "code_extensions": {".py": 8}, "github_disk_usage_kb": 69, "github_languages": {"Python": 82660}, "github_archived": false, "github_pushed_at": "2022-10-19T12:22:57Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/clocs-contrastive-learning-of-cardiac-signals"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "H1g4M0EtPS", "year": 2020, "status": "rejected", "title": "Gaussian MRF Covariance Modeling for Efficient Black-Box Adversarial Attacks", "authors": ["Anit Kumar Sahu", "J. Zico Kolter", "Satya Narayan Shukla"], "authorids": ["anit.sahu@gmail.com", "zkolter@cs.cmu.edu", "snshukla@cs.umass.edu"], "authors_source": "OpenReview API", "abstract": "We study the problem of generating adversarial examples in a black-box setting, where we only have access to a zeroth order oracle, providing us with loss function evaluations. We employ Markov Random Fields (MRF) to exploit the structure of input data to systematically model the covariance structure of the gradients. The MRF structure in addition to Bayesian inference for the gradients facilitates one-step attacks akin to Fast Gradient Sign Method (FGSM) albeit in the black-box setting. The resulting method uses fewer queries than the current state of the art to achieve comparable performance. In particular, in the regime of lower query budgets, we show that our method is particularly effective in terms of fewer average queries with high attack accuracy while employing one-step attacks.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=H1g4M0EtPS", "arxiv_id": "2010.04205", "paper_pdf": "papers/H1g4M0EtPS.pdf", "paper_pdf_sha256": "abb700b6ed17cb5325d5f68ae345ffb45f3a3eca64fdb2b6260a613c29a9bade", "paper_pdf_bytes": 587358, "paper_pdf_source": "openreview", "code_url": "https://github.com/anitksahu/GMRF", "code_repository": "anitksahu/GMRF", "code_commit": "2ae642589cfefac06bafbe0325154b1066fd367f", "code_archive": "repos/H1g4M0EtPS.zip", "code_archive_sha256": "8469a3f16d9fd5a4eae193295dfafe36beefafafc2f8e02784297ec7abc7c5ed", "code_archive_bytes": 94175, "code_file_count": 7, "code_extensions": {".py": 7}, "github_disk_usage_kb": 102, "github_languages": {"Python": 30924}, "github_archived": false, "github_pushed_at": "2020-10-09T07:07:56Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/gaussian-mrf-covariance-modeling-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "HkGmDsR9YQ", "year": 2019, "status": "rejected", "title": "Generalization and Regularization in DQN", "authors": ["Jesse Farebrother", "Marlos C. Machado", "Michael Bowling"], "authorids": ["jfarebro@ualberta.ca", "machado@ualberta.ca", "mbowling@ualberta.ca"], "authors_source": "OpenReview API", "abstract": "Deep reinforcement learning (RL) algorithms have shown an impressive ability to learn complex control policies in high-dimensional environments. However, despite the ever-increasing performance on popular benchmarks like the Arcade Learning Environment (ALE), policies learned by deep RL algorithms can struggle to generalize when evaluated in remarkably similar environments. These results are unexpected given the fact that, in supervised learning, deep neural networks often learn robust features that generalize across tasks. In this paper, we study the generalization capabilities of DQN in order to aid in understanding this mismatch between generalization in deep RL and supervised learning methods. We provide evidence suggesting that DQN overspecializes to the domain it is trained on. We then comprehensively evaluate the impact of traditional methods of regularization from supervised learning, $\\ell_2$ and dropout, and of reusing learned representations to improve the generalization capabilities of DQN. We perform this study using different game modes of Atari 2600 games, a recently introduced modification for the ALE which supports slight variations of the Atari 2600 games used for benchmarking in the field. Despite regularization being largely underutilized in deep RL, we show that it can, in fact, help DQN learn more general features. These features can then be reused and fine-tuned on similar tasks, considerably improving the sample efficiency of DQN.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=HkGmDsR9YQ", "arxiv_id": "1810.00123", "paper_pdf": "papers/HkGmDsR9YQ.pdf", "paper_pdf_sha256": "9bdf26769a0e0e1bf2596c7940165a52d2fa4792a14d54a17a067c8a77c8bf93", "paper_pdf_bytes": 649929, "paper_pdf_source": "openreview", "code_url": "https://github.com/JesseFarebro/dqn-ale", "code_repository": "JesseFarebro/dqn-ale", "code_commit": "a39b3ba0f5e9fd8da05c9dd39f5264e9e4df5575", "code_archive": "repos/HkGmDsR9YQ.zip", "code_archive_sha256": "1f5cd4f16c4215fee020f915f0b22d3e57037b7da13f9513e96a17280f0319e5", "code_archive_bytes": 609319, "code_file_count": 7, "code_extensions": {".py": 7}, "github_disk_usage_kb": 594, "github_languages": {"Python": 30937}, "github_archived": false, "github_pushed_at": "2019-04-16T17:06:51Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/generalization-and-regularization-in-dqn"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "PwaUsAk4i6", "year": 2026, "status": "rejected", "title": "Diffusion Guidance Is a Controllable Policy Improvement Operator", "authors": ["Kevin Frans", "Seohong Park", "Pieter Abbeel", "Sergey Levine"], "authorids": ["~Kevin_Frans1", "~Seohong_Park1", "~Pieter_Abbeel2", "~Sergey_Levine1"], "authors_source": "OpenReview API", "abstract": "At the core of reinforcement learning is the idea of learning beyond the performance in the data. However, scaling such systems has proven notoriously tricky. In contrast, techniques from generative modeling have shown to be remarkably scalable and are simple to train. \nIn this work, we combine these strengths, by deriving a direct relation between policy improvement and guidance of diffusion models. The resulting framework, CFGRL, is a policy improvement operator that is trained with the simplicity of supervised learning, yet is more effective than typically-used weighted policy extraction strategies. On offline RL tasks, we observe a reliable trend---increased guidance weighting leads to increased performance. Additionally, the CFGRL framework can be adapted to \"directly'' extract policies from offline data *without* running a full end-to-end RL algorithm, allowing us to generalize simple supervised methods (e.g. goal-conditioned behavior cloning) to further prioritize optimality, gaining performance across the board without additional cost.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=PwaUsAk4i6", "arxiv_id": "2505.23458", "paper_pdf": "papers/PwaUsAk4i6.pdf", "paper_pdf_sha256": "e78e3498ca4539c1394a48a0dda7a1c517cb3bfce95f914fe8a25347b8409275", "paper_pdf_bytes": 3346170, "paper_pdf_source": "openreview", "code_url": "https://github.com/kvfrans/cfgrl", "code_repository": "kvfrans/cfgrl", "code_commit": "14bd57b53bea27a3951687387520f11fbbe44abb", "code_archive": "repos/PwaUsAk4i6.zip", "code_archive_sha256": "4a0ea7596a89ca9db733f3ec8fc01ba55d44e321738092b31df6405bd756af09", "code_archive_bytes": 440989, "code_file_count": 42, "code_extensions": {".py": 42}, "github_disk_usage_kb": 412, "github_languages": {"Python": 308759}, "github_archived": false, "github_pushed_at": "2025-05-31T01:00:54Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/diffusion-guidance-is-a-controllable-policy"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "aya06N6R4W", "year": 2025, "status": "rejected", "title": "Counterfactual Causal Inference in Natural Language with Large Language Models", "authors": ["Gael Gendron", "Joze M. Rozanec", "Michael Witbrock", "Gillian Dobbie"], "authorids": ["~Gael_Gendron1", "~Joze_M._Rozanec1", "~Michael_Witbrock1", "~Gillian_Dobbie1"], "authors_source": "OpenReview API", "abstract": "Causal structure discovery methods are commonly applied to structured data where the causal variables are known and where statistical testing can be used to assess the causal relationships. By contrast, recovering a causal structure from unstructured natural language data such as news articles contains numerous challenges due to the absence of known variables or counterfactual data to estimate the causal links. Large Language Models (LLMs) have shown promising results in this direction but also exhibit limitations. This work investigates LLM's abilities to build causal graphs from text documents and perform counterfactual causal inference. We propose an end-to-end causal structure discovery and causal inference method from natural language: we first use an LLM to extract the instantiated causal variables from text data and build a causal graph. We merge causal graphs from multiple data sources to represent the most exhaustive set of causes possible. We then conduct counterfactual inference on the estimated graph. The causal graph conditioning allows reduction of LLM biases and better represents the causal estimands. We use our method to show that the limitations in the counterfactual causal reasoning abilities come from prediction errors and propose directions to mitigate them. We demonstrate the applicability of our method on real-world news articles.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=aya06N6R4W", "arxiv_id": "2410.06392", "paper_pdf": "papers/aya06N6R4W.pdf", "paper_pdf_sha256": "6651cc456ba0319d57b39b1a2d483c3e95b465ce4f07f676ef14805cc9df62bb", "paper_pdf_bytes": 389421, "paper_pdf_source": "openreview", "code_url": "https://github.com/Strong-AI-Lab/counterfactual-llm-inference", "code_repository": "Strong-AI-Lab/counterfactual-llm-inference", "code_commit": "0b25e48ac59bdf22417b449310d481616f96d54c", "code_archive": "repos/aya06N6R4W.zip", "code_archive_sha256": "f84cebb7aad9aabd847289d7a5c3c13867227712bb3224b614fa9f7d21fb8c79", "code_archive_bytes": 42383, "code_file_count": 19, "code_extensions": {".py": 19}, "github_disk_usage_kb": 52, "github_languages": {"Python": 116979}, "github_archived": false, "github_pushed_at": "2024-10-02T08:34:10Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/counterfactual-causal-inference-in-natural"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "uJPWeZffgl", "year": 2024, "status": "rejected", "title": "Convex and Bilevel Optimization for Neuro-Symbolic Inference and Learning", "authors": ["Charles Andrew Dickens", "Changyu Gao", "Connor Pryor", "Stephen Wright", "Lise Getoor"], "authorids": ["~Charles_Andrew_Dickens1", "~Changyu_Gao1", "~Connor_Pryor1", "~Stephen_Wright1", "~Lise_Getoor1"], "authors_source": "OpenReview API", "abstract": "We address a key challenge for neuro-symbolic (NeSy) systems by leveraging convex and bilevel optimization techniques to develop a general first-order gradient-based framework for end-to-end neural and symbolic parameter learning.\nSpecifically, we formulate NeSy learning as a bilevel program, and we employ Moreau smoothing and a graduated value-function approach to support learning with a constrained lower-level inference problem.\nThe applicability of our learning framework is demonstrated with NeuPSL, a state-of-the-art NeSy architecture.\nTo achieve this, we propose a primal and dual formulation of NeuPSL inference as a strongly convex linearly constrained quadratic program and show learning gradients are functions of the optimal dual variables.\nBased on this formulation, we develop a corresponding dual block coordinate descent algorithm that naturally exploits warm-starts. \nThis leads to over $100 \\times$ learning runtime improvements over the current state-of-the-art NeuPSL inference method.\nFinally, we provide extensive empirical evaluations across $8$ datasets covering a range of prediction tasks and demonstrate our learning framework achieves up to a $16$% point prediction performance improvement over the current standard learning process.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=uJPWeZffgl", "arxiv_id": "2401.09651", "paper_pdf": "papers/uJPWeZffgl.pdf", "paper_pdf_sha256": "0f8cea504930db62a2f94cd7d27380b8c58edd68dba8d58ee4171fa90ea4790c", "paper_pdf_bytes": 651254, "paper_pdf_source": "openreview", "code_url": "https://github.com/convexbilevelnesylearning/experimentscripts", "code_repository": "convexbilevelnesylearning/experimentscripts", "code_commit": "8d01210737c69c00ef1f867ead21dc22540a6244", "code_archive": "repos/uJPWeZffgl.zip", "code_archive_sha256": "6a641e389fee9fb08e464074fb1d0e105cb3142236d9bb5b1a862ee93ab7d05a", "code_archive_bytes": 83587, "code_file_count": 29, "code_extensions": {".py": 25, ".sh": 4}, "github_disk_usage_kb": 43, "github_languages": {"Python": 302497, "Shell": 13191}, "github_archived": false, "github_pushed_at": "2023-09-28T16:29:07Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/convex-and-bilevel-optimization-for-neuro"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "SZYXyhE2c6f", "year": 2023, "status": "rejected", "title": "A Probabilistic Framework For Modular Continual Learning", "authors": ["Lazar Valkov", "Akash Srivastava", "Dipak Chaudhari", "Swarat Chaudhuri", "Charles Sutton"], "authorids": ["~Lazar_Valkov1", "~Akash_Srivastava1", "~Dipak_Chaudhari1", "~Swarat_Chaudhuri1", "~Charles_Sutton1"], "authors_source": "OpenReview API", "abstract": "Continual learning (CL) algorithms seek to accumulate and transfer knowledge across a sequence of tasks and achieve better performance on each successive task. Modular approaches, which use a different composition of modules for each task and avoid forgetting by design, have been shown to be a promising direction to CL. However, searching through the large space of possible module compositions remains a challenge. In this work, we develop a scalable probabilistic search framework as a solution to this challenge. Our framework has two distinct components. The first is designed to transfer knowledge across similar input domains. To this end, it models each module’s training input distribution and uses a Bayesian model to find the most promising module compositions for a new task. The second component targets transfer across tasks with disparate input distributions or different input spaces and uses Bayesian optimisation to explore the space of module compositions. We show that these two methods can be easily combined and evaluate the resulting approach on two benchmark suites designed to capture different desiderata of CL techniques. The experiments show that our framework offers superior performance compared to state-of-the-art CL baselines.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=SZYXyhE2c6f", "arxiv_id": "2306.06545", "paper_pdf": "papers/SZYXyhE2c6f.pdf", "paper_pdf_sha256": "2c0dbf34d6206d73e553d655ae2956328acf4aa5aa4c0120afbd97672444ea17", "paper_pdf_bytes": 1316703, "paper_pdf_source": "openreview", "code_url": "https://github.com/LazarValkov/PICLE", "code_repository": "LazarValkov/PICLE", "code_commit": "80b262825d39a8a54d8bc82bda3b89d8c0742eb2", "code_archive": "repos/SZYXyhE2c6f.zip", "code_archive_sha256": "2cd0485d172103c103bfc4cfa0f491b658b2ae3e09decf40067925e9c31ba7d1", "code_archive_bytes": 74043, "code_file_count": 44, "code_extensions": {".py": 44}, "github_disk_usage_kb": 62, "github_languages": {"Python": 253415}, "github_archived": false, "github_pushed_at": "2024-05-05T06:08:57Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/a-probabilistic-framework-for-modular"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "LLHwQh9zEb", "year": 2022, "status": "rejected", "title": "Permutation invariant graph-to-sequence model for template-free retrosynthesis and reaction prediction", "authors": ["Zhengkai Tu", "Connor W. Coley"], "authorids": ["~Zhengkai_Tu1", "~Connor_W._Coley1"], "authors_source": "OpenReview API", "abstract": "Synthesis planning and reaction outcome prediction are two fundamental problems in computer-aided organic chemistry for which a variety of data-driven approaches have emerged. Natural language approaches that model each problem as a SMILES-to-SMILES translation lead to a simple end-to-end formulation, reduce the need for data preprocessing, and enable the use of well-optimized machine translation model architectures. However, SMILES representations are not an efficient representation for capturing information about molecular structure, as evidenced by the success of SMILES augmentation to boost empirical performance. Here, we describe a novel Graph2SMILES model that combines the power of Transformer models for text generation with the permutation invariance of molecular graph encoders. As an end-to-end architecture, Graph2SMILES can be used as a drop-in replacement for the Transformer in any task involving molecule(s)-to-molecule(s) transformations. In our encoder, an attention-augmented directed message passing neural network (D-MPNN) captures local chemical environments, and the global attention encoder allows for long-range and intermolecular interactions, enhanced by graph-aware positional embedding.  Graph2SMILES improves the top-1 accuracy of the Transformer baselines by $1.7\\%$ and $1.9\\%$ for reaction outcome prediction on USPTO_480k and USPTO_STEREO datasets respectively, and by $9.8\\%$ for one-step retrosynthesis on the USPTO_50k dataset.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=LLHwQh9zEb", "arxiv_id": "2110.09681", "paper_pdf": "papers/LLHwQh9zEb.pdf", "paper_pdf_sha256": "c6823d4c31d1c81225617915ad117900ed90fa877beac11bd15595ce4093ddc1", "paper_pdf_bytes": 577501, "paper_pdf_source": "openreview", "code_url": "https://github.com/coleygroup/Graph2SMILES", "code_repository": "coleygroup/Graph2SMILES", "code_commit": "1b87f9ffdb739346b94853dfe8e5d63ac0198a38", "code_archive": "repos/LLHwQh9zEb.zip", "code_archive_sha256": "9b8495d262054c4af0f5bd697d6b2232642db0c4db32ece1f2b55eed8d34556a", "code_archive_bytes": 50516, "code_file_count": 23, "code_extensions": {".py": 18, ".sh": 5}, "github_disk_usage_kb": 67, "github_languages": {"Python": 139440, "Shell": 4780}, "github_archived": false, "github_pushed_at": "2023-11-07T03:01:34Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/permutation-invariant-graph-to-sequence-model-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "I-VfjSBzi36", "year": 2021, "status": "rejected", "title": "EarlyBERT: Efficient BERT Training via Early-bird Lottery Tickets", "authors": ["Xiaohan Chen", "Yu Cheng", "Shuohang Wang", "Zhe Gan", "Zhangyang Wang", "Jingjing Liu"], "authorids": ["~Xiaohan_Chen1", "~Yu_Cheng1", "~Shuohang_Wang1", "~Zhe_Gan1", "~Zhangyang_Wang1", "~Jingjing_Liu2"], "authors_source": "OpenReview API", "abstract": "Deep, heavily overparameterized language models such as BERT, XLNet and T5 have achieved impressive success in many NLP tasks. However, their high model complexity requires enormous computation resources and extremely long training time for both pre-training and fine-tuning. Many works have studied model compression on large NLP models, but only focus on reducing inference cost/time, while still requiring expensive training process. Other works use extremely large batch sizes to shorten the pre-training time at the expense of high demand for computation resources. In this paper, inspired by the Early-Bird Lottery Tickets studied for computer vision tasks, we propose EarlyBERT, a general computationally-efficient training algorithm applicable to both pre-training and fine-tuning of large-scale language models. We are the first to identify structured winning tickets in the early stage of BERT training, and use them for efficient training. Comprehensive pre-training and fine-tuning experiments on GLUE and SQuAD downstream tasks show that EarlyBERT easily achieves comparable performance to standard BERT with 35~45% less training time.", "decision": "Reject", "meta_review": null, "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=I-VfjSBzi36", "arxiv_id": "2101.00063", "paper_pdf": "papers/I-VfjSBzi36.pdf", "paper_pdf_sha256": "50e16d340f463235f2dedd0a5f59f01464523ad0768145cc059c3878a530865a", "paper_pdf_bytes": 2139091, "paper_pdf_source": "openreview", "code_url": "https://github.com/VITA-Group/EarlyBERT", "code_repository": "VITA-Group/EarlyBERT", "code_commit": "593624240d17fb8e1dee9b7ccf83641d603d5ce9", "code_archive": "repos/I-VfjSBzi36.zip", "code_archive_sha256": "3ba5735d02c31421d99d77fa743c0e36e7dbb80bb211da08fbf52818fe06e08b", "code_archive_bytes": 94512, "code_file_count": 13, "code_extensions": {".py": 11, ".sh": 2}, "github_disk_usage_kb": 71, "github_languages": {"Python": 379037, "Shell": 5680}, "github_archived": false, "github_pushed_at": "2021-12-30T05:58:20Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/earlybert-efficient-bert-training-via-early-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "B1lsXREYvr", "year": 2020, "status": "rejected", "title": "One-Shot Neural Architecture Search via Compressive Sensing", "authors": ["Minsu Cho", "Mohammadreza Soltani", "Chinmay Hegde"], "authorids": ["chomd90@iastate.edu", "mohammadreza.soltani@duke.edu", "chinmay@iastate.edu"], "authors_source": "OpenReview API", "abstract": "Neural architecture search (NAS), or automated design of neural network models, remains a very challenging meta-learning problem. Several recent works (called \"one-shot\" approaches) have focused on dramatically reducing NAS running time by leveraging proxy models that still provide architectures with competitive performance. In our work, we propose a new meta-learning algorithm that we call CoNAS, or Compressive sensing-based Neural Architecture Search. Our approach merges ideas from one-shot NAS approaches with iterative techniques for learning low-degree sparse Boolean polynomial functions. We validate our approach on several standard test datasets, discover novel architectures hitherto unreported, and achieve competitive (or better) results in both performance and search time compared to existing NAS approaches. Further, we provide theoretical analysis via upper bounds on the number of validation error measurements needed to perform reliable meta-learning; to our knowledge, these analysis tools are novel to the NAS literature and may be of independent interest.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=B1lsXREYvr", "arxiv_id": "1906.02869", "paper_pdf": "papers/B1lsXREYvr.pdf", "paper_pdf_sha256": "3138e5c9053cc6294421860b58a08677aba87909d765229335699d81757b80fa", "paper_pdf_bytes": 658471, "paper_pdf_source": "openreview", "code_url": "https://github.com/chomd90/CoNAS_release", "code_repository": "chomd90/CoNAS_release", "code_commit": "1df9f24f13fa5c8624f4fdcdf5e4e937851baf9d", "code_archive": "repos/B1lsXREYvr.zip", "code_archive_sha256": "a794179f5a79aee5af39cb2442fd9e707423a6e830fede32c8c0776c2f2b07de", "code_archive_bytes": 107890, "code_file_count": 13, "code_extensions": {".py": 13}, "github_disk_usage_kb": 105, "github_languages": {"Python": 64031}, "github_archived": false, "github_pushed_at": "2021-02-26T18:29:22Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/one-shot-neural-architecture-search-via"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "BJfRpoA9YX", "year": 2019, "status": "rejected", "title": "Adversarial Information Factorization", "authors": ["Antonia Creswell", "Yumnah Mohamied", "Biswa Sengupta", "Anil Bharath"], "authorids": ["ac2211@ic.ac.uk", "ym1008@ic.ac.uk", "biswasengupta@gmail.com", "aab01@ic.ac.uk"], "authors_source": "OpenReview API", "abstract": "We propose a novel generative model architecture designed to learn representations for images that factor out a single attribute from the rest of the representation. A single object may have many attributes which when altered do not change the identity of the object itself. Consider the human face; the identity of a particular person is independent of whether or not they happen to be wearing glasses. The attribute of wearing glasses can be changed without changing the identity of the person. However, the ability to manipulate and alter image attributes without altering the object identity is not a trivial task. Here, we are interested in learning a representation of the image that separates the identity of an object (such as a human face) from an attribute (such as 'wearing glasses'). We demonstrate the success of our factorization approach by using the learned representation to synthesize the same face with and without a chosen attribute. We refer to this specific synthesis process as image attribute manipulation. We further demonstrate that our model achieves competitive scores, with state of the art, on a facial attribute classification task.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=BJfRpoA9YX", "arxiv_id": "1711.05175", "paper_pdf": "papers/BJfRpoA9YX.pdf", "paper_pdf_sha256": "e43c338d64b4ac679357d74af313f4938e2a2ee4229d90b82bb679e84b0fbb3f", "paper_pdf_bytes": 2942774, "paper_pdf_source": "openreview", "code_url": "https://github.com/ToniCreswell/attribute-cVAEGAN", "code_repository": "ToniCreswell/attribute-cVAEGAN", "code_commit": "b8d543a962e68ffb4f5153470c1778a15f5508e4", "code_archive": "repos/BJfRpoA9YX.zip", "code_archive_sha256": "bb39a5c192d789606a0eeb8c4b71902f2e29cf900ccf22a50138a3fb6bfb3c76", "code_archive_bytes": 498734, "code_file_count": 7, "code_extensions": {".py": 6, ".ipynb": 1}, "github_disk_usage_kb": 684, "github_languages": {"Jupyter Notebook": 253095, "Python": 38052}, "github_archived": false, "github_pushed_at": "2018-11-27T09:20:45Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/adversarial-information-factorization"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "DBqOInhRkG", "year": 2026, "status": "rejected", "title": "RARE: Retrieval-Aware Robustness Evaluation for Retrieval-Augmented Generation Systems", "authors": ["Yixiao Zeng", "Tianyu Cao", "Danqing Wang", "Xinran Zhao", "Zimeng Qiu", "Morteza Ziyadi", "Tongshuang Wu", "Lei Li"], "authorids": ["~Yixiao_Zeng1", "~Tianyu_Cao2", "~Danqing_Wang1", "~Xinran_Zhao1", "~Zimeng_Qiu1", "~Morteza_Ziyadi2", "~Tongshuang_Wu1", "~Lei_Li11"], "authors_source": "OpenReview API", "abstract": "Retrieval-Augmented Generation (RAG) enhances recency and factuality in answers. However, existing evaluations rarely test how well these systems cope with real-world noise, conflicting between internal and external retrieved contexts, or fast-changing facts. \nWe introduce Retrieval-Aware Robustness Evaluation (RARE), a unified framework and large-scale benchmark that jointly stress-tests query and document perturbations over dynamic, time-sensitive corpora. One of the central features of RARE is a knowledge-graph-driven synthesis pipeline (RARE-Get) that automatically extracts single and multi-hop relations from the customized corpus and generates multi-level question sets without manual intervention. Leveraging this pipeline, we construct a dataset (RARE-Set) spanning 527 expert-level time-sensitive finance, economics, and policy documents and 48295 questions whose distribution evolves as the underlying sources change. To quantify resilience, we formalize retrieval-conditioned robustness metrics (RARE-Met) that capture a model’s ability to remain correct or recover when queries, documents, or real-world retrieval results are systematically altered. Our findings reveal that RAG systems are unexpectedly sensitive to perturbations. Moreover, they consistently demonstrate lower robustness on multi-hop queries compared to single-hop queries across all domains.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=DBqOInhRkG", "arxiv_id": "2506.00789", "paper_pdf": "papers/DBqOInhRkG.pdf", "paper_pdf_sha256": "ff00589f9b45328c693d4e8a46e357683706ea6b24ed698a91a8b47cb1964591", "paper_pdf_bytes": 848980, "paper_pdf_source": "openreview", "code_url": "https://github.com/LeiLiLab/RARE", "code_repository": "LeiLiLab/RARE", "code_commit": "2496f95c695eb23dcb197191e72c692f14d53781", "code_archive": "repos/DBqOInhRkG.zip", "code_archive_sha256": "a3744cdb1785c5639a609d6613173aa0e57e5c4913c3445d92ac37a70ffc6932", "code_archive_bytes": 368880, "code_file_count": 21, "code_extensions": {".py": 19, ".ipynb": 2}, "github_disk_usage_kb": 441, "github_languages": {"Jupyter Notebook": 4133839, "Python": 362344}, "github_archived": false, "github_pushed_at": "2026-05-03T06:25:28Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/rare-retrieval-aware-robustness-evaluation"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "F9JZiGradI", "year": 2025, "status": "rejected", "title": "MLP-KAN: Unifying Deep Representation and Function Learning", "authors": ["Yunhong He", "Zhengqing Yuan", "Yifeng Xie", "Lichao Sun"], "authorids": ["~Yunhong_He1", "~Zhengqing_Yuan2", "~Yifeng_Xie1", "~Lichao_Sun1"], "authors_source": "OpenReview API", "abstract": "Recent advancements in both representation learning and function learning have demonstrated substantial promise across diverse domains of artificial intelligence. However, the effective integration of these paradigms poses a significant challenge, particularly in cases where users must manually decide whether to apply a representation learning or function learning model based on dataset characteristics. To address this issue, we introduce MLP-KAN, a unified method designed to eliminate the need for manual model selection. By integrating Multi-Layer Perceptrons (MLPs) for representation learning and Kolmogorov-Arnold Networks (KANs) for function learning within a Mixture-of-Experts (MoE) architecture, MLP-KAN dynamically adapts to the specific characteristics of the task at hand, ensuring optimal performance. Embedded within a transformer-based framework, our work achieves remarkable results on four widely-used datasets across diverse domains. Extensive experimental evaluation demonstrates its superior versatility, delivering competitive performance across both deep representation and function learning tasks. These findings highlight the potential of MLP-KAN to simplify the model selection process, offering a comprehensive, adaptable solution across various domains.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=F9JZiGradI", "arxiv_id": "2410.03027", "paper_pdf": "papers/F9JZiGradI.pdf", "paper_pdf_sha256": "0e8953a3e56c1180178dd9428e3f6f1373f75800ef86343fb5f0bad7943ea92a", "paper_pdf_bytes": 798828, "paper_pdf_source": "openreview", "code_url": "https://github.com/DLYuanGod/MLP-KAN", "code_repository": "DLYuanGod/MLP-KAN", "code_commit": "827569a7c04d4ce9d41e219e27d51b285735f165", "code_archive": "repos/F9JZiGradI.zip", "code_archive_sha256": "f768a03a32f83c52b5f9cd6957b977175a58e13bccc7c40138543133522d28eb", "code_archive_bytes": 57967, "code_file_count": 23, "code_extensions": {".py": 19, ".sh": 4}, "github_disk_usage_kb": 53, "github_languages": {"Python": 230546, "Shell": 2154}, "github_archived": false, "github_pushed_at": "2024-10-08T16:34:22Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/mlp-kan-unifying-deep-representation-and"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "CXjz7p4qha", "year": 2024, "status": "rejected", "title": "Rotation Invariant Quantization for Model Compression", "authors": ["Joseph Kampeas", "Yury Nahshan", "Hanoch Kremer", "Gil Lederman", "Shira Zaloshinski", "Zheng Li", "Emir Haleva"], "authorids": ["~Joseph_Kampeas1", "~Yury_Nahshan1", "~Hanoch_Kremer1", "~Gil_Lederman1", "~Shira_Zaloshinski1", "~Zheng_Li16", "~Emir_Haleva1"], "authors_source": "OpenReview API", "abstract": "Post-training Neural Network (NN) model compression is an attractive approach for deploying large, memory-consuming models on devices with limited memory resources. \nIn this study, we investigate the rate-distortion tradeoff for NN model compression. First, we suggest a Rotation-Invariant Quantization (RIQ) technique that utilizes a single parameter to quantize the entire NN model, yielding a different rate at each layer, i.e., mixed-precision quantization. Then, we prove that our rotation-invariant approach is optimal in terms of compression. We rigorously evaluate RIQ and demonstrate its capabilities on various models and tasks. For example, RIQ facilitates $\\times 19.4$ and $\\times 52.9$ compression ratios on pre-trained VGG dense and pruned models, respectively,  with $<0.4\\%$ accuracy degradation.\nThe code is available in the supplementary material.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=CXjz7p4qha", "arxiv_id": "2303.03106", "paper_pdf": "papers/CXjz7p4qha.pdf", "paper_pdf_sha256": "cf0709aeeac86d3c7658505f8a13909a4f76f1b6e146419a3bee84006bbf8b02", "paper_pdf_bytes": 559904, "paper_pdf_source": "openreview", "code_url": "https://github.com/ehaleva/RIQ", "code_repository": "ehaleva/RIQ", "code_commit": "2b4dfd28b8e1debe562e26cd11f612da8725d20a", "code_archive": "repos/CXjz7p4qha.zip", "code_archive_sha256": "6d1bf11d6979d940749180ab893a2ffbeb911106f91047a6cd157c21f74c1dd7", "code_archive_bytes": 24376, "code_file_count": 18, "code_extensions": {".py": 16, ".ipynb": 1, ".sh": 1}, "github_disk_usage_kb": 44, "github_languages": {"Python": 41133, "Jupyter Notebook": 8153, "Shell": 4482}, "github_archived": false, "github_pushed_at": "2024-01-23T17:24:06Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/rotation-invariant-quantization-for-model"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ooqH4D9Xys", "year": 2023, "status": "rejected", "title": "LatentAugment: Dynamically Optimized Latent Probabilities of Data Augmentation", "authors": ["Koichi Kuriyama"], "authorids": ["~Koichi_Kuriyama1"], "authors_source": "OpenReview API", "abstract": "Although data augmentation is a powerful technique for improving the performance of image classification tasks, it is difficult to\nidentify the best augmentation policy. The optimal augmentation policy, which is the latent variable, cannot be directly observed. To address this problem, this study proposes \\textit{LatentAugment}, which estimates the latent probability of optimal augmentation. The proposed method is appealing in that it can dynamically optimize the augmentation strategies for each input and model parameter in\nlearning iterations. Theoretical analysis shows that LatentAugment is a general model that includes other augmentation methods as special cases, and it is simple and computationally efficient in comparison with existing augmentation methods. Experimental results show that the proposed LatentAugment has higher test accuracy than previous augmentation methods on the CIFAR-10, CIFAR-100, SVHN, and ImageNet datasets. \n", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=ooqH4D9Xys", "arxiv_id": "2305.02668", "paper_pdf": "papers/ooqH4D9Xys.pdf", "paper_pdf_sha256": "42462877839a0a83bd1447d62726516b439641fe4ea1ef64c36de774430c2e96", "paper_pdf_bytes": 810362, "paper_pdf_source": "openreview", "code_url": "https://github.com/KoichiKuriyama/LatentAugment", "code_repository": "KoichiKuriyama/LatentAugment", "code_commit": "09179d794499fac8fc9de8584bda3ad3c0139da0", "code_archive": "repos/ooqH4D9Xys.zip", "code_archive_sha256": "5234e5ece051889c35f72bd72962b5e5f92bf61e8c7f6e28b5da87a2b4a2054d", "code_archive_bytes": 46775, "code_file_count": 21, "code_extensions": {".sh": 11, ".py": 10}, "github_disk_usage_kb": 63, "github_languages": {"Python": 65130, "Shell": 4156}, "github_archived": false, "github_pushed_at": "2023-05-05T01:07:51Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/latentaugment-dynamically-optimized-latent"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "zXne1klXIQ", "year": 2022, "status": "rejected", "title": "Improving Out-of-Distribution Robustness via Selective Augmentation", "authors": ["Huaxiu Yao", "Yu Wang", "Sai Li", "Linjun Zhang", "Weixin Liang", "James Zou", "Chelsea Finn"], "authorids": ["~Huaxiu_Yao1", "~Yu_Wang24", "saili@ruc.edu.cn", "~Linjun_Zhang1", "~Weixin_Liang1", "~James_Zou1", "~Chelsea_Finn1"], "authors_source": "OpenReview API", "abstract": "Machine learning algorithms typically assume that training and test examples are drawn from the same distribution. However, distribution shifts is a common problem in real-world applications and can cause models to perform dramatically worse at test time. In this paper, we specifically consider the problems of domain shifts and subpopulation shifts, where learning invariant representations by aligning domain-specific representations or balancing the risks across domains with regularizers are popular solutions. However, designing regularizers that are suitable for diverse real-world datasets is challenging. Instead, we shed new light on addressing distribution shifts by directly eliminating domain-related spurious correlations with augmentation, leading to a simple technique based on mixup, called LISA (Learning Invariant Representations via Selective Augmentation). LISA selectively interpolates samples either with the same labels but different domains or with the same domain but different labels. Empirically, we study the effectiveness of LISA on nine benchmarks ranging from subpopulation shifts to domain shifts. The results indicate that LISA consistently outperforms other state-of-the-art methods with superior invariant representations. The empirical findings are further strengthened by our theoretical analysis.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=zXne1klXIQ", "arxiv_id": "2201.00299", "paper_pdf": "papers/zXne1klXIQ.pdf", "paper_pdf_sha256": "e795abdd0efa5aee990b496a6fb2fe9938ec5e90cc79b2bbfa3cc552c1cc462b", "paper_pdf_bytes": 484630, "paper_pdf_source": "openreview", "code_url": "https://github.com/huaxiuyao/LISA", "code_repository": "huaxiuyao/LISA", "code_commit": "c857a6b296b5d130898f0d51a6d693411c39e651", "code_archive": "repos/zXne1klXIQ.zip", "code_archive_sha256": "2a456d77d55efb4562682f9d0c124a91477c8d41260a87bc2d4b8c9adc0f745d", "code_archive_bytes": 65661, "code_file_count": 38, "code_extensions": {".py": 37, ".sh": 1}, "github_disk_usage_kb": 72, "github_languages": {"Python": 196014, "Shell": 3428}, "github_archived": false, "github_pushed_at": "2023-04-12T19:08:01Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/improving-out-of-distribution-robustness-via-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "le9LIliDOG", "year": 2021, "status": "rejected", "title": "Efficient Long-Range Convolutions for Point Clouds", "authors": ["Yifan Peng", "Lin Lin", "Lexing Ying", "Leonardo Zepeda-Nunez"], "authorids": ["pyf04142017@sjtu.edu.cn", "~Lin_Lin1", "~Lexing_Ying1", "~Leonardo_Zepeda-Nunez1"], "authors_source": "OpenReview API", "abstract": "The efficient treatment of long-range interactions for point clouds is a challenging problem in many scientific machine learning applications. To extract global information, one usually needs a large window size, a large number of layers, and/or a large number of channels. This can often significantly increase the computational cost. In this work, we present a novel neural network layer that directly incorporates long-range information for a point cloud. This layer, dubbed the long-range convolutional (LRC)-layer, leverages the convolutional theorem coupled with the non-uniform Fourier transform. In a nutshell, the LRC-layer mollifies the point cloud to an adequately sized regular grid, computes its Fourier transform, multiplies the result by a set of trainable Fourier multipliers, computes the inverse Fourier transform, and finally interpolates the result back to the point cloud. The resulting global all-to-all convolution operation can be performed in nearly-linear time asymptotically with respect to the number of input points. The LRC-layer is a particularly powerful tool when combined with local convolution as together they offer efficient and seamless treatment of both short and long range interactions. We showcase this framework by introducing a neural network architecture that combines LRC-layers with short-range convolutional layers to accurately learn the energy and force associated with a $N$-body potential.  We also exploit the induced two-level decomposition and propose an efficient strategy to train the combined architecture with a reduced number of samples.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=le9LIliDOG", "arxiv_id": "2010.05295", "paper_pdf": "papers/le9LIliDOG.pdf", "paper_pdf_sha256": "5e8fbe55c69c7e2b9b31010936d602d3cb02ed5aa541d14ca0a730ef5dd6d702", "paper_pdf_bytes": 1514152, "paper_pdf_source": "openreview", "code_url": "https://github.com/Forgotten/Efficient_Long-Range_Convolutions_for_Point_Clouds", "code_repository": "Forgotten/Efficient_Long-Range_Convolutions_for_Point_Clouds", "code_commit": "1fe364052eca9330edeaeb32c59d0ec5195c12c4", "code_archive": "repos/le9LIliDOG.zip", "code_archive_sha256": "3134b04565d12849475e67c4b80bcfb0fd3a0d3f9ef0d61ca57acb91291e7ac5", "code_archive_bytes": 45981, "code_file_count": 16, "code_extensions": {".py": 16}, "github_disk_usage_kb": 73, "github_languages": {"Python": 49242}, "github_archived": false, "github_pushed_at": "2020-10-12T15:16:29Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/efficient-long-range-convolutions-for-point-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "r1e8WTEYPB", "year": 2020, "status": "rejected", "title": "Sparse and Structured Visual Attention", "authors": ["Pedro Henrique Martins", "Vlad Niculae", "Zita Marinho", "André F.T. Martins"], "authorids": ["pedrohenriqueamartins@gmail.com", "vlad@vene.ro", "zita.marinho@priberam.pt", "andre.martins@unbabel.com"], "authors_source": "OpenReview API", "abstract": "Visual attention mechanisms have been widely used in image captioning models. In this paper, to better link the image structure with the generated text, we replace the traditional softmax attention mechanism by two alternative sparsity-promoting transformations: sparsemax and Total-Variation Sparse Attention (TVmax). With sparsemax, we obtain sparse attention weights, selecting relevant features.  In order to promote sparsity and encourage fusing of the related adjacent spatial locations, we propose TVmax.  By selecting relevant groups of features, the TVmax transformation improves interpretability. We present results in the Microsoft COCO and Flickr30k datasets, obtaining gains in comparison to softmax.  TVmax outperforms the other compared attention mechanisms in terms of human-rated caption quality and attention relevance.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=r1e8WTEYPB", "arxiv_id": "2002.05556", "paper_pdf": "papers/r1e8WTEYPB.pdf", "paper_pdf_sha256": "cfa8a50ed83609a2c05f43e6b7f0e9873bd3c591c15efa99edf78b2af89b535d", "paper_pdf_bytes": 1566959, "paper_pdf_source": "openreview", "code_url": "https://github.com/deep-spin/TVmax", "code_repository": "deep-spin/TVmax", "code_commit": "2b812dc1b78fe8869bbd2f11dca89aba5341bcc5", "code_archive": "repos/r1e8WTEYPB.zip", "code_archive_sha256": "fd95b4c318f14b6769d38fb43033436ebdc56d223eb10d73e16d61974efd2e10", "code_archive_bytes": 115789, "code_file_count": 16, "code_extensions": {".py": 16}, "github_disk_usage_kb": 113, "github_languages": {"Python": 90932}, "github_archived": false, "github_pushed_at": "2021-12-21T23:41:02Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/sparse-and-structured-visual-attention-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "B1lfHhR9tm", "year": 2019, "status": "rejected", "title": "The Natural Language Decathlon: Multitask Learning as Question Answering", "authors": ["Bryan McCann", "Nitish Shirish Keskar", "Caiming Xiong", "Richard Socher"], "authorids": ["bmccann@salesforce.com", "nkeskar@salesforce.com", "cxiong@salesforce.com", "rsocher@salesforce.com"], "authors_source": "OpenReview API", "abstract": "Deep learning has improved performance on many natural language processing (NLP) tasks individually.\nHowever, general NLP models cannot emerge within a paradigm that focuses on the particularities of a single metric, dataset, and task.\nWe introduce the Natural Language Decathlon (decaNLP), a challenge that spans ten tasks:\nquestion answering, machine translation, summarization, natural language inference, sentiment analysis, semantic role labeling, relation extraction, goal-oriented dialogue, semantic parsing, and commonsense pronoun resolution.\nWe cast all tasks as question answering over a context.\nFurthermore, we present a new multitask question answering network (MQAN) that jointly learns all tasks in decaNLP without any task-specific modules or parameters more effectively than sequence-to-sequence and reading comprehension baselines.\nMQAN shows improvements in transfer learning for machine translation and named entity recognition, domain adaptation for sentiment analysis and natural language inference, and zero-shot capabilities for text classification.\nWe demonstrate that the MQAN's multi-pointer-generator decoder is key to this success and that performance further improves with an anti-curriculum training strategy.\nThough designed for decaNLP, MQAN also achieves state of the art results on the WikiSQL semantic parsing task in the single-task setting. \nWe also release code for procuring and processing data, training and evaluating models, and reproducing all experiments for decaNLP.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=B1lfHhR9tm", "arxiv_id": "1806.08730", "paper_pdf": "papers/B1lfHhR9tm.pdf", "paper_pdf_sha256": "16f82a5a5549b2c93afc4664e34d376ae7cf874f7e74f326cc62b7d5c703b708", "paper_pdf_bytes": 1268248, "paper_pdf_source": "openreview", "code_url": "https://github.com/salesforce/decaNLP", "code_repository": "salesforce/decaNLP", "code_commit": "f1d474b0ff7c7c45a325177401d46b8d0dd16b38", "code_archive": "repos/B1lfHhR9tm.zip", "code_archive_sha256": "c3bad1fc03bc37bbe341c4f6c3b5998c60b9936470c8794526eb709b844c504e", "code_archive_bytes": 738498, "code_file_count": 59, "code_extensions": {".py": 56, ".sh": 3}, "github_disk_usage_kb": 849, "github_languages": {"Python": 381337, "Shell": 2794, "CSS": 2303, "Makefile": 1036, "Batchfile": 817}, "github_archived": true, "github_pushed_at": "2025-05-01T17:29:03Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/the-natural-language-decathlon-multitask"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "dw2vxWVrA9", "year": 2026, "status": "rejected", "title": "Learning to Dissipate Energy in Oscillatory State-Space Models", "authors": ["Jared Boyer", "T. Konstantin Rusch", "Daniela Rus"], "authorids": ["~Jared_Boyer1", "~T._Konstantin_Rusch1", "~Daniela_Rus1"], "authors_source": "OpenReview API", "abstract": "State-space models (SSMs) are a class of networks for sequence learning that benefit from fixed state size and linear complexity with respect to sequence length, contrasting the quadratic scaling of typical attention mechanisms.  Inspired from observations in neuroscience, Linear Oscillatory State-Space models (LinOSS) are a recently proposed class of SSMs constructed from layers of discretized forced harmonic oscillators.  Although these models perform competitively, leveraging fast parallel scans over diagonal recurrent matrices and achieving state-of-the-art performance on tasks with sequence length up to 50k, LinOSS models rely on rigid energy dissipation (``forgetting'') mechanisms that are inherently coupled to the time scale of state evolution.  As forgetting is a crucial mechanism for long-range reasoning, we demonstrate the representational limitations of these models and introduce Damped Linear Oscillatory State-Space models (D-LinOSS), a more general class of oscillatory SSMs that learn to dissipate latent state energy on arbitrary time scales.  We analyze the spectral distribution of the model's recurrent matrices and prove that the SSM layers exhibit stable dynamics under a simple, flexible parameterization. Without additional complexity, D-LinOSS consistently outperforms previous LinOSS methods on long-range learning tasks, achieves faster convergence, and relinquishes the need for multiple discretization schemes.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=dw2vxWVrA9", "arxiv_id": "2505.12171", "paper_pdf": "papers/dw2vxWVrA9.pdf", "paper_pdf_sha256": "b96d986d489a6dcacc251e02dede1519b338c4b5e3e1c0e3ed878bc364d26ff3", "paper_pdf_bytes": 5603773, "paper_pdf_source": "openreview", "code_url": "https://github.com/jaredbmit/damped-linoss", "code_repository": "jaredbmit/damped-linoss", "code_commit": "450b546f693918fe7cfe44082e88538fb29fbd64", "code_archive": "repos/dw2vxWVrA9.zip", "code_archive_sha256": "e5c7b629bccac05837c9e605269110141ae866b62ea8e491f110a2a2de1bbcc1", "code_archive_bytes": 136532, "code_file_count": 18, "code_extensions": {".py": 18}, "github_disk_usage_kb": 453, "github_languages": {"Python": 156887}, "github_archived": false, "github_pushed_at": "2026-02-13T19:58:21Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/learning-to-dissipate-energy-in-oscillatory"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ZtvRqm6oBu", "year": 2025, "status": "rejected", "title": "Applying Sparse Autoencoders to Unlearn Knowledge in Language Models", "authors": ["Eoin Farrell", "Yeu-Tong Lau", "Arthur Conmy"], "authorids": ["~Eoin_Farrell1", "~Yeu-Tong_Lau1", "~Arthur_Conmy1"], "authors_source": "OpenReview API", "abstract": "We investigate whether sparse autoencoders (SAEs) can be used to remove knowledge from language models. We use the biology subset of the Weapons of Mass Destruction Proxy dataset and test on the gemma-2b-it and gemma-2-2b-it language models. We demonstrate that individual interpretable biology-related SAE features can be used to unlearn a subset of WMDP-Bio questions with minimal side-effects in domains other than biology. Our results suggest that negative scaling of feature activations is necessary and that zero ablating features is ineffective. We find that intervening using multiple SAE features simultaneously can unlearn multiple different topics, but with similar or larger unwanted side-effects than the existing Representation Misdirection for Unlearning technique. Current SAE quality or intervention techniques would need to improve to make SAE-based unlearning comparable to the existing fine-tuning based techniques.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=ZtvRqm6oBu", "arxiv_id": "2410.19278", "paper_pdf": "papers/ZtvRqm6oBu.pdf", "paper_pdf_sha256": "f057083a13568fd4348195f56362be03be40e9af1e96bd669c22e14723800cf8", "paper_pdf_bytes": 2540107, "paper_pdf_source": "openreview", "code_url": "https://github.com/efarrell1/train_sparse_autoencoder", "code_repository": "efarrell1/train_sparse_autoencoder", "code_commit": "60c756d9bd2eb796736ab6c2d74ffc0acc65b717", "code_archive": "repos/ZtvRqm6oBu.zip", "code_archive_sha256": "ab947d64e46984af824033f39e5128c09850beb0ae9ad0a225a5e4c75415d00c", "code_archive_bytes": 65912, "code_file_count": 18, "code_extensions": {".ipynb": 11, ".py": 7}, "github_disk_usage_kb": 54, "github_languages": {"Jupyter Notebook": 153502, "Python": 131912}, "github_archived": false, "github_pushed_at": "2024-10-07T02:24:43Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/applying-sparse-autoencoders-to-unlearn"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ICuUgRLp4C", "year": 2024, "status": "rejected", "title": "Learning High-Order Relationships of Brain Regions", "authors": ["Weikang Qiu", "Huangrui Chu", "Selena Wang", "Haolan Zuo", "Xiaoxiao Li", "Yize Zhao", "Zhitao Ying"], "authorids": ["~Weikang_Qiu1", "~Huangrui_Chu1", "~Selena_Wang1", "~Haolan_Zuo1", "~Xiaoxiao_Li1", "~Yize_Zhao1", "~Zhitao_Ying1"], "authors_source": "OpenReview API", "abstract": "Discovering reliable and informative interactions among brain regions from functional magnetic resonance imaging (fMRI) signals is essential in neuroscientific predictions of cognition. Most of the current methods fail to accurately characterize those interactions because they only focus on pairwise connections and overlook the high-order relationships of brain regions. We delve into this problem and argue that these high-order relationships should be maximally informative and minimally redundant (MIMR). However, identifying such high-order relationships is challenging and highly under-explored. Methods that can be tailored to our context are also non-existent. In response to this gap, we propose a novel method named HyBRiD that aims to extract MIMR high-order relationships from fMRI data. HyBRiD employs a Constructor to identify hyperedge structures, and a Weighter to compute a weight for each hyperedge. HyBRiD achieves the MIMR objective through an innovative information bottleneck framework named multi-head drop-bottleneck with theoretical guarantees. Our comprehensive experiments demonstrate the effectiveness of our model. In terms of the quality of hyperedges measured by the CPM metric, our model outperforms the state-of-the-art predictive model by an average of 12.1%.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=ICuUgRLp4C", "arxiv_id": "2312.02203", "paper_pdf": "papers/ICuUgRLp4C.pdf", "paper_pdf_sha256": "55b8b34bd73c861cc819ee501e0a319a97770673ffa4a6b569d06de8a44f2bdf", "paper_pdf_bytes": 2038438, "paper_pdf_source": "openreview", "code_url": "https://github.com/Graph-and-Geometric-Learning/HyBRiD", "code_repository": "Graph-and-Geometric-Learning/HyBRiD", "code_commit": "973d937a38449409264218e380ec6593ee30568c", "code_archive": "repos/ICuUgRLp4C.zip", "code_archive_sha256": "6a588170a4f396e7d6efee6b063e9204381193925cdd11c05646a878545d0f74", "code_archive_bytes": 27411, "code_file_count": 7, "code_extensions": {".py": 7}, "github_disk_usage_kb": 44, "github_languages": {"Python": 35629}, "github_archived": false, "github_pushed_at": "2025-05-29T02:39:58Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/learning-high-order-relationships-of-brain"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "6dtI3qPVwVp", "year": 2023, "status": "rejected", "title": "Link Prediction without Graph Neural Networks", "authors": ["Zexi Huang", "Mert Kosan", "Arlei Lopes da Silva", "Ambuj Singh"], "authorids": ["~Zexi_Huang1", "~Mert_Kosan1", "~Arlei_Lopes_da_Silva1", "~Ambuj_Singh1"], "authors_source": "OpenReview API", "abstract": "Link prediction, which consists of predicting edges based on graph features, is a fundamental task in many graph applications. As for several related problems, Graph Neural Networks (GNNs), which are based on an attribute-centric message-passing paradigm, have become the predominant framework for link prediction. GNNs have consistently outperformed traditional topology-based heuristics, but what contributes to their performance? Are there simpler approaches that achieve comparable or better results? To answer these questions, we first identify important limitations in how GNN-based link prediction methods handle the intrinsic class imbalance of the problem---due to the graph sparsity---in their training and evaluation. Moreover, we propose Gelato, a novel topology-centric framework that applies a topological heuristic to a graph enhanced by attribute information via graph learning. Our model is trained end-to-end with an N-pair loss on an unbiased training set to address class imbalance. Experiments show that Gelato is 145% more accurate, trains 11 times faster, infers 6,000 times faster, and has less than half of the trainable parameters compared to state-of-the-art GNNs for link prediction.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=6dtI3qPVwVp", "arxiv_id": "2305.13656", "paper_pdf": "papers/6dtI3qPVwVp.pdf", "paper_pdf_sha256": "7640862273a54a6ff4d9a4ca398f6944d02e81d9a177af508a80286c6bdc5b3a", "paper_pdf_bytes": 14523542, "paper_pdf_source": "openreview", "code_url": "https://github.com/facebookresearch/SEAL_OGB", "code_repository": "facebookresearch/SEAL_OGB", "code_commit": "ea01ef509df1a90bc246e4e9828e192eeec4289c", "code_archive": "repos/6dtI3qPVwVp.zip", "code_archive_sha256": "c46325538a3d8b48d413996fae2f332f952c816a209f4cf30c04f20c3ea47e67", "code_archive_bytes": 18330, "code_file_count": 3, "code_extensions": {".py": 3}, "github_disk_usage_kb": 65, "github_languages": {"Python": 55932}, "github_archived": true, "github_pushed_at": "2023-06-24T04:27:09Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/link-prediction-without-graph-neural-networks"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Kvbr8NicKq", "year": 2022, "status": "rejected", "title": "Fast and Reliable Evaluation of Adversarial Robustness with Minimum-Margin Attack", "authors": ["Ruize Gao", "Jiongxiao Wang", "Kaiwen Zhou", "Feng Liu", "Binghui Xie", "Gang Niu", "Bo Han", "James Cheng"], "authorids": ["~Ruize_Gao1", "~Jiongxiao_Wang1", "~Kaiwen_Zhou2", "~Feng_Liu2", "~Binghui_Xie1", "~Gang_Niu1", "~Bo_Han1", "~James_Cheng2"], "authors_source": "OpenReview API", "abstract": "The AutoAttack (AA) has been the most reliable method to evaluate adversarial robustness when considerable computational resources are available. However, the high computational cost (e.g., 100 times more than that of the project gradient descent attack) makes AA infeasible for practitioners with limited computational resources, and also hinders applications of AA in the adversarial training (AT). In this paper, we propose a novel method, minimum-margin (MM) attack, to fast and reliably evaluate adversarial robustness. Compared with AA, our method achieves comparable performance but only costs 3% of the computational time in extensive experiments. The reliability of our method lies in that we evaluate the quality of adversarial examples using the margin between two targets that can precisely identify the most adversarial example. The computational efficiency of our method lies in an effective Sequential TArget Ranking Selection (STARS) method, ensuring that the cost of the MM attack is independent of the number of classes. The MM attack opens a new way for evaluating adversarial robustness and contributes a feasible and reliable method to generate high-quality adversarial examples in AT.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=Kvbr8NicKq", "arxiv_id": "2206.07314", "paper_pdf": "papers/Kvbr8NicKq.pdf", "paper_pdf_sha256": "ea19558a51fceac6a52522dff9be361328184b52864424f76f0a8867be17caf3", "paper_pdf_bytes": 685991, "paper_pdf_source": "openreview", "code_url": "https://github.com/Sjtubrian/MM-attack", "code_repository": "Sjtubrian/MM-attack", "code_commit": "0a9883eb3cf06b8800540b841f70424302f24012", "code_archive": "repos/Kvbr8NicKq.zip", "code_archive_sha256": "0429fc55ec909b316e9bd53c66c27ffbfa618a2c005f741d6b17a51cbcc561dc", "code_archive_bytes": 55767, "code_file_count": 27, "code_extensions": {".py": 27}, "github_disk_usage_kb": 76, "github_languages": {"Python": 215771}, "github_archived": false, "github_pushed_at": "2022-10-20T05:03:48Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/fast-and-reliable-evaluation-of-adversarial-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "KJSC_AsN14", "year": 2021, "status": "rejected", "title": "Contrastive Learning with Stronger Augmentations", "authors": ["Xiao Wang", "Guo-Jun Qi"], "authorids": ["~Xiao_Wang6", "~Guo-Jun_Qi1"], "authors_source": "OpenReview API", "abstract": "Representation learning has been greatly improved with the advance of contrastive learning methods with the performance being closer to their supervised learning counterparts. Those methods have greatly benefited from various data augmentations that are carefully designated to maintain their identities so that the images transformed from the same instance can still be retrieved. Although stronger augmentations could expose novel patterns of representations to improve their generalizability, directly using stronger augmentations in instance discrimination-based contrastive learning may even deteriorate the performance, because the distortions induced from the stronger augmentations could ridiculously change the image structures and thus the transformed images cannot be viewed as the same as the original ones any more. Additional efforts are needed for us to explore the role of the stronger augmentations in further pushing the performance of unsupervised learning to the fully supervised upper bound. Instead of applying the stronger augmentations directly to minimize the contrastive loss, we propose to minimize the distribution divergence between the weakly and strongly augmented images over the representation bank to supervise the retrieval of strongly augmented queries from a pool of candidates. This avoids an overoptimistic assumption that could overfit the strongly augmented queries containing distorted visual structures into the positive targets in the representation bank, while still being able to distinguish them from the negative samples by leveraging the distributions of weakly augmented counterparts. The proposed method achieves top-1 accuracy of 76.2% on ImageNet with a standard ResNet-50 architecture with a single-layer classifier fine-tuned. This is almost the same as 76.5% of top-1 accuracy with a fully supervised ResNet-50. Moreover, it outperforms the previous self-supervised and supervised methods on both the transfer learning and object detection tasks.\n", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=KJSC_AsN14", "arxiv_id": "2104.07713", "paper_pdf": "papers/KJSC_AsN14.pdf", "paper_pdf_sha256": "fc0db3bc2416967f913beeaaed593dcbf48da5a26af2643d5727fe740381eb71", "paper_pdf_bytes": 1749257, "paper_pdf_source": "openreview", "code_url": "https://github.com/maple-research-lab/CLSA", "code_repository": "maple-research-lab/CLSA", "code_commit": "37df76cf5cb032683e57b70a3a4090f0d524c8fd", "code_archive": "repos/KJSC_AsN14.zip", "code_archive_sha256": "d6404572a58a7b1a5074853099730c7b61d2033524096619a3d70133ac1dc34c", "code_archive_bytes": 45948, "code_file_count": 21, "code_extensions": {".py": 19, ".sh": 2}, "github_disk_usage_kb": 80, "github_languages": {"Python": 117550, "Shell": 928}, "github_archived": false, "github_pushed_at": "2021-08-02T13:24:09Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/contrastive-learning-with-stronger-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ryga2CNKDH", "year": 2020, "status": "rejected", "title": "Evaluating Lossy Compression Rates of Deep Generative Models", "authors": ["Sicong Huang", "Alireza Makhzani", "Yanshuai Cao", "Roger Grosse"], "authorids": ["huang@cs.toronto.edu", "a.makhzani@gmail.com", "yanshuai.cao@borealisai.com", "rgrosse@cs.toronto.edu"], "authors_source": "OpenReview API", "abstract": "Deep generative models have achieved remarkable progress in recent years. Despite this progress, quantitative evaluation and comparison of generative models remains as one of the important challenges. One of the most popular metrics for evaluating generative models is the log-likelihood. While the direct computation of log-likelihood can be intractable, it has been recently shown that the log-likelihood of some of the most interesting generative models such as variational autoencoders (VAE) or generative adversarial networks (GAN) can be efficiently estimated using annealed importance sampling (AIS). In this work, we argue that the log-likelihood metric by itself cannot represent all the different performance characteristics of generative models, and propose to use rate distortion curves to evaluate and compare deep generative models. We show that we can approximate the entire rate distortion curve using one single run of AIS for roughly the same computational cost as a single log-likelihood estimate. We evaluate lossy compression rates of different deep generative models such as VAEs, GANs (and its variants) and adversarial autoencoders (AAE) on MNIST and CIFAR10, and arrive at a number of insights not obtainable from log-likelihoods alone.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=ryga2CNKDH", "arxiv_id": "2008.06653", "paper_pdf": "papers/ryga2CNKDH.pdf", "paper_pdf_sha256": "4353dceb005eae52a14bd303432db1cbce657e8a2b93c598d3185dc925f9bd3b", "paper_pdf_bytes": 1447534, "paper_pdf_source": "openreview", "code_url": "https://github.com/BorealisAI/rate_distortion", "code_repository": "BorealisAI/rate_distortion", "code_commit": "13b60b94c65199bd228292f301151a3fd54dcd89", "code_archive": "repos/ryga2CNKDH.zip", "code_archive_sha256": "3dd1201d2cf753888f2d71c421fc7cd39a12fbb4baba26820f9b7279606678d7", "code_archive_bytes": 95209, "code_file_count": 57, "code_extensions": {".py": 49, ".sh": 8}, "github_disk_usage_kb": 122, "github_languages": {"Python": 275869, "Shell": 16034}, "github_archived": false, "github_pushed_at": "2021-01-20T07:31:25Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/evaluating-lossy-compression-rates-of-deep"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "0LwWImlpCP", "year": 2026, "status": "rejected", "title": "Electrostatics from Laplacian Eigenbasis for Neural Network Interatomic Potentials", "authors": ["Maksim Zhdanov", "Vladislav Kurenkov"], "authorids": ["~Maksim_Zhdanov2", "~Vladislav_Kurenkov1"], "authors_source": "OpenReview API", "abstract": "In this work, we introduce $\\Phi$-Module, a universal plugin module that enforces Poisson’s equation within the message-passing framework to learn electrostatic interactions in a self-supervised manner. Specifically, each atom-wise representation is encouraged to satisfy a discretized Poisson's equation, making it possible to acquire a potential $\\boldsymbol{\\phi}$ and a corresponding charges $\\boldsymbol{\\rho}$ linked to the learnable Laplacian eigenbasis coefficients of a given molecular graph. We then derive an electrostatic energy term, crucial for improved total energy predictions. This approach integrates seamlessly into any existing neural potential with insignificant computational overhead. Our results underscore how embedding a first-principles constraint in neural interatomic potentials can significantly improve performance while remaining hyperparameter-friendly, memory-efficient and lightweight in training.", "decision": "Reject", "meta_review": "", "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=0LwWImlpCP", "arxiv_id": "2505.14606", "paper_pdf": "papers/0LwWImlpCP.pdf", "paper_pdf_sha256": "3e2df0f85baedac9ddc8c127d36eebcb1fde559e7d5c386ee4499806d100dfa5", "paper_pdf_bytes": 818079, "paper_pdf_source": "openreview", "code_url": "https://github.com/dunnolab/phi-module", "code_repository": "dunnolab/phi-module", "code_commit": "730c4ddb5be07d6f67a5a42db77f923e2de5fac0", "code_archive": "repos/0LwWImlpCP.zip", "code_archive_sha256": "47edc2cf26fa67e6c5a9e4f0d1470a53b56bba6c4d904a289c09cecc6a313bfb", "code_archive_bytes": 505301, "code_file_count": 66, "code_extensions": {".py": 66}, "github_disk_usage_kb": 473, "github_languages": {"Python": 546165, "Dockerfile": 831}, "github_archived": false, "github_pushed_at": "2025-06-12T23:37:13Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/electrostatics-from-laplacian-eigenbasis-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "3Xfa63ggsq", "year": 2025, "status": "rejected", "title": "AlignIQL: Policy Alignment in Implicit Q-Learning through Constrained Optimization", "authors": ["Longxiang He", "Li Shen", "Junbo Tan", "Xueqian Wang"], "authorids": ["~Longxiang_He2", "~Li_Shen1", "~Junbo_Tan1", "~Xueqian_Wang1"], "authors_source": "OpenReview API", "abstract": "Implicit Q-learning (IQL) serves as a strong baseline for offline RL, which never needs to evaluate actions outside of the dataset through quantile regression. However, it is unclear how to recover the implicit policy from the learned implicit Q-function and whether IQL can utilize weighted regression for policy extraction. IDQL reinterprets IQL as an actor-critic method and gets weights of implicit policy, however, this weight only holds for the optimal value function under certain critic loss functions. In this work, we introduce a different way to solve the $\\textit{implicit policy-finding problem}$ (IPF) by formulating this problem as an optimization problem. Based on this optimization problem, we further propose two practical algorithms AlignIQL and AlignIQL-hard, which inherit the advantages of decoupling actor from critic in IQL and provide insights into why IQL can use weighted regression for policy extraction. Compared with IQL and IDQL, we find that our method keeps the simplicity of IQL and solves the implicit policy-finding problem.  Experimental results on D4RL datasets show that our method achieves competitive or superior results compared with other SOTA offline RL methods. Especially in complex sparse reward tasks like AntMaze and Adroit, our method outperforms IQL and IDQL by a significant margin.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=3Xfa63ggsq", "arxiv_id": "2405.18187", "paper_pdf": "papers/3Xfa63ggsq.pdf", "paper_pdf_sha256": "990b2578d66c04189767d06767dd79e96efccb2253e5990c79f96c52c61846eb", "paper_pdf_bytes": 571327, "paper_pdf_source": "openreview", "code_url": "https://github.com/felix-thu/AlignIQL", "code_repository": "felix-thu/AlignIQL", "code_commit": "857f90b5fa355bb979f3ef6d997cbb20fc4b5ded", "code_archive": "repos/3Xfa63ggsq.zip", "code_archive_sha256": "e5081f67d032ef9b80ee2412f5109872c0406307f106f23e81d466b0c7bfe71e", "code_archive_bytes": 83667, "code_file_count": 71, "code_extensions": {".py": 71}, "github_disk_usage_kb": 54, "github_languages": {"Python": 168878}, "github_archived": false, "github_pushed_at": "2024-05-29T05:28:42Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/aligniql-policy-alignment-in-implicit-q"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "sVl1KO5K76", "year": 2024, "status": "rejected", "title": "Momentum-SAM: Sharpness Aware Minimization without Computational Overhead", "authors": ["Marlon Becker", "Frederick Altrock", "Benjamin Risse"], "authorids": ["~Marlon_Becker1", "~Frederick_Altrock1", "~Benjamin_Risse1"], "authors_source": "OpenReview API", "abstract": "The recently proposed optimization algorithm for deep neural networks Sharpness Aware Minimization (SAM) suggests perturbing parameters before gradient calculation by a gradient ascent step to guide the optimization into parameter space regions of flat loss.\nWhile significant generalization improvements and thus reduction of overfitting could be demonstrated, the computational costs are doubled due to the additionally needed gradient calculation, making SAM unfeasible in case of limited computationally capacities.\nMotivated by Nesterov Accelerated Gradient (NAG) we propose Momentum-SAM (MSAM), which perturbs parameters in the direction of the accumulated momentum vector to achieve low sharpness without significant computational overhead or memory demands over SGD or Adam.\nWe evaluate MSAM in detail and reveal insights on separable mechanisms of NAG, SAM and MSAM regarding training optimization and generalization.\nCode is available at https://XXXXXXXX.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=sVl1KO5K76", "arxiv_id": "2401.12033", "paper_pdf": "papers/sVl1KO5K76.pdf", "paper_pdf_sha256": "760e595b58f33cad182cb99a9a98e1314b7592c41c258cd6ef7ad04b3ab52c33", "paper_pdf_bytes": 1820284, "paper_pdf_source": "openreview", "code_url": "https://github.com/MarlonBecker/MSAM", "code_repository": "MarlonBecker/MSAM", "code_commit": "780fa4f623fe0c99760fd7cfb1a94d14f00179ee", "code_archive": "repos/sVl1KO5K76.zip", "code_archive_sha256": "478747d058bad76ab7ad4e6ab9d2ad05aa9dddea508ac45d837c02ee309f5231", "code_archive_bytes": 57887, "code_file_count": 27, "code_extensions": {".py": 27}, "github_disk_usage_kb": 44, "github_languages": {"Python": 123241}, "github_archived": false, "github_pushed_at": "2024-01-23T15:00:30Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/momentum-sam-sharpness-aware-minimization"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "VBB4fh45HF", "year": 2023, "status": "rejected", "title": "Learning Interpretable Dynamics from Images of a Freely Rotating 3D Rigid Body", "authors": ["Justice Mason", "Christine Allen-Blanchette", "Nicholas F Zolman", "Elizabeth Davison", "Naomi Leonard"], "authorids": ["~Justice_Mason1", "~Christine_Allen-Blanchette1", "~Nicholas_F_Zolman1", "~Elizabeth_Davison1", "~Naomi_Leonard1"], "authors_source": "OpenReview API", "abstract": "In many real-world settings, image observations of freely rotating 3D rigid bodies, such as satellites, may be available when low-dimensional measurements are not. However, the high-dimensionality of image data precludes the use of classical estimation techniques to learn the dynamics and a lack of interpretability reduces the usefulness of standard deep learning methods. In this work, we present a physics-informed neural network model to estimate and predict 3D rotational dynamics from image sequences. We achieve this using a multi-stage prediction pipeline that maps individual images to a latent representation homeomorphic to $\\mathbf{SO}(3)$, computes angular velocities from latent pairs, and predicts future latent states using the Hamiltonian equations of motion with a learned representation of the Hamiltonian. We demonstrate the efficacy of our approach on a new rotating rigid-body dataset with sequences of rotating cubes and rectangular prisms with uniform and non-uniform density.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=VBB4fh45HF", "arxiv_id": "2209.11355", "paper_pdf": "papers/VBB4fh45HF.pdf", "paper_pdf_sha256": "b8304673157153ef2d7945831ab4964519f371c2b6d6c55fa9eec6926b513386", "paper_pdf_bytes": 11568674, "paper_pdf_source": "openreview", "code_url": "https://github.com/jjmason687/LearningSO3fromImages", "code_repository": "jjmason687/LearningSO3fromImages", "code_commit": "357128e2eccd0dd1c1788152380e78b2ec4b4087", "code_archive": "repos/VBB4fh45HF.zip", "code_archive_sha256": "43f37f804d388398194758b13a4c514525c02a179841b67d0bcceefdac25d6de", "code_archive_bytes": 65302, "code_file_count": 18, "code_extensions": {".py": 18}, "github_disk_usage_kb": 71, "github_languages": {"Python": 245876}, "github_archived": false, "github_pushed_at": "2022-10-15T01:31:10Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/learning-interpretable-dynamics-from-images"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "TKrlyiqKWB", "year": 2022, "status": "rejected", "title": "Prototype Based Classification from Hierarchy to Fairness", "authors": ["Mycal Tucker", "Julie Shah"], "authorids": ["~Mycal_Tucker1", "~Julie_Shah1"], "authors_source": "OpenReview API", "abstract": "Artificial neural nets can represent and classify many types of high-dimensional data but are often tailored to particular applications -- e.g., for ``fair'' or ``hierarchical'' classification. Once an architecture has been selected, it is often difficult for humans to adjust models for a new task; for example, a hierarchical classifier cannot be easily transformed into a fair classifier that shields a protected field. Our contribution in this work is a new neural network architecture, the concept subspace network (CSN), which generalizes existing specialized classifiers to produce a unified model capable of learning a spectrum of multi-concept relationships. We demonstrate that CSNs reproduce state-of-the-art results in fair classification when enforcing concept independence, may be transformed into hierarchical classifiers, or may even reconcile fairness and hierarchy within a single classifier. The CSN is inspired by and matches the performance of existing prototype-based classifiers that promote interpretability.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=TKrlyiqKWB", "arxiv_id": "2205.13997", "paper_pdf": "papers/TKrlyiqKWB.pdf", "paper_pdf_sha256": "901ff0aee3df2d2f350f3fbb77e2f901a9c809681ee2915534ced49e1f0a7f52", "paper_pdf_bytes": 963722, "paper_pdf_source": "openreview", "code_url": "https://github.com/mycal-tucker/csn", "code_repository": "mycal-tucker/csn", "code_commit": "0628a7c76d0d1a1f07b96d02d050780f2bd2cf06", "code_archive": "repos/TKrlyiqKWB.zip", "code_archive_sha256": "972660a1f3e6b1e752907f294fad9d16d93878a0499dfc6231f0d531c5413bc2", "code_archive_bytes": 91987, "code_file_count": 38, "code_extensions": {".py": 38}, "github_disk_usage_kb": 78, "github_languages": {"Python": 250389}, "github_archived": false, "github_pushed_at": "2022-12-05T19:06:35Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/prototype-based-classification-from-hierarchy-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "PdauS7wZBfC", "year": 2021, "status": "rejected", "title": "Predictive Coding Approximates Backprop along Arbitrary Computation Graphs", "authors": ["Beren Millidge", "Alexander Tschantz", "Christopher Buckley"], "authorids": ["~Beren_Millidge1", "~Alexander_Tschantz1", "~Christopher_Buckley1"], "authors_source": "OpenReview API", "abstract": "The backpropagation of error (backprop) is a powerful algorithm for training machine learning architectures through end-to-end differentiation. Recently it has been shown that backprop in multilayer-perceptrons (MLPs) can be approximated using predictive coding, a biologically-plausible process theory of cortical computation which relies solely on local and Hebbian updates. The power of backprop, however, lies not in its instantiation in MLPs, but rather in the concept of automatic differentiation which allows for the optimisation of any differentiable program expressed as a computation graph. Here, we demonstrate that predictive coding converges asymptotically (and in practice rapidly) to exact backprop gradients on arbitrary computation graphs using only local learning rules. We apply this result to develop a straightforward strategy to translate core machine learning architectures into their predictive coding equivalents. We construct predictive coding CNNs, RNNs, and the more complex LSTMs, which include a non-layer-like branching internal graph structure and multiplicative interactions. Our models perform equivalently to backprop on challenging machine learning benchmarks, while utilising only local and (mostly) Hebbian plasticity. Our method raises the potential that standard machine learning algorithms could in principle be directly implemented in neural circuitry, and may also contribute to the development of completely distributed neuromorphic architectures.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=PdauS7wZBfC", "arxiv_id": "2006.04182", "paper_pdf": "papers/PdauS7wZBfC.pdf", "paper_pdf_sha256": "11083ec53b9e00b29830dc252523378b10b6ec0e4ce0bb743501352ee8726d4c", "paper_pdf_bytes": 2116262, "paper_pdf_source": "openreview", "code_url": "https://github.com/BerenMillidge/PredictiveCodingBackprop", "code_repository": "BerenMillidge/PredictiveCodingBackprop", "code_commit": "2154d7829a8f786e0a7a7b1bd8bb33c39d5af9ba", "code_archive": "repos/PdauS7wZBfC.zip", "code_archive_sha256": "09f15e47efddbe5b651f5077a36c3f86f2d6c98b0d98a280b4766cd5daf9c846", "code_archive_bytes": 34865, "code_file_count": 18, "code_extensions": {".py": 16, ".sh": 2}, "github_disk_usage_kb": 82, "github_languages": {"Python": 131154, "Shell": 3713}, "github_archived": false, "github_pushed_at": "2020-11-23T11:12:31Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/predictive-coding-approximates-backprop-along"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Skltqh4KvB", "year": 2020, "status": "rejected", "title": "Are there any 'object detectors' in the hidden layers of CNNs trained to identify objects or scenes?", "authors": ["Ella M. Gale", "Nicholas Martin", "Ryan Blything", "Anh Nguyen", "Jeffrey S. Bowers"], "authorids": ["ella.gale@bristol.ac.uk", "nm13850@bristol.ac.uk", "ryan.blything@bristol.ac.uk", "anhnguyen@auburn.edu", "j.bowers@bristol.ac.uk"], "authors_source": "OpenReview API", "abstract": "Various methods of measuring unit selectivity have been developed with the aim of better understanding how neural networks work.  But the different measures provide divergent estimates of selectivity, and this has led to different conclusions regarding the conditions in which selective object representations are learned and the functional relevance of these representations. In an attempt to better characterize object selectivity, we undertake a comparison of various selectivity measures on a large set of units in AlexNet, including localist selectivity, precision, class-conditional mean activity selectivity (CCMAS), network dissection, the human interpretation of activation maximization (AM) images, and standard signal-detection measures.  We find that the different measures provide different estimates of object selectivity, with precision and CCMAS measures providing misleadingly high estimates. Indeed, the most selective units had a poor hit-rate or a high false-alarm rate (or both) in object classification, making them poor object detectors.  We fail to find any units that are even remotely as selective as the 'grandmother cell' units reported in recurrent neural networks. In order to generalize these results, we compared selectivity measures on a few units in VGG-16 and GoogLeNet trained on the ImageNet or Places-365 datasets that have been described as 'object detectors'. Again, we find poor hit-rates and high false-alarm rates for object classification. ", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=Skltqh4KvB", "arxiv_id": "2007.01062", "paper_pdf": "papers/Skltqh4KvB.pdf", "paper_pdf_sha256": "a67a7902d5d5014d891296dee8ea77c07dd0e98c152f59b755408ab77ef44ff3", "paper_pdf_bytes": 39476601, "paper_pdf_source": "openreview", "code_url": "https://github.com/ellagale/testing_object_detectors_in_deepCNNs", "code_repository": "ellagale/testing_object_detectors_in_deepCNNs", "code_commit": "fbc15e0b8f7e023c1e3e139feec185b36f68e6f6", "code_archive": "repos/Skltqh4KvB.zip", "code_archive_sha256": "67ce0cfe94cdf6de63c06bbfd04ceeee2492fb0cca729f5d6b39cb61d44327ef", "code_archive_bytes": 161205, "code_file_count": 36, "code_extensions": {".py": 36}, "github_disk_usage_kb": 125, "github_languages": {"Python": 671935}, "github_archived": false, "github_pushed_at": "2020-05-28T10:09:47Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/are-there-any-object-detectors-in-the-hidden-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "r1gR2sC9FX", "year": 2019, "status": "rejected", "title": "On the Spectral Bias of Neural Networks", "authors": ["Nasim Rahaman", "Aristide Baratin", "Devansh Arpit", "Felix Draxler", "Min Lin", "Fred Hamprecht", "Yoshua Bengio", "Aaron Courville"], "authorids": ["nasim.rahaman@iwr.uni-heidelberg.de", "aristidebaratin@hotmail.com", "devansharpit@gmail.com", "felix.draxler@iwr.uni-heidelberg.de", "mavenlin@gmail.com", "yoshua.umontreal@gmail.com", "aaron.courville@umontreal.ca"], "authors_source": "OpenReview API", "abstract": "Neural networks are known to be a class of highly expressive functions able to fit even random input-output mappings with 100% accuracy. In this work we present properties of neural networks that complement this aspect of expressivity. By using tools from Fourier analysis, we show that deep ReLU networks are biased towards low frequency functions, meaning that they cannot have local fluctuations without affecting their global behavior. Intuitively, this property is in line with the observation that over-parameterized networks find simple patterns that generalize across data samples. We also investigate how the shape of the data manifold affects expressivity by showing evidence that learning high frequencies gets easier with increasing manifold complexity, and present a theoretical understanding of this behavior. Finally, we study the robustness of the frequency components with respect to parameter perturbation, to develop the intuition that the parameters must be finely tuned to express high frequency functions.", "decision": null, "meta_review": "Reject", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=r1gR2sC9FX", "arxiv_id": "1806.08734", "paper_pdf": "papers/r1gR2sC9FX.pdf", "paper_pdf_sha256": "c9e0b9df78ebfe4c5e22e13a3fe6cd80635ff080ff569a0276ad24c94c3b2060", "paper_pdf_bytes": 2656003, "paper_pdf_source": "openreview", "code_url": "https://github.com/nasimrahaman/SpectralBias", "code_repository": "nasimrahaman/SpectralBias", "code_commit": "edfe5150c648414f6f4e7404626f834b26a0286c", "code_archive": "repos/r1gR2sC9FX.zip", "code_archive_sha256": "e8cef0b146e2214ab50b074c3244f74cb3683d554d8ed1b7caa54d1281edefe9", "code_archive_bytes": 944902, "code_file_count": 5, "code_extensions": {".ipynb": 5}, "github_disk_usage_kb": 922, "github_languages": {"Jupyter Notebook": 1311346}, "github_archived": false, "github_pushed_at": "2019-05-06T14:06:56Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/on-the-spectral-bias-of-neural-networks"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "AeHZWzDjTk", "year": 2026, "status": "rejected", "title": "Any-Order GPT as Masked Diffusion Model: Decoupling Formulation and Architecture", "authors": ["Shuchen Xue", "Tianyu Xie", "Tianyang Hu", "Zijin Feng", "Jiacheng Sun", "Kenji Kawaguchi", "Zhenguo Li", "Zhi-Ming Ma"], "authorids": ["~Shuchen_Xue1", "~Tianyu_Xie1", "~Tianyang_Hu1", "~Zijin_Feng1", "~Jiacheng_Sun1", "~Kenji_Kawaguchi1", "~Zhenguo_Li1", "~Zhi-Ming_Ma1"], "authors_source": "OpenReview API", "abstract": "Large language models (LLMs) predominantly use autoregressive (AR) approaches, but masked diffusion models (MDMs) are emerging as viable alternatives. A key challenge in comparing AR and MDM paradigms is their typical architectural difference: AR models are often decoder-only, while MDMs have largely been encoder-only. This practice of changing both the modeling paradigm and architecture simultaneously makes direct comparisons unfair, as it's hard to distinguish whether observed differences stem from the paradigm itself or the architectural shift. This research evaluates MDMs within a decoder-only framework to: (1) equitably compare MDM (as Any-Order AR, or AO-AR) and standard AR paradigms. Our investigation suggests that the standard AO-AR objective, which averages over all token permutations, may benefit from refinement, as many permutations appear less informative compared to the language's inherent left-to-right structure. (2) Investigate architectural influences (decoder-only vs. encoder-only) within MDMs. We demonstrate that while encoder-only MDMs model a simpler conditional probability space, decoder-only MDMs can achieve dramatic generation speedups ($\\sim25\\times$) and comparable perplexity with temperature annealing despite modeling a vastly larger space, highlighting key trade-offs. This work thus decouples core paradigm differences from architectural influences, offering insights for future model design.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=AeHZWzDjTk", "arxiv_id": "2506.19935", "paper_pdf": "papers/AeHZWzDjTk.pdf", "paper_pdf_sha256": "41cd8050bfe9d7931934454d9e43bf2784ec214d5e789da4de0af6871149ec69", "paper_pdf_bytes": 680030, "paper_pdf_source": "openreview", "code_url": "https://github.com/scxue/AO-GPT-MDM", "code_repository": "scxue/AO-GPT-MDM", "code_commit": "3a02e3bc60c7514f93601de6a39034bbefa8555b", "code_archive": "repos/AeHZWzDjTk.zip", "code_archive_sha256": "4388e0819cb0b3932d7fd61ed825ae8d451a15f4faf4723c03f51d0c7cce46c8", "code_archive_bytes": 69704, "code_file_count": 30, "code_extensions": {".py": 25, ".sh": 5}, "github_disk_usage_kb": 485, "github_languages": {"Python": 197240, "Shell": 5594}, "github_archived": false, "github_pushed_at": "2025-06-23T20:52:49Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/any-order-gpt-as-masked-diffusion-model"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Ng1r9kTep4", "year": 2025, "status": "rejected", "title": "Inverted Activations: Reducing Memory Footprint in Neural Network Training", "authors": ["Georgii Sergeevich Novikov", "Ivan Oseledets"], "authorids": ["~Georgii_Sergeevich_Novikov1", "~Ivan_Oseledets1"], "authors_source": "OpenReview API", "abstract": "The scaling of neural networks with increasing data and model sizes necessitates the development of more efficient deep learning algorithms. \n    A significant challenge in neural network training is the memory footprint associated with activation tensors, particularly in pointwise nonlinearity layers that traditionally save the entire input tensor for the backward pass, leading to substantial memory consumption.\n    \n    In this paper, we propose a modification to the handling of activation tensors in pointwise nonlinearity layers. \n    Our method involves saving the output tensor instead of the input tensor during the forward pass. Since the subsequent layer typically also saves its input tensor, this approach reduces the total memory required by storing only one tensor between layers instead of two. This optimization is especially beneficial for transformer-based architectures like GPT, BERT, Mistral, and Llama.\n\n    To enable this approach, we utilize the inverse function of the nonlinearity during the backward pass. As the inverse cannot be computed analytically for most nonlinearities, we construct accurate approximations using simpler functions. \n    Experimental results demonstrate that our method significantly reduces memory usage without affecting training accuracy or computational performance.\n\n    Our implementation is provided as a drop-in replacement for standard nonlinearity layers in the PyTorch framework, facilitating easy adoption without requiring architectural modifications. The code is available at \\url{https://github.com/removed/for/anonimity}.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=Ng1r9kTep4", "arxiv_id": "2407.15545", "paper_pdf": "papers/Ng1r9kTep4.pdf", "paper_pdf_sha256": "0792c8e015f8d0e201ea646afb644276c0b7cd8c497438d653370221d91e3421", "paper_pdf_bytes": 227385, "paper_pdf_source": "openreview", "code_url": "https://github.com/PgLoLo/optiacts", "code_repository": "PgLoLo/optiacts", "code_commit": "1d7a1e4173bf98f1db9fa11d6e2409b149491f9f", "code_archive": "repos/Ng1r9kTep4.zip", "code_archive_sha256": "5875d2117ed6bbabddf16ef0e8b6817fa73b6115193464907a639979ddd362a8", "code_archive_bytes": 57115, "code_file_count": 10, "code_extensions": {".py": 7, ".ipynb": 3}, "github_disk_usage_kb": 55, "github_languages": {"Python": 5973}, "github_archived": false, "github_pushed_at": "2024-07-19T08:10:18Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/inverted-activations"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "WpQbM1kBuy", "year": 2024, "status": "rejected", "title": "Prodigy: An Expeditiously Adaptive Parameter-Free Learner", "authors": ["Konstantin Mishchenko", "Aaron Defazio"], "authorids": ["~Konstantin_Mishchenko1", "~Aaron_Defazio1"], "authors_source": "OpenReview API", "abstract": "We consider the problem of estimating the learning rate in adaptive methods, such as Adagrad and Adam. We describe two techniques, Prodigy and Resetting, to provably estimate the distance to the solution $D$, which is needed to set the learning rate optimally. Our techniques are modifications of the D-Adaptation method for learning-rate-free learning. Our methods improve upon the convergence rate of D-Adaptation by a factor of $\\mathcal{O}(\\sqrt{\\log(D/d_0)})$, where $d_0$ is the initial estimate of $D$. We test our methods on 12 common logistic-regression benchmark datasets, VGG11 and ResNet-50 training on CIFAR10, ViT training on Imagenet, LSTM training on IWSLT14, DLRM training on Criteo dataset, VarNet on Knee MRI dataset, as well as RoBERTa and GPT transformer training on BookWiki. Our experimental results show that our approaches consistently outperform D-Adaptation and reach test accuracy values close to that of hand-tuned Adam.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=WpQbM1kBuy", "arxiv_id": "2306.06101", "paper_pdf": "papers/WpQbM1kBuy.pdf", "paper_pdf_sha256": "0420c6fcd569f0a2c5eae1140832d961d449631b829204fb587fb32532947186", "paper_pdf_bytes": 2518028, "paper_pdf_source": "openreview", "code_url": "https://github.com/konstmish/prodigy", "code_repository": "konstmish/prodigy", "code_commit": "3efb213ee8af5a6bf76f28726398433a847b38e9", "code_archive": "repos/WpQbM1kBuy.zip", "code_archive_sha256": "3da66d295cfc3628cdd98ada112ecce4f333ce3a90cfa9f6ec892be111a5a0e0", "code_archive_bytes": 12537, "code_file_count": 4, "code_extensions": {".py": 4}, "github_disk_usage_kb": 47, "github_languages": {"Python": 21063}, "github_archived": false, "github_pushed_at": "2025-01-16T14:57:40Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/prodigy-an-expeditiously-adaptive-parameter"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "UFKW7EVrJAm", "year": 2023, "status": "rejected", "title": "Generating Adversarial Examples with Task Oriented Multi-Objective Optimization", "authors": ["Anh Tuan Bui", "Trung Le", "He Zhao", "Quan Hung Tran", "Paul Montague", "Dinh Phung"], "authorids": ["~Anh_Tuan_Bui2", "~Trung_Le2", "~He_Zhao1", "~Quan_Hung_Tran1", "~Paul_Montague1", "~Dinh_Phung2"], "authors_source": "OpenReview API", "abstract": "Deep learning models, even the-state-of-the-art ones, are highly vulnerable to adversarial examples. Adversarial training is one of the most efficient methods to improve the model's robustness. The key factor for the success of adversarial training is the capability to generate qualified and divergent adversarial examples which satisfy some objectives/goals (e.g., finding adversarial examples that maximize the model losses for simultaneously attacking multiple models). Therefore, multi-objective optimization (MOO) is a natural tool for adversarial example generation, where we search adversarial examples simultaneously maximizing some objectives/goals. However, we observe that a naive application of MOO tends to maximize all objectives/goals equally, without caring if an objective/goal has been achieved yet. This leads to useless effort to further improve the goal-achieved tasks, while putting less focus on the goal-unachieved tasks. In this paper, we propose \\emph{Task Oriented MOO} to address this issue, in the context where we can explicitly define the goal achievement for a task. Our principle is to only maintain the goal-achieved tasks, while letting the optimizer spend more effort on improving the goal-unachieved tasks. We conduct comprehensive experiments for our Task Oriented MOO on various adversarial example generation schemes. The experimental results firmly demonstrate the merit of our proposed approach.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=UFKW7EVrJAm", "arxiv_id": "2304.13229", "paper_pdf": "papers/UFKW7EVrJAm.pdf", "paper_pdf_sha256": "5f6bf05e4a5abac0e7f338d71b33698729b3abdf69a59c81b9d5644620ffce0c", "paper_pdf_bytes": 4862642, "paper_pdf_source": "openreview", "code_url": "https://github.com/tuananhbui89/TAMOO", "code_repository": "tuananhbui89/TAMOO", "code_commit": "1485ca64e879f9bf333957a4176929eacadc99ee", "code_archive": "repos/UFKW7EVrJAm.zip", "code_archive_sha256": "ac08f33003ebdc5ba0d8e5fb67f2df1ad78af5748185af2930780c0196d0b605", "code_archive_bytes": 99592, "code_file_count": 47, "code_extensions": {".py": 38, ".sh": 9}, "github_disk_usage_kb": 79, "github_languages": {"Python": 363104, "Shell": 46110}, "github_archived": false, "github_pushed_at": "2023-06-01T19:12:26Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/generating-adversarial-examples-with-task"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "B7O85qTDgU4", "year": 2022, "status": "rejected", "title": "Meta-Learning Dynamics Forecasting Using Task Inference", "authors": ["Rui Wang", "Robin Walters", "Rose Yu"], "authorids": ["~Rui_Wang11", "~Robin_Walters1", "~Rose_Yu1"], "authors_source": "OpenReview API", "abstract": "Current deep learning models for dynamics forecasting struggle with generalization. They can only forecast in a specific domain and fail when applied to systems with different parameters, external forces, or boundary conditions.  We propose a model-based meta-learning method called DyAd which can generalize across heterogeneous domains by partitioning them into different tasks.  DyAd has two parts: an encoder which infers the time-invariant hidden features of the task with weak supervision, and a forecaster which learns the shared dynamics of the entire domain. The encoder adapts and controls the forecaster during inference using adaptive instance normalization and adaptive padding.  Theoretically, we prove that the generalization error of such procedure is related to the task relatedness in the source domain, as well as the domain differences between source and target. Experimentally, we demonstrate that our model outperforms state-of-the-art approaches on both turbulent flow and real-world ocean data forecasting tasks.     ", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=B7O85qTDgU4", "arxiv_id": "2102.10271", "paper_pdf": "papers/B7O85qTDgU4.pdf", "paper_pdf_sha256": "8cdf1f89d50b36d2aaaa75ca6dd95d35ab4283ac0a33d8cf78fb8685595dccd2", "paper_pdf_bytes": 3999225, "paper_pdf_source": "openreview", "code_url": "https://github.com/Rose-STL-Lab/Dynamic-Adaptation-Network", "code_repository": "Rose-STL-Lab/Dynamic-Adaptation-Network", "code_commit": "2afd6e09db1cce65102a7b8d55c24668cd220ab6", "code_archive": "repos/B7O85qTDgU4.zip", "code_archive_sha256": "fe5fefec5fac4d63cf0cb23c16f4edadccb4512cce852011cbd1dc63671e27ae", "code_archive_bytes": 49545, "code_file_count": 29, "code_extensions": {".py": 29}, "github_disk_usage_kb": 78, "github_languages": {"Python": 149813}, "github_archived": false, "github_pushed_at": "2023-10-30T14:05:47Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/meta-learning-dynamics-forecasting-using-task"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Ip195saXqIX", "year": 2021, "status": "rejected", "title": "Knowledge Distillation By Sparse Representation Matching", "authors": ["Dat Thanh Tran", "Moncef Gabbouj", "Alexandros Iosifidis"], "authorids": ["~Dat_Thanh_Tran1", "~Moncef_Gabbouj1", "~Alexandros_Iosifidis2"], "authors_source": "OpenReview API", "abstract": "Knowledge Distillation refers to a class of methods that transfers the knowledge from a teacher network to a student network. In this paper, we propose Sparse Representation Matching (SRM), a method to transfer intermediate knowledge obtained from one Convolutional Neural Network (CNN) to another by utilizing sparse representation learning. SRM first extracts sparse representations of the hidden features of the teacher CNN, which are then used to generate both pixel-level and image-level labels for training intermediate feature maps of the student network. We formulate SRM as a neural processing block, which can be efficiently optimized using stochastic gradient descent and integrated into any CNN in a plug-and-play manner. Our experiments demonstrate that SRM is robust to architectural differences between the teacher and student networks, and outperforms other KD techniques across several datasets. ", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=Ip195saXqIX", "arxiv_id": "2103.17012", "paper_pdf": "papers/Ip195saXqIX.pdf", "paper_pdf_sha256": "447743ef5767e211c317c60c7869733051dbfe5a1b6064c519bac43d4d8f8b6c", "paper_pdf_bytes": 237873, "paper_pdf_source": "openreview", "code_url": "https://github.com/viebboy/SRM", "code_repository": "viebboy/SRM", "code_commit": "77a12ba790c799e64624d155c8faff3a5e47c477", "code_archive": "repos/Ip195saXqIX.zip", "code_archive_sha256": "935175d781aff22a6f1ba8328058f20dabdbad6e4be0109146055bf14e08814d", "code_archive_bytes": 109690, "code_file_count": 71, "code_extensions": {".py": 68, ".sh": 3}, "github_disk_usage_kb": 83, "github_languages": {"Python": 520853, "Shell": 3444}, "github_archived": false, "github_pushed_at": "2021-03-31T09:37:42Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/knowledge-distillation-by-sparse-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "HyeqPJHYvH", "year": 2020, "status": "rejected", "title": "Stochastic Latent Residual Video Prediction", "authors": ["Jean-Yves Franceschi", "Edouard Delasalles", "Mickael Chen", "Sylvain Lamprier", "Patrick Gallinari"], "authorids": ["jean-yves.franceschi@lip6.fr", "edouard.delasalles@lip6.fr", "mickael.chen@lip6.fr", "sylvain.lamprier@lip6.fr", "patrick.gallinari@lip6.fr"], "authors_source": "OpenReview API", "abstract": "Video prediction is a challenging task: models have to account for the inherent uncertainty of the future. Most works in the literature are based on stochastic image-autoregressive recurrent networks, raising several performance and applicability issues. An alternative is to use fully latent temporal models which untie frame synthesis and dynamics. However, no such model for video prediction has been proposed in the literature yet, due to design and training difficulties. In this paper, we overcome these difficulties by introducing a novel stochastic temporal model. It is based on residual updates of a latent state, motivated by discretization schemes of differential equations. This first-order principle naturally models video dynamics as it allows our simpler, lightweight, interpretable, latent model to outperform prior state-of-the-art methods on challenging datasets.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=HyeqPJHYvH", "arxiv_id": "2002.09219", "paper_pdf": "papers/HyeqPJHYvH.pdf", "paper_pdf_sha256": "1c3f295587273043f7d8aea2ab845dded749683bdc1ab9cd67e7c09c66c8bd80", "paper_pdf_bytes": 5806910, "paper_pdf_source": "openreview", "code_url": "https://github.com/edouardelasalles/srvp", "code_repository": "edouardelasalles/srvp", "code_commit": "05661faf767cdb33d40fc328679bbe50c3a1f938", "code_archive": "repos/HyeqPJHYvH.zip", "code_archive_sha256": "63bdeb12c2fe10117b52e9613f4729d25a13e1990a4f2b3883bf08a2dcdeb43b", "code_archive_bytes": 77526, "code_file_count": 29, "code_extensions": {".py": 27, ".sh": 2}, "github_disk_usage_kb": 130, "github_languages": {"Python": 199873, "Shell": 551}, "github_archived": false, "github_pushed_at": "2022-03-29T09:25:25Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/stochastic-latent-residual-video-prediction-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "SkGpW3C5KX", "year": 2019, "status": "rejected", "title": "Heated-Up Softmax Embedding", "authors": ["Xu Zhang", "Felix Xinnan Yu", "Svebor Karaman", "Wei Zhang", "Shih-Fu Chang"], "authorids": ["xu.zhang@columbia.edu", "felixyu@google.com", "svebor.karaman@gmail.com", "wz2363@columbia.edu", "sc250@columbia.edu"], "authors_source": "OpenReview API", "abstract": "Metric learning aims at learning a distance which is consistent with the semantic meaning of the samples. The problem is generally solved by learning an embedding, such that the samples of the same category are close (compact) while samples from different categories are far away (spread-out) in the embedding space. One popular way of generating such embeddings is to use the second-to-last layer of a deep neural network trained as a classifier with the softmax cross-entropy loss. In this paper, we show that training classifiers with different temperatures of the softmax function lead to different distributions of the embedding space. And finding a balance between the compactness, 'spread-out' and the generalization ability of the feature is critical in metric learning. Leveraging these insights, we propose a 'heating-up' strategy to train a classifier with increasing temperatures. Extensive experiments show that the proposed method achieves state-of-the-art embeddings on a variety of metric learning benchmarks. ", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=SkGpW3C5KX", "arxiv_id": "1809.04157", "paper_pdf": "papers/SkGpW3C5KX.pdf", "paper_pdf_sha256": "dce246b2e8149e0d5e2b825e6a6ecaa1211b18ee8c9d9ac76f9fdaf3cc8aaace", "paper_pdf_bytes": 642781, "paper_pdf_source": "openreview", "code_url": "https://github.com/ColumbiaDVMM/Heated_Up_Softmax_Embedding", "code_repository": "ColumbiaDVMM/Heated_Up_Softmax_Embedding", "code_commit": "cb62d28e5faaf7fdb134b31c461125e3fef50d06", "code_archive": "repos/SkGpW3C5KX.zip", "code_archive_sha256": "f2cd6c095403753c7de604ad92b602c7c1554615c0d59f73ad6f703364fdc33e", "code_archive_bytes": 765171, "code_file_count": 17, "code_extensions": {".py": 17}, "github_disk_usage_kb": 951, "github_languages": {"Python": 116371}, "github_archived": false, "github_pushed_at": "2018-09-13T15:13:41Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/heated-up-softmax-embedding"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "vBZXJzFV6x", "year": 2026, "status": "rejected", "title": "Rethinking Gating Mechanism in Sparse MoE: Handling Arbitrary Modality Inputs with Confidence-Guided Gate", "authors": ["Liangwei Nathan Zheng", "Wei Emma Zhang", "Mingyu Guo", "Miao Xu", "Olaf Maennel", "Weitong Chen"], "authorids": ["~Liangwei_Nathan_Zheng1", "~Wei_Emma_Zhang1", "~Mingyu_Guo1", "~Miao_Xu3", "~Olaf_Maennel1", "~Weitong_Chen2"], "authors_source": "OpenReview API", "abstract": "Effectively managing missing modalities is a fundamental challenge in real-world multimodal learning scenarios, where data incompleteness often results from systematic collection errors or sensor failures. Sparse Mixture-of-Experts (SMoE) architectures have the potential to naturally handle multimodal data, with individual experts specializing in different modalities. However, existing SMoE approach often lacks proper ability to handle missing modality, leading to performance degradation and poor generalization in real-world applications. We propose ConfSMoE to introduce a two-stage imputation module to handle the missing modality problem for the SMoE architecture by taking the opinion of experts and reveal the insight of expert collapse from theoretical analysis with strong empirical evidence. Inspired by our theoretical analysis, ConfSMoE propose a novel expert gating mechanism by detaching the softmax routing score to task confidence score w.r.t ground truth signal. This naturally relieves expert collapse without introducing additional load balance loss function. We show that the insights of expert collapse aligns with other gating mechanism such as Gaussian and Laplacian gate. The proposed method is evaluated on four different real world dataset with three distinct experiment settings to conduct comprehensive analysis of ConfSMoE on resistance to missing modality and the impacts of proposed gating mechanism.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=vBZXJzFV6x", "arxiv_id": "2505.19525", "paper_pdf": "papers/vBZXJzFV6x.pdf", "paper_pdf_sha256": "2810fd185a75d35414c8ea9bab94e6f341db230cc633f923b5b271c5c80ac8ea", "paper_pdf_bytes": 1442306, "paper_pdf_source": "openreview", "code_url": "https://github.com/IcurasLW/ICML2026-Official-Repository-of-ConfSMoE", "code_repository": "IcurasLW/ICML2026-Official-Repository-of-ConfSMoE", "code_commit": "b73c5167aff2c6d53507ddfee2d1a98b749511e5", "code_archive": "repos/vBZXJzFV6x.zip", "code_archive_sha256": "ea4918a61f89f127285dfa85186390d5b43ca6962b71a8ebed0539349b162589", "code_archive_bytes": 496042, "code_file_count": 26, "code_extensions": {".py": 17, ".sh": 9}, "github_disk_usage_kb": 490, "github_languages": {"Python": 188459, "Shell": 6729}, "github_archived": false, "github_pushed_at": "2026-09-01T02:40:14Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/rethinking-gating-mechanism-in-sparse-moe"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "xSSo8kCA9G", "year": 2025, "status": "rejected", "title": "FRUGAL: Memory-Efficient Optimization by Reducing State Overhead for Scalable Training", "authors": ["Philip Zmushko", "Aleksandr Beznosikov", "Martin Takáč", "Samuel Horváth"], "authorids": ["~Philip_Zmushko1", "~Aleksandr_Beznosikov1", "~Martin_Takáč1", "~Samuel_Horváth1"], "authors_source": "OpenReview API", "abstract": "With the increase in the number of parameters in large language models, the process of pre-training and fine-tuning increasingly demands larger volumes of GPU memory. A significant portion of this memory is typically consumed by the optimizer state. To overcome this challenge, recent approaches such as low-rank adaptation (LoRA (Hu et al., 2021)), low-rank gradient projection (GaLore (Zhao\net al., 2024)), and block-wise optimization (BAdam (Luo et al., 2024)) have been proposed. However, in all these algorithms, the effective rank of the weight updates remains low-rank, which can lead to a substantial loss of information from the gradient. This loss can be critically important, especially during the pre-training stage. In this paper, we introduce **FRUGAL**; (**F**ull-**R**ank **U**pdates with **G**r**A**dient sp**L**itting),, a new memory-efficient optimization framework. The framework leverages gradient splitting to perform low-rank updates using advanced optimization algorithms (such as Adam), while updates along the remaining directions are\nexecuted via state-free methods like SGD or signSGD. Our framework can be integrated with various low-rank update selection techniques, including GaLore and BAdam. We provide theoretical convergence guarantees for our framework when\nusing SGDM for low-rank updates and SGD for state-free updates. Additionally, our method consistently outperforms concurrent approaches across various fixed memory budgets, achieving state-of-the-art results in pre-training and fine-tuning\ntasks while balancing memory efficiency and perplexity targets.", "decision": "Reject", "meta_review": "", "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=xSSo8kCA9G", "arxiv_id": "2411.07837", "paper_pdf": "papers/xSSo8kCA9G.pdf", "paper_pdf_sha256": "a39a5fff17c0f5996b516fda3478e72295fd4ef145777e3ce6d50fdd4cc4f093", "paper_pdf_bytes": 608202, "paper_pdf_source": "openreview", "code_url": "https://github.com/fzmushko/FRUGAL", "code_repository": "fzmushko/FRUGAL", "code_commit": "c7ebe4c242f7908e036c287e3a10512a84634648", "code_archive": "repos/xSSo8kCA9G.zip", "code_archive_sha256": "b98977abc6b484b9a5bec1aa6ea04a19057d4b03e4c2b5192d35dacf9e3e5936", "code_archive_bytes": 74185, "code_file_count": 36, "code_extensions": {".sh": 21, ".py": 13, ".ipynb": 2}, "github_disk_usage_kb": 57, "github_languages": {"Python": 178820, "Shell": 17200, "Jupyter Notebook": 12467}, "github_archived": false, "github_pushed_at": "2024-11-13T11:07:39Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/frugal-memory-efficient-optimization-by"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "vlQ56aWJhl", "year": 2024, "status": "rejected", "title": "S-TLLR: STDP-inspired Temporal Local Learning Rule for Spiking Neural Networks", "authors": ["Marco Paul E. Apolinario", "Kaushik Roy"], "authorids": ["~Marco_Paul_E._Apolinario1", "~Kaushik_Roy1"], "authors_source": "OpenReview API", "abstract": "Spiking Neural Networks (SNNs) are biologically plausible models that have been identified as potentially apt for deploying energy-efficient intelligence at the edge, particularly for sequential learning tasks. However, training of SNNs poses significant challenges due to the necessity for precise temporal and spatial credit assignments. Back-propagation through time (BPTT) algorithm, whilst the most widely used method for addressing these issues, incurs a high computational cost due to its temporal dependency. In this work, we propose S-TLLR, a novel three-factor temporal local learning rule inspired by the Spike-Timing Dependent Plasticity (STDP) mechanism, aimed at training deep SNNs on event-based learning tasks. Furthermore, S-TLLR is designed to have low memory and time complexities, which are independent of the number of time steps, rendering it suitable for online learning on low-power edge devices. To demonstrate the scalability of our proposed method, we have conducted extensive evaluations on event-based datasets spanning a wide range of applications, such as image and gesture recognition, audio classification, and optical flow estimation. In all the experiments, S-TLLR achieved high accuracy, comparable to BPTT, with reduction in the number of computations between $1.1-10\\times$.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=vlQ56aWJhl", "arxiv_id": "2306.15220", "paper_pdf": "papers/vlQ56aWJhl.pdf", "paper_pdf_sha256": "1ead20f31432e8866b4336f287e529f4617f55f405f0507643d6c912c1e116da", "paper_pdf_bytes": 1078947, "paper_pdf_source": "openreview", "code_url": "https://github.com/mapolinario94/S-TLLR", "code_repository": "mapolinario94/S-TLLR", "code_commit": "29c575529617f6c384b0a045e5f32ea01102845d", "code_archive": "repos/vlQ56aWJhl.zip", "code_archive_sha256": "a0aa92bb9fe816d8e0c18a2646cca5397b8cd0261b7fa7c5aac89e665acb1c44", "code_archive_bytes": 46660, "code_file_count": 22, "code_extensions": {".py": 22}, "github_disk_usage_kb": 50, "github_languages": {"Python": 169171}, "github_archived": false, "github_pushed_at": "2025-02-14T00:06:53Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/s-tllr-stdp-inspired-temporal-local-learning"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "9qgOs_IwRS3", "year": 2023, "status": "rejected", "title": "Neighborhood Gradient Clustering: An Efficient Decentralized Learning Method for Non-IID Data Distributions", "authors": ["Sai Aparna Aketi", "Sangamesh Kodge", "Kaushik Roy"], "authorids": ["~Sai_Aparna_Aketi1", "~Sangamesh_Kodge1", "~Kaushik_Roy1"], "authors_source": "OpenReview API", "abstract": "Decentralized learning algorithms enable the training of deep learning models over large distributed datasets generated at different devices and locations, without the need for a central server. In practical scenarios, the distributed datasets can have significantly different data distributions across the agents. The current state-of-the-art decentralized algorithms mostly assume the data distributions to be Independent and Identically Distributed (IID). This paper focuses on improving decentralized learning over non-IID data distributions with minimal compute and memory overheads. We propose Neighborhood Gradient Clustering (NGC), a novel decentralized learning algorithm that modifies the local gradients of each agent using self- and cross-gradient information. Cross-gradients for a pair of neighboring agents are the derivatives of the model parameters of an agent with respect to the dataset of the other agent. In particular, the proposed method replaces the local gradients of the model with the weighted mean of the self-gradients, model-variant cross-gradients (derivatives of the received neighbors’ model parameters with respect to the local dataset - computed locally), and data-variant cross-gradients (derivatives of the local model with respect to its neighbors’ datasets - received through communication). The data-variant cross-gradients are aggregated through an additional communication round without breaking the privacy constraints of the decentralized setting. Further, we present CompNGC, a compressed version of NGC that reduces the communication overhead by $32 \\times$ by compressing the cross-gradients. We demonstrate the empirical convergence and efficiency of the proposed technique over non-IID data distributions sampled from the CIFAR-10 dataset on various model architectures and graph topologies. Our experiments demonstrate that NGC and CompNGC outperform the existing state-of-the-art (SoTA) decentralized learning algorithm over non-IID data by $1-5\\%$ with significantly less compute and memory requirements. Further, we also show that the proposed NGC method outperforms the baseline by $5-40\\%$ with no additional communication. ", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=9qgOs_IwRS3", "arxiv_id": "2209.14390", "paper_pdf": "papers/9qgOs_IwRS3.pdf", "paper_pdf_sha256": "43589f0371a8d45a93eac13d7e6cd8501bcd785a0b3d67df34fe76ac85c2405d", "paper_pdf_bytes": 696053, "paper_pdf_source": "openreview", "code_url": "https://github.com/aparna-aketi/neighborhood_gradient_clustering", "code_repository": "aparna-aketi/neighborhood_gradient_clustering", "code_commit": "2e6e3676f335b79dcb2cb53722c4f8cd80073862", "code_archive": "repos/9qgOs_IwRS3.zip", "code_archive_sha256": "9b401b95185efe31d3b0fe305134369596f881a1bc24e100c2306646816b19ec", "code_archive_bytes": 64907, "code_file_count": 38, "code_extensions": {".py": 37, ".sh": 1}, "github_disk_usage_kb": 82, "github_languages": {"Python": 181638, "Shell": 1825}, "github_archived": false, "github_pushed_at": "2023-11-15T01:31:11Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/neighborhood-gradient-clustering-an-efficient"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "WN2Sup7qLdw", "year": 2022, "status": "rejected", "title": "Multi-Resolution Continuous Normalizing Flows", "authors": ["Vikram Voleti", "Chris Finlay", "Adam M Oberman", "Christopher Pal"], "authorids": ["~Vikram_Voleti1", "~Chris_Finlay1", "~Adam_M_Oberman1", "~Christopher_Pal1"], "authors_source": "OpenReview API", "abstract": "Recent work has shown that Neural Ordinary Differential Equations (ODEs) can serve as generative models of images using the perspective of Continuous Normalizing Flows (CNFs). Such models offer exact likelihood calculation, and invertible generation/density estimation. In this work we introduce a Multi-Resolution variant of such models (MRCNF), by characterizing the conditional distribution over the additional information required to generate a fine image that is consistent with the coarse image. We introduce a transformation between resolutions that allows for no change in the log likelihood. We show that this approach yields comparable likelihood values for various image datasets, using orders of magnitude fewer parameters than the prior methods, in significantly less training time, using only one GPU.", "decision": "Reject", "meta_review": null, "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=WN2Sup7qLdw", "arxiv_id": "2106.08462", "paper_pdf": "papers/WN2Sup7qLdw.pdf", "paper_pdf_sha256": "cc211cb96e7585a2094675b2f3533761e6afc94fbd0cc81660a069650045bb63", "paper_pdf_bytes": 1541495, "paper_pdf_source": "openreview", "code_url": "https://github.com/voletiv/mrcnf", "code_repository": "voletiv/mrcnf", "code_commit": "a564ca562909a0749a7c83f680a4a3b857b8961c", "code_archive": "repos/WN2Sup7qLdw.zip", "code_archive_sha256": "fd7b975d6cb916d3607a9960d38de8dc49868090cf39a794ebbbf2be0112677c", "code_archive_bytes": 89074, "code_file_count": 43, "code_extensions": {".py": 42, ".sh": 1}, "github_disk_usage_kb": 79, "github_languages": {"Python": 322462, "Shell": 1588}, "github_archived": false, "github_pushed_at": "2021-08-22T21:51:30Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/multi-resolution-continuous-normalizing-flows"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "iQxS0S9ir1a", "year": 2021, "status": "rejected", "title": "Distributional Generalization: A New Kind of Generalization", "authors": ["Preetum Nakkiran", "Yamini Bansal"], "authorids": ["~Preetum_Nakkiran1", "~Yamini_Bansal1"], "authors_source": "OpenReview API", "abstract": "We introduce a new notion of generalization--- Distributional Generalization--- which roughly states that outputs of a classifier at train and test time are close as distributions, as opposed to close in just their average error. For example, if we mislabel 30% of dogs as cats in the train set of CIFAR-10, then a ResNet trained to interpolation will in fact mislabel roughly 30% of dogs as cats on the test set as well, while leaving other classes unaffected. This behavior is not captured by classical generalization, which would only consider the average error and not the distribution of errors over the input domain. Our formal conjectures, which are much more general than this example, characterize the form of distributional generalization that can be expected in terms of problem parameters: model architecture, training procedure, number of samples, and data distribution. We give empirical evidence for these conjectures across a variety of domains in machine learning, including neural networks, kernel machines, and decision trees. Our results thus advance our understanding of interpolating classifiers.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=iQxS0S9ir1a", "arxiv_id": "2009.08092", "paper_pdf": "papers/iQxS0S9ir1a.pdf", "paper_pdf_sha256": "ad589072d553ac2cf4b41f43ed6e426f0af7787bd5900a51b60a53a6f7c9d3bf", "paper_pdf_bytes": 5339815, "paper_pdf_source": "openreview", "code_url": "https://github.com/kuangliu/pytorch-cifar", "code_repository": "kuangliu/pytorch-cifar", "code_commit": "49b7aa97b0c12fe0d4054e670403a16b6b834ddd", "code_archive": "repos/iQxS0S9ir1a.zip", "code_archive_sha256": "9a807c8c6c68a1842346f41d7f0a18effa36a28179d86fc56dee8cf05f1fb41c", "code_archive_bytes": 26050, "code_file_count": 21, "code_extensions": {".py": 21}, "github_disk_usage_kb": 85, "github_languages": {"Python": 73999}, "github_archived": false, "github_pushed_at": "2023-02-24T10:06:53Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/distributional-generalization-a-new-kind-of"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "rJgDb1SFwB", "year": 2020, "status": "rejected", "title": "MGP-AttTCN: An Interpretable Machine Learning Model for the Prediction of Sepsis", "authors": ["Margherita Rosnati", "Vincent Fortuin"], "authorids": ["mrosnati@ethz.ch", "fortuin@inf.ethz.ch"], "authors_source": "OpenReview API", "abstract": "With a mortality rate of 5.4 million lives worldwide every year and a healthcare cost of more than 16 billion dollars in the USA alone, sepsis is one of the leading causes of hospital mortality and an increasing concern in the ageing western world. Recently, medical and technological advances have helped re-define the illness criteria of this disease, which is otherwise poorly understood by the medical society. Together with the rise of widely accessible Electronic Health Records, the advances in data mining and complex nonlinear algorithms are a promising avenue for the early detection of sepsis. This work contributes to the research effort in the field of automated sepsis detection with an open-access labelling of the medical MIMIC-III data set. Moreover, we propose MGP-AttTCN: a joint multitask Gaussian Process and attention-based deep learning model to early predict the occurrence of sepsis in an interpretable manner. We show that our model outperforms the current state-of-the-art and present evidence that different labelling heuristics lead to discrepancies in task difficulty.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=rJgDb1SFwB", "arxiv_id": "1909.12637", "paper_pdf": "papers/rJgDb1SFwB.pdf", "paper_pdf_sha256": "967ee4ec0bbe2f2f0019fe50f4bc7cd3d091ab0e3cc5d8c3fb58a13ceab7f466", "paper_pdf_bytes": 425124, "paper_pdf_source": "openreview", "code_url": "https://github.com/mmr12/MGP-AttTCN", "code_repository": "mmr12/MGP-AttTCN", "code_commit": "57c8f623d7deee4085a0ac58e795c0264d2f0e49", "code_archive": "repos/rJgDb1SFwB.zip", "code_archive_sha256": "275b19da07b5c4a0a463c1a6729ec9256acf6c26a676df94116fc25e119fd56f", "code_archive_bytes": 112897, "code_file_count": 59, "code_extensions": {".py": 31, ".sql": 28}, "github_disk_usage_kb": 130, "github_languages": {"Python": 174580}, "github_archived": false, "github_pushed_at": "2023-03-25T01:05:09Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/mgp-atttcn-an-interpretable-machine-learning"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "rJlMBjAcYX", "year": 2019, "status": "rejected", "title": "Optimizing for Generalization in Machine Learning with Cross-Validation Gradients", "authors": ["Barratt", "Shane", "Sharma", "Rishi"], "authorids": ["sbarratt@stanford.edu", "rsh@stanford.edu"], "authors_source": "OpenReview API", "abstract": "Cross-validation is the workhorse of modern applied statistics and machine learning, as it provides a principled framework for selecting the model that maximizes generalization performance. In this paper, we show that the cross-validation risk is differentiable with respect to the hyperparameters and training data for many common machine learning algorithms, including logistic regression, elastic-net regression, and support vector machines. Leveraging this property of differentiability, we propose a cross-validation gradient method (CVGM) for hyperparameter optimization. Our method enables efficient optimization in high-dimensional hyperparameter spaces of the cross-validation risk, the best surrogate of the true generalization ability of our learning algorithm.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=rJlMBjAcYX", "arxiv_id": "1805.07072", "paper_pdf": "papers/rJlMBjAcYX.pdf", "paper_pdf_sha256": "117605edac73b0b9385521ced3b7618a821c27299b58b0f2695af87dbf712967", "paper_pdf_bytes": 417525, "paper_pdf_source": "openreview", "code_url": "https://github.com/sbarratt/crossval", "code_repository": "sbarratt/crossval", "code_commit": "d275d40cd8c6109c844fdb9242bd588266971275", "code_archive": "repos/rJlMBjAcYX.zip", "code_archive_sha256": "9bade9e20b5f52f712df25fc8df3d65c88b0f0638cf82193274b593d96f877ca", "code_archive_bytes": 1105379, "code_file_count": 3, "code_extensions": {".ipynb": 3}, "github_disk_usage_kb": 1078, "github_languages": {"Jupyter Notebook": 1580579}, "github_archived": false, "github_pushed_at": "2018-05-21T00:53:09Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/optimizing-for-generalization-in-machine-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Sm8yTdWjai", "year": 2026, "status": "rejected", "title": "RedDebate: Safer Responses Through Multi-Agent Red Teaming Debates", "authors": ["Ali Asad", "Stephen Obadinma", "Radin Shayanfar", "Xiaodan Zhu"], "authorids": ["~Ali_Asad1", "~Stephen_Obadinma1", "~Radin_Shayanfar1", "~Xiaodan_Zhu1"], "authors_source": "OpenReview API", "abstract": "We introduce RedDebate, a novel multi-agent debate framework that provides the foundation for Large Language Models (LLMs) to identify and mitigate their own unsafe behaviors. Existing AI safety approaches often rely on costly human evaluation or isolated single-model assessment, both constrained by scalability and prone to oversight failures. RedDebate employs collaborative argumentation among multiple LLMs across diverse debate scenarios, enabling them to critically evaluate one another’s reasoning and systematically uncover unsafe failure modes through fully automated red-teaming. We further integrate distinct long-term memory modules that preserve safety-relevant insights from debate interactions and leverage them during subsequent inference, facilitating continuous refinement of model behavior. Empirical evaluation on safety benchmarks across a diverse set of models demonstrates that RedDebate substantially reduces unsafe outputs. While debate alone allows LLMs to refine their behavior, the addition of memory modules yields further significant reductions. To the best of our knowledge, RedDebate is the first fully automated framework to unify multi-agent debate and red-teaming to progressively enhance LLM safety without human intervention.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=Sm8yTdWjai", "arxiv_id": "2506.11083", "paper_pdf": "papers/Sm8yTdWjai.pdf", "paper_pdf_sha256": "a6a38a07c70025b7780108c8f02baaa92689d52bb5808415b672a888178dc690", "paper_pdf_bytes": 2279544, "paper_pdf_source": "openreview", "code_url": "https://github.com/aliasad059/RedDebate", "code_repository": "aliasad059/RedDebate", "code_commit": "713bcf09a7ea63247453d8df644b53c833ce13d2", "code_archive": "repos/Sm8yTdWjai.zip", "code_archive_sha256": "7fc7a97a6032d93a48e249cdeeabcb4f1dda18e13d268055f72eedc30df19170", "code_archive_bytes": 183869, "code_file_count": 14, "code_extensions": {".py": 14}, "github_disk_usage_kb": 491, "github_languages": {"Python": 148182}, "github_archived": false, "github_pushed_at": "2026-07-04T19:01:44Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/reddebate-safer-responses-through-multi-agent"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "nD5tbHBfut", "year": 2025, "status": "rejected", "title": "LRVS-Fashion: Extending Visual Search with Referring Instructions", "authors": ["Simon Lepage", "Jeremie Mary", "David Picard"], "authorids": ["~Simon_Lepage1", "~Jeremie_Mary1", "~David_Picard1"], "authors_source": "OpenReview API", "abstract": "This paper introduces a new challenge for image similarity search in the context of fashion, addressing the inherent ambiguity in this domain stemming from complex images. We present Referred Visual Search (RVS), a task allowing users to define more precisely the desired similarity, following recent interest in the industry. We release a new large public dataset, LRVS-Fashion, consisting of 272k fashion products with 842k images extracted from fashion catalogs, designed explicitly for this task. However, unlike traditional visual search methods in the industry, we demonstrate that superior performance can be achieved by bypassing explicit object detection and adopting weakly-supervised conditional contrastive learning on image tuples. Our method is lightweight and demonstrates robustness, reaching Recall at one superior to strong detection-based baselines against 2M distractors.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=nD5tbHBfut", "arxiv_id": "2306.02928", "paper_pdf": "papers/nD5tbHBfut.pdf", "paper_pdf_sha256": "137cad4fbb0efdea90430a7f59ad782fb9dd7ed09d857176d74d330486912b72", "paper_pdf_bytes": 7803827, "paper_pdf_source": "openreview", "code_url": "https://github.com/Simon-Lepage/LRVSF-Benchmark", "code_repository": "Simon-Lepage/LRVSF-Benchmark", "code_commit": "46e5e943b3db0962daa843fc6dfa2a78c8182382", "code_archive": "repos/nD5tbHBfut.zip", "code_archive_sha256": "324495c2d7fa17e52e3893b48fdfa0061619e09f6d2b991e6b2216b76e631d93", "code_archive_bytes": 57740, "code_file_count": 6, "code_extensions": {".py": 6}, "github_disk_usage_kb": 58, "github_languages": {"Python": 14949, "Makefile": 382}, "github_archived": false, "github_pushed_at": "2024-05-15T15:04:24Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/weakly-supervised-conditional-embedding-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Tpk0p9QBM6", "year": 2024, "status": "rejected", "title": "Computing Low-Entropy Couplings for Large-Support Distributions", "authors": ["Samuel Sokota", "Dylan Sam", "Christian Schroeder de Witt", "Spencer Compton", "Jakob Nicolaus Foerster", "J Zico Kolter"], "authorids": ["~Samuel_Sokota1", "~Dylan_Sam1", "~Christian_Schroeder_de_Witt1", "~Spencer_Compton1", "~Jakob_Nicolaus_Foerster1", "~J_Zico_Kolter1"], "authors_source": "OpenReview API", "abstract": "A minimum-entropy coupling is a joint probability distribution having minimum joint entropy among all joint distributions with given pre-specified marginals. While provable approximation algorithms for a minimum-entropy coupling exist, they take log-linear time in the size of the support of the marginal distributions. Thus, applications involving very large-support distributions instead use a class of heuristic iterative minimum-entropy coupling (IMEC) algorithms. Unfortunately, existing IMEC algorithms are limited to specific classes of distributions, prohibiting applications involving general large-support distributions. In this work, we resolve this issue by making three main contributions: 1) We unify existing IMEC algorithms under a single formalism using sets of partitions. 2) We derive a new IMEC instance from this formalism, which we call ARIMEC, that, unlike existing IMEC algorithms, can be applied to arbitrary discrete distributions; furthermore, we introduce associated operations that make ARIMEC efficient in practice. 3) We empirically illustrate the utility of ARIMEC for both Markov coding games and steganography.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=Tpk0p9QBM6", "arxiv_id": "2405.19540", "paper_pdf": "papers/Tpk0p9QBM6.pdf", "paper_pdf_sha256": "a15fd57eae9b2344d5edfa50c2b509f6944ad48244d0b95257eb9a3d4c44f48e", "paper_pdf_bytes": 1153300, "paper_pdf_source": "openreview", "code_url": "https://github.com/ssokota/mec", "code_repository": "ssokota/mec", "code_commit": "e56ad6619498914fe289d91516febb1fa3be050f", "code_archive": "repos/Tpk0p9QBM6.zip", "code_archive_sha256": "b9724a7ec296565ffe0ac486f6492d3ca314b52e8313366e650e954d7edc735e", "code_archive_bytes": 43750, "code_file_count": 27, "code_extensions": {".py": 25, ".rs": 2}, "github_disk_usage_kb": 50, "github_languages": {"Python": 86017, "Rust": 13793}, "github_archived": false, "github_pushed_at": "2026-01-06T05:19:11Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/computing-low-entropy-couplings-for-large"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "7qSpaOSbRVO", "year": 2023, "status": "rejected", "title": "Data Poisoning Attacks Against Multimodal Encoders", "authors": ["Ziqing Yang", "Xinlei He", "Zheng Li", "Michael Backes", "Mathias Humbert", "Pascal Berrang", "Yang Zhang"], "authorids": ["~Ziqing_Yang3", "~Xinlei_He1", "~Zheng_Li17", "~Michael_Backes1", "~Mathias_Humbert2", "~Pascal_Berrang1", "~Yang_Zhang15"], "authors_source": "OpenReview API", "abstract": "Traditional machine learning (ML) models, e.g., image classifiers, usually rely on large-scale labeled datasets to achieve strong performance. However, such labeled datasets are often challenging and expensive to obtain. Also, the predefined categories limit the model's ability to generalize to other visual concepts as additional labeled data is required. On the contrary, the newly emerged multimodal model, which contains both visual and linguistic modalities, learns the concept of images from the raw text. It is a promising way to solve the above problems as it can use easy-to-collect image-text pairs to construct the training dataset and the raw texts contain almost unlimited categories according to their semantics. However, learning from a  large-scale unlabeled dataset also exposes the model to the risk of potential poisoning attacks, whereby the adversary aims to perturb the model's training dataset to trigger malicious behaviors in it. Previous work mainly focuses on the visual modality. In this paper, we instead focus on answering two questions: (1) Is the linguistic modality also vulnerable to poisoning attacks? and (2) Which modality is most vulnerable? To answer the two questions, we conduct three types of poisoning attacks against CLIP, the most representative multimodal contrastive learning framework. Extensive evaluations on different datasets and model architectures show that all three attacks can perform well on the linguistic modality with only a relatively low poisoning rate and limited epochs. Also, we observe that the poisoning effect differs between different modalities, i.e., with lower MinRank in the visual modality and with higher Hit@K when K is small in the linguistic modality. To mitigate the attacks, we propose both pre-training and post-training defenses. We empirically show that both defenses can significantly reduce the attack performance while preserving the model's utility.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=7qSpaOSbRVO", "arxiv_id": "2209.15266", "paper_pdf": "papers/7qSpaOSbRVO.pdf", "paper_pdf_sha256": "614d4150779cd439bf5c9e517e4a48a35639143c870b6cc25bc541d988d4592a", "paper_pdf_bytes": 985510, "paper_pdf_source": "openreview", "code_url": "https://github.com/zqypku/mm_poison", "code_repository": "zqypku/mm_poison", "code_commit": "b11a503035db543cccd944c4b01efc6f9dcc5d11", "code_archive": "repos/7qSpaOSbRVO.zip", "code_archive_sha256": "a94727d488eab2fe75273783e10381eaf8e3187cd4f0f53fd49d7d8cb9ddd077", "code_archive_bytes": 86713, "code_file_count": 35, "code_extensions": {".py": 35}, "github_disk_usage_kb": 83, "github_languages": {"Python": 308268}, "github_archived": false, "github_pushed_at": "2023-10-25T12:28:05Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/data-poisoning-attacks-against-multimodal"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "6PTUd_zPdHL", "year": 2022, "status": "rejected", "title": "Differentiable Top-k Classification Learning", "authors": ["Felix Petersen", "Hilde Kuehne", "Christian Borgelt", "Oliver Deussen"], "authorids": ["~Felix_Petersen1", "~Hilde_Kuehne5", "~Christian_Borgelt1", "~Oliver_Deussen1"], "authors_source": "OpenReview API", "abstract": "The top-k classification accuracy is one of the core metrics in machine learning. Here, k is conventionally a positive integer, such as 1 or 5. In this work, we relax this assumption and propose to draw k from a probability distribution for training. Combining this with recent advances in differentiable sorting and ranking, we propose a new family of differentiable top-k cross-entropy classification losses. We find that relaxing k does not only produce better top-5 accuracies, but also makes models more robust, which leads to top-1 accuracy improvements. When fine-tuning publicly available ImageNet models, we achieve a new state-of-the-art on ImageNet for publicly available models with an 88.36% top-1 and a 98.71% top-5 accuracy. ", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=6PTUd_zPdHL", "arxiv_id": "2206.07290", "paper_pdf": "papers/6PTUd_zPdHL.pdf", "paper_pdf_sha256": "be2738bba9d8af9e65d0d7779cb95a2e42e7ea57b42ad9945900996fe99d627f", "paper_pdf_bytes": 548857, "paper_pdf_source": "openreview", "code_url": "https://github.com/Felix-Petersen/difftopk", "code_repository": "Felix-Petersen/difftopk", "code_commit": "76ef96db648058a73571628f1db5e6a9f4478bfd", "code_archive": "repos/6PTUd_zPdHL.zip", "code_archive_sha256": "70e8ef6e5492eea07f2be84ca12a5f9fdfc6806f83933992b3ae81314fe73d44", "code_archive_bytes": 76608, "code_file_count": 13, "code_extensions": {".py": 13}, "github_disk_usage_kb": 80, "github_languages": {"Python": 80416}, "github_archived": false, "github_pushed_at": "2023-01-03T11:59:44Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/differentiable-top-k-classification-learning-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "3Wp8HM2CNdR", "year": 2021, "status": "rejected", "title": "Whitening for Self-Supervised Representation Learning", "authors": ["Aleksandr Ermolov", "Aliaksandr Siarohin", "Enver Sangineto", "Nicu Sebe"], "authorids": ["~Aleksandr_Ermolov1", "~Aliaksandr_Siarohin1", "~Enver_Sangineto1", "~Nicu_Sebe1"], "authors_source": "OpenReview API", "abstract": "Most of the self-supervised representation learning methods are based on the contrastive loss and the instance-discrimination task, where augmented versions of the same image instance  (\"positives\") are contrasted with instances extracted from other images (\"negatives\"). For the learning to be effective, a lot of negatives should be compared with a positive pair, which is computationally demanding. In this paper, we propose a different direction and a new loss function for self-supervised representation learning which is based on the whitening of the latent-space features. The whitening operation has a \"scattering\" effect on the batch samples, which compensates the use of negatives, avoiding degenerate solutions where all the sample representations collapse to a single point. Our Whitening MSE (W-MSE) loss does not require special heuristics (e.g. additional networks) and it is conceptually simple. Since negatives are not needed, we can extract multiple positive pairs from the same image instance. We  empirically show that W-MSE is competitive with respect to popular, more complex self-supervised methods. The source code of the method and all the experiments is included in the Supplementary Material.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=3Wp8HM2CNdR", "arxiv_id": "2007.06346", "paper_pdf": "papers/3Wp8HM2CNdR.pdf", "paper_pdf_sha256": "241282b4a3aed827fd56034b81ecb8cbe4e035d212a72e57b5461dd980d56cd3", "paper_pdf_bytes": 1720986, "paper_pdf_source": "openreview", "code_url": "https://github.com/htdt/self-supervised", "code_repository": "htdt/self-supervised", "code_commit": "d9662c8d07dafd194a9045375f4f6aa09f5b03e9", "code_archive": "repos/3Wp8HM2CNdR.zip", "code_archive_sha256": "7f8bdde022d80048d6fc16e7e693b44d3fb8c3bf19e6bf8a6d764026e1a0cb26", "code_archive_bytes": 20021, "code_file_count": 24, "code_extensions": {".py": 24}, "github_disk_usage_kb": 87, "github_languages": {"Python": 30903, "Dockerfile": 206}, "github_archived": false, "github_pushed_at": "2023-02-05T10:14:12Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/whitening-for-self-supervised-representation"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Skl3SkSKDr", "year": 2020, "status": "rejected", "title": "Generating valid Euclidean distance matrices", "authors": ["Moritz Hoffmann", "Frank Noe"], "authorids": ["moritz.hoffmann@fu-berlin.de", "frank.noe@fu-berlin.de"], "authors_source": "OpenReview API", "abstract": "Generating point clouds, e.g., molecular structures, in arbitrary rotations, translations, and enumerations remains a challenging task. Meanwhile, neural networks\nutilizing symmetry invariant layers have been shown to be able to optimize their\ntraining objective in a data-efficient way. In this spirit, we present an architecture\nwhich allows to produce valid Euclidean distance matrices, which by construction are already invariant under rotation and translation of the described object.\nMotivated by the goal to generate molecular structures in Cartesian space, we use\nthis architecture to construct a Wasserstein GAN utilizing a permutation invariant critic network. This makes it possible to generate molecular structures in a\none-shot fashion by producing Euclidean distance matrices which have a three-\ndimensional embedding.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=Skl3SkSKDr", "arxiv_id": "1910.03131", "paper_pdf": "papers/Skl3SkSKDr.pdf", "paper_pdf_sha256": "969e8ccab5a960b005d1db73848a6c3f0ac02a15094960d580f0183ac31572fe", "paper_pdf_bytes": 1475096, "paper_pdf_source": "openreview", "code_url": "https://github.com/noegroup/EDMnets", "code_repository": "noegroup/EDMnets", "code_commit": "c73eedd31248166ce739e9a196f08ea0ec23e54f", "code_archive": "repos/Skl3SkSKDr.zip", "code_archive_sha256": "21abb8ba0f7afb1c68382fb206a569b35d673b37eb57c8a5e8960c785af9617c", "code_archive_bytes": 131239, "code_file_count": 19, "code_extensions": {".py": 13, ".h": 5, ".cpp": 1}, "github_disk_usage_kb": 140, "github_languages": {"Python": 111314, "C++": 17508, "C": 87}, "github_archived": false, "github_pushed_at": "2019-10-18T16:52:18Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/generating-valid-euclidean-distance-matrices-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "B1x33sC9KQ", "year": 2019, "status": "rejected", "title": "ACIQ: Analytical Clipping for Integer Quantization of neural networks", "authors": ["Ron Banner", "Yury Nahshan", "Elad Hoffer", "Daniel Soudry"], "authorids": ["ron.banner@intel.com", "yury.nahshan@intel.com", "daniel.soudry@gmail.com", "elad.hoffer@gmail.com"], "authors_source": "OpenReview API", "abstract": "We analyze the trade-off between quantization noise and clipping distortion in low precision networks. We identify the statistics of various tensors, and derive exact expressions for the mean-square-error degradation due to clipping. By optimizing these expressions, we show marked improvements over standard quantization schemes that normally avoid clipping. For example, just by choosing the accurate clipping values, more than 40\\% accuracy improvement is obtained for the quantization of VGG-16 to 4-bits of precision. Our results have many applications for the quantization of neural networks at both training and inference time. \n", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=B1x33sC9KQ", "arxiv_id": null, "paper_pdf": "papers/B1x33sC9KQ.pdf", "paper_pdf_sha256": "0ca44443e9315e4d24d17e48138263ad121803955018e1f33320b81ded23b959", "paper_pdf_bytes": 740828, "paper_pdf_source": "openreview", "code_url": "https://github.com/submission2019/AnalyticalScaleForIntegerQuantization", "code_repository": "submission2019/AnalyticalScaleForIntegerQuantization", "code_commit": "3246ee8cbfb747d7ef821c8cecc50283a73eaf92", "code_archive": "repos/B1x33sC9KQ.zip", "code_archive_sha256": "9a3fc3214025cd06281a04708bc55f868c06017f50d6be496a9c0d4b721f0847", "code_archive_bytes": 957719, "code_file_count": 22, "code_extensions": {".py": 18, ".sh": 1, ".cu": 1, ".cpp": 1, ".ipynb": 1}, "github_disk_usage_kb": 1120, "github_languages": {"Jupyter Notebook": 413088, "Python": 87304, "Cuda": 1661, "C++": 337, "Shell": 328}, "github_archived": false, "github_pushed_at": "2019-01-20T07:53:21Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/aciq-analytical-clipping-for-integer"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "oGlgHjYKBi", "year": 2026, "status": "rejected", "title": "Model-Preserving Adaptive Rounding", "authors": ["Albert Tseng", "Zhaofeng Sun", "Christopher De Sa"], "authorids": ["~Albert_Tseng1", "~Zhaofeng_Sun1", "~Christopher_De_Sa2"], "authors_source": "OpenReview API", "abstract": "The goal of quantization is to produce a compressed model whose output distribution is as close to the original model's as possible. \nTo do this tractably, most quantization algorithms minimize the immediate activation error of each layer as a proxy for the end-to-end error.\nHowever, this ignores the effect of future layers, making it a poor proxy.\nIn this work, we introduce Yet Another Quantization Algorithm (YAQA), an adaptive rounding algorithm that directly considers the error at the network's output.\nYAQA introduces a series of theoretical results that culminate in the first end-to-end error bounds for quantization algorithms.\nFirst, we characterize the convergence time of adaptive rounding algorithms via the structure of their Hessian approximations.\nWe then show that the end-to-end error can be bounded by the approximation's cosine similarity to the true Hessian.\nThis admits a natural Kronecker-factored approximation with corresponding near-optimal Hessian sketches.\nYAQA is provably better than GPTQ/LDLQ and empirically reduces the error by $\\approx 30\\%$ over these methods.\nYAQA even achieves a lower error than quantization aware training.\nThis translates to state of the art performance on downstream tasks, all while adding no inference overhead.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=oGlgHjYKBi", "arxiv_id": "2505.22988", "paper_pdf": "papers/oGlgHjYKBi.pdf", "paper_pdf_sha256": "e641e4e0e1cca650357d6e4ea5d1287d3a9fe06e4d46f78a4c91fd1a999e7215", "paper_pdf_bytes": 480975, "paper_pdf_source": "openreview", "code_url": "https://github.com/Cornell-RelaxML/yaqa-quantization", "code_repository": "Cornell-RelaxML/yaqa-quantization", "code_commit": "f9508723251ad839f0162326569f17fe70486fcc", "code_archive": "repos/oGlgHjYKBi.zip", "code_archive_sha256": "c98cf680caef3e7b48c4685f8e76f2e8b6f345d42389be6d42442b6a74e3ffa7", "code_archive_bytes": 379085, "code_file_count": 42, "code_extensions": {".py": 36, ".cu": 3, ".sh": 1, ".h": 1, ".cpp": 1}, "github_disk_usage_kb": 547, "github_languages": {"Python": 2045846, "Cuda": 45186, "C++": 26529, "Makefile": 432, "Shell": 328, "HTML": 53}, "github_archived": false, "github_pushed_at": "2025-06-20T14:03:31Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/model-preserving-adaptive-rounding"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "WDheQxWAo4", "year": 2025, "status": "rejected", "title": "Simple synthetic data reduces sycophancy in large language models", "authors": ["Jerry Wei", "Da Huang", "Yifeng Lu", "Denny Zhou", "Quoc V Le"], "authorids": ["~Jerry_Wei1", "~Da_Huang2", "~Yifeng_Lu1", "~Denny_Zhou1", "~Quoc_V_Le1"], "authors_source": "OpenReview API", "abstract": "Sycophancy is an undesirable behavior where models tailor their responses to follow a human user's view even when that view is not objectively correct (e.g., adapting liberal views once a user reveals that they are liberal). In this paper, we study the prevalence of sycophancy in language models and propose a simple synthetic-data intervention to reduce this behavior.\n\nFirst, on a set of three sycophancy tasks where models are asked for an opinion on statements with no correct answers (e.g., politics), we observe that both model scaling and instruction tuning significantly increase sycophancy for large language models up to 540B parameters. Second, we extend sycophancy evaluations to simple addition statements that are objectively incorrect, finding that despite knowing that these statements are wrong, language models will still agree with them if the user does as well.\n\nTo reduce sycophancy, we present a straightforward synthetic-data intervention that takes public NLP tasks and encourages models to be robust to user opinions on these tasks. Adding these data in a lightweight finetuning step can significantly reduce sycophantic behavior on held-out prompts. Code for generating synthetic data for intervention can be found at https://anonymous.4open.science/r/sycophancy-intervention-F0D1/.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=WDheQxWAo4", "arxiv_id": "2308.03958", "paper_pdf": "papers/WDheQxWAo4.pdf", "paper_pdf_sha256": "2cf4859df7cc557c2e3f823fbc909e9a25d61aed44635bee906329ecc22076ec", "paper_pdf_bytes": 426892, "paper_pdf_source": "openreview", "code_url": "https://github.com/google/sycophancy-intervention", "code_repository": "google/sycophancy-intervention", "code_commit": "fb75986c759abdff7fe4d4b09fd7102383dc71aa", "code_archive": "repos/WDheQxWAo4.zip", "code_archive_sha256": "45c7d65e00424e451f378cf5f63a43fd742a94d0d48a34c3d940f01bd17e5465", "code_archive_bytes": 61856, "code_file_count": 4, "code_extensions": {".py": 4}, "github_disk_usage_kb": 60, "github_languages": {"Python": 15465}, "github_archived": true, "github_pushed_at": "2023-08-16T17:34:42Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/simple-synthetic-data-reduces-sycophancy-in"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ro4CgvfUKy", "year": 2024, "status": "rejected", "title": "Latent Noise Segmentation: How Neural Noise Leads to the Emergence of Segmentation and Grouping", "authors": ["Ben Lonnqvist", "Frank Zhengqing Wu", "Michael Herzog"], "authorids": ["~Ben_Lonnqvist1", "~Frank_Zhengqing_Wu1", "~Michael_Herzog1"], "authors_source": "OpenReview API", "abstract": "Deep Neural Networks (DNNs) that achieve human-level performance in general tasks like object segmentation typically require supervised labels. In contrast, humans are able to perform these tasks effortlessly without supervision.\nTo accomplish this, the human visual system makes use of perceptual grouping: for example, the black and white stripes of a zebra are perceptually grouped together despite their vastly different colors.\nUnderstanding how perceptual grouping arises in an unsupervised manner is critical for improving both models of the visual system, and computer vision models. In this work, we propose a counterintuitive approach to unsupervised perceptual grouping and segmentation: that they arise because of neural noise, rather than in spite of it. We (1) mathematically demonstrate that under realistic assumptions, neural noise can be used to separate objects from each other, and (2) show that adding noise in a DNN enables the network to segment images even though it was never trained on any segmentation labels. Interestingly, we find that (3) segmenting objects using noise results in segmentation performance that aligns with the perceptual grouping phenomena observed in humans. We introduce the Good Gestalt (GG) datasets --- six datasets designed to specifically test perceptual grouping, and show that our DNN models reproduce many important phenomena in human perception, such as illusory contours, closure, continuity, proximity, and occlusion. Finally, we (4) demonstrate the ecological plausibility of the method by analyzing the sensitivity of the DNN to different magnitudes of noise. We find that some model variants consistently succeed with remarkably low levels of neural noise ($\\sigma<0.001$), and surprisingly, that segmenting this way requires as few as a handful of samples. Together, our results suggest a novel unsupervised segmentation method requiring few assumptions, a new explanation for the formation of perceptual grouping, and a potential benefit of neural noise in the visual system.", "decision": "Reject", "meta_review": "", "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=ro4CgvfUKy", "arxiv_id": "2309.16515", "paper_pdf": "papers/ro4CgvfUKy.pdf", "paper_pdf_sha256": "9475886aeba18c6f1930e9f9b7fc2aa86c1a3c683f895d8b20954915a4d64917", "paper_pdf_bytes": 10063675, "paper_pdf_source": "openreview", "code_url": "https://github.com/ZhengqingUUU/LatentNoiseSegmentation", "code_repository": "ZhengqingUUU/LatentNoiseSegmentation", "code_commit": "eb79511f66c6e2f32c80aa2a35ef804bc30637c3", "code_archive": "repos/ro4CgvfUKy.zip", "code_archive_sha256": "c88c03e654dd54a8240d843551190f2a429bda77cdec50368824bf4ff344ff7c", "code_archive_bytes": 52462, "code_file_count": 12, "code_extensions": {".py": 10, ".sh": 2}, "github_disk_usage_kb": 52, "github_languages": {"Python": 165266, "Shell": 578}, "github_archived": false, "github_pushed_at": "2024-05-14T16:37:37Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/latent-noise-segmentation-how-neural-noise"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "D4JQEKlTyG", "year": 2023, "status": "rejected", "title": "Individual Privacy Accounting for Differentially Private Stochastic Gradient Descent", "authors": ["Da Yu", "Gautam Kamath", "Janardhan Kulkarni", "Tie-Yan Liu", "Jian Yin", "Huishuai Zhang"], "authorids": ["~Da_Yu1", "~Gautam_Kamath1", "~Janardhan_Kulkarni1", "~Tie-Yan_Liu1", "~Jian_Yin3", "~Huishuai_Zhang3"], "authors_source": "OpenReview API", "abstract": "Differentially private stochastic gradient descent (DP-SGD) is the workhorse algorithm for recent advances in private deep learning. It provides a single privacy guarantee to all datapoints in the dataset. We propose an efficient algorithm to compute privacy guarantees for individual examples when releasing models trained by DP-SGD. We use our algorithm to investigate individual privacy parameters across a number of datasets. We find that most examples enjoy stronger privacy guarantees than the worst-case bound. We further discover that the training loss and the privacy parameter of an example are well-correlated. This implies groups that are underserved in terms of model utility are simultaneously underserved in terms of privacy guarantee. For example, on CIFAR-10, the average $\\varepsilon$ of the class with the lowest test accuracy is 43.6% higher than that of the class with the highest accuracy. We also run membership inference attacks to show this reflects disparate empirical privacy risks.\n", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=D4JQEKlTyG", "arxiv_id": "2206.02617", "paper_pdf": "papers/D4JQEKlTyG.pdf", "paper_pdf_sha256": "473fc3d8ec4c4b341e7b322369d2deb3615fbec51e9a1704f12829d0bec4fdbb", "paper_pdf_bytes": 1484045, "paper_pdf_source": "openreview", "code_url": "https://github.com/dayu11/individual_privacy_of_DPSGD", "code_repository": "dayu11/individual_privacy_of_DPSGD", "code_commit": "ba43551e4ec8673efeb6191f9b243bcae12ffbd2", "code_archive": "repos/D4JQEKlTyG.zip", "code_archive_sha256": "aa8ec89b88296fe1e1d7e0ac929c898188171b747f14af0cc267181395d14992", "code_archive_bytes": 58839, "code_file_count": 8, "code_extensions": {".py": 8}, "github_disk_usage_kb": 86, "github_languages": {"Python": 36419}, "github_archived": false, "github_pushed_at": "2023-09-04T21:52:27Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/per-instance-privacy-accounting-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "IptBMO1AR5g", "year": 2022, "status": "rejected", "title": "Regularizing Deep Neural Networks with Stochastic Estimators of Hessian Trace", "authors": ["Yucong Liu", "Tong Lin"], "authorids": ["~Yucong_Liu1", "~Tong_Lin1"], "authors_source": "OpenReview API", "abstract": "In this paper we develop a novel regularization method for deep neural networks by penalizing the trace of Hessian. This regularizer is motivated by a recent guarantee bound of the generalization error. Hutchinson method is a classical unbiased estimator for the trace of a matrix, but it is very time-consuming on deep learning models. Hence a dropout scheme is proposed to efficiently implements the Hutchinson method. Then we discuss a connection to linear stability of a nonlinear dynamical system. Experiments demonstrate that our method outperforms existing regularizers such as Jacobian, confidence penalty, and label smoothing. Our regularization method is also orthogonal to data augmentation methods, achieving the best performance when our method is combined with data augmentation. ", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=IptBMO1AR5g", "arxiv_id": "2208.05924", "paper_pdf": "papers/IptBMO1AR5g.pdf", "paper_pdf_sha256": "c2017a0a918c16c906d6f83c15f081d87afc2c55910b0d97c29e2c222abe7505", "paper_pdf_bytes": 256462, "paper_pdf_source": "openreview", "code_url": "https://github.com/ICLRsubmission1596/Regularizing-Deep-Neural-Networks-with-Stochastic-Estimators-of-Hessian-Trace", "code_repository": "ICLRsubmission1596/Regularizing-Deep-Neural-Networks-with-Stochastic-Estimators-of-Hessian-Trace", "code_commit": "cd721ae5c569bc9b9e73a35c5f10c87bbd97febb", "code_archive": "repos/IptBMO1AR5g.zip", "code_archive_sha256": "d69b21fd66585c6fc38344d391c8448955bf0a536023def040fa4f7cabc50492", "code_archive_bytes": 76671, "code_file_count": 21, "code_extensions": {".py": 21}, "github_disk_usage_kb": 86, "github_languages": {"Python": 235849}, "github_archived": false, "github_pushed_at": "2021-10-05T20:54:20Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/regularizing-deep-neural-networks-with-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Twm9LnWK-zt", "year": 2021, "status": "rejected", "title": "Searching towards Class-Aware Generators for Conditional Generative Adversarial Networks", "authors": ["Peng Zhou", "Lingxi Xie", "XIAOPENG ZHANG", "Bingbing Ni", "Qi Tian"], "authorids": ["~Peng_Zhou2", "~Lingxi_Xie1", "~XIAOPENG_ZHANG7", "~Bingbing_Ni3", "~Qi_Tian3"], "authors_source": "OpenReview API", "abstract": "Conditional Generative Adversarial Networks (cGAN) were designed to generate images based on the provided conditions, e.g., class-level distributions. However, existing methods have used the same generating architecture for all classes. This paper presents a novel idea that adopts NAS to find a distinct architecture for each class. The search space contains regular and class-modulated convolutions, where the latter is designed to introduce class-specific information while avoiding the reduction of training data for each class generator. The search algorithm follows a weight-sharing pipeline with mixed-architecture optimization so that the search cost does not grow with the number of classes. To learn the sampling policy, a Markov decision process is embedded into the search algorithm and a moving average is applied for better stability. We evaluate our approach on CIFAR10 and CIFAR100. Besides achieving better image generation quality in terms of FID scores, we discover several insights that are helpful in designing cGAN models.", "decision": "Reject", "meta_review": null, "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=Twm9LnWK-zt", "arxiv_id": "2006.14208", "paper_pdf": "papers/Twm9LnWK-zt.pdf", "paper_pdf_sha256": "a63d0d3afd49b4abda88b0e68437a8f7730ea351ddfe82512a114c201942dc79", "paper_pdf_bytes": 1401870, "paper_pdf_source": "openreview", "code_url": "https://github.com/PeterouZh/NAS_cGAN", "code_repository": "PeterouZh/NAS_cGAN", "code_commit": "2f7819aee0ddc0f97e4d8084099b656b7c7a36a9", "code_archive": "repos/Twm9LnWK-zt.zip", "code_archive_sha256": "8938b4216b3254d10e47868e5abd08b9696dc038f30d4bbcd69826a4ac0d017e", "code_archive_bytes": 81268, "code_file_count": 21, "code_extensions": {".py": 21}, "github_disk_usage_kb": 92, "github_languages": {"Python": 305756}, "github_archived": false, "github_pushed_at": "2022-07-02T07:36:17Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/searching-towards-class-aware-generators-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "S1lXnhVKPr", "year": 2020, "status": "rejected", "title": "Variance Reduced Local SGD with Lower Communication Complexity", "authors": ["Xianfeng Liang", "Shuheng Shen", "Jingchang Liu", "Zhen Pan", "Yifei Cheng", "Enhong Chen"], "authorids": ["zeroxf@mail.ustc.edu.cn", "vaip@mail.ustc.edu.cn", "jliude@cse.ust.hk", "pzhen@mail.ustc.edu.cn", "chengyif@mail.ustc.edu.cn", "cheneh@ustc.edu.cn"], "authors_source": "OpenReview API", "abstract": "To accelerate the training of machine learning models, distributed stochastic gradient descent (SGD) and its variants have been widely adopted, which apply multiple workers in parallel to speed up training. Among them, Local SGD has gained much attention due to its lower communication cost. Nevertheless, when the data distribution on workers is non-identical, Local SGD requires $O(T^{\\frac{3}{4}} N^{\\frac{3}{4}})$ communications to maintain its \\emph{linear iteration speedup} property, where $T$ is the total number of iterations and $N$ is the number of workers. In this paper, we propose Variance Reduced Local SGD (VRL-SGD) to further reduce the communication complexity. Benefiting from eliminating the dependency on the gradient variance among workers, we theoretically prove that VRL-SGD achieves a \\emph{linear iteration speedup} with a lower communication complexity $O(T^{\\frac{1}{2}} N^{\\frac{3}{2}})$ even if workers access non-identical datasets. We conduct experiments on three machine learning tasks, and the experimental results demonstrate that VRL-SGD performs impressively better than Local SGD when the data among workers are quite diverse.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=S1lXnhVKPr", "arxiv_id": "1912.12844", "paper_pdf": "papers/S1lXnhVKPr.pdf", "paper_pdf_sha256": "6dbdfdb83c30b49a906328f033c803db678cf03f290ed8f6c225c4fd54c585c2", "paper_pdf_bytes": 734475, "paper_pdf_source": "openreview", "code_url": "https://github.com/zerolxf/VRL-SGD", "code_repository": "zerolxf/VRL-SGD", "code_commit": "8fb56dad507239f760b2756887c23404f0985fca", "code_archive": "repos/S1lXnhVKPr.zip", "code_archive_sha256": "5eecbcc041643e2da825035a5729e9dba4d2ef37e07f52898acaac2f82dce6b4", "code_archive_bytes": 200321, "code_file_count": 12, "code_extensions": {".py": 10, ".sh": 2}, "github_disk_usage_kb": 142, "github_languages": {"Python": 60210, "Shell": 3657}, "github_archived": false, "github_pushed_at": "2020-05-20T03:30:48Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/variance-reduced-local-sgd-with-lower-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "HJxfm2CqKm", "year": 2019, "status": "rejected", "title": "Discovering General-Purpose Active Learning Strategies", "authors": ["Ksenia Konyushkova", "Raphael Sznitman", "Pascal Fua"], "authorids": ["ksenia.konyushkova@epfl.ch", "raphael.sznitman@artorg.unibe.ch", "pascal.fua@epfl.ch"], "authors_source": "OpenReview API", "abstract": "We propose a general-purpose approach to discovering active learning (AL) strategies from data. These strategies are transferable from one domain to another and can be used in conjunction with many machine learning models. To this end, we formalize the annotation process as a Markov decision process, design universal state and action spaces and introduce a new reward function that precisely reflects the AL objective of minimizing the annotation cost We seek to find an optimal (non-myopic) AL strategy using reinforcement learning. We evaluate the learned strategies on multiple unrelated domains and show that they consistently outperform state-of-the-art baselines.", "decision": null, "meta_review": "Reject", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=HJxfm2CqKm", "arxiv_id": "1810.04114", "paper_pdf": "papers/HJxfm2CqKm.pdf", "paper_pdf_sha256": "f51dad5ce5325b214273c3f0c8441de6c539d85451f575a00994ec574bd27783", "paper_pdf_bytes": 424128, "paper_pdf_source": "openreview", "code_url": "https://github.com/ksenia-konyushkova/LAL-RL", "code_repository": "ksenia-konyushkova/LAL-RL", "code_commit": "6377fb0e5b6b64b62b0c3bc6e26ed4459138e1b8", "code_archive": "repos/HJxfm2CqKm.zip", "code_archive_sha256": "773efb64e773d732bfd1138c5c13aae1a511440cb3daec5cbec8e5236312a67d", "code_archive_bytes": 1458041, "code_file_count": 59, "code_extensions": {".py": 51, ".c": 3, ".ipynb": 2, ".h": 2, ".cpp": 1}, "github_disk_usage_kb": 1398, "github_languages": {"Python": 208230, "Jupyter Notebook": 41641, "C++": 37716, "C": 24287, "Cython": 5194}, "github_archived": false, "github_pushed_at": "2022-05-17T21:20:06Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/discovering-general-purpose-active-learning"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "knc1y6jkTL", "year": 2026, "status": "rejected", "title": "Accelerating Newton-Schulz Iteration for Orthogonalization via Chebyshev-type Polynomials", "authors": ["Ekaterina Grishina", "Matvey Smirnov", "Maxim Rakhuba"], "authorids": ["~Ekaterina_Grishina1", "~Matvey_Smirnov1", "~Maxim_Rakhuba1"], "authors_source": "OpenReview API", "abstract": "The problem of computing optimal orthogonal approximation to a given matrix has attracted growing interest in machine learning. Notable applications include the recent Muon optimizer or Riemannian optimization on the Stiefel manifold. Among existing approaches, the Newton-Schulz iteration has emerged as a particularly effective solution, as it relies solely on matrix multiplications and thus achieves high computational efficiency on GPU hardware. Despite its efficiency, the method has inherent limitations—its coefficients are fixed and thus not optimized for a given matrix. In this paper we address this issue by proposing a Chebyshev-optimized version of Newton-Schulz (CANS). Based on the Chebyshev's alternance theorem, we theoretically derive optimal coefficients for the 3-rd order Newton-Schulz iteration and apply a Remez algorithm to compute optimal higher-degree polynomials. We leverage these polynomials to construct controlled approximate orthogonalization schemes, which is of interest in deep learning applications. Practically, we demonstrate the method's effectiveness in two key applications: orthogonalization in the Muon optimizer, and providing an efficient retraction alternative for Riemannian optimization on the Stiefel manifold.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=knc1y6jkTL", "arxiv_id": "2506.10935", "paper_pdf": "papers/knc1y6jkTL.pdf", "paper_pdf_sha256": "3b4f2aaf95bf1eca282365aed1c87d4e1594ed87561f582fa2d1cfdfe6aeb67d", "paper_pdf_bytes": 847943, "paper_pdf_source": "openreview", "code_url": "https://github.com/GrishKate/accelerating_orthogonalization", "code_repository": "GrishKate/accelerating_orthogonalization", "code_commit": "3eca916ffa6f46ab3b313dd34c48235322747565", "code_archive": "repos/knc1y6jkTL.zip", "code_archive_sha256": "6e2f0ab727e6f821e077e4c557caa6d1646a2aa97433e7e5ba8e3215c81fb5a8", "code_archive_bytes": 599122, "code_file_count": 10, "code_extensions": {".py": 8, ".sh": 1, ".ipynb": 1}, "github_disk_usage_kb": 582, "github_languages": {"Jupyter Notebook": 791598, "Python": 54029, "Shell": 54}, "github_archived": false, "github_pushed_at": "2025-06-13T15:01:00Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/accelerating-newton-schulz-iteration-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "UFzE9njwMG", "year": 2025, "status": "rejected", "title": "Mitigating Time Discretization Challenges with WeatherODE: A Sandwich Physics-Driven Neural ODE for Weather Forecasting", "authors": ["Peiyuan Liu", "Tian Zhou", "Liang Sun", "Rong Jin"], "authorids": ["~Peiyuan_Liu1", "~Tian_Zhou2", "~Liang_Sun2", "~Rong_Jin1"], "authors_source": "OpenReview API", "abstract": "In the field of weather forecasting, traditional models often grapple with discretization errors and time-dependent source discrepancies, which limit their predictive performance. In this paper, we present WeatherODE, a novel one-stage, physics-driven ordinary differential equation (ODE) model designed to enhance weather forecasting accuracy. By leveraging wave equation theory and integrating a time-dependent source model, WeatherODE effectively addresses the challenges associated with time-discretization error and dynamic atmospheric processes. Moreover, we design a CNN-ViT-CNN sandwich structure, facilitating efficient learning dynamics tailored for distinct yet interrelated tasks with varying optimization biases in advection equation estimation. Through rigorous experiments, WeatherODE demonstrates superior performance in both global and regional weather forecasting tasks, outperforming recent state-of-the-art approaches by significant margins of over 40.0\\% and 31.8\\% in root mean square error (RMSE), respectively. The source code is available at https://anonymous.4open.science/r/WeatherODE-5C13/.", "decision": "Reject", "meta_review": "", "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=UFzE9njwMG", "arxiv_id": "2410.06560", "paper_pdf": "papers/UFzE9njwMG.pdf", "paper_pdf_sha256": "b6718d50007b5853993e21b0514a35227952333421ced7867644453b720b5636", "paper_pdf_bytes": 15165138, "paper_pdf_source": "openreview", "code_url": "https://github.com/DAMO-DI-ML/WeatherODE", "code_repository": "DAMO-DI-ML/WeatherODE", "code_commit": "056b961ea0bc3dbcae47b88c79cdbb257a8c8d47", "code_archive": "repos/UFzE9njwMG.zip", "code_archive_sha256": "4e025cbc7dc26fa56d55a1744075a50966487c9575df370d155026fb5dca2fe7", "code_archive_bytes": 83053, "code_file_count": 44, "code_extensions": {".py": 28, ".sh": 16}, "github_disk_usage_kb": 61, "github_languages": {"Python": 200463, "Shell": 17503}, "github_archived": false, "github_pushed_at": "2024-10-23T05:58:20Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/mitigating-time-discretization-challenges"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "sSWGqY2qNJ", "year": 2024, "status": "rejected", "title": "Indeterminate Probability Theory", "authors": ["Tao Yang", "Chuang Liu", "Xiaofeng Ma", "Weijia Lu", "Ning Wu", "Bingyang Li", "ZHIFEI YANG", "Peng Liu", "Lin Sun", "xiaodong Zhang", "Can Zhang"], "authorids": ["~Tao_Yang19", "~Chuang_Liu5", "~Xiaofeng_Ma1", "~Weijia_Lu1", "~Ning_Wu3", "~Bingyang_Li1", "~ZHIFEI_YANG2", "~Peng_Liu17", "~Lin_Sun11", "~xiaodong_Zhang6", "~Can_Zhang4"], "authors_source": "OpenReview API", "abstract": "Currently, there is no mathematical analytical form for a general posterior. We have discovered a new theory to address this issue, which is called Indeterminate Probability Theory. This is a big discovery in the field of probability, and it is an extension of classical probability theory, and makes classical probability theory a special case to our theory. In this paper, we propose a new perspective for understanding probability theory by introducing Observers and treating the outcome of each random experiment as an indeterminate probability distribution, which leads to probability calculations being a combination of ground truth and observation errors. We then discover three conditional mutual independent assumptions as Candidate Axioms and divide the probability process into two phase: observation phase and inference phase. In the observation phase, a general equation for any complex posterior is derived. In the inference phase, the inference probability equation with the posterior is derived. Base on this theory, we propose a new general model called IPNN - Indeterminate Probability Neural Network to validate our theory. Furthermore, in one of our another papers, this new theory is successfully applied to the task of multivariate time series (MTS) forecasting without relying on any neural models, and it outperforms LSTM models as well as some transformer-based models. In addition, further applications of this new theory are also discussed in this paper. Validations of this theory are reflected in experimental results.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=sSWGqY2qNJ", "arxiv_id": "2303.11536", "paper_pdf": "papers/sSWGqY2qNJ.pdf", "paper_pdf_sha256": "3b6ee3a5860119d31030839d96a66511286f45a5efda551896e11968e43a488d", "paper_pdf_bytes": 1180207, "paper_pdf_source": "openreview", "code_url": "https://github.com/Starfruit007/ipnn", "code_repository": "Starfruit007/ipnn", "code_commit": "987666de7d4bead745d498aa17a80f05bad3fe06", "code_archive": "repos/sSWGqY2qNJ.zip", "code_archive_sha256": "d3134f21e683002d35f966b950d36aca435fc77188cd44b5788cea9531912ce9", "code_archive_bytes": 68900, "code_file_count": 12, "code_extensions": {".py": 12}, "github_disk_usage_kb": 54, "github_languages": {"Python": 84780}, "github_archived": false, "github_pushed_at": "2023-10-05T05:37:33Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/indeterminate-probability-neural-network"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "dfPuLye6RvY", "year": 2023, "status": "rejected", "title": "Light-weight probing of unsupervised representations for Reinforcement Learning", "authors": ["Wancong Zhang", "Anthony GX-Chen", "Vlad Sobal", "Yann LeCun", "Nicolas Carion"], "authorids": ["~Wancong_Zhang1", "~Anthony_GX-Chen1", "~Vlad_Sobal1", "~Yann_LeCun1", "~Nicolas_Carion1"], "authors_source": "OpenReview API", "abstract": "Unsupervised visual representation learning offers the opportunity to leverage large corpora of unlabeled trajectories to form useful visual representations, which can benefit the training of reinforcement learning (RL) algorithms. However, evaluating the fitness of such representations requires training RL algorithms which is both computationally intensive and has high variance outcomes. To alleviate this issue, we design an evaluation protocol for unsupervised RL representations with lower variance and up to 600x lower computational cost. Inspired by the vision community, we propose two linear probing tasks: predicting the reward observed in a given state, and predicting the action of an expert in a given state. These two tasks are generally applicable to many RL domains, and we show through rigorous experimentation that they correlate strongly with the actual downstream control performance on the Atari100k Benchmark. This provides a better method for exploring the space of pretraining algorithms without the need of running RL evaluations for every setting. Leveraging this framework, we further improve existing self-supervised learning (SSL) recipes for RL, highlighting the importance of the forward model, the size of the visual backbone, and the precise formulation of the unsupervised objective.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=dfPuLye6RvY", "arxiv_id": "2208.12345", "paper_pdf": "papers/dfPuLye6RvY.pdf", "paper_pdf_sha256": "71b212d02add11fa5b302f115318ea399f09a8404dfe9fac8943cdba9fbc3fae", "paper_pdf_bytes": 595693, "paper_pdf_source": "openreview", "code_url": "https://github.com/kevinghst/lightweight_rl_probe", "code_repository": "kevinghst/lightweight_rl_probe", "code_commit": "0467b1f0f5aecafa780a52c85be724b493fd6a4e", "code_archive": "repos/dfPuLye6RvY.zip", "code_archive_sha256": "267b2bdafe934d3e416b35e949ef679a24b09c356ca552af2bc636a2cf8c251a", "code_archive_bytes": 80419, "code_file_count": 37, "code_extensions": {".sh": 22, ".py": 15}, "github_disk_usage_kb": 93, "github_languages": {"Python": 182716, "Shell": 41459}, "github_archived": false, "github_pushed_at": "2022-08-13T15:24:49Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/light-weight-probing-of-unsupervised"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "e_FK_rDajEv", "year": 2022, "status": "rejected", "title": "Learning Neural Causal Models with Active Interventions", "authors": ["Nino Scherrer", "Olexa Bilaniuk", "Yashas Annadani", "Anirudh Goyal", "Patrick Schwab", "Bernhard Schölkopf", "Michael Curtis Mozer", "Yoshua Bengio", "Stefan Bauer", "Nan Rosemary Ke"], "authorids": ["~Nino_Scherrer1", "~Olexa_Bilaniuk1", "~Yashas_Annadani1", "~Anirudh_Goyal1", "~Patrick_Schwab1", "~Bernhard_Schölkopf1", "~Michael_Curtis_Mozer1", "~Yoshua_Bengio1", "~Stefan_Bauer1", "~Nan_Rosemary_Ke1"], "authors_source": "OpenReview API", "abstract": "Discovering causal structures from data is a challenging inference problem of fundamental importance in all areas of science. The appealing scaling properties of neural networks have recently led to a surge of interest in differentiable neural network-based methods for learning causal structures from data. So far, differentiable causal discovery has focused on static datasets of observational or interventional origin. In this work, we introduce an active intervention-targeting mechanism which enables quick identification of the underlying causal structure of the data-generating process. Our method significantly reduces the required number of interactions compared with random intervention targeting and is applicable for both discrete and continuous optimization formulations of learning the underlying directed acyclic graph (DAG) from data. We examine the proposed method across multiple frameworks in a wide range of settings and demonstrate superior performance on multiple benchmarks from simulated to real-world data. ", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=e_FK_rDajEv", "arxiv_id": "2109.02429", "paper_pdf": "papers/e_FK_rDajEv.pdf", "paper_pdf_sha256": "fc766fc4f6b331923d57c293ffc6373f87655de76f179edcdb08931e7cd069ad", "paper_pdf_bytes": 3777007, "paper_pdf_source": "openreview", "code_url": "https://github.com/nke001/causal_learning_unknown_interventions", "code_repository": "nke001/causal_learning_unknown_interventions", "code_commit": "e41302aad36b0a5fedf6211d18cecca202c5e6b1", "code_archive": "repos/e_FK_rDajEv.zip", "code_archive_sha256": "4535c5d17fc2b03f7d2a9f9a5d8d02d4a96e2cc83e0ec9fdd979aa0b54367281", "code_archive_bytes": 61747, "code_file_count": 14, "code_extensions": {".py": 12, ".c": 1, ".m": 1}, "github_disk_usage_kb": 88, "github_languages": {"C": 95203, "Python": 91114, "MATLAB": 1633}, "github_archived": false, "github_pushed_at": "2020-07-08T01:59:19Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/learning-neural-causal-models-with-active"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "r1x_DaVKwH", "year": 2020, "status": "rejected", "title": "Is Deep Reinforcement Learning Really Superhuman on Atari? Leveling the playing field", "authors": ["Marin Toromanoff", "Emilie Wirbel", "Fabien Moutarde"], "authorids": ["marin.toromanoff@mines-paristech.fr", "emilie.wirbel@valeo.com", "fabien.moutarde@mines-paristech.fr"], "authors_source": "OpenReview API", "abstract": "Consistent and reproducible evaluation of Deep Reinforcement Learning (DRL) is not straightforward. In the Arcade Learning Environment (ALE), small changes in environment parameters such as stochasticity or the maximum allowed play time can lead to very different performance. In this work, we discuss the difficulties of comparing different agents trained on ALE. In order to take a step further towards reproducible and comparable DRL, we introduce SABER, a Standardized Atari BEnchmark for general Reinforcement learning algorithms. Our methodology extends previous recommendations and contains a complete set of environment parameters as well as train and test procedures. We then use SABER to evaluate the current state of the art, Rainbow. Furthermore, we introduce a human world records baseline, and argue that previous claims of expert or superhuman performance of DRL might not be accurate. Finally, we propose Rainbow-IQN by extending Rainbow with Implicit Quantile Networks (IQN) leading to new state-of-the-art performance. Source code is available for reproducibility.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=r1x_DaVKwH", "arxiv_id": "1908.04683", "paper_pdf": "papers/r1x_DaVKwH.pdf", "paper_pdf_sha256": "97bdef21666bc1ff43ec061d8272c8dc2122a19cfa1357789c228a96458f5d61", "paper_pdf_bytes": 685654, "paper_pdf_source": "openreview", "code_url": "https://github.com/valeoai/rainbow-iqn-apex", "code_repository": "valeoai/rainbow-iqn-apex", "code_commit": "7e71986703e112eecb943104fb2a9403c7cdb88a", "code_archive": "repos/r1x_DaVKwH.zip", "code_archive_sha256": "313b3551d5a10273acda2bb1779361df8a147600a3ae6c5d50dd9d263229bb06", "code_archive_bytes": 69382, "code_file_count": 17, "code_extensions": {".py": 15, ".sh": 2}, "github_disk_usage_kb": 149, "github_languages": {"Python": 114228, "Shell": 2656, "Dockerfile": 856}, "github_archived": false, "github_pushed_at": "2019-12-17T10:13:06Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/is-deep-reinforcement-learning-really"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "BygMAiRqK7", "year": 2019, "status": "rejected", "title": "Entropic GANs meet VAEs: A Statistical Approach to Compute Sample Likelihoods in GANs", "authors": ["Yogesh Balaji", "Hamed Hasani", "Rama Chellappa", "Soheil Feizi"], "authorids": ["yogesh@cs.umd.edu", "hassani@seas.upenn.edu", "rama@umiacs.umd.edu", "sfeizi@cs.umd.edu"], "authors_source": "OpenReview API", "abstract": "Building on the success of deep learning, two modern approaches to learn a probability model of the observed data are Generative Adversarial Networks (GANs) and Variational AutoEncoders (VAEs). VAEs consider an explicit probability model for the data and compute a generative distribution by maximizing a variational lower-bound on the log-likelihood function. GANs, however, compute a generative model by minimizing a distance between observed and generated probability distributions without considering an explicit model for the observed data. The lack of having explicit probability models in GANs prohibits computation of sample likelihoods in their frameworks and limits their use in statistical inference problems. In this work, we show that an optimal transport GAN with the entropy regularization can be viewed as a generative model that maximizes a lower-bound on average sample likelihoods, an approach that VAEs are based on. In particular, our proof constructs an explicit probability model for GANs that can be used to compute likelihood statistics within GAN's framework. Our numerical results on several datasets demonstrate consistent trends with the proposed theory. ", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=BygMAiRqK7", "arxiv_id": "1810.04147", "paper_pdf": "papers/BygMAiRqK7.pdf", "paper_pdf_sha256": "6e4c3a3c5646afaf4c11523d5357c16ed99d3e338e7e43d15a12993e406440f5", "paper_pdf_bytes": 9986289, "paper_pdf_source": "openreview", "code_url": "https://github.com/yogeshbalaji/EntropicGANs_meet_VAEs", "code_repository": "yogeshbalaji/EntropicGANs_meet_VAEs", "code_commit": "8021a601fdbc3be04d2197b5b4201f75435ae739", "code_archive": "repos/BygMAiRqK7.zip", "code_archive_sha256": "d46c78704e653c6244741822a3951f05356978fcc55604dd340e6be45a67fbbf", "code_archive_bytes": 328019, "code_file_count": 22, "code_extensions": {".py": 22}, "github_disk_usage_kb": 1413, "github_languages": {"Python": 78205}, "github_archived": false, "github_pushed_at": "2019-11-22T22:38:13Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/entropic-gans-meet-vaes-a-statistical"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "97kU2E3UKU", "year": 2026, "status": "rejected", "title": "LLM-First Search: Self-Guided Exploration of the Solution Space", "authors": ["Nathan Herr", "Tim Rocktäschel", "Roberta Raileanu"], "authorids": ["~Nathan_Herr1", "~Tim_Rocktäschel1", "~Roberta_Raileanu2"], "authors_source": "OpenReview API", "abstract": "Large Language Models (LLMs) have demonstrated remarkable improvements in reasoning and planning through increased test-time compute, often by framing problem-solving as a search process. While methods like Monte Carlo Tree Search (MCTS) have proven effective in some domains, their reliance on fixed exploration hyperparameters limits their adaptability across tasks of varying difficulty, rendering them impractical or expensive in certain settings. In this paper, we propose \\textbf{LLM-First Search (LFS)}, a novel \\textit{LLM Self-Guided Search} method that removes the need for pre-defined search strategies by empowering the LLM to autonomously control the search process via self-guided exploration. Rather than relying on external heuristics or hardcoded policies, the LLM evaluates whether to pursue the current search path or explore alternative branches based on its internal scoring mechanisms. This enables more flexible and context-sensitive reasoning without requiring manual tuning or task-specific adaptation. We evaluate LFS on Countdown and Sudoku against three classic widely-used search algorithms, Tree-of-Thoughts' Breadth First Search (ToT-BFS), Best First Search (BestFS), and MCTS, each of which have been used to achieve SotA results on a range of challenging reasoning tasks. We found that LFS (1) performs better on more challenging tasks without additional tuning, (2) is more computationally efficient compared to the other methods, especially when powered by a stronger model, (3) scales better with stronger models, due to its LLM-First design, and (4) scales better with increased compute budget. Our code will become publicly available upon acceptance.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=97kU2E3UKU", "arxiv_id": "2506.05213", "paper_pdf": "papers/97kU2E3UKU.pdf", "paper_pdf_sha256": "fcf6ee94339c2ef0a81bb64bffdbda4b4d2740acc77a9691f54fab46a2662724", "paper_pdf_bytes": 1215370, "paper_pdf_source": "openreview", "code_url": "https://github.com/NathanHerr/LLM-First-Search", "code_repository": "NathanHerr/LLM-First-Search", "code_commit": "3025bdaa3add6f41388c1d5a6d354522489d312e", "code_archive": "repos/97kU2E3UKU.zip", "code_archive_sha256": "089f4ed6ec86e9455027f34e33e46a7b3f9e59af1fd230b84157b57ebb1ba822", "code_archive_bytes": 533322, "code_file_count": 21, "code_extensions": {".py": 18, ".sh": 3}, "github_disk_usage_kb": 616, "github_languages": {"Python": 337568, "Shell": 10071}, "github_archived": false, "github_pushed_at": "2025-06-09T15:11:59Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/llm-first-search-self-guided-exploration-of"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Hxm0hOxph2", "year": 2025, "status": "rejected", "title": "On Provable Length and Compositional Generalization", "authors": ["Kartik Ahuja", "Amin Mansouri"], "authorids": ["~Kartik_Ahuja1", "~Amin_Mansouri1"], "authors_source": "OpenReview API", "abstract": "Out-of-distribution generalization capabilities of sequence-to-sequence models can be studied from the lens of two crucial forms of generalization: length generalization --  the ability to generalize to longer sequences than ones seen during training, and compositional generalization: the ability to generalize to token combinations not seen during training. In this work, we provide first provable guarantees on length and compositional generalization for common sequence-to-sequence models -- deep sets, transformers, state space models, and recurrent neural nets -- trained to minimize the prediction error. Taking a first principles perspective, we study the realizable case, i.e., the labeling function is realizable on the architecture.  We show that simple limited capacity versions of these different architectures achieve  both length and compositional generalization. Across different architectures, we also find that a linear relationship between the learned representation and the representation in the labeling function is necessary for length and compositional generalization.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=Hxm0hOxph2", "arxiv_id": "2402.04875", "paper_pdf": "papers/Hxm0hOxph2.pdf", "paper_pdf_sha256": "97cb08d76fd8c04dcc5212dc922ab907556a628cd687f8a3c4e0e4f3d0e251b0", "paper_pdf_bytes": 1863816, "paper_pdf_source": "openreview", "code_url": "https://github.com/facebookresearch/Length-and-Compositional-Generalization", "code_repository": "facebookresearch/Length-and-Compositional-Generalization", "code_commit": "5a77eb14d7b378bce0d0c6c69716fea8a038de05", "code_archive": "repos/Hxm0hOxph2.zip", "code_archive_sha256": "1a82e6860437e1d53716d77968f3fd360506dc4be93c9ca97105341d45a9953a", "code_archive_bytes": 72778, "code_file_count": 15, "code_extensions": {".py": 15}, "github_disk_usage_kb": 63, "github_languages": {"Python": 93691}, "github_archived": true, "github_pushed_at": "2024-07-19T15:24:15Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/on-provable-length-and-compositional"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "MpWRCiw8g5", "year": 2024, "status": "rejected", "title": "JOSENet: A Joint Stream Embedding Network for Violence Detection in Surveillance Videos", "authors": ["Pietro Nardelli", "Danilo Comminiello"], "authorids": ["~Pietro_Nardelli1", "~Danilo_Comminiello1"], "authors_source": "OpenReview API", "abstract": "Due to the ever-increasing availability of video surveillance cameras and the growing need for crime prevention, the violence detection task is attracting greater attention from the research community. With respect to other action recognition tasks, violence detection in surveillance videos shows additional issues. Indeed, violence detection requires real-world fights from surveillance cameras and available datasets seem to be very small compared with other action recognition datasets. Moreover, in surveillance applications, people in the scenes always differ for each video and the background of the footage differs for each camera. Also, violent actions in real-life surveillance videos must be detected quickly to prevent unwanted consequences, thus models would definitely benefit from a reduction in memory usage and computational costs. Such problems make classical action recognition methods difficult to be adopted. To tackle such problems, we introduce JOSENet, a novel self-supervised framework that provides outstanding performance for violence detection in surveillance videos. The proposed model receives two spatiotemporal video streams, i.e., RGB frames and optical flows, and involves a new regularized self-supervised learning approach for videos. JOSENet provides improved performance while requiring one-fourth of the number of frames per video segment and a reduced frame rate compared to state-of-the-art methods.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=MpWRCiw8g5", "arxiv_id": "2405.02961", "paper_pdf": "papers/MpWRCiw8g5.pdf", "paper_pdf_sha256": "7b0208b7c9f39304499239d936852dfccbfedeae8904035f81d0232131213fda", "paper_pdf_bytes": 3995719, "paper_pdf_source": "openreview", "code_url": "https://github.com/ispamm/JOSENet", "code_repository": "ispamm/JOSENet", "code_commit": "3f9512e9e08e6c4e1cd983993fa57369996bc79e", "code_archive": "repos/MpWRCiw8g5.zip", "code_archive_sha256": "ec68abe8eb5ca4bf2b712adbff0531d15077181bb31dacaa846742f3d2c2888a", "code_archive_bytes": 28501, "code_file_count": 10, "code_extensions": {".py": 10}, "github_disk_usage_kb": 56, "github_languages": {"Python": 135093}, "github_archived": false, "github_pushed_at": "2025-06-05T08:15:26Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/josenet-a-joint-stream-embedding-network-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "MCe881WzBr0", "year": 2023, "status": "rejected", "title": "Variational Classification", "authors": ["Shehzaad Zuzar Dhuliawala", "Mrinmaya Sachan", "Carl Allen"], "authorids": ["~Shehzaad_Zuzar_Dhuliawala3", "~Mrinmaya_Sachan3", "~Carl_Allen1"], "authors_source": "OpenReview API", "abstract": "Classification tasks, ubiquitous across machine learning, are commonly tackled by a suitably designed neural network with a softmax output layer, mapping each data point to a categorical distribution over class labels. \nWe extend this familiar model from a latent variable perspective to variational classification (VC), analogous to how the variational auto-encoder relates to its deterministic counterpart. We derive a training objective based on the ELBO together with an \\textit{adversarial} approach for optimising it.\n\nWithin this framework, we identify design choices made implicitly in off-the-shelf softmax functions and can instead include domain-specific assumptions, such as class-conditional latent priors. We demonstrate benefits of the VC model in image classification. We show on several standard datasets, that treating inputs to the softmax layer as latent variables under a mixture of Gaussians prior, improves several desirable aspects of a classifier, such as prediction accuracy, calibration, out-of-domain calibration and adversarial robustness.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=MCe881WzBr0", "arxiv_id": "2305.10406", "paper_pdf": "papers/MCe881WzBr0.pdf", "paper_pdf_sha256": "c5b40d01fc7e7e96c8f49ffda7db42285d62f702c5ebdb946b102015af38851a", "paper_pdf_bytes": 851388, "paper_pdf_source": "openreview", "code_url": "https://github.com/shehzaadzd/variational-classification", "code_repository": "shehzaadzd/variational-classification", "code_commit": "e7463a33d97a60497671b06167d26f813b61d67d", "code_archive": "repos/MCe881WzBr0.zip", "code_archive_sha256": "b40f637c7cced1bf34a2bdfbacd66f69055fe89fb95ce226e5ae94554f8bdf0e", "code_archive_bytes": 22906, "code_file_count": 8, "code_extensions": {".py": 5, ".sh": 3}, "github_disk_usage_kb": 96, "github_languages": {"Python": 82521, "Shell": 269}, "github_archived": false, "github_pushed_at": "2023-12-22T18:11:03Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/variational-classification"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "7vXQJ2QW8hR", "year": 2022, "status": "rejected", "title": "Max-Affine Spline Insights Into Deep Network Pruning", "authors": ["Randall Balestriero", "Haoran You", "Zhihan Lu", "Yutong Kou", "Huihong Shi", "Yingyan Lin", "Richard Baraniuk"], "authorids": ["~Randall_Balestriero1", "~Haoran_You1", "~Zhihan_Lu1", "~Yutong_Kou1", "~Huihong_Shi1", "~Yingyan_Lin1", "~Richard_Baraniuk1"], "authors_source": "OpenReview API", "abstract": "State-of-the-art (SOTA) approaches to deep network (DN) training overparametrize the model and then prune a posteriori to obtain a \"winning ticket'' subnetwork that can be trained from scratch to achieve high accuracy. To date, the literature has remained largely empirical and hence provides little insights into how pruning affects a DN's decision boundary and no guidance regarding how to design a principled pruning technique. Using a recently developed spline interpretation of DNs, we develop new theory and visualization tools that provide new insights into how pruning DN nodes affects the decision boundary. We discover that a DN's spline mappings exhibit an early-bird (EB) phenomenon whereby the spline's partition converges at early training stages, bridging the recently developed max-affine spline theory and lottery ticket hypothesis of DNs. We leverage this new insight to develop a principled and efficient pruning strategy that focuses on a tiny fraction of DN nodes whose corresponding spline partition regions actually contribute to the final decision boundary. Extensive experiments on four networks and three datasets validate that our new spline-based DN pruning approach reduces training FLOPs by up to 3.5x while achieving similar or even better accuracy than state-of-the-art methods. All the codes will be released publicly upon acceptance.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=7vXQJ2QW8hR", "arxiv_id": "2101.02338", "paper_pdf": "papers/7vXQJ2QW8hR.pdf", "paper_pdf_sha256": "c37f967fb01869cb33073086d253227ac408f24746f8219f15fb9e8c76c2ba6d", "paper_pdf_bytes": 2933925, "paper_pdf_source": "openreview", "code_url": "https://github.com/GATECH-EIC/Spline-EB", "code_repository": "GATECH-EIC/Spline-EB", "code_commit": "f26909ec902cbb752d1192ad6c7bcb8a3c835fb2", "code_archive": "repos/7vXQJ2QW8hR.zip", "code_archive_sha256": "e88575856ef856aab097fff70b6bb4b14fc673cdde371d7e94d4d81c9b30ab51", "code_archive_bytes": 105140, "code_file_count": 28, "code_extensions": {".py": 28}, "github_disk_usage_kb": 89, "github_languages": {"Python": 183056}, "github_archived": false, "github_pushed_at": "2022-11-12T03:22:13Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/max-affine-spline-insights-into-deep-network-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "RgDq8-AwvtN", "year": 2021, "status": "rejected", "title": "Model-Based Robust Deep Learning: Generalizing to Natural, Out-of-Distribution Data", "authors": ["Alexander Robey", "Hamed Hassani", "George J. Pappas"], "authorids": ["~Alexander_Robey1", "~Hamed_Hassani2", "~George_J._Pappas1"], "authors_source": "OpenReview API", "abstract": "While deep learning (DL) has resulted in major breakthroughs in many applications, the frameworks commonly used in DL remain fragile to seemingly innocuous changes in the data. In response, adversarial training has emerged as a principled approach for improving the robustness of DL against norm-bounded perturbations.  Despite this progress, DL is also known to be fragile to unbounded shifts in the data distribution due to many forms of natural variation, including changes in weather or lighting in images.  However, there are remarkably few techniques that can address robustness to natural, out-of-distribution shifts in the data distribution in a general context.  To address this gap, we propose a paradigm shift from perturbation-based adversarial robustness to model-based robust deep learning.  Critical to our paradigm is to obtain models of natural variation, which vary data over a range of natural conditions.  Then by exploiting these models, we develop three novel model-based robust training algorithms that improve the robustness of DL with respect to natural variation. Our extensive experiments show that across a variety of natural conditions in twelve distinct datasets, classifiers trained with our algorithms significantly outperform classifiers trained via ERM, adversarial training, and domain adaptation techniques.  Specifically, when training on ImageNet and testing on various subsets of ImageNet-c, our algorithms improve over baseline methods by up to 30 percentage points in top-1 accuracy.  Further, we show that our methods provide robustness (1) against natural, out-of-distribution data, (2) against multiple simultaneous distributional shifts, and (3) to domains entirely unseen during training.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=RgDq8-AwvtN", "arxiv_id": "2005.10247", "paper_pdf": "papers/RgDq8-AwvtN.pdf", "paper_pdf_sha256": "7446633b0f2e826d700e4ec66f4e3c49730bf220e37cd00790e6545678f86f96", "paper_pdf_bytes": 3137186, "paper_pdf_source": "openreview", "code_url": "https://github.com/arobey1/mbrdl", "code_repository": "arobey1/mbrdl", "code_commit": "f49bd0514850b109689a371334f585dc85d172a3", "code_archive": "repos/RgDq8-AwvtN.zip", "code_archive_sha256": "571fc65ef1ef57fd6a65adf6a858d5e424f5621197af99143f73260dab805925", "code_archive_bytes": 55987, "code_file_count": 28, "code_extensions": {".py": 24, ".sh": 4}, "github_disk_usage_kb": 113, "github_languages": {"Python": 139936, "Shell": 5775}, "github_archived": false, "github_pushed_at": "2021-01-22T08:14:05Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/model-based-robust-deep-learning"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "S1x0CnEtvB", "year": 2020, "status": "rejected", "title": "AutoGrow: Automatic Layer Growing in Deep Convolutional Networks", "authors": ["Wei Wen", "Feng Yan", "Hai Li"], "authorids": ["wei.wen@duke.edu", "fyan@unr.edu", "hai.li@duke.edu"], "authors_source": "OpenReview API", "abstract": "Depth is a key component of Deep Neural Networks (DNNs), however, designing depth is heuristic and requires many human efforts. We propose AutoGrow to automate depth discovery in DNNs: starting from a shallow seed architecture, AutoGrow grows new layers if the growth improves the accuracy; otherwise, stops growing and thus discovers the depth. We propose robust growing and stopping policies to generalize to different network architectures and datasets. Our experiments show that by applying the same policy to different network architectures, AutoGrow can always discover near-optimal depth on various datasets of MNIST, FashionMNIST, SVHN, CIFAR10, CIFAR100 and ImageNet. For example, in terms of accuracy-computation trade-off, AutoGrow discovers a better depth combination in ResNets than human experts. Our AutoGrow is efficient. It discovers depth within similar time of training a single DNN.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=S1x0CnEtvB", "arxiv_id": "1906.02909", "paper_pdf": "papers/S1x0CnEtvB.pdf", "paper_pdf_sha256": "f67b3aed2d1282528c779172ebe52e274de7fe0ca6c19144324bbec0476a7a06", "paper_pdf_bytes": 1987332, "paper_pdf_source": "openreview", "code_url": "https://github.com/wenwei202/autogrow", "code_repository": "wenwei202/autogrow", "code_commit": "4435ee057126bacfc85063791ee2008b32c75cad", "code_archive": "repos/S1x0CnEtvB.zip", "code_archive_sha256": "5047b39fe379f14078e06f2413dddf382b2d989a94f507599f883b804ba0b67e", "code_archive_bytes": 61234, "code_file_count": 31, "code_extensions": {".py": 31}, "github_disk_usage_kb": 152, "github_languages": {"Python": 202941}, "github_archived": false, "github_pushed_at": "2019-06-10T05:25:41Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/autogrow-automatic-layer-growing-in-deep"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "SkxxIs0qY7", "year": 2019, "status": "rejected", "title": "CoT: Cooperative Training for Generative Modeling of Discrete Data", "authors": ["Sidi Lu", "Lantao Yu", "Siyuan Feng", "Yaoming Zhu", "Weinan Zhang", "Yong Yu"], "authorids": ["steve_lu@apex.sjtu.edu.cn", "yulantao@apex.sjtu.edu.cn", "siyuanfeng@apex.sjtu.edu", "ymzhu@apex.sjtu.edu.cn", "wnzhang@apex.sjtu.edu.cn", "yyu@apex.sjtu.edu.cn"], "authors_source": "OpenReview API", "abstract": "We propose Cooperative Training (CoT) for training generative models that measure a tractable density for discrete data. CoT coordinately trains a generator G and an auxiliary predictive mediator M. The training target of M is to estimate a mixture density of the learned distribution G and the target distribution P, and that of G is to minimize the Jensen-Shannon divergence estimated through M. CoT achieves independent success without the necessity of pre-training via Maximum Likelihood Estimation or involving high-variance algorithms like REINFORCE. This low-variance algorithm is theoretically proved to be superior for both sample generation and likelihood prediction. We also theoretically and empirically show the superiority of CoT over most previous algorithms in terms of generative quality and diversity, predictive generalization ability and computational cost.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=SkxxIs0qY7", "arxiv_id": "1804.03782", "paper_pdf": "papers/SkxxIs0qY7.pdf", "paper_pdf_sha256": "f5f03e901fe5739c0d7abbf8bce2874ba9af8ff7268313a74f4cf2c9e1b2b34c", "paper_pdf_bytes": 800823, "paper_pdf_source": "openreview", "code_url": "https://github.com/desire2020/CoT", "code_repository": "desire2020/CoT", "code_commit": "05f390817e11733b7e4328ed2677e63287ed1290", "code_archive": "repos/SkxxIs0qY7.zip", "code_archive_sha256": "91e3088ad8fbe773cd6a989db6135b4dd01ecc04c03047927d9bca56d3e8a2a2", "code_archive_bytes": 1279051, "code_file_count": 5, "code_extensions": {".py": 5}, "github_disk_usage_kb": 1513, "github_languages": {"Python": 37310}, "github_archived": false, "github_pushed_at": "2019-05-25T14:01:57Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/cot-cooperative-training-for-generative"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "QDLGcnYdxe", "year": 2026, "status": "rejected", "title": "Rrls: Robust reinforcement learning suite", "authors": ["Adil Zouitine", "David Bertoin", "Pierre Clavier", "Matthieu Geist", "Emmanuel Rachelson"], "authorids": ["~Adil_Zouitine1", "~David_Bertoin1", "~Pierre_Clavier1", "~Matthieu_Geist1", "~Emmanuel_Rachelson1"], "authors_source": "OpenReview API", "abstract": "Robust reinforcement learning is the problem of learning control policies that provide optimal worst-case performance against a span of adversarial environments. It is a crucial ingredient for deploying algorithms in real-world scenarios with prevalent environmental uncertainties and has been a long-standing object of attention in the community, without a standardized set of benchmarks. This contribution endeavors to fill this gap. We introduce the Robust Reinforcement Learning Suite (RRLS), a benchmark suite based on Mujoco environments. RRLS provides six continuous control tasks with two types of uncertainty sets for training and evaluation. Our benchmark aims to standardize robust reinforcement learning tasks, facilitating reproducible and comparable experiments, in particular those from recent state-of-the-art contributions, for which we demonstrate the use of RRLS. It is also designed to be easily expandable to new environments.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=QDLGcnYdxe", "arxiv_id": "2406.08406", "paper_pdf": "papers/QDLGcnYdxe.pdf", "paper_pdf_sha256": "43a7f18574ae397f1b4dadb16f9f1273ef025c0d3668c35949ca9ce7bd5e2277", "paper_pdf_bytes": 1728928, "paper_pdf_source": "openreview", "code_url": "https://github.com/SuReLI/RRLS", "code_repository": "SuReLI/RRLS", "code_commit": "0c43dc981c0e892a604c596e2bc399df187a87ac", "code_archive": "repos/QDLGcnYdxe.zip", "code_archive_sha256": "44ddf88d1894f811c6988f16bfcdcfe26fae28feed725dd73c86d9126c29648a", "code_archive_bytes": 682944, "code_file_count": 27, "code_extensions": {".py": 27}, "github_disk_usage_kb": 736, "github_languages": {"Python": 142354}, "github_archived": false, "github_pushed_at": "2024-12-24T11:20:18Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/rrls-robust-reinforcement-learning-suite"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "nwZHFKrYTB", "year": 2025, "status": "rejected", "title": "How to Train Long-Context Language Models (Effectively)", "authors": ["Tianyu Gao", "Alexander Wettig", "Howard Yen", "Danqi Chen"], "authorids": ["~Tianyu_Gao1", "~Alexander_Wettig1", "~Howard_Yen1", "~Danqi_Chen1"], "authors_source": "OpenReview API", "abstract": "We study the problem of adapting a language model (LM) to make effective use of long-context information. We first establish a reliable evaluation protocol to guide model development---instead of perplexity, we use a broad set of long-context tasks, and we evaluate models after supervised fine-tuning (SFT) with instruction data as this better reveals long-context abilities. Supported by our robust evaluations, we run thorough experiments to decide the data mix for continued pre-training, the instruction tuning dataset, and other design choices such as position extrapolation. We find that (1) code repositories and books are excellent sources of long data, but it is crucial to combine them with high-quality short data; (2) training with a sequence length beyond the evaluation length boosts long-context performance; (3) for SFT, using only short instruction datasets yields strong performance on long-context tasks. Our final model, ProLong-8B, which is initialized from Llama-3 and trained on 40B tokens, demonstrates state-of-the-art long-context performance among similarly sized models at a length of 128K, outperforming Llama-3.1-8B on the majority of long-context tasks despite having seen 5% as many tokens during long-context training. Additionally, ProLong can effectively process up to 512K tokens, one of the longest context windows of publicly available LMs.", "decision": "Reject", "meta_review": "", "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=nwZHFKrYTB", "arxiv_id": "2410.02660", "paper_pdf": "papers/nwZHFKrYTB.pdf", "paper_pdf_sha256": "9c00e32fddd31390bb4fc30c6486f51ac28d5821179e93d45213e07268aa9bb6", "paper_pdf_bytes": 412185, "paper_pdf_source": "openreview", "code_url": "https://github.com/princeton-nlp/ProLong", "code_repository": "princeton-nlp/ProLong", "code_commit": "499fa2994fa727b4c8b53b706ccd5c23f77d7f89", "code_archive": "repos/nwZHFKrYTB.zip", "code_archive_sha256": "ca1ffa9dbaa9002e596244eb11e481237cbfb1597e857a355173b14e901717f9", "code_archive_bytes": 43146, "code_file_count": 8, "code_extensions": {".py": 5, ".sh": 3}, "github_disk_usage_kb": 64, "github_languages": {"Python": 105806, "Shell": 14272}, "github_archived": false, "github_pushed_at": "2025-09-12T19:56:21Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/how-to-train-long-context-language-models"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "5oJlyJXUxK", "year": 2024, "status": "rejected", "title": "Beyond Concept Bottleneck Models: How to Make Black Boxes Intervenable?", "authors": ["Ričards Marcinkevičs", "Sonia Laguna", "Moritz Vandenhirtz", "Julia E Vogt"], "authorids": ["~Ričards_Marcinkevičs1", "~Sonia_Laguna1", "~Moritz_Vandenhirtz1", "~Julia_E_Vogt1"], "authors_source": "OpenReview API", "abstract": "Recently, interpretable machine learning has re-explored concept bottleneck models (CBM), comprising step-by-step prediction of the high-level concepts from the raw features and the target variable from the predicted concepts. A compelling advantage of this model class is the user's ability to intervene on the predicted concept values, consequently affecting the model's downstream output. In this work, we introduce a method to perform such concept-based interventions on already-trained neural networks, which are not interpretable by design. Furthermore, we formalise the model's *intervenability* as a measure of the effectiveness of concept-based interventions and leverage this definition to fine-tune black-box models. Empirically, we explore the intervenability of black-box classifiers on synthetic tabular and natural image benchmarks.  We demonstrate that fine-tuning improves intervention effectiveness and often yields better-calibrated predictions. To showcase the practical utility of the proposed techniques, we apply them to deep chest X-ray classifiers and show that fine-tuned black boxes can be as intervenable and more performant than CBMs.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=5oJlyJXUxK", "arxiv_id": "2401.13544", "paper_pdf": "papers/5oJlyJXUxK.pdf", "paper_pdf_sha256": "3a61726319456c88ec4d3597a1501751c4203d32b858a5e7804ae5c79593b348", "paper_pdf_bytes": 12053856, "paper_pdf_source": "openreview", "code_url": "https://github.com/sonialagunac/Beyond-CBM", "code_repository": "sonialagunac/Beyond-CBM", "code_commit": "d1c24659496558c2a79b9b51bf9077a2a157f12e", "code_archive": "repos/5oJlyJXUxK.zip", "code_archive_sha256": "e854cc92cac9672fde19b5f0c07895d10ea6f8c0df5b10a27c115a1dcca3948f", "code_archive_bytes": 69894, "code_file_count": 25, "code_extensions": {".py": 25}, "github_disk_usage_kb": 56, "github_languages": {"Python": 179737, "Shell": 1928}, "github_archived": false, "github_pushed_at": "2024-10-26T12:15:31Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/beyond-concept-bottleneck-models-how-to-make"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "lYZZl2hp1gp", "year": 2023, "status": "rejected", "title": "Robustness of Unsupervised Representation Learning without Labels", "authors": ["Aleksandar Petrov", "Marta Kwiatkowska"], "authorids": ["~Aleksandar_Petrov1", "~Marta_Kwiatkowska1"], "authors_source": "OpenReview API", "abstract": "Unsupervised representation learning leverages large unlabeled datasets and is competitive with supervised learning. But non-robust encoders may affect downstream task robustness. Recently, robust representation encoders have become of interest. Still, all prior work evaluates robustness using a downstream classification task. Instead, we propose a family of unsupervised robustness measures, which are model- and task-agnostic and label-free. We benchmark state-of-the-art representation encoders and show that none dominates the rest. We offer unsupervised extensions to the FGSM and PGD attacks. When used in adversarial training, they improve most unsupervised robustness measures, including certified robustness. We validate our results against a linear probe and show that, for MOCOv2, adversarial training results in 3 times higher certified accuracy, a 2-fold decrease in impersonation attack success rate and considerable improvements in certified robustness.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=lYZZl2hp1gp", "arxiv_id": "2210.04076", "paper_pdf": "papers/lYZZl2hp1gp.pdf", "paper_pdf_sha256": "372ad962d3643bff026649ca555969f5445934e26af1f142dfafe8f2cec8f3c0", "paper_pdf_bytes": 24651918, "paper_pdf_source": "openreview", "code_url": "https://github.com/AleksandarPetrov/unsupervised-robustness", "code_repository": "AleksandarPetrov/unsupervised-robustness", "code_commit": "b8606562c397516c2e68d9318861d0cd6dbbf312", "code_archive": "repos/lYZZl2hp1gp.zip", "code_archive_sha256": "127cb0ab9f20fcfa0c26c89f6a690946e420f5d7b0dd67681622f0533aeb938d", "code_archive_bytes": 117639, "code_file_count": 39, "code_extensions": {".py": 34, ".sh": 5}, "github_disk_usage_kb": 97, "github_languages": {"Python": 263016, "Shell": 7195}, "github_archived": false, "github_pushed_at": "2022-10-18T05:29:10Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/robustness-of-unsupervised-representation"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "tG8QrhMwEqS", "year": 2022, "status": "rejected", "title": "Adaptive Activation-based Structured Pruning", "authors": ["Kaiqi Zhao", "Animesh Jain", "Ming Zhao"], "authorids": ["~Kaiqi_Zhao2", "~Animesh_Jain1", "~Ming_Zhao2"], "authors_source": "OpenReview API", "abstract": "Pruning is a promising approach to compress complex deep learning models in order to deploy them on resource-constrained edge devices. However, many existing pruning solutions are based on unstructured pruning, which yield models that cannot efficiently run on commodity hardware, and require users to manually explore and tune the pruning process, which is time consuming and often leads to sub-optimal results. To address these limitations, this paper presents an adaptive, activation-based, structured pruning approach to automatically and efficiently generate small, accurate, and hardware-efficient models that meet user requirements. First, it proposes iterative structured pruning using activation-based attention feature maps to effectively identify and prune unimportant filters. Then, it proposes adaptive pruning policies for automatically meeting the pruning objectives of accuracy-critical, memory-constrained, and latency-sensitive tasks. A comprehensive evaluation shows that the proposed method can substantially outperform the state-of-the-art structured pruning works on CIFAR-10 and ImageNet datasets. For example, on ResNet-56 with CIFAR-10, without any accuracy drop, our method achieves the largest parameter reduction (79.11%), outperforming the related works by 22.81% to 66.07%, and the largest FLOPs reduction (70.13%), outperforming the related works by 14.13% to 26.53%.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=tG8QrhMwEqS", "arxiv_id": "2201.10520", "paper_pdf": "papers/tG8QrhMwEqS.pdf", "paper_pdf_sha256": "929e7656e34e04e658a6f1be45b178e0f979f6023e9e8d12ac6bafabd11d0d21", "paper_pdf_bytes": 3102856, "paper_pdf_source": "openreview", "code_url": "https://github.com/kaiqi123/Automatic-Attention-Pruning", "code_repository": "kaiqi123/Automatic-Attention-Pruning", "code_commit": "96f71d3eec651097b1c09dc0cb8412db94c3c8a6", "code_archive": "repos/tG8QrhMwEqS.zip", "code_archive_sha256": "ebf13c14f4ba78b14d0d12e3518b4a244bc6347eb361321e43c45b336952dce1", "code_archive_bytes": 56223, "code_file_count": 26, "code_extensions": {".py": 18, ".sh": 8}, "github_disk_usage_kb": 107, "github_languages": {"Python": 176097, "Shell": 9581, "Dockerfile": 228}, "github_archived": false, "github_pushed_at": "2023-10-25T18:07:13Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/adaptive-activation-based-structured-pruning-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "edku48LG0pT", "year": 2021, "status": "rejected", "title": "A Neural Network MCMC sampler that maximizes Proposal Entropy", "authors": ["ZENGYI LI", "Yubei Chen", "Friedrich Sommer"], "authorids": ["~ZENGYI_LI1", "~Yubei_Chen1", "~Friedrich_Sommer1"], "authors_source": "OpenReview API", "abstract": "Markov Chain Monte Carlo (MCMC) methods sample from unnormalized probability distributions and offer guarantees of exact sampling. However, in the continuous case, unfavorable geometry of the target distribution can greatly limit the efficiency of MCMC methods. Augmenting samplers with neural networks can potentially improve their efficiency. Previous neural network based samplers were trained with objectives that either did not explicitly encourage exploration, or used a L2 jump objective which could only be applied to well structured distributions. Thus it seems promising to instead maximize the proposal entropy for adapting the proposal to distributions of any shape. To allow direct optimization of the proposal entropy, we propose a neural network MCMC sampler that has a flexible and tractable proposal distribution. Specifically, our network architecture utilizes the gradient of the target distribution for generating proposals. Our model achieves significantly higher efficiency than previous neural network MCMC techniques in a variety of sampling tasks. Further, the sampler is applied on training of a convergent energy-based model of natural images. The learned sampler achieves significantly higher proposal entropy and sample quality compared to Langevin dynamics sampler.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=edku48LG0pT", "arxiv_id": "2010.03587", "paper_pdf": "papers/edku48LG0pT.pdf", "paper_pdf_sha256": "c17fb80293c6a7fa758d1a6c6c6f9a393d3cfc3c51356feed80224f13c682c0b", "paper_pdf_bytes": 1161195, "paper_pdf_source": "openreview", "code_url": "https://github.com/NEM-MC/Entropy-NeuralMC", "code_repository": "NEM-MC/Entropy-NeuralMC", "code_commit": "067dee6795d8403df1e5214f8e859c7c90c21b65", "code_archive": "repos/edku48LG0pT.zip", "code_archive_sha256": "7a18c1816a9daa537c6fabc0662ff068ec81199761c0c9720fd871925979eea6", "code_archive_bytes": 62520, "code_file_count": 4, "code_extensions": {".py": 3, ".ipynb": 1}, "github_disk_usage_kb": 136, "github_languages": {"Jupyter Notebook": 77677, "Python": 34163}, "github_archived": false, "github_pushed_at": "2020-10-01T00:58:44Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/a-neural-network-mcmc-sampler-that-maximizes-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "S1x1IkHtPr", "year": 2020, "status": "rejected", "title": "A Generative Model for Molecular Distance Geometry", "authors": ["Gregor N. C. Simm", "José Miguel Hernández-Lobato"], "authorids": ["gncsimm@gmail.com", "jmh233@cam.ac.uk"], "authors_source": "OpenReview API", "abstract": "Computing equilibrium states for many-body systems, such as molecules, is a long-standing challenge. In the absence of methods for generating statistically independent samples, great computational effort is invested in simulating these systems using, for example, Markov chain Monte Carlo. We present a probabilistic model that generates such samples for molecules from their graph representations. Our model learns a low-dimensional manifold that preserves the geometry of local atomic neighborhoods through a principled learning representation that is based on Euclidean distance geometry. We create a new dataset for molecular conformation generation with which we show experimentally that our generative model achieves state-of-the-art accuracy. Finally, we show how to use our model as a proposal distribution in an importance sampling scheme to compute molecular properties.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=S1x1IkHtPr", "arxiv_id": "1909.11459", "paper_pdf": "papers/S1x1IkHtPr.pdf", "paper_pdf_sha256": "69b96be331dcd0003e3063de83e05c5c16edeb6c710ab5cc23d6e43c8e96f01d", "paper_pdf_bytes": 5856860, "paper_pdf_source": "openreview", "code_url": "https://github.com/gncs/graphdg", "code_repository": "gncs/graphdg", "code_commit": "a138448de24fcc1ead3284e380ace5e88a2c6643", "code_archive": "repos/S1x1IkHtPr.zip", "code_archive_sha256": "f7b02985374294fa06c3269ebf1790feb2d8a311c64efbfdf29dab302d8a3428", "code_archive_bytes": 164480, "code_file_count": 32, "code_extensions": {".py": 32}, "github_disk_usage_kb": 169, "github_languages": {"Python": 101020}, "github_archived": false, "github_pushed_at": "2023-03-24T22:10:50Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/a-generative-model-for-molecular-distance-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "BkxAUjRqY7", "year": 2019, "status": "rejected", "title": "An Information-Theoretic Metric of Transferability for Task Transfer Learning", "authors": ["Yajie Bao", "Yang Li", "Shao-Lun Huang", "Lin Zhang", "Amir R. Zamir", "Leonidas J. Guibas"], "authorids": ["byjem123@163.com", "tori2011@gmail.com", "shaolun.huang@sz.tsinghua.edu.cn", "linzhang@tsinghua.edu.cn", "zamir@cs.stanford.edu", "guibas@cs.stanford.edu"], "authors_source": "OpenReview API", "abstract": "An important question in task transfer learning is to determine task transferability, i.e. given a common input domain, estimating to what extent representations learned from a source task can help in learning a target task. Typically, transferability is either measured experimentally or inferred through task relatedness, which is often defined without a clear operational meaning. In this paper, we present a novel metric, H-score, an easily-computable evaluation function that estimates the performance of transferred representations from one task to another in classification problems. Inspired by a principled information theoretic approach, H-score has a direct connection to the asymptotic error probability of the decision function based on the transferred feature. This formulation of transferability can further be used to select a suitable set of source tasks in task transfer learning problems or to devise efficient transfer learning policies. Experiments using both synthetic and real image data show that not only our formulation of transferability is meaningful in practice, but also it can generalize to inference problems beyond classification, such as recognition tasks for 3D indoor-scene understanding.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=BkxAUjRqY7", "arxiv_id": null, "paper_pdf": "papers/BkxAUjRqY7.pdf", "paper_pdf_sha256": "1b1717abdb36ac9eb325635441a0321092c26d241d2b3962d330c95959188195", "paper_pdf_bytes": 3466947, "paper_pdf_source": "openreview", "code_url": "https://github.com/YaojieBao/An-Information-theoretic-Metric-of-Transferability", "code_repository": "YaojieBao/An-Information-theoretic-Metric-of-Transferability", "code_commit": "9a3f2d3f2511b6825a9d18c5fd7becbb06cae0b1", "code_archive": "repos/BkxAUjRqY7.zip", "code_archive_sha256": "834820c1d2f042baf175e0d3140190a8dd160583ca325d69a57922c1590f8648", "code_archive_bytes": 1793811, "code_file_count": 13, "code_extensions": {".py": 13}, "github_disk_usage_kb": 1752, "github_languages": {"Python": 47878}, "github_archived": false, "github_pushed_at": "2019-03-01T06:38:44Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/an-information-theoretic-metric-of"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "MYqAKKsjF9", "year": 2026, "status": "rejected", "title": "LifelongAgentBench: Evaluating LLM Agents as Lifelong Learners", "authors": ["Junhao Zheng", "Xidi Cai", "Qiuke Li", "Duzhen Zhang", "Zhong-Zhi Li", "Yingying Zhang", "Le Song", "Qianli Ma"], "authorids": ["~Junhao_Zheng3", "~Xidi_Cai1", "~Qiuke_Li1", "~Duzhen_Zhang1", "~Zhong-Zhi_Li1", "~Yingying_Zhang2", "~Le_Song1", "~Qianli_Ma3"], "authors_source": "OpenReview API", "abstract": "Lifelong learning is essential for intelligent agents operating in dynamic environments. Current large language model (LLM)-based agents, however, remain stateless and unable to accumulate or transfer knowledge over time. Existing benchmarks treat agents as static systems and fail to evaluate lifelong learning capabilities. We present LifelongAgentBench, the first unified benchmark designed to systematically assess the lifelong learning ability of LLM agents. It provides skill-grounded, interdependent tasks across three interactive environments—Database, Operating System, and Knowledge Graph—with automatic label verification, reproducibility, and modular extensibility. Extensive experiments reveal that conventional experience replay has limited effectiveness for LLM agents due to irrelevant information and context length constraints. We further introduce a group self-consistency mechanism that significantly improves lifelong learning performance. We hope LifelongAgentBench will advance the development of adaptive, memory-capable LLM agents.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=MYqAKKsjF9", "arxiv_id": "2505.11942", "paper_pdf": "papers/MYqAKKsjF9.pdf", "paper_pdf_sha256": "d829274949b3bf7f51b3ae449c7a186f2323f4626418df4a24ea06d59a19a8b4", "paper_pdf_bytes": 6526450, "paper_pdf_source": "openreview", "code_url": "https://github.com/caixd-220529/LifelongAgentBench", "code_repository": "caixd-220529/LifelongAgentBench", "code_commit": "d6f19b42eb358d9150379f0c68c2985c5a867520", "code_archive": "repos/MYqAKKsjF9.zip", "code_archive_sha256": "29a0a57ec64dbe6913cc80bc5ba601031165205f0a93e492733399f0876fd1bd", "code_archive_bytes": 563670, "code_file_count": 99, "code_extensions": {".py": 95, ".sh": 4}, "github_disk_usage_kb": 739, "github_languages": {"Python": 685603, "Shell": 9205}, "github_archived": false, "github_pushed_at": "2025-05-30T02:43:44Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/lifelongagentbench-evaluating-llm-agents-as"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "2L1OxhQCwS", "year": 2025, "status": "rejected", "title": "Transformers versus LSTMs for electronic trading", "authors": ["Paul Alexander Bilokon", "Yitao Qiu"], "authorids": ["~Paul_Alexander_Bilokon1", "~Yitao_Qiu1"], "authors_source": "OpenReview API", "abstract": "The rapid advancement of artificial intelligence has seen widespread application of long short-term memory (LSTM), a type of recurrent neural network (RNN), in time series forecasting. Despite the success of Transformers in natural language processing (NLP), which prompted interest in their efficacy for time series prediction, their application in financial time series forecasting is less explored compared to the dominant LSTM models. This study investigates whether Transformer-based models can outperform LSTMs in financial time series forecasting. It involves a comparative analysis of various LSTM-based and Transformer-based models on multiple financial prediction tasks using high-frequency limit order book data. A novel LSTM-based model named DLSTM is introduced alongside a newly designed Transformer-based model tailored for financial predictions. The findings indicate that Transformer-based models exhibit only a marginal advantage in predicting absolute price sequences, whereas LSTM-based models demonstrate superior and more consistent performance in predicting differential sequences such as price differences and movements.", "decision": "Reject", "meta_review": "", "num_reviews": 6, "openreview_url": "https://openreview.net/forum?id=2L1OxhQCwS", "arxiv_id": "2309.11400", "paper_pdf": "papers/2L1OxhQCwS.pdf", "paper_pdf_sha256": "a7d60bfaa0fcc03d864026a9baa0bd1a09d2d9f07d2757fb456b335c114be055", "paper_pdf_bytes": 2383241, "paper_pdf_source": "openreview", "code_url": "https://github.com/772435284/transformers_versus_lstms_for_electronic_trading", "code_repository": "772435284/transformers_versus_lstms_for_electronic_trading", "code_commit": "65d691a52c4ebc5646048cb1a183e2b1748efedd", "code_archive": "repos/2L1OxhQCwS.zip", "code_archive_sha256": "a8ce17af9be0073966659adf9dce10f9649b73bc51be7c5bb0efbd80be93a09e", "code_archive_bytes": 73701, "code_file_count": 52, "code_extensions": {".py": 31, ".sh": 18, ".ipynb": 3}, "github_disk_usage_kb": 65, "github_languages": {"Python": 144742, "Jupyter Notebook": 49889, "Shell": 42520}, "github_archived": false, "github_pushed_at": "2023-12-04T13:14:26Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/transformers-versus-lstms-for-electronic-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "e1IMBXiDhW", "year": 2024, "status": "rejected", "title": "Matrix Information Theory for Self-Supervised Learning", "authors": ["Yifan Zhang", "Zhiquan Tan", "Jingqin Yang", "Weiran Huang", "Yang Yuan"], "authorids": ["~Yifan_Zhang16", "~Zhiquan_Tan1", "~Jingqin_Yang2", "~Weiran_Huang1", "~Yang_Yuan4"], "authors_source": "OpenReview API", "abstract": "Contrastive learning often relies on comparing a single positive anchor sample with multiple negative samples to perform Self-Supervised Learning (SSL). However, non-contrastive approaches like BYOL, SimSiam, and Barlow Twins achieve SSL without explicit negative samples. In this paper, we introduce a unified matrix information-theoretic framework that explains many contrastive and non-contrastive learning methods. We then propose a novel method Matrix-SSL based on matrix information theory. Experimental results reveal that Matrix-SSL significantly outperforms state-of-the-art methods on the ImageNet dataset under linear evaluation settings and on MS-COCO for transfer learning tasks. Specifically, when performing 100 epochs pre-training, our method outperforms SimCLR by 4.6\\%, and when performing transfer learning tasks on MS-COCO, our method outperforms previous SOTA methods such as MoCo v2 and BYOL up to 3.3\\% with only 400 epochs compared to 800 epochs pre-training.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=e1IMBXiDhW", "arxiv_id": "2305.17326", "paper_pdf": "papers/e1IMBXiDhW.pdf", "paper_pdf_sha256": "f4bf75d7b9f0d1c4242d007bb875987b2189287c8ede929e33d813b38e68a51e", "paper_pdf_bytes": 977142, "paper_pdf_source": "openreview", "code_url": "https://github.com/yifanzhang-pro/Matrix-SSL", "code_repository": "yifanzhang-pro/Matrix-SSL", "code_commit": "b8164dfcb279f673654d94d544f0b73e26f152d8", "code_archive": "repos/e1IMBXiDhW.zip", "code_archive_sha256": "6b12b87d4db0a36a697e74aa18e9297f71ad4bd90d28365675b0815d0c55bf6b", "code_archive_bytes": 19843, "code_file_count": 6, "code_extensions": {".py": 4, ".sh": 2}, "github_disk_usage_kb": 57, "github_languages": {"Python": 47611, "Shell": 383}, "github_archived": false, "github_pushed_at": "2025-09-21T05:02:27Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/kernel-ssl-kernel-kl-divergence-for-self"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "zfiYcbeQkH", "year": 2023, "status": "rejected", "title": "SciRepEval: A Multi-Format Benchmark for Scientific Document Representations", "authors": ["Amanpreet Singh", "Mike D'Arcy", "Arman Cohan", "Doug Downey", "Sergey Feldman"], "authorids": ["~Amanpreet_Singh2", "~Mike_D'Arcy1", "~Arman_Cohan1", "~Doug_Downey1", "~Sergey_Feldman1"], "authors_source": "OpenReview API", "abstract": "Learned representations of scientific documents can serve as valuable input features for downstream tasks, without the need for further fine-tuning. However, existing benchmarks for evaluating these representations fail to capture the diversity of relevant tasks. In response, we introduce SciRepEval, the first comprehensive benchmark for training and evaluating scientific document representations. It includes 25 challenging and realistic tasks across four formats: classification, regression, ranking and search. We then use the benchmark to study and improve the generalization ability of scientific document representation models.  We show how state-of-the-art models struggle to generalize across task formats, and that simple multi-task training fails to improve them.  However, a new approach that learns multiple embeddings per document, each tailored to a different task format, can improve performance.\nWe experiment with task-format-specific control codes and adapters in a multi-task setting and find that they outperform the existing single-embedding state-of-the-art by up to 1.5 points absolute. ", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=zfiYcbeQkH", "arxiv_id": "2211.13308", "paper_pdf": "papers/zfiYcbeQkH.pdf", "paper_pdf_sha256": "a87de0a9e40d306f5de4efa34c156738b197d2317a007df350b26f76a0a74a2e", "paper_pdf_bytes": 616170, "paper_pdf_source": "openreview", "code_url": "https://github.com/allenai/SPECTER2", "code_repository": "allenai/SPECTER2", "code_commit": "fac1cb0940fe7bd3c0db00973f6889ca0b481b63", "code_archive": "repos/zfiYcbeQkH.zip", "code_archive_sha256": "3ab2e8bb6096640cf33859f6ba3ec89290360540653a644a337b511acf1e8f39", "code_archive_bytes": 59393, "code_file_count": 4, "code_extensions": {".py": 4}, "github_disk_usage_kb": 97, "github_languages": {"Python": 12946}, "github_archived": false, "github_pushed_at": "2026-02-24T21:56:36Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/scirepeval-a-multi-format-benchmark-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "MTsBazXmX00", "year": 2022, "status": "rejected", "title": "Target Propagation via Regularized Inversion", "authors": ["Vincent Roulet", "Zaid Harchaoui"], "authorids": ["~Vincent_Roulet1", "~Zaid_Harchaoui1"], "authors_source": "OpenReview API", "abstract": "Target Propagation (TP) algorithms compute targets instead of gradients along neural networks and propagate them backward in a way that is similar yet different than gradient back-propagation (BP). The idea was first presented as a perturbative alternative to back-propagation that may achieve greater accuracy in gradient evaluation when training multi-layer neural networks (LeCun et al., 1989). However, TP has remained more of a template algorithm with many variations than a well-identified algorithm. Revisiting insights of LeCun et al., (1989) and more recently of Lee et al. (2015), we present a simple version of target propagation based on regularized inversion of network layers, easily implementable in a differentiable programming framework. We compare its computational complexity to the one of BP and delineate the regimes in which TP can be attractive compared to BP. We show how our TP can be used to train recurrent neural networks with long sequences on various sequence modeling problems. The experimental results underscore the importance of regularization in TP in practice.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=MTsBazXmX00", "arxiv_id": "2112.01453", "paper_pdf": "papers/MTsBazXmX00.pdf", "paper_pdf_sha256": "9dd652243c6f080ea6441f3048c1715b9ec1c3585052fdd33b062310a17ffbfa", "paper_pdf_bytes": 1097086, "paper_pdf_source": "openreview", "code_url": "https://github.com/vroulet/tpri", "code_repository": "vroulet/tpri", "code_commit": "c7e7c10e3382773b634c183a977f125147aa9419", "code_archive": "repos/MTsBazXmX00.zip", "code_archive_sha256": "cb33997af6569f6e38029d1e5ed6c9e952325379fda7b7f346268173bbfe726b", "code_archive_bytes": 56869, "code_file_count": 22, "code_extensions": {".py": 22}, "github_disk_usage_kb": 109, "github_languages": {"Python": 133096}, "github_archived": false, "github_pushed_at": "2023-01-09T22:06:00Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/target-propagation-via-regularized-inversion-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Rld-9OxQ6HU", "year": 2021, "status": "rejected", "title": "MC-LSTM: Mass-conserving LSTM", "authors": ["Pieter-Jan Hoedt", "Frederik Kratzert", "Daniel Klotz", "Christina Halmich", "Markus Holzleitner", "Grey Nearing", "Sepp Hochreiter", "Günter Klambauer"], "authorids": ["~Pieter-Jan_Hoedt1", "~Frederik_Kratzert1", "~Daniel_Klotz1", "halmich@ml.jku.at", "~Markus_Holzleitner1", "gsnearing@google.com", "~Sepp_Hochreiter1", "~Günter_Klambauer1"], "authors_source": "OpenReview API", "abstract": "The success of Convolutional Neural Networks (CNNs) in computer vision is mainly driven by their strong inductive bias,  which is strong enough to allow CNNs to solve vision-related tasks with random weights, meaning without learning. Similarly, Long Short-Term Memory (LSTM) has a strong inductive bias towards storing information over time. However, many real-world systems are governed by conservation laws, which lead to the redistribution of particular quantities —e.g. in physical and economical systems.   Our novel Mass-Conserving LSTM (MC-LSTM) adheres to these conservation laws by extending the inductive bias of LSTM to model the redistribution of those stored quantities. MC-LSTMs set a new state-of-the-art for neural arithmetic units at learning arithmetic operations, such as addition tasks, which have a strong conservation law, as the sum is constant overtime. Further, MC-LSTM is applied to traffic forecasting, modeling a pendulum, and a large benchmark dataset in hydrology, where it sets a new state-of-the-art for predicting peak flows. In the hydrology example, we show that MC-LSTM states correlate with real world processes and are therefore interpretable.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=Rld-9OxQ6HU", "arxiv_id": "2101.05186", "paper_pdf": "papers/Rld-9OxQ6HU.pdf", "paper_pdf_sha256": "04c9f65f5c64ac5258fdf5b37e4adc380f86ad0780ac4430cf0203c658b0b46a", "paper_pdf_bytes": 2045609, "paper_pdf_source": "openreview", "code_url": "https://github.com/ml-jku/mc-lstm", "code_repository": "ml-jku/mc-lstm", "code_commit": "8bbaece3ecb4187a76c6318d4c6e40c1dcc71303", "code_archive": "repos/Rld-9OxQ6HU.zip", "code_archive_sha256": "6e1b4ff407b3d943a3080eafc41417841e47421767bfa02467af1aba5f588ea6", "code_archive_bytes": 100843, "code_file_count": 36, "code_extensions": {".py": 29, ".sh": 6, ".ipynb": 1}, "github_disk_usage_kb": 137, "github_languages": {"Python": 187499, "Jupyter Notebook": 42130, "Shell": 9385}, "github_archived": false, "github_pushed_at": "2021-07-20T14:47:30Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/mc-lstm-mass-conserving-lstm-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ByxCrerKvS", "year": 2020, "status": "rejected", "title": "Set Functions for Time Series", "authors": ["Max Horn", "Michael Moor", "Christian Bock", "Bastian Rieck", "Karsten Borgwardt"], "authorids": ["max.horn@bsse.ethz.ch", "michael.moor@bsse.ethz.ch", "christian.bock@bsse.ethz.ch", "bastian.rieck@bsse.ethz.ch", "karsten.borgwardt@bsse.ethz.ch"], "authors_source": "OpenReview API", "abstract": "Despite the eminent successes of deep neural networks, many architectures are often hard to transfer to irregularly-sampled and asynchronous time series that occur in many real-world datasets, such as healthcare applications. This paper proposes a novel framework for classifying irregularly sampled time series with unaligned measurements, focusing on high scalability and data efficiency.\nOur method SeFT (Set Functions for Time Series) is based on recent advances in differentiable set function learning, extremely parallelizable, and scales well to very large datasets and online monitoring scenarios.\nWe extensively compare our method to competitors on multiple healthcare time series datasets and show that it performs competitively whilst significantly reducing runtime.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=ByxCrerKvS", "arxiv_id": "1909.12064", "paper_pdf": "papers/ByxCrerKvS.pdf", "paper_pdf_sha256": "e9e77a4880f1eda442783a8068787ec52908d19e08990a45bb1282f6cdd09729", "paper_pdf_bytes": 308714, "paper_pdf_source": "openreview", "code_url": "https://github.com/BorgwardtLab/Set_Functions_for_Time_Series", "code_repository": "BorgwardtLab/Set_Functions_for_Time_Series", "code_commit": "6abd69c265b719f9703902af2c6a3254b6d2a779", "code_archive": "repos/ByxCrerKvS.zip", "code_archive_sha256": "21e0f85de43ece10e491631b0bd11250b7547df7824c615bb9516663b8b469f0", "code_archive_bytes": 87547, "code_file_count": 23, "code_extensions": {".py": 23}, "github_disk_usage_kb": 171, "github_languages": {"Python": 173486, "Makefile": 3709}, "github_archived": false, "github_pushed_at": "2021-02-25T09:32:37Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/set-functions-for-time-series-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "H1gFuiA9KX", "year": 2019, "status": "rejected", "title": "Skip-gram word embeddings in hyperbolic space", "authors": ["Matthias Leimeister", "Benjamin J. Wilson"], "authorids": ["matthias@lateral.io", "benjamin@lateral.io"], "authors_source": "OpenReview API", "abstract": "Embeddings of tree-like graphs in hyperbolic space were recently shown to surpass their Euclidean counterparts in performance by a large margin.\nInspired by these results, we present an algorithm for learning word embeddings in hyperbolic space from free text. An objective function based on the hyperbolic distance is derived and included in the skip-gram negative-sampling architecture from word2vec. The hyperbolic word embeddings are then evaluated on word similarity and analogy benchmarks. The results demonstrate the potential of hyperbolic word embeddings, particularly in low dimensions, though without clear superiority over their Euclidean counterparts. We further discuss subtleties in the formulation of the analogy task in curved spaces.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=H1gFuiA9KX", "arxiv_id": "1809.01498", "paper_pdf": "papers/H1gFuiA9KX.pdf", "paper_pdf_sha256": "087364bb6cb2edb131a8f25fc2a36f5bd23302d2cc95142433f34a0116171e69", "paper_pdf_bytes": 351264, "paper_pdf_source": "openreview", "code_url": "https://github.com/lateral/minkowski", "code_repository": "lateral/minkowski", "code_commit": "7d312b750dc7104a1c7699336903f541956e3aaf", "code_archive": "repos/H1gFuiA9KX.zip", "code_archive_sha256": "39b2a907e1e630538b71abc7399f777cbce2f42cb3177194bf00ce17aac3d2d0", "code_archive_bytes": 1903752, "code_file_count": 151, "code_extensions": {".cc": 77, ".h": 41, ".py": 30, ".sh": 2, ".cmake": 1}, "github_disk_usage_kb": 1770, "github_languages": {"C++": 2169503, "Python": 282420, "CMake": 24793, "M4": 19093, "Shell": 15028, "Makefile": 14479, "C": 13067}, "github_archived": false, "github_pushed_at": "2018-09-28T09:33:56Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/skip-gram-word-embeddings-in-hyperbolic-space"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "REthH7hinP", "year": 2026, "status": "rejected", "title": "On Finetuning Tabular Foundation Models", "authors": ["Ivan Rubachev", "Akim Kotelnikov", "Nikolay Kartashev", "Artem Babenko"], "authorids": ["~Ivan_Rubachev1", "~Akim_Kotelnikov1", "~Nikolay_Kartashev1", "~Artem_Babenko1"], "authors_source": "OpenReview API", "abstract": "Foundation models are an emerging research direction in tabular deep learning. Notably, TabPFNv2 recently claimed superior performance over traditional GBDT-based methods on small-scale datasets using an in-context learning paradigm, which does not adapt model parameters to target datasets. However, the optimal finetuning approach for adapting tabular foundational models, and how this adaptation reshapes their internal mechanisms, remains underexplored. While prior works studied finetuning for earlier foundational models, inconsistent findings and TabPFNv2's unique architecture necessitate fresh investigation. To address these questions, we first systematically evaluate various finetuning strategies on diverse datasets. Our findings establish full finetuning as the most practical solution for TabPFNv2 in terms of time-efficiency and effectiveness. We then investigate how finetuning alters TabPFNv2's inner mechanisms, drawing an analogy to retrieval-augmented models. We reveal that the success of finetuning stems from the fact that after gradient-based adaptation, the dot products of the query-representations of test objects and the key-representations of in-context training objects more accurately reflect their target similarity. This improved similarity allows finetuned TabPFNv2 to better approximate target dependency by appropriately weighting relevant in-context samples, improving the retrieval-based prediction logic. From the practical perspective, we managed to finetune TabPFNv2 on datasets with up to 50K objects, observing performance improvements on almost all tasks. More precisely, on academic datasets with I.I.D. splits, finetuning allows TabPFNv2 to achieve state-of-the-art results, while on datasets with gradual temporal shifts and rich feature sets, TabPFNv2 is less stable and prior methods remain better.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=REthH7hinP", "arxiv_id": "2506.08982", "paper_pdf": "papers/REthH7hinP.pdf", "paper_pdf_sha256": "d713ba870923c2e49284b08db68677610dbc9597979271c8d3f994e1e2715762", "paper_pdf_bytes": 1179675, "paper_pdf_source": "openreview", "code_url": "https://github.com/yandex-research/tabpfn-finetuning", "code_repository": "yandex-research/tabpfn-finetuning", "code_commit": "c3ce0f74b8501de2c5f52abe2c03f5b4b382c628", "code_archive": "repos/REthH7hinP.zip", "code_archive_sha256": "b61ab048b1e3a66cd90ed0e741ff90002d49240004aa7d4b0a73e1da94dfe312", "code_archive_bytes": 1595905, "code_file_count": 32, "code_extensions": {".py": 32}, "github_disk_usage_kb": 845, "github_languages": {"Python": 541084}, "github_archived": false, "github_pushed_at": "2025-09-03T10:39:40Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/on-finetuning-tabular-foundation-models"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "UfczlMudN6", "year": 2025, "status": "rejected", "title": "GRAM: Generalization in Deep RL with a Robust Adaptation Module", "authors": ["James Queeney", "Xiaoyi Cai", "Mouhacine Benosman", "JONATHAN P HOW"], "authorids": ["~James_Queeney1", "~Xiaoyi_Cai2", "~Mouhacine_Benosman1", "~JONATHAN_P_HOW1"], "authors_source": "OpenReview API", "abstract": "The reliable deployment of deep reinforcement learning in real-world settings requires the ability to generalize across a variety of conditions, including both in-distribution scenarios seen during training as well as novel out-of-distribution scenarios. In this work, we present a framework for dynamics generalization in deep reinforcement learning that unifies these two distinct types of generalization within a single architecture. We introduce a robust adaptation module that provides a mechanism for identifying and reacting to both in-distribution and out-of-distribution environment dynamics, along with a joint training pipeline that combines the goals of in-distribution adaptation and out-of-distribution robustness. Our algorithm GRAM achieves strong generalization performance across in-distribution and out-of-distribution scenarios upon deployment, which we demonstrate on a variety of realistic simulated locomotion tasks with a quadruped robot.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=UfczlMudN6", "arxiv_id": "2412.04323", "paper_pdf": "papers/UfczlMudN6.pdf", "paper_pdf_sha256": "68bc51b302afc7118dc7bc9edc60b5b762bc646f0398ee2d02b7943780b0952b", "paper_pdf_bytes": 1208532, "paper_pdf_source": "openreview", "code_url": "https://github.com/merlresearch/gram", "code_repository": "merlresearch/gram", "code_commit": "ea3f26ce898d7fdae4cc51497b22e8361a1c3f02", "code_archive": "repos/UfczlMudN6.zip", "code_archive_sha256": "30487eeff83c146517db59b94b0d706ff64ddccd3858fedd05d7d64621e942e8", "code_archive_bytes": 56237, "code_file_count": 19, "code_extensions": {".py": 19}, "github_disk_usage_kb": 70, "github_languages": {"Python": 209080}, "github_archived": false, "github_pushed_at": "2026-07-16T20:57:31Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/gram-generalization-in-deep-rl-with-a-robust"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "pm7O7gJObtk", "year": 2023, "status": "rejected", "title": "NIERT: Accurate Numerical Interpolation through Unifying Scattered Data Representations using Transformer Encoder", "authors": ["Shizhe Ding", "Dongbo Bu"], "authorids": ["~Shizhe_Ding2", "~Dongbo_Bu1"], "authors_source": "OpenReview API", "abstract": "Numerical interpolation for scattered data, i.e., estimating values for target points based on those of some observed points, is widely used in  computational science and engineering. The existing approaches either require explicitly pre-defined basis functions, which makes them inflexible and limits their performance in practical scenarios, or train neural networks as interpolators, which still have limited interpolation accuracy as they treat observed and target points separately and cannot effectively exploit the correlations among data points. Here, we present a learning-based approach to numerical interpolation for scattered data using encoder representation of Transformers (called NIERT). Unlike the recent learning-based approaches, NIERT treats observed and target points in a unified fashion through embedding them into the same representation space, thus gaining the advantage of  effectively exploiting the correlations among them. The specially-designed partial self-attention mechanism used by NIERT makes it escape from the unexpected interference of target points on observed points. We further show that the partial self-attention is essentially a learnable interpolation module combining multiple neural basis functions, which provides interpretability of NIERT. Through pre-training on large-scale synthetic datasets,  NIERT achieves considerable improvement in interpolation accuracy for practical tasks. On both synthetic and real-world datasets, NIERT  outperforms the existing approaches, e.g., on the TFRD-ADlet dataset for temperature field reconstruction, NIERT achieves an MAE of $1.897\\times 10^{-3}$, substantially better than the state-of-the-art  approach (MAE: $27.074\\times 10^{-3}$).  The source code of NIERT is available at  https://anonymous.4open.science/r/NIERT-2BCF.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=pm7O7gJObtk", "arxiv_id": "2209.09078", "paper_pdf": "papers/pm7O7gJObtk.pdf", "paper_pdf_sha256": "cee365554b4ab7b58dee3d6bf866b9a12639926f582dae8587ea3e12a4286129", "paper_pdf_bytes": 4251714, "paper_pdf_source": "openreview", "code_url": "https://github.com/DingShizhe/NIERT", "code_repository": "DingShizhe/NIERT", "code_commit": "053e7edd4ff0270a2205a78ba07ef1460ea8a354", "code_archive": "repos/pm7O7gJObtk.zip", "code_archive_sha256": "c61a0033d69fdc9bdcf68322f485a3fcf93154550e2e9ccec2676bfd649a6639", "code_archive_bytes": 77290, "code_file_count": 29, "code_extensions": {".py": 29}, "github_disk_usage_kb": 101, "github_languages": {"Python": 250843}, "github_archived": false, "github_pushed_at": "2024-11-14T11:08:21Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/niert-accurate-numerical-interpolation"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "vyn49BUAkoD", "year": 2022, "status": "rejected", "title": "Bayesian Active Learning with Fully Bayesian Gaussian Processes", "authors": ["Christoffer Riis", "Francisco Antunes", "Frederik Boe Hüttel", "Carlos Lima Azevedo", "Francisco C. Pereira"], "authorids": ["~Christoffer_Riis1", "~Francisco_Antunes1", "~Frederik_Boe_Hüttel1", "~Carlos_Lima_Azevedo1", "~Francisco_C._Pereira1"], "authors_source": "OpenReview API", "abstract": "The bias-variance trade-off is a well-known problem in machine learning that only gets more pronounced the less available data there is. When data is scarce, such as in metamodeling, active learning, and Bayesian optimization, neglecting this trade-off can cause inefficient and non-optimal querying, leading to unnecessary data labeling. In this paper, we focus on metamodeling with active learning and the canonical Gaussian Process (GP). We recognize that, for the GP, the bias-variance trade-off regulation is made by optimization of the two hyperparameters: the length scale and noise-term. Considering that the optimal mode of the joint posterior of the hyperparameters is equivalent to the optimal bias-variance trade-off, we approximate this joint posterior and utilize it to design two new acquisition functions. The first one is a mode-seeking Bayesian variant of Query-by-Committee (B-QBC), and the second is simultaneously mode-seeking and minimizing the predictive variance through a Query by Mixture Gaussian Processes (QB-MGP) formulation. Across seven simulators, we empirically show that B-QBC outperforms the benchmark functions, whereas QB-MGP is the most robust acquisition function and achieves the best accuracy with the fewest iterations. We generally show that incorporating the bias-variance trade-off in the acquisition functions mitigates unnecessary and expensive data labeling.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=vyn49BUAkoD", "arxiv_id": "2205.10186", "paper_pdf": "papers/vyn49BUAkoD.pdf", "paper_pdf_sha256": "641da463c7ee1ac220fffa83cb7eda2a7866921cdc2845457129df78789db8d6", "paper_pdf_bytes": 910925, "paper_pdf_source": "openreview", "code_url": "https://github.com/CoRiis/active-learning-fbgp", "code_repository": "CoRiis/active-learning-fbgp", "code_commit": "3dbf09f2ec709981036f2ea7d018401ff9dac5be", "code_archive": "repos/vyn49BUAkoD.zip", "code_archive_sha256": "56da7e1247bb56f27d14fbea4adfe9e5cfafef70f818b69b65e632c0cd177a0b", "code_archive_bytes": 113716, "code_file_count": 26, "code_extensions": {".py": 25, ".sh": 1}, "github_disk_usage_kb": 110, "github_languages": {"Python": 90217, "Shell": 1735}, "github_archived": false, "github_pushed_at": "2022-09-29T20:45:09Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/bayesian-active-learning-with-fully-bayesian-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "TJSOfuZEd1B", "year": 2021, "status": "rejected", "title": "GeDi: Generative Discriminator Guided Sequence Generation", "authors": ["Ben Krause", "Akhilesh Deepak Gotmare", "Bryan McCann", "Nitish Shirish Keskar", "Shafiq Joty", "richard socher", "Nazneen Rajani"], "authorids": ["~Ben_Krause1", "~Akhilesh_Deepak_Gotmare1", "~Bryan_McCann1", "~Nitish_Shirish_Keskar1", "~Shafiq_Joty1", "~richard_socher1", "~Nazneen_Rajani1"], "authors_source": "OpenReview API", "abstract": "While large-scale language models (LMs) are able to imitate the distribution of natural language well enough to generate realistic text, it is difficult to control which regions of the distribution they generate. This is especially problematic because datasets used for training large LMs usually contain significant toxicity, hate, bias, and negativity. We propose GeDi as an efficient method for using smaller LMs as generative discriminators to guide generation from large LMs to make them safer and more controllable. GeDi guides generation at each step by computing classification probabilities for all possible next tokens via Bayes rule by normalizing over two class-conditional distributions; one conditioned on the desired attribute, or control code, and another conditioned on the undesired attribute, or anti control code. We find that GeDi gives controllability on par with or better than the state of the art method in a variety of settings, while also achieving generation speeds more than $30$ times faster. Additionally, training GeDi on only three topics allows us to controllably generate new topics zero-shot from just a keyword. Lastly, we show that GeDi can make GPT-2 and GPT-3 significantly less toxic without sacrificing on linguistic fluency, making it by far the most practical existing method for detoxifying large language models while maintaining a fast generation speed.", "decision": "Reject", "meta_review": null, "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=TJSOfuZEd1B", "arxiv_id": "2009.06367", "paper_pdf": "papers/TJSOfuZEd1B.pdf", "paper_pdf_sha256": "f576582972d25ff322637851c708384dde33ff02b57dcfc28b8156c0d6299c7b", "paper_pdf_bytes": 1190201, "paper_pdf_source": "openreview", "code_url": "https://github.com/salesforce/GeDi", "code_repository": "salesforce/GeDi", "code_commit": "2346c7ee99cd9b2f434b088b98803a90e370ca49", "code_archive": "repos/TJSOfuZEd1B.zip", "code_archive_sha256": "b29f9f4e251b421afe9d00f2ccc4cb0b418ccdf025fccbe15f948aa1180a27b0", "code_archive_bytes": 119113, "code_file_count": 11, "code_extensions": {".py": 5, ".sh": 5, ".ipynb": 1}, "github_disk_usage_kb": 150, "github_languages": {"Python": 178636, "Jupyter Notebook": 10407, "Shell": 2192}, "github_archived": true, "github_pushed_at": "2025-06-16T14:46:30Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/gedi-generative-discriminator-guided-sequence"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "SJeHwJSYvH", "year": 2020, "status": "rejected", "title": "Learning De-biased Representations with Biased Representations", "authors": ["Hyojin Bahng", "Sanghyuk Chun", "Sangdoo Yun", "Jaegul Choo", "Seong Joon Oh"], "authorids": ["hjj552@korea.ac.kr", "sanghyuk.c@navercorp.com", "sangdoo.yun@navercorp.com", "jchoo@korea.ac.kr", "coallaoh@linecorp.com"], "authors_source": "OpenReview API", "abstract": "Many machine learning algorithms are trained and evaluated by splitting data from a single source into training and test sets. While such focus on in-distribution learning scenarios has led interesting advances, it has not been able to tell if models are relying on dataset biases as shortcuts for successful prediction (e.g., using snow cues for recognising snowmobiles). Such biased models fail to generalise when the bias shifts to a different class. The cross-bias generalisation problem has been addressed by de-biasing training data through augmentation or re-sampling, which are often prohibitive due to the data collection cost (e.g., collecting images of snowmobile on a desert) and the difficulty of quantifying or expressing biases in the first place. In this work, we propose a novel framework to train a de-biased representation by encouraging it to be different from a set of representations that are biased by design. This tactic is feasible in many scenarios where it is much easier to define a set of biased representations than to define and quantify bias. Our experiments and analyses show that our method discourages models from taking bias shortcuts, resulting in improved performances on de-biased test data.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=SJeHwJSYvH", "arxiv_id": "1910.02806", "paper_pdf": "papers/SJeHwJSYvH.pdf", "paper_pdf_sha256": "2e49b9447992376b1fd2a1ca431054322c0e8d3bd6b03ca647d7152841183470", "paper_pdf_bytes": 5120080, "paper_pdf_source": "openreview", "code_url": "https://github.com/clovaai/rebias", "code_repository": "clovaai/rebias", "code_commit": "79d1d1d892412a5e3b07fbea8b776461bed0e096", "code_archive": "repos/SJeHwJSYvH.zip", "code_archive_sha256": "4b4ebd6391c239dce9907b2403a0d2791a651e27bcf062a1dddb8877963ab1c4", "code_archive_bytes": 171862, "code_file_count": 36, "code_extensions": {".py": 36}, "github_disk_usage_kb": 171, "github_languages": {"Python": 208634}, "github_archived": false, "github_pushed_at": "2024-07-25T10:17:00Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/learning-de-biased-representations-with-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Sy4lojC9tm", "year": 2019, "status": "rejected", "title": "Dataset Distillation", "authors": ["Tongzhou Wang", "Jun-Yan Zhu", "Antonio Torralba", "Alexei A. Efros"], "authorids": ["tongzhou.wang.1994@gmail.com", "junyanz@mit.edu", "torralba@mit.edu", "efros@eecs.berkeley.edu"], "authors_source": "OpenReview API", "abstract": "Model distillation aims to distill the knowledge of a complex model into a simpler one.  In this paper, we consider an alternative formulation called {\\em dataset distillation}: we keep the model fixed and instead attempt to distill the knowledge from a large training dataset into a small one.  The idea is to {\\em synthesize} a small number of data points that do not need to come from the correct data distribution, but will, when given to the learning algorithm as training data, approximate the model trained on the original data.  For example, we show that it is possible to compress $60,000$ MNIST training images into just $10$ synthetic {\\em distilled images} (one per class) and achieve close to original performance with only a few steps of gradient descent, given a particular fixed network initialization. We evaluate our method in a wide range of initialization settings and with different learning objectives.  Experiments on multiple datasets show the advantage of our approach compared to alternative methods in most settings. ", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=Sy4lojC9tm", "arxiv_id": "1811.10959", "paper_pdf": "papers/Sy4lojC9tm.pdf", "paper_pdf_sha256": "ad01e938ef806e9031e08a809eaab74354f85cc023c0f93b19c7cd9ef1c24608", "paper_pdf_bytes": 2555599, "paper_pdf_source": "openreview", "code_url": "https://github.com/ssnl/dataset-distillation", "code_repository": "ssnl/dataset-distillation", "code_commit": "98e4894d108df23f386303e0b98fe4e0b3d15ba8", "code_archive": "repos/Sy4lojC9tm.zip", "code_archive_sha256": "53634ef2e4fb1355e3212125a2635ca0f1600684a8aec1ab7381f4152ad6c7b1", "code_archive_bytes": 1862739, "code_file_count": 19, "code_extensions": {".py": 19}, "github_disk_usage_kb": 1845, "github_languages": {"Python": 132281}, "github_archived": false, "github_pushed_at": "2025-06-17T22:44:27Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/dataset-distillation"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "XeZ5WBIRvz", "year": 2026, "status": "rejected", "title": "When Ethics and Payoffs Diverge: LLM Agents in Morally Charged Social Dilemmas", "authors": ["Steffen Backmann", "David Guzman Piedrahita", "Emanuel Tewolde", "Rada Mihalcea", "Bernhard Schölkopf", "Zhijing Jin"], "authorids": ["~Steffen_Backmann1", "~David_Guzman_Piedrahita1", "~Emanuel_Tewolde1", "~Rada_Mihalcea1", "~Bernhard_Schölkopf1", "~Zhijing_Jin1"], "authors_source": "OpenReview API", "abstract": "Recent advances in large language models (LLMs) have enabled their use in complex agentic roles, involving decision-making with humans or other agents, making ethical alignment a key AI safety concern. While prior work has examined both LLMs' moral judgment and strategic behavior in social dilemmas, there is limited understanding of how they act when moral imperatives directly conflict with rewards or incentives. To investigate this, we introduce MORAL Behavior in Social Dilemma SIMulation (MoralSim) and evaluate how LLMs behave in the prisoner's dilemma and public goods game with morally charged contexts. In MoralSim, we test a range of frontier models across both game structures and three distinct moral framings, enabling a systematic examination of how LLMs navigate social dilemmas in which ethical norms conflict with payoff-maximizing strategies. Our results show substantial variation across models in both their general tendency to act morally and the consistency of their behavior across game types, the specific moral framing, and situational factors such as opponent behavior and survival risks. Crucially, no model exhibits consistently moral behavior in MoralSim, highlighting the need for caution when deploying LLMs in agentic roles where the agent's \"self-interest\" may conflict with ethical expectations.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=XeZ5WBIRvz", "arxiv_id": "2505.19212", "paper_pdf": "papers/XeZ5WBIRvz.pdf", "paper_pdf_sha256": "b496485458c8df743bc31687d488b21db87e0c702ad7deea0c88a0a83c595e4a", "paper_pdf_bytes": 787306, "paper_pdf_source": "openreview", "code_url": "https://github.com/sbackmann/moralsim", "code_repository": "sbackmann/moralsim", "code_commit": "6e7daa2fd393ebee7563de2944442353fa04414e", "code_archive": "repos/XeZ5WBIRvz.zip", "code_archive_sha256": "04aef98cdcd59817b61e0ebc3cd357dc1b35e4f9ef799e4aacfed35c6187309e", "code_archive_bytes": 942025, "code_file_count": 61, "code_extensions": {".py": 56, ".ipynb": 5}, "github_disk_usage_kb": 866, "github_languages": {"Jupyter Notebook": 838848, "Python": 207797}, "github_archived": false, "github_pushed_at": "2025-06-03T17:19:08Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/when-ethics-and-payoffs-diverge-llm-agents-in"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "CSZKElOtG5", "year": 2025, "status": "rejected", "title": "MeanSparse: Post-Training Robustness Enhancement Through Mean-Centered Feature Sparsification", "authors": ["Sajjad Amini", "Mohammadreza Teymoorianfard", "Shiqing Ma", "Amir Houmansadr"], "authorids": ["~Sajjad_Amini1", "~Mohammadreza_Teymoorianfard1", "~Shiqing_Ma2", "~Amir_Houmansadr1"], "authors_source": "OpenReview API", "abstract": "We present a simple yet effective method to improve the robustness of both Convolutional and attention-based Neural Networks against adversarial examples by post-processing an adversarially trained model. \nOur technique, MeanSparse,  cascades the activation functions of a trained model with novel operators that sparsify mean-centered feature vectors. \nThis is equivalent to reducing feature variations around the mean, and we show that such reduced variations merely affect the model's utility, yet they strongly attenuate the adversarial perturbations and decrease the attacker's success rate.\nOur experiments show that, when applied to the top models in the RobustBench leaderboard, MeanSparse achieves a new robustness record of $75.28$% (from $73.71$%), $44.78$% (from $42.67$%) and $62.12$% (from $59.56$%) on CIFAR-10, CIFAR-100 and ImageNet, respectively, in terms of AutoAttack accuracy. \nCode: https://anonymous.4open.science/r/MeanSparse-84B0/", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=CSZKElOtG5", "arxiv_id": "2406.05927", "paper_pdf": "papers/CSZKElOtG5.pdf", "paper_pdf_sha256": "8f6f42c37182e448b8f35d5c830d235b7e3189c51d791cc83316f9b5357c69c9", "paper_pdf_bytes": 2082384, "paper_pdf_source": "openreview", "code_url": "https://github.com/SPIN-UMass/MeanSparse", "code_repository": "SPIN-UMass/MeanSparse", "code_commit": "4d7eb1a15e39bcfd88236b379e2496a289f9e66c", "code_archive": "repos/CSZKElOtG5.zip", "code_archive_sha256": "f660aa76974b4eb2aaaeb827dd7c014856aa2d2e0a37ecb3821547f5235031df", "code_archive_bytes": 106941, "code_file_count": 40, "code_extensions": {".py": 40}, "github_disk_usage_kb": 70, "github_languages": {"Python": 351891}, "github_archived": false, "github_pushed_at": "2025-10-16T00:36:24Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/meansparse-post-training-robustness"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "TYMeXb6PAw", "year": 2024, "status": "rejected", "title": "Adaptive Compression of the Latent Space in Variational Autoencoders", "authors": ["Gabriela Sejnova", "Michal Vavrecka", "Karla Stepanova"], "authorids": ["~Gabriela_Sejnova1", "~Michal_Vavrecka1", "~Karla_Stepanova1"], "authors_source": "OpenReview API", "abstract": "Variational Autoencoders (VAEs) are powerful generative models that have been widely used in various fields, including image and text generation. However, one of the known challenges in using VAEs is the model's sensitivity to its hyperparameters, such as the latent space size. This paper presents a simple extension of VAEs for automatically determining the optimal latent space size during the training process by gradually decreasing the latent size through neuron removal and observing the model performance.  The proposed method is compared to traditional hyperparameter grid search and is shown to be significantly faster while still achieving the best optimal dimensionality on four image datasets. Furthermore, we show that the final performance of our method is comparable to training on the optimal latent size from scratch, and might thus serve as a convenient substitute.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=TYMeXb6PAw", "arxiv_id": "2312.06280", "paper_pdf": "papers/TYMeXb6PAw.pdf", "paper_pdf_sha256": "6066150ba43be18e301ea40665e9877eb83066f153ca2312d6dddcf73f30097d", "paper_pdf_bytes": 1013259, "paper_pdf_source": "openreview", "code_url": "https://github.com/gabinsane/ald-vae", "code_repository": "gabinsane/ald-vae", "code_commit": "cdf2e8b71470092761daae01adabe3fac781f42c", "code_archive": "repos/TYMeXb6PAw.zip", "code_archive_sha256": "5c2b785cef8c31b4c8a8f88f70db28b46499a47d811df0c244afa02bea5deec5", "code_archive_bytes": 63873, "code_file_count": 17, "code_extensions": {".py": 17}, "github_disk_usage_kb": 59, "github_languages": {"Python": 196713}, "github_archived": false, "github_pushed_at": "2024-10-13T17:52:17Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/adaptive-compression-of-the-latent-space-in"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "i2e2wqt0nAI", "year": 2023, "status": "rejected", "title": "Rethinking Symbolic Regression Datasets and Benchmarks for Scientific Discovery", "authors": ["Yoshitomo Matsubara", "Naoya Chiba", "Ryo Igarashi", "Tatsunori Taniai", "Yoshitaka Ushiku"], "authorids": ["~Yoshitomo_Matsubara1", "~Naoya_Chiba1", "~Ryo_Igarashi1", "~Tatsunori_Taniai4", "~Yoshitaka_Ushiku3"], "authors_source": "OpenReview API", "abstract": "This paper revisits datasets and evaluation criteria for Symbolic Regression, a task of expressing given data using mathematical equations, specifically focused on its potential for scientific discovery. Focused on a set of formulas used in the existing datasets based on Feynman Lectures on Physics, we recreate 120 datasets to discuss the performance of symbolic regression for scientific discovery (SRSD). For each of the 120 SRSD datasets, we carefully review the properties of the formula and its variables to design reasonably realistic sampling range of values so that our new SRSD datasets can be used for evaluating the potential of SRSD such as whether or not an SR method can (re)discover physical laws from such datasets. As an evaluation metric, we also propose to use normalized edit distances between a predicted equation and the ground-truth equation trees. While existing metrics are either binary or errors between the target values and an SR model's predicted values for a given input, normalized edit distances evaluate a sort of similarity between the ground-truth and predicted equation trees. We have conducted experiments on our new SRSD datasets using five state-of-the-art SR methods in SRBench and a simple baseline based on a recent Transformer architecture. The results show that we provide a more realistic performance evaluation and open up a new machine learning-based approach for scientific discovery. We provide our datasets and code as part of the supplementary material.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=i2e2wqt0nAI", "arxiv_id": "2206.10540", "paper_pdf": "papers/i2e2wqt0nAI.pdf", "paper_pdf_sha256": "e5c84c0fd791d9ffb63c348cf04e7aafca6d7dcf0afe4b9bc64b1620eb1bfbd5", "paper_pdf_bytes": 665887, "paper_pdf_source": "openreview", "code_url": "https://github.com/omron-sinicx/srsd-benchmark", "code_repository": "omron-sinicx/srsd-benchmark", "code_commit": "7d00b45d56250717ac08a256dc9f79b836d61027", "code_archive": "repos/i2e2wqt0nAI.zip", "code_archive_sha256": "eb46e4fcdb853a2741d424d1c1303462cc2b1e7e0b16b005b6eaad6e05063287", "code_archive_bytes": 102944, "code_file_count": 31, "code_extensions": {".py": 30, ".ipynb": 1}, "github_disk_usage_kb": 109, "github_languages": {"Python": 220721, "Jupyter Notebook": 14834}, "github_archived": false, "github_pushed_at": "2024-03-05T07:26:52Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/rethinking-symbolic-regression-datasets-and"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Le8fg2ppDSv", "year": 2022, "status": "rejected", "title": "HydraSum - Disentangling Stylistic Features in Text Summarization using Multi-Decoder Models", "authors": ["Tanya Goyal", "Nazneen Rajani", "Wenhao Liu", "Wojciech Maciej Kryscinski"], "authorids": ["~Tanya_Goyal1", "~Nazneen_Rajani1", "~Wenhao_Liu1", "~Wojciech_Maciej_Kryscinski1"], "authors_source": "OpenReview API", "abstract": "Existing abstractive summarization models lack explicit control mechanisms that would allow users to influence the stylistic features of the model outputs. This results in generating generic summaries that do not cater to the users needs or preferences. To address this issue we introduce HydraSum, a new summarization architecture that extends the single decoder framework of current models, e.g. BART, to a mixture-of-experts version consisting of multiple decoders. Our proposed model encourages each expert, i.e. decoder, to learn and generate stylistically-distinct summaries along dimensions such as abstractiveness, length, specificity, and others. At each time step, HydraSum employs a gating mechanism that decides the contribution of each individual decoder to the next token's output probability distribution. Through experiments on three summarization datasets (CNN, Newsroom, XSum), we demonstrate that this gating mechanism automatically learns to assign contrasting summary styles to different HydraSum decoders under the standard training objective without the need for additional supervision. We further show that a guided version of the training process can explicitly govern which summary style is partitioned between decoders, e.g. high abstractiveness vs. low abstractiveness or high specificity vs. low specificity, and also increase the stylistic-difference between individual decoders. Finally, our experiments demonstrate that our decoder framework is highly flexible: during inference, we can sample from individual decoders or mixtures of different subsets of the decoders to yield a diverse set of summaries and enforce single- and multi-style control over summary generation.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=Le8fg2ppDSv", "arxiv_id": "2110.04400", "paper_pdf": "papers/Le8fg2ppDSv.pdf", "paper_pdf_sha256": "fec894c1926bfd28787f3706a089341c9e3cb01edabc31e630aea07246e1ce04", "paper_pdf_bytes": 2375098, "paper_pdf_source": "openreview", "code_url": "https://github.com/salesforce/hydra-sum", "code_repository": "salesforce/hydra-sum", "code_commit": "0cb4ea1be9a18fc227406e3a23a41f8c05853267", "code_archive": "repos/Le8fg2ppDSv.zip", "code_archive_sha256": "9688b9d5fc8313897cdfd1a973ed2f6b63dba0e421110d33526ba1d24bf94fc3", "code_archive_bytes": 74103, "code_file_count": 16, "code_extensions": {".py": 15, ".sh": 1}, "github_disk_usage_kb": 120, "github_languages": {"Python": 391463, "Shell": 860}, "github_archived": true, "github_pushed_at": "2025-05-01T17:29:59Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/hydrasum-disentangling-stylistic-features-in-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "gW8n0uD6rl", "year": 2021, "status": "rejected", "title": "Amortized Causal Discovery: Learning to Infer Causal Graphs from Time-Series Data", "authors": ["Sindy Löwe", "David Madras", "Richard Zemel", "Max Welling"], "authorids": ["~Sindy_Löwe1", "~David_Madras1", "~Richard_Zemel1", "~Max_Welling1"], "authors_source": "OpenReview API", "abstract": "Standard causal discovery methods must fit a new model whenever they encounter samples from a new underlying causal graph. However, these samples often share relevant information - for instance, the dynamics describing the effects of causal relations - which is lost when following this approach. We propose Amortized Causal Discovery, a novel framework that leverages such shared dynamics to learn to infer causal relations from time-series data. This enables us to train a single, amortized model that infers causal relations across samples with different underlying causal graphs, and thus makes use of the information that is shared. We demonstrate experimentally that this approach, implemented as a variational model, leads to significant improvements in causal discovery performance, and show how it can be extended to perform well under hidden confounding.\n", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=gW8n0uD6rl", "arxiv_id": "2006.10833", "paper_pdf": "papers/gW8n0uD6rl.pdf", "paper_pdf_sha256": "ae0d46e9fccc33f0ffa29757c4805699a176c44b366639e1730c2b92c9478475", "paper_pdf_bytes": 547682, "paper_pdf_source": "openreview", "code_url": "https://github.com/loeweX/AmortizedCausalDiscovery", "code_repository": "loeweX/AmortizedCausalDiscovery", "code_commit": "4d1661118c9fcfbefbbd00226bc41f7a76743e11", "code_archive": "repos/gW8n0uD6rl.zip", "code_archive_sha256": "f213392592897a594b0a2f35cba2c9a7d86ad2ab66832a0cc3331923d6956dc8", "code_archive_bytes": 127313, "code_file_count": 24, "code_extensions": {".py": 23, ".sh": 1}, "github_disk_usage_kb": 156, "github_languages": {"Python": 142492, "Shell": 155}, "github_archived": false, "github_pushed_at": "2022-03-25T13:03:02Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/amortized-causal-discovery-learning-to-infer"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "BJlnmgrFvS", "year": 2020, "status": "rejected", "title": "BAIL: Best-Action Imitation Learning for Batch Deep Reinforcement Learning", "authors": ["Xinyue Chen", "Zijian Zhou", "Zheng Wang", "Che Wang", "Yanqiu Wu", "Qing Deng", "Keith Ross"], "authorids": ["xc1305@nyu.edu", "zz1435@nyu.edu", "zw1454@nyu.edu", "cw1681@nyu.edu", "yanqiu.wu@nyu.edu", "qd319@nyu.edu", "keithwross@nyu.edu"], "authors_source": "OpenReview API", "abstract": "The field of Deep Reinforcement Learning (DRL) has recently seen a surge in research in batch reinforcement learning, which aims for sample-efficient learning from a given data set without additional interactions with the environment. In the batch DRL setting, commonly employed off-policy DRL algorithms can perform poorly and sometimes even fail to learn altogether.  In this paper we propose anew algorithm, Best-Action Imitation Learning (BAIL), which unlike many off-policy DRL algorithms does not involve maximizing Q functions over the action space. Striving for simplicity as well as performance, BAIL first selects from the batch the actions it believes to be high-performing actions for their corresponding states; it then uses those state-action pairs to train a policy network using imitation learning.  Although BAIL is simple, we demonstrate that BAIL achieves state of the art performance on the Mujoco benchmark, typically outperforming BatchConstrained deep Q-Learning (BCQ) by a wide margin.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=BJlnmgrFvS", "arxiv_id": "1910.12179", "paper_pdf": "papers/BJlnmgrFvS.pdf", "paper_pdf_sha256": "ed7f835dd2972eb29461e62ffe7b28a7d7cea605fb8f5b596e1f1c961b89e307", "paper_pdf_bytes": 20584286, "paper_pdf_source": "openreview", "code_url": "https://github.com/lanyavik/BAIL", "code_repository": "lanyavik/BAIL", "code_commit": "ca1cbc6ff60368a8cb1734a0a67e5478dceba732", "code_archive": "repos/BJlnmgrFvS.zip", "code_archive_sha256": "357547d03c47e73ee2d21ef8d9d7b1cf0cf8f715b3e083ccd9cb0020853db54c", "code_archive_bytes": 168052, "code_file_count": 33, "code_extensions": {".py": 33}, "github_disk_usage_kb": 186, "github_languages": {"Python": 196841}, "github_archived": false, "github_pushed_at": "2022-07-13T12:59:40Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/bail-best-action-imitation-learning-for-batch-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "BJxbYoC9FQ", "year": 2019, "status": "rejected", "title": "Classifier-agnostic saliency map extraction", "authors": ["Konrad Zolna", "Krzysztof J. Geras", "Kyunghyun Cho"], "authorids": ["konrad.zolna@gmail.com", "k.j.geras@nyu.edu", "kyunghyun.cho@nyu.edu"], "authors_source": "OpenReview API", "abstract": "Extracting saliency maps, which indicate parts of the image important to classification, requires many tricks to achieve satisfactory performance when using classifier-dependent methods. Instead, we propose classifier-agnostic saliency map extraction, which finds all parts of the image that any classifier could use, not just one given in advance. We observe that the proposed approach extracts higher quality saliency maps and outperforms existing weakly-supervised localization techniques, setting the new state of the art result on the ImageNet dataset.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=BJxbYoC9FQ", "arxiv_id": "1805.08249", "paper_pdf": "papers/BJxbYoC9FQ.pdf", "paper_pdf_sha256": "9af0cffc8eaccd2205d4ea4726ae42fbdb450b7342c33c2dfdcd339f54ad35e5", "paper_pdf_bytes": 19268307, "paper_pdf_source": "openreview", "code_url": "https://github.com/kondiz/casme", "code_repository": "kondiz/casme", "code_commit": "52cd5769bb7c1089934dd528a331afaf3882eff2", "code_archive": "repos/BJxbYoC9FQ.zip", "code_archive_sha256": "325307e1e60da717ae00008d137f9c7165c88aff5f7e759a1244df48e777c0c4", "code_archive_bytes": 1881246, "code_file_count": 9, "code_extensions": {".py": 9}, "github_disk_usage_kb": 1848, "github_languages": {"Python": 52297}, "github_archived": false, "github_pushed_at": "2020-05-12T20:40:27Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/classifier-agnostic-saliency-map-extraction"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "4jMeUvcO26", "year": 2026, "status": "rejected", "title": "Surrogate Modeling of 3D Rayleigh-Bénard Convection with Equivariant Autoencoders", "authors": ["Fynn Fromme", "Hans Harder", "Christine Allen-Blanchette", "Sebastian Peitz"], "authorids": ["~Fynn_Fromme1", "~Hans_Harder1", "~Christine_Allen-Blanchette1", "~Sebastian_Peitz1"], "authors_source": "OpenReview API", "abstract": "The use of machine learning for modeling, understanding, and controlling large-scale physics systems is quickly gaining in popularity, with examples ranging from electromagnetism over nuclear fusion reactors and magneto-hydrodynamics to fluid mechanics and climate modeling. These systems — governed by partial differential equations — present unique challenges regarding the large number of degrees of freedom and the complex dynamics over many scales both in space and time, and additional measures to improve accuracy and sample efficiency are highly desirable. We present an end-to-end equivariant surrogate model consisting of an equivariant convolutional autoencoder and an equivariant convolutional LSTM using $G$-steerable kernels. As a case study, we consider the three-dimensional Rayleigh-Bénard convection, which describes the buoyancy-driven fluid flow between a heated bottom and a cooled top plate. While the system is E(2)-equivariant in the horizontal plane, the boundary conditions break the translational equivariance in the vertical direction. Our architecture leverages vertically stacked layers of $D_4$-steerable kernels, with additional partial kernel sharing in the vertical direction for further efficiency improvement. We demonstrate significant gains in sample and parameter efficiency, as well as a better scaling to more complex dynamics.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=4jMeUvcO26", "arxiv_id": "2505.13569", "paper_pdf": "papers/4jMeUvcO26.pdf", "paper_pdf_sha256": "078526e509a839ac74972091b87b293a136c5b5abc3cab09602c766d8588b7b0", "paper_pdf_bytes": 6681465, "paper_pdf_source": "openreview", "code_url": "https://github.com/FynnFromme/equivariant-rb-forecasting", "code_repository": "FynnFromme/equivariant-rb-forecasting", "code_commit": "9a5b424834ab345e6e7654580930f7e99688930c", "code_archive": "repos/4jMeUvcO26.zip", "code_archive_sha256": "d9d38455987d615ab0f88fa48d63b52023bd764dad8b7cfb41357975e09fbaf3", "code_archive_bytes": 1110099, "code_file_count": 25, "code_extensions": {".py": 23, ".ipynb": 1, ".jl": 1}, "github_disk_usage_kb": 877, "github_languages": {"Python": 328557, "Jupyter Notebook": 33220, "Julia": 12727}, "github_archived": false, "github_pushed_at": "2025-05-21T11:54:34Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/surrogate-modeling-of-3d-rayleigh-benard"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "8r8H4gbFXf", "year": 2025, "status": "rejected", "title": "Uncertainty Quantification in Retrieval Augmented Question Answering", "authors": ["Laura Perez-Beltrachini", "Mirella Lapata"], "authorids": ["~Laura_Perez-Beltrachini1", "~Mirella_Lapata1"], "authors_source": "OpenReview API", "abstract": "Retrieval augmented Question Answering (QA) enables QA models to overcome knowledge gaps when answering questions at test time by taking as input the question together with retrieved evidence, that is usually a set of passages.  Previous studies show that this approach has numerous benefits such as improving QA performance and reducing hallucinations, without, however, qualifying whether the retrieved passages are indeed useful at answering correctly. In this work, we evaluate existing uncertainty quantification approaches and propose an approach that predicts answer correctness based on utility judgements on individual input passages. We train a small neural model that predicts passage utility for a target QA model. We find that simple information theoretic metrics can predict answer correctness up to a certain extent, more expensive sampling based approaches perform better, while our lightweight approach can efficiently approximate or improve upon sampling-based approaches.", "decision": "Reject", "meta_review": "", "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=8r8H4gbFXf", "arxiv_id": "2502.18108", "paper_pdf": "papers/8r8H4gbFXf.pdf", "paper_pdf_sha256": "9814bc713f59f0fd4ce2ca8fe33a803161dc7fff07f8aab156f4535c41988759", "paper_pdf_bytes": 467337, "paper_pdf_source": "openreview", "code_url": "https://github.com/lauhaide/ragu", "code_repository": "lauhaide/ragu", "code_commit": "15c317f8a070ac9073c850bda09d181aaf7be1f8", "code_archive": "repos/8r8H4gbFXf.zip", "code_archive_sha256": "7cbf7a31393b4b32bd32155a7462b3aea24ea19d56ecd7139195190f685f5222", "code_archive_bytes": 85301, "code_file_count": 29, "code_extensions": {".py": 29}, "github_disk_usage_kb": 74, "github_languages": {"Python": 259896, "Dockerfile": 506}, "github_archived": false, "github_pushed_at": "2025-02-25T10:36:13Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/uncertainty-quantification-in-retrieval"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "20KYsQ8Q4Z", "year": 2024, "status": "rejected", "title": "High-dimensional Bayesian Optimization with Group Testing", "authors": ["Erik Orm Hellsten", "Carl Hvarfner", "Leonard Papenmeier", "Luigi Nardi"], "authorids": ["~Erik_Orm_Hellsten1", "~Carl_Hvarfner1", "~Leonard_Papenmeier1", "~Luigi_Nardi1"], "authors_source": "OpenReview API", "abstract": "Bayesian optimization is an effective method for optimizing expensive-to-evaluate black-box functions. \nHigh-dimensional problems are particularly challenging as the surrogate model of the objective suffers from the curse of dimensionality, which makes accurate modeling difficult. \nWe propose a group testing approach to identify active variables to facilitate efficient optimization in these domains. \nThe proposed algorithm, Group Testing Bayesian Optimization (GTBO), first runs a testing phase where groups of variables are systematically selected and tested on whether they influence the objective. \nTo that end, we extend the well-established theory of group testing to functions of continuous ranges.\nIn the second phase, GTBO guides optimization by placing more importance on the active dimensions.\nBy exploiting the axis-aligned subspace assumption, GTBO is competitive against state-of-the-art methods on several synthetic and real-world high-dimensional optimization tasks. \nFurthermore, GTBO aids in the discovery of active parameters in applications, thereby enhancing practitioners' understanding of the problem at hand.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=20KYsQ8Q4Z", "arxiv_id": "2310.03515", "paper_pdf": "papers/20KYsQ8Q4Z.pdf", "paper_pdf_sha256": "e6170406d850d4ddfc14c4bae8100bf1629f1c09624c3689da86367b44fe8751", "paper_pdf_bytes": 2957796, "paper_pdf_source": "openreview", "code_url": "https://github.com/gtboauthors/gtbo", "code_repository": "gtboauthors/gtbo", "code_commit": "fe3dca9d179e46b4d9cc7bc6954b09e1078919e8", "code_archive": "repos/20KYsQ8Q4Z.zip", "code_archive_sha256": "8b7954aefdeb96db9318d460e31ca1ce54c917d58583c965c5a8d6a6de039bd7", "code_archive_bytes": 67575, "code_file_count": 20, "code_extensions": {".py": 20}, "github_disk_usage_kb": 59, "github_languages": {"Python": 117055}, "github_archived": false, "github_pushed_at": "2023-10-04T08:05:30Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/high-dimensional-bayesian-optimization-with-6"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "e1u9PVnwNr", "year": 2023, "status": "rejected", "title": "BiBench: Benchmarking and Analyzing Network Binarization", "authors": ["Haotong Qin", "Mingyuan Zhang", "Yifu Ding", "Aoyu Li", "Zhongang Cai", "Ziwei Liu", "Fisher Yu", "Xianglong Liu"], "authorids": ["~Haotong_Qin1", "~Mingyuan_Zhang1", "~Yifu_Ding2", "s220059@e.ntu.edu.sg", "~Zhongang_Cai1", "~Ziwei_Liu1", "~Fisher_Yu2", "~Xianglong_Liu3"], "authors_source": "OpenReview API", "abstract": "Neural network binarization emerges as one of the most promising compression approaches with extraordinary computation and memory savings by minimizing the bit-width of weight and activation. However, despite being a generic technique, recent works reveal that applying binarization in a wide range of realistic scenarios involving diverse tasks, architectures, and hardware is not trivial. Moreover, common challenges, such as severe degradation in accuracy and limited efficiency gains, suggest that specific attributes of binarization are not thoroughly studied and adequately understood. To close this gap, we present BiBench, a rigorously designed benchmark with in-depth analysis for network binarization. We first carefully scrutinize the requirements of binarization in the actual production setting. We thus define the evaluation tracks and metrics for a fair and systematic investigation. We then perform a comprehensive evaluation with a rich collection of milestone binarization algorithms. Our benchmark results show binarization still faces severe accuracy challenges but diminishing improvements brought by newer state-of-the-art binarization algorithms, even at the expense of efficiency. Moreover, the actual deployment of certain binarization operations reveals a surprisingly large deviation from their theoretical consumption. Finally, we provide suggestions based on our benchmark results and analysis, devoted to establishing a paradigm for accurate and efficient binarization among existing techniques. We hope BiBench paves the way towards more extensive adoption of network binarization and serves as a foundation for future research.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=e1u9PVnwNr", "arxiv_id": "2301.11233", "paper_pdf": "papers/e1u9PVnwNr.pdf", "paper_pdf_sha256": "ebb42417997fc95d53977274f363d91e041a5891188722fc30a0542360aa02f4", "paper_pdf_bytes": 701606, "paper_pdf_source": "openreview", "code_url": "https://github.com/AI-Efficiency/BiBench", "code_repository": "AI-Efficiency/BiBench", "code_commit": "a6662179aff8051a3605e771fc076b35080b5be7", "code_archive": "repos/e1u9PVnwNr.zip", "code_archive_sha256": "0b60acd60517ffecdb7ddbf1ab0177973c1fe4062c93c0a88405962a0146cbbb", "code_archive_bytes": 245865, "code_file_count": 369, "code_extensions": {".py": 367, ".sh": 2}, "github_disk_usage_kb": 113, "github_languages": {"Python": 400727, "Shell": 1265}, "github_archived": false, "github_pushed_at": "2024-03-04T15:37:56Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/bibench-benchmarking-and-analyzing-network"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "rryJiPXifr", "year": 2021, "status": "rejected", "title": "Optimization Planning for 3D ConvNets", "authors": ["Zhaofan Qiu", "Ting Yao", "Chong-wah Ngo", "Tao Mei"], "authorids": ["~Zhaofan_Qiu2", "~Ting_Yao1", "~Chong-wah_Ngo2", "~Tao_Mei3"], "authors_source": "OpenReview API", "abstract": "3D Convolutional Neural Networks (3D ConvNets) have been regarded as a powerful class of models for video recognition. Nevertheless, it is not trivial to optimally learn a 3D ConvNets due to high complexity and various options of the training scheme. The most common hand-tuning process starts from learning 3D ConvNets using short video clips and then is followed by learning long-term temporal dependency using lengthy clips, while gradually decaying the learning rate from high to low as training progresses. The fact that such process comes along with several heuristic settings motivates the study to seek an optimal ``path'' to automate the entire training. In this paper, we decompose the path into a series of training ``states'' and specify the hyper-parameters, e.g., learning rate and the length of input clips, in each state. The estimation of the knee point on the performance-epoch curve triggers the transition from one state to another. We perform dynamic programming over all the candidate states to plan the optimal permutation of states, i.e., optimization path. Furthermore, we devise a new 3D ConvNets with a unique design of dual-head classifier to improve the spatial and temporal discrimination. Extensive experiments conducted on seven public video recognition benchmarks demonstrate the advantages of our proposal. With the optimization planning, our 3D ConvNets achieves superior results when comparing to the state-of-the-art video recognition approaches. More remarkably, we obtain the top-1 accuracy of 82.5% and 84.3% on the large-scale Kinetics-400 and Kinetics-600 datasets, respectively.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=rryJiPXifr", "arxiv_id": "2201.04021", "paper_pdf": "papers/rryJiPXifr.pdf", "paper_pdf_sha256": "ff22c8572b0beef4bd3caf469e03d23c9cd3e7344e040c4b6936613887438600", "paper_pdf_bytes": 6527274, "paper_pdf_source": "openreview", "code_url": "https://github.com/ZhaofanQiu/Optimization-Planning-for-3D-ConvNets", "code_repository": "ZhaofanQiu/Optimization-Planning-for-3D-ConvNets", "code_commit": "d9f1b777811ca0d8f462798ca2efcea39b96fcc5", "code_archive": "repos/rryJiPXifr.zip", "code_archive_sha256": "ccf1dd4f51f10d8c6e75241cd2aef37df1093ffe020b08d1a0c3042441a6e029", "code_archive_bytes": 218724, "code_file_count": 46, "code_extensions": {".py": 45, ".sh": 1}, "github_disk_usage_kb": 156, "github_languages": {"Python": 467759, "Shell": 585}, "github_archived": false, "github_pushed_at": "2022-01-12T02:05:07Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/optimization-planning-for-3d-convnets-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "BJlbo6VtDH", "year": 2020, "status": "rejected", "title": "A Generalized Framework of Sequence Generation with Application to Undirected Sequence Models", "authors": ["Elman Mansimov", "Alex Wang", "Kyunghyun Cho"], "authorids": ["elman.mansimov@gmail.com", "wangalexc@gmail.com", "kyunghyun.cho@nyu.edu"], "authors_source": "OpenReview API", "abstract": "Undirected neural sequence models such as BERT (Devlin et al., 2019) have received renewed interest due to their success on discriminative natural language understanding tasks such as question-answering and natural language inference. \nThe problem of generating sequences directly from these models has received relatively little attention, in part because generating from such models departs significantly from the conventional approach of monotonic generation in directed sequence models. We investigate this problem by first proposing a generalized model of sequence generation that unifies decoding in directed and undirected models. The proposed framework models the process of generation rather than a resulting sequence, and under this framework, we derive various neural sequence models as special cases, such as autoregressive, semi-autoregressive, and refinement-based non-autoregressive models. This unification enables us to adapt decoding algorithms originally developed for directed sequence models to undirected models. We demonstrate this by evaluating various decoding strategies for a cross-lingual masked translation model (Lample and Conneau, 2019). Our experiments show that generation from undirected sequence models, under our framework, is competitive with the state of the art on WMT'14 English-German translation. We also demonstrate that the proposed approach enables constant-time translation with similar performance to linear-time translation from the same model by rescoring hypotheses with an autoregressive model.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=BJlbo6VtDH", "arxiv_id": "1905.12790", "paper_pdf": "papers/BJlbo6VtDH.pdf", "paper_pdf_sha256": "59fdc2a025a91b4a470be9bcc1e797969e5309b334bb6fa7e0dc13cc8a25c7ee", "paper_pdf_bytes": 644135, "paper_pdf_source": "openreview", "code_url": "https://github.com/nyu-dl/dl4mt-seqgen", "code_repository": "nyu-dl/dl4mt-seqgen", "code_commit": "dd5a08182f1a32386e6203cbeaa1cee5a0b83994", "code_archive": "repos/BJlbo6VtDH.zip", "code_archive_sha256": "a396533a744ac9329db47336b6b2909a4f22168ee41009a2af87414c9ecb24cb", "code_archive_bytes": 210726, "code_file_count": 26, "code_extensions": {".py": 19, ".sh": 7}, "github_disk_usage_kb": 189, "github_languages": {"Python": 275812, "Shell": 18347, "Perl": 5233}, "github_archived": false, "github_pushed_at": "2019-06-13T21:40:41Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/a-generalized-framework-of-sequence"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Bye9LiR9YX", "year": 2019, "status": "rejected", "title": "Remember and Forget for Experience Replay", "authors": ["Guido Novati", "Petros Koumoutsakos"], "authorids": ["novatig@ethz.ch", "petros@ethz.ch"], "authors_source": "OpenReview API", "abstract": "Experience replay (ER) is crucial for attaining high data-efficiency in off-policy deep reinforcement learning (RL).  ER entails the recall of experiences obtained in past iterations to compute gradient estimates for the current policy. However, the accuracy of such updates may deteriorate when the policy diverges from past behaviors, possibly undermining the effectiveness of ER. Previous off-policy RL algorithms mitigated this issue by tuning their hyper-parameters in order to abate policy changes. We propose ReF-ER, a method for active management of experiences in the Replay Memory (RM). ReF-ER forgets experiences that would be too unlikely with the current policy and constrains policy changes within a trust region of the behaviors in the RM. We couple ReF-ER with Q-learning, deterministic policy gradient and off-policy gradient methods to show that ReF-ER reliably improves the performance of continuous-action off-policy RL. We complement ReF-ER with a novel off-policy actor-critic algorithm (RACER) for continuous-action control. RACER employs a computationally efficient closed-form approximation of the action values and is shown to be highly competitive with state-of-the-art algorithms on benchmark problems, while being robust to large hyper-parameter variations.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=Bye9LiR9YX", "arxiv_id": "1807.05827", "paper_pdf": "papers/Bye9LiR9YX.pdf", "paper_pdf_sha256": "9c3c123f4725c967172069e8397b923fa06a0d26a532a67c59c2be0e14d94adb", "paper_pdf_bytes": 3524359, "paper_pdf_source": "openreview", "code_url": "https://github.com/cselab/smarties", "code_repository": "cselab/smarties", "code_commit": "34c53d1ef03a324b20b418eaac6343ba44b6c966", "code_archive": "repos/Bye9LiR9YX.zip", "code_archive_sha256": "b808af657a87bf4813004836622e52ecec212b130f7c9209faf464cbc2e57040", "code_archive_bytes": 1463248, "code_file_count": 223, "code_extensions": {".h": 80, ".cpp": 62, ".sh": 47, ".py": 21, ".m": 5, ".f90": 3, ".hpp": 3, ".c": 2}, "github_disk_usage_kb": 2535, "github_languages": {"C++": 810341, "C": 136407, "Python": 103875, "Fortran": 22040, "Shell": 18886, "MATLAB": 13301, "Makefile": 12975, "CMake": 7959}, "github_archived": false, "github_pushed_at": "2023-12-20T16:53:43Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/remember-and-forget-for-experience-replay"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "z55Fut1CRw", "year": 2026, "status": "rejected", "title": "UniFork: Exploring Modality Alignment for Unified Multimodal Understanding and Generation", "authors": ["Teng Li", "Quanfeng Lu", "Lirui Zhao", "Hao Li", "Xizhou Zhu", "Yu Qiao", "Jun Zhang", "Wenqi Shao"], "authorids": ["~Teng_Li5", "~Quanfeng_Lu1", "~Lirui_Zhao1", "~Hao_Li13", "~Xizhou_Zhu1", "~Yu_Qiao1", "~Jun_Zhang25", "~Wenqi_Shao2"], "authors_source": "OpenReview API", "abstract": "Unified image understanding and generation has emerged as a promising paradigm in multimodal artificial intelligence. Despite recent progress, the optimal architectural design for such unified models remains an open challenge. In this work, we start by analyzing the modality alignment behaviors of task-specific expert models for understanding and generation, as well as current unified models. Our analysis reveals a crucial observation: understanding tasks benefit from a progressively increasing modality alignment across network depth, which helps build up semantic information for better comprehension; In contrast, generation tasks follow a different trend—modality alignment increases in the early layers but decreases in the deep layers to recover spatial details. These divergent alignment patterns create a fundamental conflict in fully shared Transformer backbones, where a uniform representational flow often leads to performance compromises across two tasks. Motivated by this finding, we introduce UniFork, a novel Y-shaped architecture that shares the shallow layers for cross-task representation learning, while employing task-specific branches in deeper layers to avoid task interference. This design effectively balances shared learning and task specialization. Through extensive ablation experiments, we demonstrate that Unifork consistently outperforms conventional fully shared Transformer architectures, and achieves performance on par with or better than task-specific models.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=z55Fut1CRw", "arxiv_id": "2506.17202", "paper_pdf": "papers/z55Fut1CRw.pdf", "paper_pdf_sha256": "26bcac7a02ff6bd6d51ba9eb1a14ee0c2d8bd003f82b4e84ee53468af0f6067a", "paper_pdf_bytes": 12781869, "paper_pdf_source": "openreview", "code_url": "https://github.com/tliby/UniFork", "code_repository": "tliby/UniFork", "code_commit": "2c5b75750e880589aacf2ab5f25efafb37477272", "code_archive": "repos/z55Fut1CRw.zip", "code_archive_sha256": "fd6e1de5f4bae70d0c6fa01daf42bb00ebd51d68380ce9be30308683258da5ee", "code_archive_bytes": 684083, "code_file_count": 77, "code_extensions": {".py": 66, ".sh": 11}, "github_disk_usage_kb": 915, "github_languages": {"Python": 599051, "Shell": 17112}, "github_archived": false, "github_pushed_at": "2025-08-26T03:58:28Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/unifork-exploring-modality-alignment-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "pbJMPo4HwR", "year": 2025, "status": "rejected", "title": "Learning Monotonic Attention in Transducer for Streaming Generation", "authors": ["Zhengrui Ma", "Yang Feng", "Min zhang"], "authorids": ["~Zhengrui_Ma1", "~Yang_Feng4", "~Min_zhang14"], "authors_source": "OpenReview API", "abstract": "Streaming generation models are increasingly utilized across various fields, with the Transducer architecture being particularly popular in industrial applications. However, its input-synchronous decoding mechanism presents challenges in tasks requiring non-monotonic alignments, such as simultaneous translation, leading to suboptimal performance in these contexts. In this research, we address this issue by tightly integrating Transducer's decoding with the history of input stream via a learnable monotonic attention mechanism. Our approach leverages the forward-backward algorithm to infer the posterior probability of alignments between the predictor states and input timestamps, which is then used to estimate the context representations of monotonic attention in training. This allows Transducer models to adaptively adjust the scope of attention based on their predictions, avoiding the need to enumerate the exponentially large alignment space. Extensive experiments demonstrate that our MonoAttn-Transducer significantly enhances the handling of non-monotonic alignments in streaming generation, offering a robust solution for Transducer-based frameworks to tackle more complex streaming generation tasks. Codes are publicly available in supplementary materials.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=pbJMPo4HwR", "arxiv_id": "2411.17170", "paper_pdf": "papers/pbJMPo4HwR.pdf", "paper_pdf_sha256": "2353813ef1371aa5a8b8eb4d5ed58806cecde4a3471fcc6e219278409fa85873", "paper_pdf_bytes": 3022089, "paper_pdf_source": "openreview", "code_url": "https://github.com/ictnlp/MonoAttn-Transducer", "code_repository": "ictnlp/MonoAttn-Transducer", "code_commit": "518d0555d020bfff543409bd1806541b40a6b496", "code_archive": "repos/pbJMPo4HwR.zip", "code_archive_sha256": "709a6cf896e6c2edff27670ef746f118dd8516ec1b6b65a7b0230ba60e422c3d", "code_archive_bytes": 96407, "code_file_count": 37, "code_extensions": {".py": 37}, "github_disk_usage_kb": 77, "github_languages": {"Python": 329821}, "github_archived": false, "github_pushed_at": "2025-05-19T17:59:09Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/learning-monotonic-attention-in-transducer"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "OlwW4ZG3Ta", "year": 2024, "status": "rejected", "title": "Reflective Policy Optimization", "authors": ["Yaozhong Gan", "Renye Yan", "Zhe Wu", "Junliang Xing"], "authorids": ["~Yaozhong_Gan1", "~Renye_Yan1", "~Zhe_Wu6", "~Junliang_Xing1"], "authors_source": "OpenReview API", "abstract": "On-policy reinforcement learning methods, such as Trust Region Policy Optimization (TRPO) and Proximal Policy Optimization (PPO), often require significant data to be collected at each update, giving rise to issues of sample inefficiency. This paper introduces a novel extension to on-policy methods called Reflective Policy Optimization (RPO). RPO's fundamental objective is amalgamating prior and subsequent state and action information from trajectory data to optimize the current policy. This approach empowers the agent to engage in introspection and introduce modifications to its actions within the current state to a certain degree. Furthermore, theoretical analyses substantiate that our proposed method not only upholds the crucial property of monotonically improving policy performance but also adeptly contracts the solution space of the optimized policy, consequently expediting the training procedure. We empirically demonstrate the feasibility and efficacy of our approach in reinforcement learning benchmarks, culminating in superior performance in terms of sample efficiency.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=OlwW4ZG3Ta", "arxiv_id": "2406.03678", "paper_pdf": "papers/OlwW4ZG3Ta.pdf", "paper_pdf_sha256": "4eddac48353f68653a656e39937acb8bf4975e311a8ed37795ee398175b678cc", "paper_pdf_bytes": 7292594, "paper_pdf_source": "openreview", "code_url": "https://github.com/Edgargan/RPO", "code_repository": "Edgargan/RPO", "code_commit": "3c43816ec410cf396a61fa2cedd83dda4ebc7907", "code_archive": "repos/OlwW4ZG3Ta.zip", "code_archive_sha256": "f3ec2bb286eed2569267182f92eca8cb9d883603a2ba4fe33f2858de3c868cdb", "code_archive_bytes": 49394, "code_file_count": 35, "code_extensions": {".py": 35}, "github_disk_usage_kb": 60, "github_languages": {"Python": 145153}, "github_archived": false, "github_pushed_at": "2024-05-31T03:02:08Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/reflective-policy-optimization"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "-SBZ8c356Oc", "year": 2023, "status": "rejected", "title": "Improving Adversarial Robustness by Putting More Regularizations on Less Robust Samples", "authors": ["Dongyoon Yang", "Insung Kong", "Yongdai Kim"], "authorids": ["~Dongyoon_Yang2", "~Insung_Kong1", "~Yongdai_Kim1"], "authors_source": "OpenReview API", "abstract": "\nAdversarial training, which is to enhance robustness against adversarial attacks, has received much attention because it is easy to generate human-imperceptible perturbations of data to deceive a given deep neural network. In this paper, we propose a new adversarial training algorithm that is theoretically well motivated and empirically superior to other existing algorithms. A novel feature of the proposed algorithm is to apply more regularization to data vulnerable to adversarial attacks than other existing regularization algorithms do. Theoretically, we show that our algorithm can be understood as an algorithm of minimizing a newly derived upper bound of the robust risk. Numerical experiments illustrate that our proposed algorithm improves the generalization (accuracy on examples) and robustness (accuracy on adversarial attacks) simultaneously to achieve the state-of-the-art performance.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=-SBZ8c356Oc", "arxiv_id": "2206.03353", "paper_pdf": "papers/-SBZ8c356Oc.pdf", "paper_pdf_sha256": "e547f032e74fb64bd976eca36f7fa2c17d516d8350ef71944bf9c27087b6c0f3", "paper_pdf_bytes": 368162, "paper_pdf_source": "openreview", "code_url": "https://github.com/dyoony/ARoW", "code_repository": "dyoony/ARoW", "code_commit": "0039411e16d1721b28aaf618e075a82cd74624a6", "code_archive": "repos/-SBZ8c356Oc.zip", "code_archive_sha256": "0fe876e06ee9199694b621af4376ca994b32ab92f213f6d92b56f97c8ed54c42", "code_archive_bytes": 129299, "code_file_count": 57, "code_extensions": {".py": 57}, "github_disk_usage_kb": 115, "github_languages": {"Python": 419030, "Jinja": 1720}, "github_archived": false, "github_pushed_at": "2024-09-14T13:20:49Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/adaptive-regularization-for-adversarial"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "6PahjGFjVG-", "year": 2022, "status": "rejected", "title": "Secure Distributed Training at Scale", "authors": ["Eduard Gorbunov", "Alexander Borzunov", "Michael Diskin", "Max Ryabinin"], "authorids": ["~Eduard_Gorbunov1", "~Alexander_Borzunov1", "~Michael_Diskin1", "~Max_Ryabinin1"], "authors_source": "OpenReview API", "abstract": "Some of the hardest problems in deep learning can be solved via pooling together computational resources of many independent parties, as is the case for scientific collaborations and volunteer computing. Unfortunately, any single participant in such systems can jeopardize the entire training run by sending incorrect updates, whether deliberately or by mistake. Training in presence of such peers requires specialized distributed training algorithms with Byzantine tolerance. These algorithms often sacrifice efficiency by introducing redundant communication or passing all updates through a trusted server. As a result, it can be infeasible to apply such algorithms to large-scale distributed deep learning, where models can have billions of parameters. In this work, we propose a novel protocol for secure (Byzantine-tolerant) decentralized training that emphasizes communication efficiency. We rigorously analyze this protocol: in particular, we provide theoretical bounds for its resistance against Byzantine and Sybil attacks and show that it has a marginal communication overhead. To demonstrate its practical effectiveness, we conduct large-scale experiments on image classification and language modeling in presence of Byzantine attackers.", "decision": "Reject", "meta_review": null, "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=6PahjGFjVG-", "arxiv_id": "2106.11257", "paper_pdf": "papers/6PahjGFjVG-.pdf", "paper_pdf_sha256": "f06b41fcde5eca3b4cff6d651cf777df7658462825fac1b3c06bf84ebb5bec51", "paper_pdf_bytes": 1546801, "paper_pdf_source": "openreview", "code_url": "https://github.com/icml-2022-paper/BTARD", "code_repository": "icml-2022-paper/BTARD", "code_commit": "a0e758ae5c5e405775144bc3f2376bfc14d007f8", "code_archive": "repos/6PahjGFjVG-.zip", "code_archive_sha256": "1b49a38c2b3f599a2d91c14fbbf589457dd06e4a1c1646b18e1e5d5e96506aec", "code_archive_bytes": 138960, "code_file_count": 52, "code_extensions": {".py": 50, ".ipynb": 2}, "github_disk_usage_kb": 123, "github_languages": {"Python": 391126, "Jupyter Notebook": 22854}, "github_archived": false, "github_pushed_at": "2021-10-07T21:15:42Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/secure-distributed-training-at-scale"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "N07ebsD-lHp", "year": 2021, "status": "rejected", "title": "Defending against black-box adversarial attacks with gradient-free trained sign activation neural networks", "authors": ["Yunzhe Xue", "Meiyan Xie", "Zhibo Yang", "Usman Roshan"], "authorids": ["yx277@njit.edu", "mx42@njit.edu", "zy328@njit.edu", "~Usman_Roshan1"], "authors_source": "OpenReview API", "abstract": "While machine learning models today can achieve high accuracies on classification tasks, they can be deceived by minor imperceptible distortions to the data. These are known as adversarial attacks and can be lethal in the black-box setting which does not require knowledge of the target model type or its parameters. Binary neural networks that have sign activation and are trained with gradient descent have been shown to be harder to attack than conventional sigmoid activation networks but their improvements are marginal. We instead train sign activation networks with a novel gradient-free stochastic coordinate descent algorithm and propose an ensemble of such networks as a defense model. We evaluate the robustness of our model (a hard problem in itself) on image, text, and medical ECG data and find it to be more robust than ensembles of binary, full precision, and convolutional neural networks, and than random forests while attaining comparable clean test accuracy. In order to explain our model's robustness we show that an adversary targeting a single network in our ensemble fails to attack (and thus non-transferable to) other networks in the ensemble. Thus a datapoint requires a large distortion to fool the majority of networks in our ensemble and is likely to be detected in advance. This property of non-transferability arises naturally from the non-convexity of sign activation networks and randomization in our gradient-free training algorithm without any adversarial defense effort.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=N07ebsD-lHp", "arxiv_id": null, "paper_pdf": "papers/N07ebsD-lHp.pdf", "paper_pdf_sha256": "aaa493edcbf910bf4c7472f4ed0dbe1ab8710dce10a77c03dc46b5610aeb8099", "paper_pdf_bytes": 284323, "paper_pdf_source": "openreview", "code_url": "https://github.com/zero-one-loss/scd_github", "code_repository": "zero-one-loss/scd_github", "code_commit": "c41fee1dd4d86f152748590887f7e4d0a95c89c8", "code_archive": "repos/N07ebsD-lHp.zip", "code_archive_sha256": "3a562a85154be834cd871ab4d9fab374a390087f20ceaddeb160f770d99ffe19", "code_archive_bytes": 211424, "code_file_count": 33, "code_extensions": {".py": 32, ".sh": 1}, "github_disk_usage_kb": 176, "github_languages": {"Python": 372079, "Shell": 40733}, "github_archived": false, "github_pushed_at": "2020-10-15T02:18:18Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/defending-against-black-box-adversarial"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ryxF80NYwS", "year": 2020, "status": "rejected", "title": "Neural Clustering Processes", "authors": ["Ari Pakman", "Yueqi Wang", "Catalin Mitelut", "JinHyung Lee", "Liam Paninski"], "authorids": ["aripakman@gmail.com", "yueqi.wang.pku@gmail.com", "mitelutco@gmail.com", "jl4303@columbia.edu", "liam@stat.columbia.edu"], "authors_source": "OpenReview API", "abstract": "Mixture models, a basic building block in countless statistical models, involve latent random variables over discrete spaces, and existing posterior inference methods can be inaccurate and/or very slow.  In this work we introduce a novel deep learning architecture for efficient amortized Bayesian inference over mixture models. While previous approaches to amortized clustering assumed a fixed or maximum number of mixture components and only amortized over the continuous parameters of each mixture component, our method amortizes over the local discrete labels of all the data points, and performs inference over an unbounded number of mixture components. The latter property makes our method natural for the challenging case of nonparametric Bayesian models, where the number of mixture components grows with the dataset. Our approach exploits the exchangeability of the generative models and is based on mapping distributed, permutation-invariant representations of discrete  arrangements into varying-size multinomial conditional probabilities. The resulting algorithm parallelizes easily, yields iid samples from the approximate posteriors along with a normalized probability estimate of each sample (a quantity generally unavailable using Markov Chain Monte Carlo) and can easily be applied to both conjugate and non-conjugate models, as training only requires samples from the generative model. We also present an extension of the method to models of random communities (such as infinite relational or stochastic block models). As a scientific application, we present a novel approach to neural spike sorting for high-density multielectrode arrays. \n", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=ryxF80NYwS", "arxiv_id": "1901.00409", "paper_pdf": "papers/ryxF80NYwS.pdf", "paper_pdf_sha256": "eb34cc78bf2fde50e3030a759de695867ef15ddf2e4f34afdc93ed8c78cabdfb", "paper_pdf_bytes": 2799517, "paper_pdf_source": "openreview", "code_url": "https://github.com/aripakman/neural_clustering_process", "code_repository": "aripakman/neural_clustering_process", "code_commit": "dd54a8793398f7cb2dc44365dcb3d9507b89ab34", "code_archive": "repos/ryxF80NYwS.zip", "code_archive_sha256": "03d604a55480e607e696db490e8ac339b8ca16776282b93f71cd8350d2f67410", "code_archive_bytes": 99309, "code_file_count": 6, "code_extensions": {".py": 6}, "github_disk_usage_kb": 190, "github_languages": {"Python": 30222}, "github_archived": false, "github_pushed_at": "2020-07-19T12:45:49Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/discrete-neural-processes"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "SkNSehA9FQ", "year": 2019, "status": "rejected", "title": "Open Vocabulary Learning on Source Code with a Graph-Structured Cache", "authors": ["Milan Cvitkovic", "Badal Singh", "Anima Anandkumar"], "authorids": ["mcvitkov@caltech.edu", "sbadal@amazon.com", "anima@caltech.edu"], "authors_source": "OpenReview API", "abstract": "Machine learning models that take computer program source code as input typically use Natural Language Processing (NLP) techniques. However, a major challenge is that code is written using an open, rapidly changing vocabulary due to, e.g., the coinage of new variable and method names.  Reasoning over such a vocabulary is not something for which most NLP methods are designed.  We introduce a Graph-Structured Cache to address this problem; this cache contains a node for each new word the model encounters with edges connecting each word to its occurrences in the code.  We find that combining this graph-structured cache strategy with recent Graph-Neural-Network-based models for supervised learning on code improves the models' performance on a code completion task and a variable naming task --- with over 100\\% relative improvement on the latter --- at the cost of a moderate increase in computation time.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=SkNSehA9FQ", "arxiv_id": "1810.08305", "paper_pdf": "papers/SkNSehA9FQ.pdf", "paper_pdf_sha256": "04e967642c6def63ebea8e6c13e7cd2e54b2476abd66bc0f0877ad682f8b30c4", "paper_pdf_bytes": 1201854, "paper_pdf_source": "openreview", "code_url": "https://github.com/mwcvitkovic/Open-Vocabulary-Learning-on-Source-Code-with-a-Graph-Structured-Cache--Code-Preprocessor", "code_repository": "mwcvitkovic/Open-Vocabulary-Learning-on-Source-Code-with-a-Graph-Structured-Cache--Code-Preprocessor", "code_commit": "9ff7c8caa7eb536bb6c36602424820e095a0f838", "code_archive": "repos/SkNSehA9FQ.zip", "code_archive_sha256": "9d4466364cc4db425b0925b77524db5237f1283af34029cbe75a455857ce42a7", "code_archive_bytes": 4360675, "code_file_count": 1207, "code_extensions": {".java": 1202, ".sh": 4, ".js": 1}, "github_disk_usage_kb": 2846, "github_languages": {"Java": 6667739, "JavaScript": 4889, "Shell": 3280, "HTML": 1149}, "github_archived": false, "github_pushed_at": "2018-10-22T15:48:55Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/open-vocabulary-learning-on-source-code-with"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "yAns9ibCaE", "year": 2026, "status": "rejected", "title": "ByteMorph: Benchmarking Instruction-Guided Image Editing with Non-Rigid Motions", "authors": ["Di Chang", "Mingdeng Cao", "Yichun Shi", "Bo Liu", "Shengqu Cai", "Shijie Zhou", "Weilin Huang", "Gordon Wetzstein", "Mohammad Soleymani", "Peng Wang"], "authorids": ["~Di_Chang1", "~Mingdeng_Cao1", "~Yichun_Shi1", "~Bo_Liu50", "~Shengqu_Cai1", "~Shijie_Zhou1", "~Weilin_Huang1", "~Gordon_Wetzstein3", "~Mohammad_Soleymani2", "~Peng_Wang2"], "authors_source": "OpenReview API", "abstract": "Editing images with instructions to reflect non-rigid motions—camera viewpoint shifts, object deformations, human articulations, and complex interactions—poses a challenging yet underexplored problem in computer vision. Existing approaches and datasets predominantly focus on static scenes or rigid transformations, limiting their capacity to handle expressive edits involving dynamic motion. To address this gap, we introduce ByteMorph, a comprehensive framework for instruction-based image editing with an emphasis on non-rigid motions. ByteMorph comprises a large-scale dataset, ByteMorph-6M, and a strong baseline model built upon the Diffusion Transformer (DiT), named ByteMorpher. ByteMorph-6M includes over 6 million high-resolution image editing pairs for training, along with a carefully curated evaluation benchmark ByteMorph-Bench. Both capture a wide variety of non-rigid motion types across diverse environments, human figures, and object categories. The dataset is constructed using motion-guided data generation, layered compositing techniques, and automated captioning to ensure diversity, realism, and semantic coherence. We further conduct a comprehensive evaluation of recent instruction-based image editing methods from both academic and commercial domains.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=yAns9ibCaE", "arxiv_id": "2506.03107", "paper_pdf": "papers/yAns9ibCaE.pdf", "paper_pdf_sha256": "8b1a59096f47982b6b5b64a3fdb0a552a6f7ea60ce7add91aba240d8658f0fa1", "paper_pdf_bytes": 24503894, "paper_pdf_source": "openreview", "code_url": "https://github.com/ByteDance-Seed/BM-code", "code_repository": "ByteDance-Seed/BM-code", "code_commit": "594d54346db33fbce09bd435e77232a32b26e616", "code_archive": "repos/yAns9ibCaE.zip", "code_archive_sha256": "c34a6d6f7ed9cacfd8741b27bb143cffae761bb49875b67229d1dfcfa8e8de86", "code_archive_bytes": 1232953, "code_file_count": 154, "code_extensions": {".py": 110, ".java": 19, ".sh": 13, ".swift": 11, ".cpp": 1}, "github_disk_usage_kb": 1076, "github_languages": {"Python": 721377, "Java": 113288, "Swift": 58465, "C++": 9891, "CMake": 6607, "Shell": 5731, "Dockerfile": 922, "Ruby": 408}, "github_archived": false, "github_pushed_at": "2025-06-11T06:14:55Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/bytemorph-benchmarking-instruction-guided"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "cho9iE9POr", "year": 2025, "status": "rejected", "title": "Low-Budget Simulation-Based Inference with Bayesian Neural Networks", "authors": ["Arnaud Delaunoy", "Maxence de la Brassinne Bonardeaux", "Siddharth Mishra-Sharma", "Gilles Louppe"], "authorids": ["~Arnaud_Delaunoy1", "~Maxence_de_la_Brassinne_Bonardeaux1", "~Siddharth_Mishra-Sharma1", "~Gilles_Louppe1"], "authors_source": "OpenReview API", "abstract": "Simulation-based inference methods have been shown to be inaccurate in the data-poor regime, when training simulations are limited or expensive.\nUnder these circumstances, the inference network is particularly prone to overfitting, and using it without accounting for the computational uncertainty arising from the lack of identifiability of the network weights can lead to unreliable results.\nTo address this issue, we propose using Bayesian neural networks in low-budget simulation-based inference, thereby explicitly accounting for the computational uncertainty of the posterior approximation.\nWe design a family of Bayesian neural network priors that are tailored for inference and show that they lead to well-calibrated posteriors on tested benchmarks, even when as few as $O(10)$ simulations are available.\nThis opens up the possibility of performing reliable simulation-based inference using very expensive simulators, as we demonstrate on a problem from the field of cosmology where single simulations are computationally expensive. We show that Bayesian neural networks produce informative and well-calibrated posterior estimates with only a few hundred simulations.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=cho9iE9POr", "arxiv_id": "2408.15136", "paper_pdf": "papers/cho9iE9POr.pdf", "paper_pdf_sha256": "b2e269634625109e5d3db155a64f8922c43bab9237b589e7500334aeb9361f66", "paper_pdf_bytes": 3258977, "paper_pdf_source": "openreview", "code_url": "https://github.com/ADelau/low_budget_sbi_with_bnn", "code_repository": "ADelau/low_budget_sbi_with_bnn", "code_commit": "413556a988453845bcc80c91ec959bf73dfc99db", "code_archive": "repos/cho9iE9POr.zip", "code_archive_sha256": "051aa5a0049272d3e830990fa13d000ae32f76a254df36bf836a982d42eaeff1", "code_archive_bytes": 87102, "code_file_count": 26, "code_extensions": {".py": 26}, "github_disk_usage_kb": 78, "github_languages": {"Python": 258561}, "github_archived": false, "github_pushed_at": "2024-09-05T11:27:50Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/low-budget-simulation-based-inference-with"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "xuxYaBMd9F", "year": 2024, "status": "rejected", "title": "Efficient Long Sequence Modeling via State Space Augmented Transformer", "authors": ["Simiao Zuo", "Xiaodong Liu", "Jian Jiao", "Denis X Charles", "Eren Manavoglu", "Tuo Zhao", "Jianfeng Gao"], "authorids": ["~Simiao_Zuo1", "~Xiaodong_Liu1", "~Jian_Jiao2", "~Denis_X_Charles1", "~Eren_Manavoglu1", "~Tuo_Zhao1", "~Jianfeng_Gao1"], "authors_source": "OpenReview API", "abstract": "Transformer models have achieved superior performance in various natural language processing tasks. However, the quadratic computational cost of the attention mechanism limits its practicality for long sequences. There are existing attention variants that improve the computational efficiency, but they have limited ability to effectively compute global information. In parallel to Transformer models, state space models (SSMs) are tailored for long sequences, but they are not flexible enough to capture complicated local information. We propose SPADE, short for State Space Augmented Transformer. Specifically, we augment a SSM into the bottom layer of SPADE, and we employ efficient local attention methods for the other layers. The SSM augments global information, which complements the lack of long-range dependency issue in local attention methods. Experimental results on the Long Range Arena benchmark and language modeling tasks demonstrate the effectiveness of the proposed method. To further demonstrate the scalability of SPADE, we pre-train large encoder-decoder models and present fine-tuning results on natural language understanding and natural language generation tasks. Our code and pre-trained model checkpoints will be publicly available.", "decision": "Reject", "meta_review": "", "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=xuxYaBMd9F", "arxiv_id": "2212.08136", "paper_pdf": "papers/xuxYaBMd9F.pdf", "paper_pdf_sha256": "e79869a8ec6bb64a4c2e49734bd2e3e5e4789efc9a95d424320c3c498c71acb9", "paper_pdf_bytes": 1419772, "paper_pdf_source": "openreview", "code_url": "https://github.com/microsoft/EfficientLongSequenceModeling", "code_repository": "microsoft/EfficientLongSequenceModeling", "code_commit": "0819294c9fbfb335eef08d54849aa8cb201de130", "code_archive": "repos/xuxYaBMd9F.zip", "code_archive_sha256": "ddbc311fe9b5ebc2417cd9cd7b0f77f04705721774599c7c4303cc55c5258aa5", "code_archive_bytes": 55936, "code_file_count": 10, "code_extensions": {".py": 9, ".sh": 1}, "github_disk_usage_kb": 60, "github_languages": {"Python": 179500, "Shell": 1079}, "github_archived": true, "github_pushed_at": "2023-01-19T02:15:53Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/efficient-long-sequence-modeling-via-state"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "UERcQuXlwy", "year": 2023, "status": "rejected", "title": "Pix2Struct: Screenshot Parsing as Pretraining for Visual Language Understanding", "authors": ["Kenton Lee", "Mandar Joshi", "Iulia Raluca Turc", "Hexiang Hu", "Fangyu Liu", "Julian Martin Eisenschlos", "Urvashi Khandelwal", "Peter Shaw", "Ming-Wei Chang", "Kristina Toutanova"], "authorids": ["~Kenton_Lee1", "~Mandar_Joshi1", "~Iulia_Raluca_Turc1", "~Hexiang_Hu1", "~Fangyu_Liu1", "~Julian_Martin_Eisenschlos1", "~Urvashi_Khandelwal1", "~Peter_Shaw1", "~Ming-Wei_Chang3", "~Kristina_Toutanova1"], "authors_source": "OpenReview API", "abstract": "Visually-situated language is ubiquitous---sources range from textbooks with diagrams to web pages with images and tables, to mobile apps with buttons and forms. Perhaps due to this diversity, previous work has typically relied on domain-specific recipes with limited sharing of the underlying data, model architectures, and objectives. We present Pix2Struct, a pretrained image-to-text model for purely visual language understanding, which can be finetuned on tasks containing visually-situated language. Pix2Struct is pretrained by learning to parse masked screenshots of web pages into simplified HTML. The web, with its richness of visual elements cleanly reflected in the HTML structure, provides a large source of pretraining data well suited to the diversity of downstream tasks. Intuitively, this objective subsumes common pretraining signals such as OCR, language modeling, image captioning. In addition to the novel pretraining strategy, we introduce a variable-resolution input representation and a more flexible integration of language and vision inputs, where language prompts such as questions are rendered directly on top of the input image. For the first time, we show that a single pretrained model can achieve state-of-the-art results in six out of nine tasks across four domains: documents, illustrations, user interfaces, and natural images. ", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=UERcQuXlwy", "arxiv_id": "2210.03347", "paper_pdf": "papers/UERcQuXlwy.pdf", "paper_pdf_sha256": "bf3b428b75eb08cb36871c0e22e98c459858f89954d2186480352db57b104642", "paper_pdf_bytes": 5865162, "paper_pdf_source": "openreview", "code_url": "https://github.com/google-research/pix2struct", "code_repository": "google-research/pix2struct", "code_commit": "6fe25c1dc8151823ee3b479519d8d5948812fee4", "code_archive": "repos/UERcQuXlwy.zip", "code_archive_sha256": "a7c676f1991d4f20a1117451df7359ab684cb2e71ab8f5fbf1d605a3fb02a7a7", "code_archive_bytes": 69627, "code_file_count": 32, "code_extensions": {".py": 32}, "github_disk_usage_kb": 118, "github_languages": {"Python": 97547, "HTML": 2552, "CSS": 253}, "github_archived": false, "github_pushed_at": "2026-09-02T05:50:49Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/pix2struct-screenshot-parsing-as-pretraining"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "s5lIqsrOu3Z", "year": 2022, "status": "rejected", "title": "Closed-Loop Data Transcription to an LDR via Minimaxing Rate Reduction", "authors": ["Xili Dai", "Shengbang Tong", "Mingyang Li", "Ziyang Wu", "Kwan Ho Ryan Chan", "Pengyuan Zhai", "Yaodong Yu", "Michael Psenka", "Xiaojun Yuan", "Heung-Yeung Shum", "Yi Ma"], "authorids": ["~Xili_Dai2", "~Shengbang_Tong1", "~Mingyang_Li3", "~Ziyang_Wu1", "~Kwan_Ho_Ryan_Chan1", "~Pengyuan_Zhai1", "~Yaodong_Yu4", "psenka@berkeley.edu", "xjyuan@uestc.edu.cn", "msraharry@hotmail.com", "~Yi_Ma4"], "authors_source": "OpenReview API", "abstract": "This work proposes a new computational framework for automatically learning a closed-loop transcription between multi-class multi-dimensional data and a linear discriminative representation (LDR) that consists of multiple multi-dimensional linear subspaces. In particular, we argue that the optimal encoding and decoding mappings sought can be formulated as the equilibrium point of a two-player minimax game between the encoder and decoder. A natural  utility function for this game is the so-called rate reduction, a simple information-theoretic measure for distances between mixtures of subspace-like Gaussians in the feature space. Our formulation avoids expensive evaluating and minimizing approximated distances between arbitrary distributions in either the data space or the feature space. To a large extent, conceptually and computationally this new formulation unifies the benefits of Auto-Encoding and GAN and naturally extends them to the settings of learning a both discriminative and generative representation for complex multi-class and multi-dimensional real-world data. Our extensive experiments on many benchmark datasets demonstrate tremendous potential of this framework: under fair comparison, visual quality of the learned decoder and classification performance of the encoder is competitive and often better than existing methods based on GAN, VAE or a combination of both. ", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=s5lIqsrOu3Z", "arxiv_id": "2111.06636", "paper_pdf": "papers/s5lIqsrOu3Z.pdf", "paper_pdf_sha256": "14357277906801d25c6b68c075eec61eee7f7b857fa4e4fc669c99f0faa5d5d0", "paper_pdf_bytes": 46561409, "paper_pdf_source": "openreview", "code_url": "https://github.com/Delay-Xili/LDR", "code_repository": "Delay-Xili/LDR", "code_commit": "4c6098a4ab03a67be108d4567709cd8ac013dfea", "code_archive": "repos/s5lIqsrOu3Z.zip", "code_archive_sha256": "eeb94b9216cb353c7c39a1471bc0b583d2e3cd2e7a317d551e30927cd8f04638", "code_archive_bytes": 101370, "code_file_count": 15, "code_extensions": {".py": 15}, "github_disk_usage_kb": 149, "github_languages": {"Python": 106107}, "github_archived": false, "github_pushed_at": "2022-11-03T08:35:04Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/closed-loop-data-transcription-to-an-ldr-via-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "D1E1h-K3jso", "year": 2021, "status": "rejected", "title": "Learning from Noisy Data with Robust Representation Learning", "authors": ["Junnan Li", "Caiming Xiong", "Steven Hoi"], "authorids": ["~Junnan_Li2", "~Caiming_Xiong1", "~Steven_Hoi2"], "authors_source": "OpenReview API", "abstract": "Learning from noisy data has attracted much attention, where most methods focus on label noise. In this work, we propose a new framework which simultaneously addresses three types of noise commonly seen in real-world data: label noise, out-of-distribution input, and input corruption. In contrast to most existing methods, we combat noise by learning robust representation. Specifically, we embed images into a low-dimensional subspace by training an autoencoder on the deep features. We regularize the geometric structure of the subspace with robust contrastive learning, which includes an unsupervised consistency loss and a supervised mixup prototypical loss. Furthermore, we leverage the structure of the learned subspace for noise cleaning, by aggregating information from neighboring samples. Experiments on multiple benchmarks demonstrate state-of-the-art performance of our method and robustness of the learned representation. Our code will be released.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=D1E1h-K3jso", "arxiv_id": null, "paper_pdf": "papers/D1E1h-K3jso.pdf", "paper_pdf_sha256": "98f2bfb372584218b2c0e92bde106169f820c26b5c296d6457148bd4515b29bc", "paper_pdf_bytes": 2071851, "paper_pdf_source": "openreview", "code_url": "https://github.com/salesforce/RRL", "code_repository": "salesforce/RRL", "code_commit": "a0d8a8fca05d040aef15b82f4870d33b65e59ce7", "code_archive": "repos/D1E1h-K3jso.zip", "code_archive_sha256": "4a9da6528cdfcf66ce4d0a678d027ff9473da7cad0c7d96068f36df3efad77fe", "code_archive_bytes": 178117, "code_file_count": 22, "code_extensions": {".py": 21, ".ipynb": 1}, "github_disk_usage_kb": 177, "github_languages": {"Python": 106457, "Jupyter Notebook": 31049}, "github_archived": true, "github_pushed_at": "2025-05-01T17:27:53Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/learning-from-noisy-data-with-robust"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "BJlXgkHYvS", "year": 2020, "status": "rejected", "title": "Information-Theoretic Local Minima Characterization and Regularization", "authors": ["Zhiwei Jia", "Hao Su"], "authorids": ["zjia@ucsd.edu", "haosu@eng.ucsd.edu"], "authors_source": "OpenReview API", "abstract": "Recent advances in deep learning theory have evoked the study of generalizability across different local minima of deep neural networks (DNNs). While current work focused on either discovering properties of good local minima or developing regularization techniques to induce good local minima, no approach exists that can tackle both problems. We achieve these two goals successfully in a unified manner. Specifically, based on the Fisher information we propose a metric both strongly indicative of generalizability of local minima and effectively applied as a practical regularizer. We provide theoretical analysis including a generalization bound and empirically demonstrate the success of our approach in both capturing and improving the generalizability of DNNs. Experiments are performed on CIFAR-10 and CIFAR-100 for various network architectures.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=BJlXgkHYvS", "arxiv_id": "1911.08192", "paper_pdf": "papers/BJlXgkHYvS.pdf", "paper_pdf_sha256": "334434fbf38a1c4fcb92a8a2c4b98e3ed73b6f5362c93cd1543e7fe26efaf219", "paper_pdf_bytes": 949579, "paper_pdf_source": "openreview", "code_url": "https://github.com/SeanJia/InfoMCR", "code_repository": "SeanJia/InfoMCR", "code_commit": "2b4760ad6ffdd98859fea1967eb1b8aa7e51be52", "code_archive": "repos/BJlXgkHYvS.zip", "code_archive_sha256": "dbc42beb7a7019f2a888cee260a8069ed4a9a34eef9a6eab33bbcf00465d4a33", "code_archive_bytes": 152435, "code_file_count": 8, "code_extensions": {".py": 8}, "github_disk_usage_kb": 193, "github_languages": {"Python": 17713}, "github_archived": false, "github_pushed_at": "2020-07-16T06:28:26Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/information-theoretic-local-minima-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Bke96sC5tm", "year": 2019, "status": "rejected", "title": "SOLAR: Deep Structured Representations for Model-Based Reinforcement Learning", "authors": ["Marvin Zhang*", "Sharad Vikram*", "Laura Smith", "Pieter Abbeel", "Matthew Johnson", "Sergey Levine"], "authorids": ["marvin@cs.berkeley.edu", "svikram@cs.ucsd.edu", "smithlaura@berkeley.edu", "pabbeel@cs.berkeley.edu", "mattjj@google.com", "svlevine@cs.berkeley.edu"], "authors_source": "OpenReview API", "abstract": "Model-based reinforcement learning (RL) methods can be broadly categorized as global model methods, which depend on learning models that provide sensible predictions in a wide range of states, or local model methods, which iteratively refit simple models that are used for policy improvement. While predicting future states that will result from the current actions is difficult, local model methods only attempt to understand system dynamics in the neighborhood of the current policy, making it possible to produce local improvements without ever learning to predict accurately far into the future. The main idea in this paper is that we can learn representations that make it easy to retrospectively infer simple dynamics given the data from the current policy, thus enabling local models to be used for policy learning in complex systems. We evaluate our approach against other model-based and model-free RL methods on a suite of robotics tasks, including manipulation tasks on a real Sawyer robotic arm directly from camera images.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=Bke96sC5tm", "arxiv_id": "1808.09105", "paper_pdf": "papers/Bke96sC5tm.pdf", "paper_pdf_sha256": "65447d057605cd4d9d71bf56795f9b7a786b6fa8b383ad976d440a77f01376bf", "paper_pdf_bytes": 1703454, "paper_pdf_source": "openreview", "code_url": "https://github.com/sharadmv/parasol", "code_repository": "sharadmv/parasol", "code_commit": "493d87829691428b7183bc6432200032ac86f835", "code_archive": "repos/Bke96sC5tm.zip", "code_archive_sha256": "cc76e2cb8a3aecd19765bcf5bdebc5f48d4be5f6170400a9b2171dc10a593d4b", "code_archive_bytes": 2603953, "code_file_count": 44, "code_extensions": {".py": 44}, "github_disk_usage_kb": 2883, "github_languages": {"Python": 133967}, "github_archived": false, "github_pushed_at": "2021-09-08T00:46:28Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/solar-deep-structured-representations-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "6hdzuMTNmX", "year": 2026, "status": "rejected", "title": "SIMSHIFT: A Benchmark for Adapting Neural Surrogates to Distribution Shifts", "authors": ["Paul Setinek", "Gianluca Galletti", "Thomas Gross", "Dominik Schnürer", "Johannes Brandstetter", "Werner Zellinger"], "authorids": ["~Paul_Setinek1", "~Gianluca_Galletti1", "~Thomas_Gross2", "~Dominik_Schnürer1", "~Johannes_Brandstetter1", "~Werner_Zellinger1"], "authors_source": "OpenReview API", "abstract": "Neural surrogates for Partial Differential Equations (PDEs) often suffer significant performance degradation when evaluated on unseen problem configurations, such as new initial conditions or structural dimensions.\nMeanwhile, Domain Adaptation (DA) techniques have been widely used in vision and language processing to generalize from limited information about unseen configurations.\nIn this work, we address this gap through two focused contributions.\nFirst, we introduce SIMSHIFT, a novel benchmark dataset and evaluation suite composed of four industrial simulation tasks spanning diverse processes and physics: _hot rolling_, _sheet metal forming_, _electric motor design_ and _heatsink design_.\nSecond, we extend established DA methods to state-of-the-art neural surrogates and systematically evaluate them.\nThese approaches use parametric descriptions and ground truth simulations from multiple source configurations, together with only parametric descriptions from target configurations.\nThe goal is to accurately predict target simulations without access to ground truth simulation data.\nExtensive experiments on SIMSHIFT highlight the challenges of out of distribution neural surrogate modeling, demonstrate the potential of DA in simulation, and reveal open problems in achieving robust neural surrogates under distribution shifts in industrially relevant scenarios.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=6hdzuMTNmX", "arxiv_id": "2506.12007", "paper_pdf": "papers/6hdzuMTNmX.pdf", "paper_pdf_sha256": "16b963705e3f90de7912fe3357b27b807f5110e86b308150eee45a2b84628e09", "paper_pdf_bytes": 11056205, "paper_pdf_source": "openreview", "code_url": "https://github.com/psetinek/simshift", "code_repository": "psetinek/simshift", "code_commit": "d30db3951e8aae422aeba4329781c9424745db37", "code_archive": "repos/6hdzuMTNmX.zip", "code_archive_sha256": "3adbd7373b26ae27f514f15c957cbcfba2f40719356afc161b28be5de77d187f", "code_archive_bytes": 880923, "code_file_count": 55, "code_extensions": {".py": 53, ".ipynb": 2}, "github_disk_usage_kb": 1095, "github_languages": {"Python": 294481, "Jupyter Notebook": 203201}, "github_archived": false, "github_pushed_at": "2026-05-19T13:44:36Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/simshift-a-benchmark-for-adapting-neural"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "4Po8d9GAfQ", "year": 2025, "status": "rejected", "title": "Language Models are Hidden Reasoners: Unlocking Latent Reasoning Capabilities via Self-Rewarding", "authors": ["Haolin Chen", "Yihao Feng", "Zuxin Liu", "Weiran Yao", "Akshara Prabhakar", "Shelby Heinecke", "Ricky Ho", "Phil L Mui", "Silvio Savarese", "Caiming Xiong", "Huan Wang"], "authorids": ["~Haolin_Chen2", "~Yihao_Feng1", "~Zuxin_Liu1", "~Weiran_Yao1", "~Akshara_Prabhakar1", "~Shelby_Heinecke1", "~Ricky_Ho2", "~Phil_L_Mui1", "~Silvio_Savarese1", "~Caiming_Xiong1", "~Huan_Wang1"], "authors_source": "OpenReview API", "abstract": "Large language models (LLMs) have shown impressive capabilities, but still struggle with complex reasoning tasks requiring multiple steps. While prompt-based methods like Chain-of-Thought (CoT) can improve LLM reasoning at inference time, optimizing reasoning capabilities during training remains challenging. We introduce LaTent Reasoning Optimization (LaTRO), a principled framework that formulates reasoning as sampling from a latent distribution and optimizes it via variational approaches. LaTRO enables LLMs to concurrently improve both their reasoning process and ability to evaluate reasoning quality, without requiring external feedback or reward models. We validate LaTRO through experiments on GSM8K and ARC-Challenge datasets using multiple model architectures. On GSM8K, LaTRO improves zero-shot accuracy by an average of 12.5\\% over base models and 9.6\\% over supervised fine-tuning across Phi-3.5-mini, Mistral-7B, and Llama-3.1-8B. Our findings suggest that pre-trained LLMs possess latent reasoning capabilities that can be unlocked and enhanced through our proposed optimization approach in a self-improvement manner.", "decision": "Reject", "meta_review": "", "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=4Po8d9GAfQ", "arxiv_id": "2411.04282", "paper_pdf": "papers/4Po8d9GAfQ.pdf", "paper_pdf_sha256": "6e4e45ee93db74ef1508625f12bb06e1064518bc109de3daf9124b4ccf17c059", "paper_pdf_bytes": 436976, "paper_pdf_source": "openreview", "code_url": "https://github.com/SalesforceAIResearch/LaTRO", "code_repository": "SalesforceAIResearch/LaTRO", "code_commit": "54f9356974cc138cd39ef79ec93eb9a421d5976f", "code_archive": "repos/4Po8d9GAfQ.zip", "code_archive_sha256": "a12ae43dd908bd38d7a25e132d2738013646df2fa1519032f53249f15c2a6f66", "code_archive_bytes": 44741, "code_file_count": 16, "code_extensions": {".py": 13, ".sh": 3}, "github_disk_usage_kb": 82, "github_languages": {"Python": 81899, "Shell": 3499}, "github_archived": false, "github_pushed_at": "2026-06-02T18:54:57Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/language-models-are-hidden-reasoners"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "pK7V0glCdj", "year": 2024, "status": "rejected", "title": "BOtied: Multi-objective Bayesian optimization with tied multivariate ranks", "authors": ["Ji Won Park", "Natasa Tagasovska", "Michael Maser", "Stephen Ra", "Kyunghyun Cho"], "authorids": ["~Ji_Won_Park1", "~Natasa_Tagasovska2", "~Michael_Maser1", "~Stephen_Ra1", "~Kyunghyun_Cho1"], "authors_source": "OpenReview API", "abstract": "Many scientific and industrial applications require the joint optimization of multiple, potentially competing objectives. Multi-objective Bayesian optimization (MOBO) is a sample-efficient framework for identifying Pareto-optimal solutions. At the heart of MOBO is the acquisition function, which determines the next candidate to evaluate by navigating the best compromises among the objectives. Multi-objective acquisition functions that rely on box decomposition of the objective space, such as the expected hypervolume improvement (EHVI) and entropy search, scale poorly to a large number of objectives. We begin by showing a natural connection between non-dominated solutions and the highest multivariate rank, which coincides with the outermost level line of the joint cumulative distribution function (CDF). Motivated by this link, we propose the CDF indicator, a Pareto-compliant metric for evaluating the quality of approximate Pareto sets that complements the popular hypervolume indicator. We then propose an acquisition function based on the CDF indicator, called BOtied. BOtied can be implemented efficiently with copulas, a statistical tool for modeling complex, high-dimensional distributions. We benchmark BOtied against common acquisition functions, including EHVI, entropy search, and random scalarization, in a series of synthetic and real-data experiments. BOtied performs on par with the baselines across datasets and metrics while being computationally efficient.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=pK7V0glCdj", "arxiv_id": "2306.00344", "paper_pdf": "papers/pK7V0glCdj.pdf", "paper_pdf_sha256": "c11503408aacc3eb4a34c674606adf53d59f31bfa6f344f0a23098cab7efa6fc", "paper_pdf_bytes": 1949876, "paper_pdf_source": "openreview", "code_url": "https://github.com/jiwoncpark/botied", "code_repository": "jiwoncpark/botied", "code_commit": "1d8eeb17b632a3b97a3a30519a939070178b4575", "code_archive": "repos/pK7V0glCdj.zip", "code_archive_sha256": "4d670880077cdee682595e67b6ab02f8d01e5dfb6e50a138a93917eae0dc1378", "code_archive_bytes": 71689, "code_file_count": 44, "code_extensions": {".py": 38, ".sh": 6}, "github_disk_usage_kb": 61, "github_languages": {"Python": 140924, "Shell": 3938}, "github_archived": false, "github_pushed_at": "2024-06-04T05:00:14Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/botied-multi-objective-bayesian-optimization"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "vDY5Y8HMNxO", "year": 2023, "status": "rejected", "title": "GMML is All you Need", "authors": ["Sara Atito", "Muhammad Awais", "Josef Kittler"], "authorids": ["~Sara_Atito1", "muhammad.awais@surrey.ac.uk", "~Josef_Kittler1"], "authors_source": "OpenReview API", "abstract": "Vision transformers have generated significant interest in the computer vision (CV) community because of their flexibility in exploiting contextual information, whether it is sharply confined local, or long range global. However, they are known to be data hungry. This has motivated the research in self-supervised transformer pretraining, which does not need to decode the semantic information conveyed by labels to link it to the image properties, but rather focuses directly on extracting a concise representation of the image data that reflects the notion of similarity and is invariant to nuisance factors. The key vehicle for the self-learning process used by the majority of self-learning methods is the generation of multiple views of the training data and the creation of pretext tasks which use these views to define the notion of image similarity and data integrity. However, this approach lacks the natural propensity to extract contextual information. We propose group mask model learning (GMML), a self-supervised learning (SSL) mechanism for pretraining vision transformers with the ability to extract the contextual information present in all the concepts in an image. GMML achieves this by manipulating random groups of connected tokens, ensuingly covering a meaningful part of a semantic concept, and then recovering the hidden semantic information from the visible part of the concept. GMML implicitly introduces a novel data augmentation process. Unlike most of the existing SSL approaches, GMML does not require momentum encoder, nor rely on careful implementation details such as large batches and gradient stopping, which are all artefacts of most of the current self-supervised learning techniques. Since its conception at the beginning of 2021, GMML maintains itself as unbeaten SSL method with several desirable benefits and marked a significant milestone in computer vision by being one of the first self-supervised pretraining methods which outperform supervised pretraining consistently with a large margin. GMML is simple, elegant, and currently the best mechanism to extract information from a given dataset and instil this information into transformer's weights. The code will be made publicly available for the community to train on bigger corpora.   ", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=vDY5Y8HMNxO", "arxiv_id": "2205.14986", "paper_pdf": "papers/vDY5Y8HMNxO.pdf", "paper_pdf_sha256": "f07fe9b92278b39fe118f4556af1996ea075966c1e7d8161e155158e60001517", "paper_pdf_bytes": 2678807, "paper_pdf_source": "openreview", "code_url": "https://github.com/Sara-Ahmed/GMML", "code_repository": "Sara-Ahmed/GMML", "code_commit": "1882712745e87166c26dd77c1dd346f84fde6064", "code_archive": "repos/vDY5Y8HMNxO.zip", "code_archive_sha256": "3d0fed49275870d485c00b45dd2ba30cf136ffb351d879e03f641a5288c5e5c6", "code_archive_bytes": 105705, "code_file_count": 7, "code_extensions": {".py": 7}, "github_disk_usage_kb": 118, "github_languages": {"Python": 76742}, "github_archived": false, "github_pushed_at": "2022-08-12T10:43:04Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/gmml-is-all-you-need"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "RNf9AgtRtL", "year": 2022, "status": "rejected", "title": "Continuous Control With Ensemble Deep Deterministic Policy Gradients", "authors": ["Piotr Januszewski", "Mateusz Olko", "Michał Królikowski", "Jakub Swiatkowski", "Marcin Andrychowicz", "Łukasz Kuciński", "Piotr Miłoś"], "authorids": ["~Piotr_Januszewski1", "~Mateusz_Olko1", "~Michał_Królikowski1", "~Jakub_Swiatkowski1", "~Marcin_Andrychowicz1", "~Łukasz_Kuciński1", "~Piotr_Miłoś1"], "authors_source": "OpenReview API", "abstract": "The growth of deep reinforcement learning (RL) has brought multiple exciting tools and methods to the field. This rapid expansion makes it important to understand the interplay between individual elements of the RL toolbox. We approach this task from an empirical perspective by conducting a study in the continuous control setting. We present multiple insights of fundamental nature, including: a commonly used additive action noise is not required for effective exploration and can even hinder training; the performance of policies trained using existing methods varies significantly across training runs, epochs of training, and evaluation runs; the critics' initialization plays the major role in ensemble-based actor-critic exploration, while the training is mostly invariant to the actors' initialization; a strategy based on posterior sampling explores better than the approximated UCB combined with the weighted Bellman backup; the weighted Bellman backup alone cannot replace the clipped double Q-Learning. As a conclusion, we show how existing tools can be brought together in a novel way, giving rise to the Ensemble Deep Deterministic Policy Gradients (ED2) method, to yield state-of-the-art results on continuous control tasks from \\mbox{OpenAI Gym MuJoCo}. From the practical side, ED2 is conceptually straightforward, easy to code, and does not require knowledge outside of the existing RL toolbox.", "decision": "Reject", "meta_review": null, "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=RNf9AgtRtL", "arxiv_id": "2111.15382", "paper_pdf": "papers/RNf9AgtRtL.pdf", "paper_pdf_sha256": "93c3ca8bcb21ff96bcfde5c2e39a47af731f5f260a87c10485e7dd3b9e6e95f6", "paper_pdf_bytes": 3435979, "paper_pdf_source": "openreview", "code_url": "https://github.com/ed2-paper/ED2", "code_repository": "ed2-paper/ED2", "code_commit": "884e5428fd47d63266dbdd6100d13235d9322595", "code_archive": "repos/RNf9AgtRtL.zip", "code_archive_sha256": "cbfd2edd960aa3c1c93be7e465aafd513148429afdcd7c4ba83bcda9d23e0692", "code_archive_bytes": 167634, "code_file_count": 18, "code_extensions": {".py": 18}, "github_disk_usage_kb": 149, "github_languages": {"Python": 105353}, "github_archived": false, "github_pushed_at": "2021-06-09T15:38:37Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/continuous-control-with-ensemble-deep-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "7qmQNB6Wn_B", "year": 2021, "status": "rejected", "title": "Diversity Actor-Critic: Sample-Aware Entropy Regularization for Sample-Efficient Exploration", "authors": ["Seungyul Han", "Youngchul Sung"], "authorids": ["~Seungyul_Han1", "~Youngchul_Sung1"], "authors_source": "OpenReview API", "abstract": "Policy entropy regularization is commonly used for better exploration in deep reinforcement learning (RL). However, policy entropy regularization is sample-inefficient in off-policy learning since it does not take the distribution of previous samples stored in the replay buffer into account. In order to take advantage of the previous sample distribution from the replay buffer for sample-efficient exploration, we propose sample-aware entropy regularization which maximizes the entropy of weighted sum of the policy action distribution and the sample action distribution from the replay buffer. We formulate the problem of sample-aware entropy regularized policy iteration,  prove its convergence, and provide a practical algorithm named diversity actor-critic (DAC) which is a generalization of soft actor-critic (SAC). Numerical results show that DAC significantly outperforms SAC baselines and other state-of-the-art RL algorithms.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=7qmQNB6Wn_B", "arxiv_id": "2006.01419", "paper_pdf": "papers/7qmQNB6Wn_B.pdf", "paper_pdf_sha256": "95b008894bc06eb7c06ee22a414224b5c2141f25039856f3f0c7f5500b8cafcc", "paper_pdf_bytes": 5198691, "paper_pdf_source": "openreview", "code_url": "https://github.com/seungyulhan/dac", "code_repository": "seungyulhan/dac", "code_commit": "3595d8724ac47486812c7067bce3459f65268af7", "code_archive": "repos/7qmQNB6Wn_B.zip", "code_archive_sha256": "945a0228cdd566c07bbfb407f5ff217e939fb6680f9b72c38ed0673313100ebe", "code_archive_bytes": 188305, "code_file_count": 36, "code_extensions": {".py": 36}, "github_disk_usage_kb": 180, "github_languages": {"Python": 103683}, "github_archived": false, "github_pushed_at": "2022-08-17T04:17:38Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/diversity-actor-critic-sample-aware-entropy"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "HklliySFDS", "year": 2020, "status": "rejected", "title": "Continual Learning with Gated Incremental Memories for Sequential Data Processing", "authors": ["Andrea Cossu", "Antonio Carta", "Davide Bacciu"], "authorids": ["cossu48@gmail.com", "antonio.carta@di.unipi.it", "bacciu@di.unipi.it"], "authors_source": "OpenReview API", "abstract": "The ability to learn over changing task distributions without forgetting previous knowledge, also known as continual learning, is a key enabler for scalable and trustworthy deployments of adaptive solutions. While the importance of continual learning is largely acknowledged in machine vision and reinforcement learning problems, this is mostly under-documented for sequence processing tasks. This work focuses on characterizing and quantitatively assessing the impact of catastrophic forgetting and task interference when dealing with sequential data in recurrent neural networks. We also introduce a general architecture, named Gated Incremental Memory, for augmenting recurrent models with continual learning skills, whose effectiveness is demonstrated through the benchmarks introduced in this paper.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=HklliySFDS", "arxiv_id": "2004.04077", "paper_pdf": "papers/HklliySFDS.pdf", "paper_pdf_sha256": "9c13a3a03a54690dc8556b27004b38ef002231d99b468a620c27bb82bf854036", "paper_pdf_bytes": 331239, "paper_pdf_source": "openreview", "code_url": "https://github.com/AndreaCossu/ContinualLearning-SequentialProcessing", "code_repository": "AndreaCossu/ContinualLearning-SequentialProcessing", "code_commit": "d1b1b6a99a0f5b294d37bca5e82cbbd80ad7722d", "code_archive": "repos/HklliySFDS.zip", "code_archive_sha256": "a639ee189c1ff0ef0d78f5d7ed79402516abe9ca6dc17a1cffa3884a8e6890ec", "code_archive_bytes": 176926, "code_file_count": 17, "code_extensions": {".py": 17}, "github_disk_usage_kb": 198, "github_languages": {"Python": 121707}, "github_archived": false, "github_pushed_at": "2021-10-13T07:35:56Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/continual-learning-with-gated-incremental-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "SJMnG2C9YX", "year": 2019, "status": "rejected", "title": "Complementary-label learning for arbitrary losses and models", "authors": ["Takashi Ishida", "Gang Niu", "Aditya Krishna Menon", "Masashi Sugiyama"], "authorids": ["ishida@ms.k.u-tokyo.ac.jp", "gang.niu@riken.jp", "aditya.menon@anu.edu.au", "sugi@k.u-tokyo.ac.jp"], "authors_source": "OpenReview API", "abstract": "In contrast to the standard classification paradigm where the true (or possibly noisy) class is given to each training pattern, complementary-label learning only uses training patterns each equipped with a complementary label. This only specifies one of the classes that the pattern does not belong to. The seminal paper on complementary-label learning proposed an unbiased estimator of the classification risk that can be computed only from complementarily labeled data. How- ever, it required a restrictive condition on the loss functions, making it impossible to use popular losses such as the softmax cross-entropy loss. Recently, another formulation with the softmax cross-entropy loss was proposed with consistency guarantee. However, this formulation does not explicitly involve a risk estimator. Thus model/hyper-parameter selection is not possible by cross-validation— we may need additional ordinarily labeled data for validation purposes, which is not available in the current setup. In this paper, we give a novel general framework of complementary-label learning, and derive an unbiased risk estimator for arbitrary losses and models. We further improve the risk estimator by non-negative correction and demonstrate its superiority through experiments.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=SJMnG2C9YX", "arxiv_id": "1810.04327", "paper_pdf": "papers/SJMnG2C9YX.pdf", "paper_pdf_sha256": "7cdefe80564d02407c5993c74ed1cf34abf6f01d5db745ef6770d76041006673", "paper_pdf_bytes": 1861159, "paper_pdf_source": "openreview", "code_url": "https://github.com/takashiishida/comp", "code_repository": "takashiishida/comp", "code_commit": "aab868d49eefe4443865686574684253fe9b2487", "code_archive": "repos/SJMnG2C9YX.zip", "code_archive_sha256": "071487c7d6a897d3070780267d0faf0a2c54634dd4e6185952204176776365e5", "code_archive_bytes": 6265, "code_file_count": 4, "code_extensions": {".py": 4}, "github_disk_usage_kb": 3370, "github_languages": {"Python": 10504}, "github_archived": false, "github_pushed_at": "2025-08-17T02:03:56Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/complementary-label-learning-for-arbitrary"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "GhM63V7z6v", "year": 2025, "status": "rejected", "title": "Temporal Source Recovery for Time-Series Source-Free Unsupervised Domain Adaptation", "authors": ["Yucheng Wang", "Peiliang Gong", "Min Wu", "Felix Ott", "Xiaoli Li", "Lihua Xie", "Zhenghua Chen"], "authorids": ["~Yucheng_Wang10", "~Peiliang_Gong1", "~Min_Wu2", "~Felix_Ott1", "~Xiaoli_Li1", "~Lihua_Xie2", "~Zhenghua_Chen2"], "authors_source": "OpenReview API", "abstract": "Source-Free Unsupervised Domain Adaptation (SFUDA) has gained popularity for its ability to adapt pretrained models to target domains without accessing source domains, ensuring source data privacy. While SFUDA is well-developed in visual tasks, its application to Time-Series SFUDA (TS-SFUDA) remains limited due to the challenge of transferring crucial temporal dependencies across domains. Although a few researchers begin to explore this area, they rely on specific source domain designs, which are impractical as source data owners cannot be expected to follow particular pretraining protocols. To solve this, we propose Temporal Source Recovery (TemSR), a framework that transfers temporal dependencies for effective TS-SFUDA without requiring source-specific designs. TemSR features a recovery process that leverages masking, recovery, and optimization to generate a source-like distribution with recovered source temporal dependencies. To ensure effective recovery, we further design segment-based regularization to restore local dependencies and anchor-based recovery diversity maximization to enhance the diversity of the source-like distribution. The source-like distribution is then adapted to the target domain using traditional UDA techniques. Extensive experiments across multiple TS tasks demonstrate the effectiveness of TemSR, even surpassing existing TS-SFUDA method that requires source domain designs.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=GhM63V7z6v", "arxiv_id": "2409.19635", "paper_pdf": "papers/GhM63V7z6v.pdf", "paper_pdf_sha256": "87906c60a1cecfd7adf8ac59fa022e93883416d799e03474b292b9333385bb46", "paper_pdf_bytes": 766888, "paper_pdf_source": "openreview", "code_url": "https://github.com/Frank-Wang-oss/TemSR", "code_repository": "Frank-Wang-oss/TemSR", "code_commit": "21e14339b14868faf270fcc10046d9fc138bf2e4", "code_archive": "repos/GhM63V7z6v.zip", "code_archive_sha256": "526e1afa913d20ab7b0486034ccbe5e3136f05e7d4fdb339ff67050da286ebfe", "code_archive_bytes": 66508, "code_file_count": 10, "code_extensions": {".py": 10}, "github_disk_usage_kb": 84, "github_languages": {"Python": 91336}, "github_archived": false, "github_pushed_at": "2024-10-02T12:39:36Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/temporal-source-recovery-for-time-series"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "23OEmHVkpq", "year": 2024, "status": "rejected", "title": "Disentanglement Learning via Topology", "authors": ["Nikita Balabin", "Daria Voronkova", "Ilya Trofimov", "Evgeny Burnaev", "Serguei Barannikov"], "authorids": ["~Nikita_Balabin1", "~Daria_Voronkova1", "~Ilya_Trofimov1", "~Evgeny_Burnaev1", "~Serguei_Barannikov1"], "authors_source": "OpenReview API", "abstract": "We propose TopDis (Topological Disentanglement), a method for learning disentangled representations via adding a multi-scale topological loss term. Disentanglement is a crucial property of data representations substantial for the explainability and robustness of deep learning models and a step towards high-level cognition. The state-of-the-art methods are based on VAE and encourage the joint distribution of latent variables to be factorized. We take a different perspective on disentanglement by analyzing topological properties of data manifolds. In particular, we optimize the topological similarity for data manifolds traversals. To the best of our knowledge, our paper is the first one to propose a differentiable topological loss for disentanglement learning. Our experiments have shown that the proposed TopDis loss improves disentanglement scores such as MIG, FactorVAE score, SAP score and DCI disentanglement score with respect to state-of-the-art results while preserving the reconstruction quality. Our method works in an unsupervised manner, permitting to apply it for problems without labeled factors of variation. The TopDis loss works even when factors of variation are correlated. Additionally, we show how  to use the proposed topological loss to find disentangled directions in a trained GAN.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=23OEmHVkpq", "arxiv_id": "2308.12696", "paper_pdf": "papers/23OEmHVkpq.pdf", "paper_pdf_sha256": "ea7e940dcbdccafa20685b3c0b38b6817e970c5e959d8b3b4d2d7bea1b79dc83", "paper_pdf_bytes": 21582958, "paper_pdf_source": "openreview", "code_url": "https://github.com/nikitabalabin/TopDis", "code_repository": "nikitabalabin/TopDis", "code_commit": "2ca9d2e7979eae4e5efed378b791091dae27f90b", "code_archive": "repos/23OEmHVkpq.zip", "code_archive_sha256": "6c2b5bc6d9cad429ac839b2f60922868915769ebdfcc560382dc8a927ce12284", "code_archive_bytes": 97068, "code_file_count": 50, "code_extensions": {".py": 45, ".sh": 5}, "github_disk_usage_kb": 61, "github_languages": {"Python": 291123, "Shell": 2335}, "github_archived": false, "github_pushed_at": "2024-05-30T16:29:31Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/disentanglement-learning-via-topology"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "cS45VNtZLW", "year": 2023, "status": "rejected", "title": "Traversing Between Modes in Function Space for Fast Ensembling", "authors": ["Eunggu Yun", "Hyungi Lee", "Giung Nam", "Juho Lee"], "authorids": ["~Eunggu_Yun1", "~Hyungi_Lee1", "~Giung_Nam1", "~Juho_Lee2"], "authors_source": "OpenReview API", "abstract": "Deep ensemble is a simple yet powerful way to improve the performance of deep neural networks. Under this motivation, recent works on mode connectivity have shown that parameters of ensembles are connected by low-loss subspaces, and one can efficiently collect ensemble parameters in those subspaces. While this provides a way to efficiently train ensembles, for inference, one should still execute multiple forward passes using all the ensemble parameters, which often becomes a serious bottleneck for real-world deployment. In this work, we propose a novel framework to reduce such costs. Given a low-loss subspace connecting two modes of a neural network, we build an additional neural network predicting outputs of the original neural network evaluated at a certain point in the low-loss subspace. The additional neural network, what we call a “ bridge”, is a lightweight network taking minimal features from the original network, and predicting outputs for the low-loss subspace without forward passes through the original network. We empirically demonstrate that we can indeed train such bridge networks and significantly reduce inference costs with the help of the bridge networks.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=cS45VNtZLW", "arxiv_id": "2306.11304", "paper_pdf": "papers/cS45VNtZLW.pdf", "paper_pdf_sha256": "16581ec478df829ef557c1bc56e957a93f7835e6d839c74a895ac1d5ea9b8570", "paper_pdf_bytes": 555293, "paper_pdf_source": "openreview", "code_url": "https://github.com/yuneg11/Bridge-Network", "code_repository": "yuneg11/Bridge-Network", "code_commit": "cba2980a512b55f8a0e171e6d8b1413c4715a038", "code_archive": "repos/cS45VNtZLW.zip", "code_archive_sha256": "663bc6781d1f38ffe332038aecb6d8b8e8bfd3b59dac0611d859ee2adae51170", "code_archive_bytes": 186666, "code_file_count": 25, "code_extensions": {".py": 25}, "github_disk_usage_kb": 126, "github_languages": {"Python": 239446}, "github_archived": false, "github_pushed_at": "2023-09-04T01:57:44Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/traversing-between-modes-in-function-space"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "DIsWHvtU7lF", "year": 2022, "status": "rejected", "title": "Composing Partial Differential Equations with Physics-Aware Neural Networks", "authors": ["Matthias Karlbauer", "Timothy Praditia", "Sebastian Otte", "Sergey Oladyshkin", "Wolfgang Nowak", "Martin V. Butz"], "authorids": ["~Matthias_Karlbauer1", "~Timothy_Praditia1", "~Sebastian_Otte1", "~Sergey_Oladyshkin1", "~Wolfgang_Nowak1", "~Martin_V._Butz2"], "authors_source": "OpenReview API", "abstract": "We introduce a compositional physics-aware neural network (FINN) for learning spatiotemporal advection-diffusion processes. FINN implements a new way of combining the learning abilities of artificial neural networks with physical and structural knowledge from numerical simulation by modeling the constituents of partial differential equations (PDEs) in a compositional manner. Results on both one- and two-dimensional PDEs (Burger's, diffusion-sorption, diffusion-reaction) demonstrate FINN's superior process modeling accuracy and excellent out-of-distribution generalization ability beyond initial and boundary conditions. With only one tenth of the number of parameters on average, FINN outperforms pure machine learning and other state-of-the-art physics-aware models in all cases---often even by multiple orders of magnitude. Moreover, FINN outperforms a calibrated physical model when approximating sparse real-world data in a diffusion-sorption scenario, confirming its generalization abilities and showing explanatory potential by revealing the unknown retardation factor of the observed process.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=DIsWHvtU7lF", "arxiv_id": "2111.11798", "paper_pdf": "papers/DIsWHvtU7lF.pdf", "paper_pdf_sha256": "a7f1e0de8ca5a64de2980c6e91197acd1d9d247fdc6e0667042bbed878a6f3b0", "paper_pdf_bytes": 2876471, "paper_pdf_source": "openreview", "code_url": "https://github.com/CognitiveModeling/finn", "code_repository": "CognitiveModeling/finn", "code_commit": "82b4eac6b55be3b44e736f139ed50256c4415f0c", "code_archive": "repos/DIsWHvtU7lF.zip", "code_archive_sha256": "e0cdc90bba3bff323ecb79e876d3a2376a0ea798773688677381c1797fd02526", "code_archive_bytes": 215241, "code_file_count": 68, "code_extensions": {".py": 68}, "github_disk_usage_kb": 155, "github_languages": {"Python": 635460}, "github_archived": false, "github_pushed_at": "2022-10-28T08:45:07Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/composing-partial-differential-equations-with-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "cgRzg1V9su", "year": 2021, "status": "rejected", "title": "Inner Ensemble Networks: Average Ensemble as an Effective Regularizer", "authors": ["Abduallah Mohamed", "Muhammed Mohaimin Sadiq", "Ehab AlBadawy", "Mohamed Elhoseiny", "Christian Claudel"], "authorids": ["~Abduallah_Mohamed1", "m.mohaiminsadiq@utexas.edu", "ealbadawy@albany.edu", "~Mohamed_Elhoseiny1", "~Christian_Claudel1"], "authors_source": "OpenReview API", "abstract": "We introduce Inner Ensemble Networks (IENs) which reduce the variance within the neural network itself without an increase in the model complexity. IENs utilize ensemble parameters during the training phase to reduce the network variance. While in the testing phase, these parameters are removed without a change in the enhanced performance. IENs reduce the variance of an ordinary deep model by a factor of $1/m^{L-1}$, where $m$ is the number of inner ensembles and $L$ is the depth of the model. Also, we show empirically and theoretically that IENs lead to a greater variance reduction in comparison with other similar approaches such as dropout and maxout. Our results show a decrease of error rates between 1.7\\% and 17.3\\% in comparison with an ordinary deep model. We also show that IEN was preferred by Neural Architecture Search (NAS) methods over prior approaches.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=cgRzg1V9su", "arxiv_id": "2006.08305", "paper_pdf": "papers/cgRzg1V9su.pdf", "paper_pdf_sha256": "b33d0c266351bda9939e3e573cbe66e5f5ed3a5ea6b414f30bf4a7d7c9e015d2", "paper_pdf_bytes": 390622, "paper_pdf_source": "openreview", "code_url": "https://github.com/abduallahmohamed/inner_ensemble_nets", "code_repository": "abduallahmohamed/inner_ensemble_nets", "code_commit": "56b06e33c74e49ffe8ab89af64322a01d5ee92ad", "code_archive": "repos/cgRzg1V9su.zip", "code_archive_sha256": "4e227e9fe064f7ead67222ff037ab9af94572d47290ed3dd87c0f66e42f87a28", "code_archive_bytes": 73683, "code_file_count": 3, "code_extensions": {".py": 3}, "github_disk_usage_kb": 194, "github_languages": {"Python": 5953}, "github_archived": false, "github_pushed_at": "2020-10-09T06:01:38Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/inner-ensemble-nets"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "HJx-akSKPS", "year": 2020, "status": "rejected", "title": "Neural Subgraph Isomorphism Counting", "authors": ["Xin Liu", "Haojie Pan", "Mutian He", "Yangqiu Song", "Xin Jiang"], "authorids": ["xliucr@cse.ust.hk", "hpanad@cse.ust.hk", "mhear@cse.ust.hk", "yqsong@cse.ust.hk", "jiang.xin@huawei.com"], "authors_source": "OpenReview API", "abstract": "In this paper, we study a new graph learning problem: learning to count subgraph isomorphisms. Although the learning based approach is inexact, we are able to generalize to count large patterns and data graphs in polynomial time compared to the exponential time of the original NP-complete problem. Different from other traditional graph learning problems such as node classification and link prediction, subgraph isomorphism counting requires more global inference to oversee the whole graph. To tackle this problem, we propose a dynamic intermedium attention memory network (DIAMNet) which augments different representation learning architectures and iteratively attends pattern and target data graphs to memorize different subgraph isomorphisms for the global counting. We develop both small graphs (<= 1,024 subgraph isomorphisms in each) and large graphs (<= 4,096 subgraph isomorphisms in each) sets to evaluate different models. Experimental results show that learning based subgraph isomorphism counting can help reduce the time complexity with acceptable accuracy. Our DIAMNet can further improve existing representation learning models for this more global problem.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=HJx-akSKPS", "arxiv_id": "1912.11589", "paper_pdf": "papers/HJx-akSKPS.pdf", "paper_pdf_sha256": "d2addbe7c82fc089459e7d507bbfe69d99f78e4782849341d2fb67f83fb1cbc7", "paper_pdf_bytes": 2885761, "paper_pdf_source": "openreview", "code_url": "https://github.com/HKUST-KnowComp/NeuralSubgraphCounting", "code_repository": "HKUST-KnowComp/NeuralSubgraphCounting", "code_commit": "22df62d59e112716a22f80db408fcddfc95da5c8", "code_archive": "repos/HJx-akSKPS.zip", "code_archive_sha256": "dbafb1d45ff98e07f711c6b11635b2b6bafac00b43736e61d2a502d9f02a47a4", "code_archive_bytes": 211240, "code_file_count": 24, "code_extensions": {".py": 24}, "github_disk_usage_kb": 199, "github_languages": {"Python": 318050}, "github_archived": false, "github_pushed_at": "2024-07-25T10:59:43Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/neural-subgraph-isomorphism-counting-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "rkxn7nR5KX", "year": 2019, "status": "rejected", "title": "Incremental Few-Shot Learning with Attention Attractor Networks", "authors": ["Mengye Ren", "Renjie Liao", "Ethan Fetaya", "Richard S. Zemel"], "authorids": ["mren@cs.toronto.edu", "rjliao@cs.toronto.edu", "ethanf@cs.toronto.edu", "zemel@cs.toronto.edu"], "authors_source": "OpenReview API", "abstract": "Machine learning classifiers are often trained to recognize a set of pre-defined classes. However,\nin many real applications, it is often desirable to have the flexibility of learning additional\nconcepts, without re-training on the full training set. This paper addresses this problem,\nincremental few-shot learning, where a regular classification network has already been trained to\nrecognize a set of base classes; and several extra novel classes are being considered, each with\nonly a few labeled examples. After learning the novel classes, the model is then evaluated on the\noverall performance of both base and novel classes. To this end, we propose a meta-learning model,\nthe Attention Attractor Network, which regularizes the learning of novel classes. In each episode,\nwe train a set of new weights to recognize novel classes until they converge, and we show that the\ntechnique of recurrent back-propagation can back-propagate through the optimization process and\nfacilitate the learning of the attractor network regularizer. We demonstrate that the learned\nattractor network can recognize novel classes while remembering old classes without the need to\nreview the original training set, outperforming baselines that do not rely on an iterative\noptimization process.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=rkxn7nR5KX", "arxiv_id": "1810.07218", "paper_pdf": "papers/rkxn7nR5KX.pdf", "paper_pdf_sha256": "2557a3b43bffa15fb16c8f1a6bb470d9e846d7fa02a65e15254a0d12d0220dba", "paper_pdf_bytes": 316347, "paper_pdf_source": "openreview", "code_url": "https://github.com/renmengye/inc-few-shot-attractor-public", "code_repository": "renmengye/inc-few-shot-attractor-public", "code_commit": "c560d5a81480cb22d903fa746ab0cfc2eb964e4c", "code_archive": "repos/rkxn7nR5KX.zip", "code_archive_sha256": "9f7bf7f8ecdcbdb5e124444a9ca7f97858d430b700f15735ba0a8851066ee39c", "code_archive_bytes": 5738270, "code_file_count": 38, "code_extensions": {".py": 37, ".sh": 1}, "github_disk_usage_kb": 5572, "github_languages": {"Python": 316748, "Shell": 278, "Makefile": 136}, "github_archived": false, "github_pushed_at": "2020-03-19T18:52:03Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/incremental-few-shot-learning-with-attention"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "tDNr5rt47g", "year": 2026, "status": "rejected", "title": "Computational Reasoning of Large Language Models", "authors": ["Haitao Wu", "Zongbo Han", "Joey Tianyi Zhou", "Huaxi Huang", "Changqing Zhang"], "authorids": ["~Haitao_Wu1", "~Zongbo_Han1", "~Joey_Tianyi_Zhou1", "~Huaxi_Huang1", "~Changqing_Zhang1"], "authors_source": "OpenReview API", "abstract": "With the rapid development and widespread application of Large Language Models (LLMs), multidimensional evaluation has become increasingly critical. However, current evaluations are often domain-specific and overly complex, limiting their effectiveness as cross-domain proxies for core capabilities. To address these limitations and enable a unified and simple evaluation framework, an ideal proxy task should target a basic capability that generalizes across tasks and is independent of domain-specific knowledge. Turing machine provides a powerful theoretical lens by reducing complex processes to basic, domain-agnostic computational operations. This perspective offers a principled framework for evaluating foundational computational abilities essential to a wide range of tasks, particularly those involving complex, multi-step reasoning such as mathematics.\nMotivated by this abstraction, we introduce \\textbf{Turing Machine Bench}, a benchmark designed to assess the ability of LLMs to \\textbf{strictly follow rules} and \\textbf{accurately manage internal states} for multi-step, referred to as \\textbf{computational reasoning}.\nTMBench incorporates four key features: self-contained and knowledge-agnostic reasoning, a minimalistic multi-step structure, controllable difficulty, and a solid theoretical foundation based on Turing machine.\nEmpirical results demonstrate that TMBench serves as an effective proxy for evaluating computational reasoning on representative LLMs. It produces clear step-wise accuracy curves, revealing LLMs' ability to execute multi-step reasoning processes. \nBy analyzing performance trends across TMBench and established reasoning benchmarks, we find strong correlations with real-world tasks, bridging real-task evaluation with basic ability assessment. \nThese findings suggest that TMBench holds potential as a cross-domain dimension for evaluating reasoning in LLMs.\nCode and data are available at \\href{https://anonymous.4open.science/r/Turing-Machine-Bench-Anonymous-EBF4/}{Repo}.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=tDNr5rt47g", "arxiv_id": "2504.20771", "paper_pdf": "papers/tDNr5rt47g.pdf", "paper_pdf_sha256": "27d4feece55de04c66fa97dba7f59380e9b72a28c6865e53e746d0523f5c14b1", "paper_pdf_bytes": 488760, "paper_pdf_source": "openreview", "code_url": "https://github.com/HaitaoWuTJU/Turing-Machine-Bench", "code_repository": "HaitaoWuTJU/Turing-Machine-Bench", "code_commit": "9e04fcaec05664aa28fc86f20d9a2bb0e17551ea", "code_archive": "repos/tDNr5rt47g.zip", "code_archive_sha256": "bf46074567abebd4586abad67ee3badd9b914c55488bfa5be3e751b6ca6c5695", "code_archive_bytes": 1212682, "code_file_count": 10, "code_extensions": {".py": 9, ".sh": 1}, "github_disk_usage_kb": 1174, "github_languages": {"Python": 30570, "Shell": 1705}, "github_archived": false, "github_pushed_at": "2025-05-16T14:14:10Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/turing-machine-evaluation-for-large-language"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "gDWkImLIKd", "year": 2025, "status": "rejected", "title": "Large Language Model Critics for Execution-Free Evaluation of Code Changes", "authors": ["Aashish Yadavally", "Hoan Nguyen", "Laurent Callot", "Gauthier Guinet"], "authorids": ["~Aashish_Yadavally1", "~Hoan_Nguyen1", "~Laurent_Callot1", "~Gauthier_Guinet1"], "authors_source": "OpenReview API", "abstract": "Large language models (LLMs) offer a promising way forward for automating software engineering tasks, such as bug fixes, feature additions, etc., via multi-step LLM-based agentic workflows. However, existing metrics for evaluating such workflows, mainly  build status and occasionally log analysis, are too sparse and limited in providing the information needed to assess the quality of changes made. In this work, we designed LLM-based critics to derive well-structured and rigorous intermediate/step-level, execution-free evaluation proxies for repo-level code changes. Importantly, we assume access to the gold patch for the problem (i.e., reference-aware) to assess both semantics and executability of generated patches. With the gold test patch as a reference, we predict executability of all editing locations with an accuracy of 91.6%, aggregating which, we can predict the build status in 82.1% of the instances in SWE-bench. In particular, such an execution-focused LLM critic outperforms other reference-free and reference-aware LLM critics by 38.9% to 72.5%. Moreover, we demonstrate the usefulness of such a reference-aware framework in comparing patches generated by different agentic workflows. Finally, we open-source the library developed for this project, which allow further usage for either other agentic workflows or other benchmarks.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=gDWkImLIKd", "arxiv_id": "2501.16655", "paper_pdf": "papers/gDWkImLIKd.pdf", "paper_pdf_sha256": "e1a32aa93d8e717baae677abb34fd45407acdfaa22f9a28020d21cd7c093dc90", "paper_pdf_bytes": 1207802, "paper_pdf_source": "openreview", "code_url": "https://github.com/amazon-science/code-agent-eval", "code_repository": "amazon-science/code-agent-eval", "code_commit": "947415316064212917033e6d06df20f2b1aaf743", "code_archive": "repos/gDWkImLIKd.zip", "code_archive_sha256": "b8dacdb2ab672a983332c7bd706dfc05857b41d4f63ea3bea58a29f9da199790", "code_archive_bytes": 94586, "code_file_count": 36, "code_extensions": {".py": 32, ".sh": 4}, "github_disk_usage_kb": 85, "github_languages": {"Python": 230439, "Shell": 5729}, "github_archived": false, "github_pushed_at": "2025-01-29T02:22:54Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/large-language-model-critics-for-execution"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "hgrZluxFC7", "year": 2024, "status": "rejected", "title": "Adversarial Machine Learning in Latent Representations of Neural Networks", "authors": ["Milin Zhang", "Mohammad Abdi", "Francesco Restuccia"], "authorids": ["~Milin_Zhang1", "~Mohammad_Abdi1", "~Francesco_Restuccia2"], "authors_source": "OpenReview API", "abstract": "Distributed deep neural networks (DNNs) have been shown to reduce the computational burden of mobile devices and decrease the end-to-end inference latency in edge computing scenarios. While distributed DNNs have been studied, to the best of our knowledge the resilience of distributed DNNs to adversarial action still remains an open problem. In this paper, we fill the existing research gap by rigorously analyzing the robustness of distributed DNNs against adversarial action. We cast this problem in the context of information theory and introduce two new measurements for distortion and robustness. Our theoretical findings indicate that (i) assuming the same level of information distortion, latent features are always more robust than input representations; (ii) the adversarial robustness is jointly determined by the feature dimension and the generalization capability of the DNN. To test our theoretical findings, we perform extensive experimental analysis by considering 6 different DNN architectures, 6 different approaches for distributed DNN and 10 different adversarial attacks to the ImageNet-1K dataset. Our experimental results support our theoretical findings by showing that the compressed latent representations can reduce the success rate of adversarial attacks by 88% in the best case and by 57% on the average compared to attacks to the input space.", "decision": "Reject", "meta_review": "", "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=hgrZluxFC7", "arxiv_id": "2309.17401", "paper_pdf": "papers/hgrZluxFC7.pdf", "paper_pdf_sha256": "56afaa049267e6e038178780a4adbb91309ebc28b8c48bd25ec4eb981e82b68c", "paper_pdf_bytes": 1759573, "paper_pdf_source": "openreview", "code_url": "https://github.com/asdfqwezxcf/AdvLatent", "code_repository": "asdfqwezxcf/AdvLatent", "code_commit": "f34e9a81fc4beebdc17f70f085140082fb3e3cfd", "code_archive": "repos/hgrZluxFC7.zip", "code_archive_sha256": "75a8992e535a23de27267bfa4df630d9202fd110194f5656a83dad598c668e0c", "code_archive_bytes": 53433, "code_file_count": 17, "code_extensions": {".py": 17}, "github_disk_usage_kb": 62, "github_languages": {"Python": 217090}, "github_archived": false, "github_pushed_at": "2024-01-26T18:19:04Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/adversarial-machine-learning-in-latent"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "UyC1dXUA-n", "year": 2023, "status": "rejected", "title": "Design of the topology for contrastive visual-textual alignment", "authors": ["Zhun Sun"], "authorids": ["~Zhun_Sun1"], "authors_source": "OpenReview API", "abstract": "Pre-training weakly related image-text pairs in the contrastive style shows great power in learning semantic aligning cross-modal models. The common choice to measure the distance between the feature representations of the image-text pairs is the cosine similarity, which can be considered as the negative inner product distance of features embedded on a sphere, mathematically. However, empirically, aligning image-text pairs on the spherical topology is vulnerable to the semantic ambiguity phenomenon resulting from the noise in the pre-training datasets. Specifically, under the noisy training data, instead of the optimal alignment-uniformity solution, the system would achieve an equilibrium (a gap between distances of positive and negative pairs), when the gradients for attraction and repulsion are neutralized.  Although intuitively, the model should always find this equilibrium given a sufficiently long training scheme, its numerical values might be out of the distance range (e.g. [-1, 1] for the cosine similarity). In the practice of former studies, this problem is partly tackled by introducing a learnable softmax temperature parameter, in other words, by explicitly scaling the range of the distance function.  In this work, we alternatively design the topology of embedding space and its endowed distance function. Motivated by studies that make use of Riemannian geometry for visual tasks, we propose a rather simple solution to address the aforementioned equilibrium problem. That is, we map the feature representations onto the oblique manifold endowed with the negative inner product as the distance function. In the experimental analysis, we show that we can improve the baseline performance by a large margin (e.g. 4\\% in the zero-shot image to text retrieval task) by changing only two lines of the training codes.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=UyC1dXUA-n", "arxiv_id": "2209.02127", "paper_pdf": "papers/UyC1dXUA-n.pdf", "paper_pdf_sha256": "2a4b6dac049ff385c47a041b62279cc45fe2ed2df44ecda6d322ddc28b29384a", "paper_pdf_bytes": 904016, "paper_pdf_source": "openreview", "code_url": "https://github.com/minogame/clip-mtob", "code_repository": "minogame/clip-mtob", "code_commit": "15fa1ec4fac43e1dd15a5040aaba60ca2e513caa", "code_archive": "repos/UyC1dXUA-n.zip", "code_archive_sha256": "003bd6baed66626366e297c2f0b05d5eac66854b00ec5be90eb5e9423689865e", "code_archive_bytes": 137574, "code_file_count": 4, "code_extensions": {".py": 4}, "github_disk_usage_kb": 131, "github_languages": {"Python": 41882}, "github_archived": false, "github_pushed_at": "2023-10-09T04:52:27Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/design-of-the-topology-for-contrastive-visual"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "QkfMWTl520U", "year": 2022, "status": "rejected", "title": "When do Convolutional Neural Networks Stop Learning?", "authors": ["SAHAN AHMAD", "Aminul Islam"], "authorids": ["~SAHAN_AHMAD1", "aminul@louisiana.edu"], "authors_source": "OpenReview API", "abstract": "Convolutional Neural Networks (CNNs) is one of the most essential architectures that has shown impressive performance in computer vision tasks such as image classification, detection, and segmentation. In training phase of CNN, an arbitrary number of epochs is used to train the neural networks. In a single epoch, the entire training data---divided by batch size---are fed to the network. However, the optimal number of epochs required to train a neural network is not well established. In practice, validation data is used to identify the generalization gap. To avoid overfitting, it is recommended to stop training when the generalization gap increases. However, this is a trial and error based approach. This raises a critical question: Is it possible to estimate when neural networks stop learning based on only the training data? In this research work, we introduce the stability property of data in layers and based on this property, we predict the near optimal epoch number of a CNN. We do not use any validation data to predict the near optimal epoch number. We experiment our hypothesis on six different CNN models and on three different datasets (CIFIR 10, CIFIR 100, SVHN). We save on average 58.49\\% computational time to train a CNN model. Our code is available at https://github.com/PaperUnderReviewDeepLearning/Optimization.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=QkfMWTl520U", "arxiv_id": "2403.02473", "paper_pdf": "papers/QkfMWTl520U.pdf", "paper_pdf_sha256": "df8543ffc73d6cdfe712e9ffe9bc108d01595f401f986a0bd399d96265f6ecda", "paper_pdf_bytes": 4504983, "paper_pdf_source": "openreview", "code_url": "https://github.com/PaperUnderReviewDeepLearning/Optimization", "code_repository": "PaperUnderReviewDeepLearning/Optimization", "code_commit": "a1029d56c090003289270c382ea81c82166ae22a", "code_archive": "repos/QkfMWTl520U.zip", "code_archive_sha256": "44b597c662f50b3001b1df408a5296496866856f016bba1760ad5b5900afa316", "code_archive_bytes": 155703, "code_file_count": 14, "code_extensions": {".py": 14}, "github_disk_usage_kb": 178, "github_languages": {"Python": 83145}, "github_archived": false, "github_pushed_at": "2022-10-04T15:30:47Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/when-do-convolutional-neural-networks-stop-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "xyGFYKIPTDJ", "year": 2021, "status": "rejected", "title": "Learning Causal Semantic Representation for Out-of-Distribution Prediction", "authors": ["Chang Liu", "Xinwei Sun", "Jindong Wang", "Tao Li", "Tao Qin", "Wei Chen", "Tie-Yan Liu"], "authorids": ["~Chang_Liu10", "~Xinwei_Sun1", "~Jindong_Wang1", "~Tao_Li9", "~Tao_Qin1", "~Wei_Chen1", "~Tie-Yan_Liu1"], "authors_source": "OpenReview API", "abstract": "Conventional supervised learning methods, especially deep ones, are found to be sensitive to out-of-distribution (OOD) examples, largely because the learned representation mixes the semantic factor with the variation factor due to their domain-specific correlation, while only the semantic factor causes the output. To address the problem, we propose a Causal Semantic Generative model (CSG) based on causality to model the two factors separately, and learn it on a single training domain for prediction without (OOD generalization) or with unsupervised data (domain adaptation) in a test domain. We prove that CSG identifies the semantic factor on the training domain, and the invariance principle of causality subsequently guarantees the boundedness of OOD generalization error and the success of adaptation. We also design novel and delicate learning methods for both effective learning and easy prediction, following the first principle of variational Bayes and the graphical structure of CSG. Empirical study demonstrates the effect of our methods to improve test accuracy for OOD generalization and domain adaptation.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=xyGFYKIPTDJ", "arxiv_id": "2011.01681", "paper_pdf": "papers/xyGFYKIPTDJ.pdf", "paper_pdf_sha256": "e48d9262f64427ff7c7a31b4dd9aae93b6deae69abbe3b1bde980595343245b3", "paper_pdf_bytes": 508925, "paper_pdf_source": "openreview", "code_url": "https://github.com/changliu00/causal-semantic-generative-model", "code_repository": "changliu00/causal-semantic-generative-model", "code_commit": "05c10d6790f3db8d4847efd98d18a6ecafb469cb", "code_archive": "repos/xyGFYKIPTDJ.zip", "code_archive_sha256": "580cde18a3378f13598d3820d20aebb7b52f2edea53350f48280b43f75f21874", "code_archive_bytes": 205268, "code_file_count": 39, "code_extensions": {".py": 28, ".sh": 10, ".ipynb": 1}, "github_disk_usage_kb": 198, "github_languages": {"Python": 212959, "Shell": 20528, "Jupyter Notebook": 10218}, "github_archived": false, "github_pushed_at": "2022-04-18T15:12:50Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/learning-causal-semantic-representation-for-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "B1em9h4KDS", "year": 2020, "status": "rejected", "title": "Generative Imputation and Stochastic Prediction", "authors": ["Mohammad Kachuee", "Kimmo Kärkkäinen", "Orpaz Goldstein", "Sajad Darabi", "Majid Sarrafzadeh"], "authorids": ["mkachuee@ucla.edu", "kimmo@cs.ucla.edu", "orpgol@cs.ucla.edu", "sajad.darabi@cs.ucla.edu", "majid@cs.ucla.edu"], "authors_source": "OpenReview API", "abstract": "In many machine learning applications, we are faced with incomplete datasets. In the literature, missing data imputation techniques have been mostly concerned with filling missing values. However, the existence of missing values is synonymous with uncertainties not only over the distribution of missing values but also over target class assignments that require careful consideration. In this paper, we propose a simple and effective method for imputing missing features and estimating the distribution of target assignments given incomplete data. In order to make imputations, we train a simple and effective generator network to generate imputations that a discriminator network is tasked to distinguish. Following this, a predictor network is trained using the imputed samples from the generator network to capture the classification uncertainties and make predictions accordingly. The proposed method is evaluated on CIFAR-10 image dataset as well as three real-world tabular classification datasets, under different missingness rates and structures. Our experimental results show the effectiveness of the proposed method in generating imputations as well as providing estimates for the class uncertainties in a classification task when faced with missing values.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=B1em9h4KDS", "arxiv_id": "1905.09340", "paper_pdf": "papers/B1em9h4KDS.pdf", "paper_pdf_sha256": "1f7c067789a0a37109bcc21d360087f10ddf331f214527b045ac3fc44fe7e17a", "paper_pdf_bytes": 4615662, "paper_pdf_source": "openreview", "code_url": "https://github.com/mkachuee/GenerativeImputationStochasticPrediction", "code_repository": "mkachuee/GenerativeImputationStochasticPrediction", "code_commit": "f62e8d30e8d2d42088255da0d754ed6b7d75dfdf", "code_archive": "repos/B1em9h4KDS.zip", "code_archive_sha256": "004803c161cf288f4a94784ee3d75c811a889db57a700919d761684e2ef7dcaa", "code_archive_bytes": 227195, "code_file_count": 31, "code_extensions": {".py": 30, ".sh": 1}, "github_disk_usage_kb": 200, "github_languages": {"Python": 213002, "Shell": 741}, "github_archived": false, "github_pushed_at": "2020-09-04T04:26:27Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/generative-imputation-and-stochastic"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "rJl-HsR9KX", "year": 2019, "status": "rejected", "title": "Discriminative Active Learning", "authors": ["Daniel Gissin", "Shai Shalev-Shwartz"], "authorids": ["daniel.gissin@mail.huji.ac.il", "shais@cs.huji.ac.il"], "authors_source": "OpenReview API", "abstract": "We propose a new batch mode active learning algorithm designed for neural networks and large query batch sizes. The method, Discriminative Active Learning (DAL), poses active learning as a binary classification task, attempting to choose examples to label in such a way as to make the labeled set and the unlabeled pool indistinguishable. Experimenting on image classification tasks, we empirically show our method to be on par with state of the art methods in medium and large query batch sizes, while being simple to implement and also extend to other domains besides classification tasks. Our experiments also show that none of the state of the art methods of today are clearly better than uncertainty sampling, negating some of the reported results in the recent literature.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=rJl-HsR9KX", "arxiv_id": "1907.06347", "paper_pdf": "papers/rJl-HsR9KX.pdf", "paper_pdf_sha256": "dfd6410a588accfe02faa0673439712494bf64f4a4d5f3f2ad534ea5afcc1a1b", "paper_pdf_bytes": 1414381, "paper_pdf_source": "openreview", "code_url": "https://github.com/dsgissin/DiscriminativeActiveLearning", "code_repository": "dsgissin/DiscriminativeActiveLearning", "code_commit": "625e9a183b9c14356c81b9073456d8142f25f906", "code_archive": "repos/rJl-HsR9KX.zip", "code_archive_sha256": "60bab3afbed30a4934bc3e5c597cf2a90a40cde369403ecfd171ba959d3a5294", "code_archive_bytes": 15207, "code_file_count": 3, "code_extensions": {".py": 3}, "github_disk_usage_kb": 6859, "github_languages": {"Python": 67802}, "github_archived": false, "github_pushed_at": "2019-09-26T11:48:11Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/discriminative-active-learning-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "uhVlvT3Pk1", "year": 2026, "status": "rejected", "title": "Probing Human Visual Robustness with Neurally-Guided Deep Neural Networks", "authors": ["Zhenan Shao", "Linjian Ma", "Yiqing Zhou", "Yibo Jacky Zhang", "Sanmi Koyejo", "Bo Li", "Diane Beck"], "authorids": ["~Zhenan_Shao1", "~Linjian_Ma1", "~Yiqing_Zhou1", "~Yibo_Jacky_Zhang1", "~Sanmi_Koyejo1", "~Bo_Li19", "~Diane_Beck1"], "authors_source": "OpenReview API", "abstract": "Humans effortlessly navigate the visual world, yet deep neural networks (DNNs), despite excelling at many visual tasks, are surprisingly vulnerable to minor image perturbations. Past theories suggest human visual robustness arises from a representational space that evolves along the ventral visual stream (VVS) of the brain to increasingly tolerate object transformations. To test whether robustness is supported by such progression as opposed to being confined to specialized higher-order regions, we trained DNNs to align their representations with human neural responses from consecutive VVS regions during visual tasks. We demonstrate a hierarchical improvement in DNN robustness: alignment to higher-order VVS regions yields greater gains. To investigate the mechanism behind this improvement, we test a prominent hypothesis that attributes human visual robustness to the unique geometry of neural category manifolds in the VVS. We show that desirable manifold properties, specifically, smaller extent and better linear separability, emerge across the human VVS. These properties are inherited by DNNs via neural guidance and can predict their subsequent robustness gains. Further, we show that supervision from neural manifolds alone, via manifold guidance, suffices to qualitatively reproduce the hierarchical robustness improvements. Together, our results highlight the evolving VVS representational space as critical for robust visual inference, with the more linearly separable category manifolds as one potential mechanism, offering insights for building more resilient AI systems.", "decision": "Reject", "meta_review": "", "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=uhVlvT3Pk1", "arxiv_id": "2405.02564", "paper_pdf": "papers/uhVlvT3Pk1.pdf", "paper_pdf_sha256": "31b1720b3facb27d4a5369b0b58d35617ccbb4a5edddce288d55b44e5b90ab54", "paper_pdf_bytes": 7409759, "paper_pdf_source": "openreview", "code_url": "https://github.com/shaox192/NeuralGuidance", "code_repository": "shaox192/NeuralGuidance", "code_commit": "6536eafe51d049b3d5d5f35659c2d9c51f10ec8d", "code_archive": "repos/uhVlvT3Pk1.zip", "code_archive_sha256": "aceaa4d0d83ddab4e5761e7813639f0906f27b92d23321dc0b2e18d4cbd391d8", "code_archive_bytes": 1053556, "code_file_count": 29, "code_extensions": {".py": 26, ".sh": 2, ".ipynb": 1}, "github_disk_usage_kb": 1240, "github_languages": {"Jupyter Notebook": 480570, "Python": 161971, "Shell": 2677}, "github_archived": false, "github_pushed_at": "2026-08-18T22:23:28Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/leveraging-the-human-ventral-visual-stream-to"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "63r2sTjkCv", "year": 2025, "status": "rejected", "title": "KinDEL: DNA-Encoded Library Dataset for Kinase Inhibitors", "authors": ["Benson Chen", "Tomasz Danel", "Patrick J. McEnaney", "Nikhil Jain", "Kirill Novikov", "Spurti Umesh Akki", "Joshua L. Turnbull", "Virja Atul Pandya", "Boris P. Belotserkovskii", "Jared Bryce Weaver", "Ankita Biswas", "Dat Nguyen", "Gabriel H. S. Dreiman", "Mohammad Sultan", "Nathaniel Stanley", "Daniel M Whalen", "Divya Kanichar", "Christoph Klein", "Emily Fox", "R. Edward Watts"], "authorids": ["~Benson_Chen1", "~Tomasz_Danel1", "~Patrick_J._McEnaney1", "~Nikhil_Jain1", "~Kirill_Novikov1", "~Spurti_Umesh_Akki1", "~Joshua_L._Turnbull1", "~Virja_Atul_Pandya1", "~Boris_P._Belotserkovskii1", "~Jared_Bryce_Weaver1", "~Ankita_Biswas1", "~Dat_Nguyen4", "~Gabriel_H._S._Dreiman1", "~Mohammad_Sultan1", "~Nathaniel_Stanley1", "~Daniel_M_Whalen1", "~Divya_Kanichar1", "~Christoph_Klein1", "~Emily_Fox2", "~R._Edward_Watts1"], "authors_source": "OpenReview API", "abstract": "DNA-Encoded Libraries (DEL) are combinatorial small molecule libraries that offer an efficient way to characterize diverse chemical spaces. Selection experiments using DELs are pivotal to drug discovery efforts, enabling high-throughput hit finding screens. However, limited availability of public DEL datasets hinders the advancement of computational techniques designed to utilize such data. To bridge this gap, we present KinDEL, one of the first large, publicly available DEL datasets on two kinases: Mitogen-Activated Protein Kinase 14 (MAPK14) and Discoidin Domain Receptor Tyrosine Kinase 1 (DDR1). Interest in this data modality is growing due to its ability to generate extensive supervised chemical data that densely samples around select molecular structures. Demonstrating one such application of the data, we benchmark different machine learning techniques to develop predictive models for hit identification; in particular, we highlight recent structure-based probabilistic approaches. Finally, we provide biophysical assay data, both on- and off-DNA, to validate our models on a smaller subset of molecules. Data and code for our benchmarks can be found at: https://kin-del-2024.s3.us-west-2.amazonaws.com/kindel.zip", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=63r2sTjkCv", "arxiv_id": "2410.08938", "paper_pdf": "papers/63r2sTjkCv.pdf", "paper_pdf_sha256": "784a61f90a2283fc60b66738522c314bee3d834ee2f8d2d023da7e0930f1df08", "paper_pdf_bytes": 10432627, "paper_pdf_source": "openreview", "code_url": "https://github.com/insitro/kindel", "code_repository": "insitro/kindel", "code_commit": "93bcc6f64735160a486a0ea3e69c2f95c71dd2c4", "code_archive": "repos/63r2sTjkCv.zip", "code_archive_sha256": "c67351d69995d74474f718c6d28d37cf68e812feed2800601a6faaeb62fb4b83", "code_archive_bytes": 84208, "code_file_count": 18, "code_extensions": {".py": 18}, "github_disk_usage_kb": 86, "github_languages": {"Python": 73422}, "github_archived": false, "github_pushed_at": "2025-09-02T15:52:15Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/kindel-dna-encoded-library-dataset-for-kinase"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Zr96FfaUGR", "year": 2024, "status": "rejected", "title": "ARIES: A Corpus of Scientific Paper Edits Made in Response to Peer Reviews", "authors": ["Mike D'Arcy", "Alexis Ross", "Erin Bransom", "Bailey Kuehl", "Jonathan Bragg", "Tom Hope", "Doug Downey"], "authorids": ["~Mike_D'Arcy1", "~Alexis_Ross1", "~Erin_Bransom1", "~Bailey_Kuehl1", "~Jonathan_Bragg1", "~Tom_Hope2", "~Doug_Downey1"], "authors_source": "OpenReview API", "abstract": "Revising scientific papers based on peer feedback is a challenging task that requires not only deep scientific knowledge and reasoning, but also the ability to recognize the implicit requests in high-level feedback and to choose the best of many possible ways to update the manuscript in response.  We introduce this task for large language models and release ARIES, a dataset of review comments and their corresponding paper edits, to enable training and evaluating models.  We study two versions of the task: comment-edit alignment and edit generation, and evaluate several baselines, including GPT-4.  We find that models struggle even to identify the edits that correspond to a comment, especially in cases where the comment is phrased in an indirect way or where the edit addresses the spirit of a comment but not the precise request.  When tasked with generating edits, GPT-4 often succeeds in addressing comments on a surface level, but it rigidly follows the wording of the feedback rather than the underlying intent, and includes fewer technical details than human-written edits.  We hope that our formalization, dataset, and analysis will form a foundation for future work in this area.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=Zr96FfaUGR", "arxiv_id": "2306.12587", "paper_pdf": "papers/Zr96FfaUGR.pdf", "paper_pdf_sha256": "f9949decc1aa5348284c77b11fed9ac2eb0a560a711f8ec4cac48e1d6eb359a7", "paper_pdf_bytes": 318531, "paper_pdf_source": "openreview", "code_url": "https://github.com/allenai/aries", "code_repository": "allenai/aries", "code_commit": "5691bae71a101225ed345d0ffc42e47609f03bbb", "code_archive": "repos/Zr96FfaUGR.zip", "code_archive_sha256": "2ebf1fa1ed59315916dc5a937262a322e4f7eb3711c7e1495fd2e7ee07b5ec85", "code_archive_bytes": 80400, "code_file_count": 22, "code_extensions": {".py": 22}, "github_disk_usage_kb": 65, "github_languages": {"Python": 174978, "Dockerfile": 700}, "github_archived": false, "github_pushed_at": "2023-07-05T18:05:09Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/aries-a-corpus-of-scientific-paper-edits-made"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "i1Z_VysEgu8", "year": 2023, "status": "rejected", "title": "Time-Myopic Go-Explore: Learning A State Representation for the Go-Explore Paradigm", "authors": ["Marc Höftmann", "Jan Robine", "Stefan Harmeling"], "authorids": ["~Marc_Höftmann1", "~Jan_Robine1", "~Stefan_Harmeling1"], "authors_source": "OpenReview API", "abstract": "Very large state spaces with a sparse reward signal are difficult to explore. The lack of a sophisticated guidance results in a poor performance for numerous reinforcement learning algorithms. In these cases, the commonly used random exploration is often not helpful. The literature shows that this kind of environments require enormous efforts to systematically explore large chunks of the state space. Learned state representations can help here to improve the search by providing semantic context and build a structure on top of the raw observations. In this work we introduce a novel time-myopic state representation that clusters temporal close states together while providing a time prediction capability between them. By adapting this model to the Go-Explore paradigm (Ecoffet et al., 2021b), we demonstrate the first learned state representation that reliably estimates novelty instead of using the hand-crafted representation heuristic. Our method shows an improved solution for the detachment problem which still remains an issue at the Go-Explore Exploration Phase. We provide evidence that our proposed method covers the entire state space with respect to all possible time trajectories — without causing disadvantageous conflict-overlaps in the cell archive. Analogous to native Go-Explore, our approach is evaluated on the hard exploration environments MontezumaRevenge, Gravitar and Frostbite (Atari) in order to validate its capabilities on difficult tasks. Our experiments show that time-myopic Go-Explore is an effective alternative for the domain-engineered heuristic while also being more general. The source code of the method is available on GitHub.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=i1Z_VysEgu8", "arxiv_id": "2301.05635", "paper_pdf": "papers/i1Z_VysEgu8.pdf", "paper_pdf_sha256": "7a738b15f7bed7bc9451b720c715e19aa9f774c286b74de5c518d57b3c78624d", "paper_pdf_bytes": 1191256, "paper_pdf_source": "openreview", "code_url": "https://github.com/Hauf3n/Time-Myopic-Go-Explore", "code_repository": "Hauf3n/Time-Myopic-Go-Explore", "code_commit": "4f116d4af7ce48f7b544b6b812428cc9473adcb1", "code_archive": "repos/i1Z_VysEgu8.zip", "code_archive_sha256": "4d0994fa4d419b7682181deae5bb8c37778a6846743133d2e900de35c7d1f27c", "code_archive_bytes": 135719, "code_file_count": 6, "code_extensions": {".py": 6}, "github_disk_usage_kb": 134, "github_languages": {"Python": 43198}, "github_archived": false, "github_pushed_at": "2024-04-16T12:33:55Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/time-myopic-go-explore-learning-a-state"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "hxitw01k_Ql", "year": 2022, "status": "rejected", "title": "How memory architecture affects learning in a simple POMDP: the two-hypothesis testing problem", "authors": ["Mario Geiger", "Christophe Eloy", "Matthieu Wyart"], "authorids": ["~Mario_Geiger1", "~Christophe_Eloy1", "~Matthieu_Wyart2"], "authors_source": "OpenReview API", "abstract": "Reinforcement learning is generally difficult for partially observable Markov decision processes (POMDPs), which occurs when the agent's observation is partial or noisy. To seek good performance in POMDPs, one strategy is to endow the agent with a finite memory, whose update is governed by the policy. However, policy optimization is non-convex in that case and can lead to poor training performance for random initialization. The performance can be empirically improved by constraining the memory architecture, then sacrificing optimality to facilitate training. Here we study this trade-off in a two-hypothesis testing problem, akin to the two-arm bandit problem. We compare two extreme cases: (i) the random access memory where any transitions between $M$ memory states are allowed and (ii) a fixed memory where the agent can access its last $m$ actions and rewards. For (i), the probability $q$ to play the worst arm is known to be exponentially small in $M$ for the optimal policy. Our main result is to show that similar performance can be reached for (ii) as well, despite the simplicity of the memory architecture: using a conjecture on Gray-ordered binary necklaces, we find policies for which $q$ is exponentially small in $2^m$, i.e. $q\\sim\\alpha^{2^m}$ with $\\alpha < 1$. In addition, we observe empirically that training from random initialization leads to very poor results for (i), and significantly better results for (ii) thanks to the constraints on the memory architecture.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=hxitw01k_Ql", "arxiv_id": "2106.08849", "paper_pdf": "papers/hxitw01k_Ql.pdf", "paper_pdf_sha256": "7492df3a27ba397b12970c33e3d7e6043420500b38a93cf660b6b3874bdc5623", "paper_pdf_bytes": 550874, "paper_pdf_source": "openreview", "code_url": "https://github.com/pcsl-epfl/two-hypothesis", "code_repository": "pcsl-epfl/two-hypothesis", "code_commit": "4199d45977b8207f4e34f84fa6832df6534aa284", "code_archive": "repos/hxitw01k_Ql.zip", "code_archive_sha256": "c4e259865a3db0673fa648cf774362dc1d84c6d5d51e2f277cb1ae04d5cfc01a", "code_archive_bytes": 36752, "code_file_count": 4, "code_extensions": {".py": 3, ".nb": 1}, "github_disk_usage_kb": 179, "github_languages": {"Mathematica": 167858, "Python": 34821}, "github_archived": false, "github_pushed_at": "2021-10-05T10:43:53Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/how-memory-architecture-affects-performance"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "l3YcqzaPlx0", "year": 2021, "status": "rejected", "title": "Neighbor2Seq: Deep Learning on Massive Graphs by Transforming Neighbors to Sequences", "authors": ["Meng Liu", "Shuiwang Ji"], "authorids": ["~Meng_Liu3", "~Shuiwang_Ji1"], "authors_source": "OpenReview API", "abstract": "Modern Graph Neural Networks (GNNs) follow a recursive neighbor-wise message passing scheme and have achieved great success in many fields. However, this recursive design brings expensive computation and huge memory usage, making it difficult to deploy on large-scale graphs. In this work, we propose Neighbor2Seq, which transforms the hierarchical neighborhood of each node into an ordered sequence and enables the subsequent utilization of general deep learning operations, such as convolution and attention. Neighbor2Seq grants our proposed models, i.e., Neighbor2Seq-Conv and Neighbor2Seq-Attn, the ability to learn on arbitrarily large graphs as long as the Neighbor2Seq step can be precomputed. Another potential advantage obtained by the way is that Neighbor2Seq can alleviate the over-squashing issue existing in modern GNNs. We conduct thorough experiments on a massive graph with more than 111 million nodes and 1.6 billion edges, as well as several medium-scale graphs, to evaluate our proposed method. Experimental results demonstrate that our proposed method is scalable to the massive graph and achieves superior performance across datasets.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=l3YcqzaPlx0", "arxiv_id": "2202.03341", "paper_pdf": "papers/l3YcqzaPlx0.pdf", "paper_pdf_sha256": "114e531c2632b2810e988fe59a24bf2c6adb977df5f362979ed2f92fc787cdfe", "paper_pdf_bytes": 306965, "paper_pdf_source": "openreview", "code_url": "https://github.com/divelab/Neighbor2Seq", "code_repository": "divelab/Neighbor2Seq", "code_commit": "327d73d872497eb3c7f70ee27a0ea696db30cb22", "code_archive": "repos/l3YcqzaPlx0.zip", "code_archive_sha256": "8af829b2c85caa513b307a8a0c69e5d0d106b8205ded88b04a0be19bbe367f81", "code_archive_bytes": 201897, "code_file_count": 8, "code_extensions": {".py": 7, ".sh": 1}, "github_disk_usage_kb": 206, "github_languages": {"Python": 36934, "Shell": 3874}, "github_archived": false, "github_pushed_at": "2022-02-08T03:20:32Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/neighbor2seq-deep-learning-on-massive-graphs-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "HklRKpEKDr", "year": 2020, "status": "rejected", "title": "Deep Coordination Graphs", "authors": ["Wendelin Boehmer", "Vitaly Kurin", "Shimon Whiteson"], "authorids": ["wendelin.boehmer@cs.ox.ac.uk", "vitaly.kurin@cs.ox.ac.uk", "shimon.whiteson@cs.ox.ac.uk"], "authors_source": "OpenReview API", "abstract": "This paper introduces the deep coordination graph (DCG) for collaborative multi-agent reinforcement learning. DCG strikes a flexible trade-off between representational capacity and generalization by factorizing the joint value function of all agents according to a coordination graph into payoffs between pairs of agents. The value can be maximized by local message passing along the graph, which allows training of the value function end-to-end with Q-learning. Payoff functions are approximated with deep neural networks and parameter sharing improves generalization over the state-action space. We show that DCG can solve challenging predator-prey tasks that are vulnerable to the relative overgeneralization pathology and in which all other known value factorization approaches fail.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=HklRKpEKDr", "arxiv_id": "1910.00091", "paper_pdf": "papers/HklRKpEKDr.pdf", "paper_pdf_sha256": "35e1bbc38eb69de2fda18bff97e384b31a92cc5930124973a32b30eccee157a6", "paper_pdf_bytes": 609522, "paper_pdf_source": "openreview", "code_url": "https://github.com/wendelinboehmer/dcg", "code_repository": "wendelinboehmer/dcg", "code_commit": "4de100cddf7c3a7035cd89a47d7c1b8a878e7428", "code_archive": "repos/HklRKpEKDr.zip", "code_archive_sha256": "5fd82dadfa13181010958dfb3c6ab2ba068ab2db82bceb427f850ab45564a0cf", "code_archive_bytes": 263363, "code_file_count": 41, "code_extensions": {".py": 38, ".sh": 3}, "github_disk_usage_kb": 219, "github_languages": {"Python": 230756, "Dockerfile": 2141, "Shell": 1439}, "github_archived": false, "github_pushed_at": "2024-06-02T10:01:29Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/deep-coordination-graphs"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "B1l8iiA9tQ", "year": 2019, "status": "rejected", "title": "Backdrop: Stochastic Backpropagation", "authors": ["Siavash Golkar", "Kyle Cranmer"], "authorids": ["siavash.golkar@gmail.com", "kyle.cranmer@nyu.edu"], "authors_source": "OpenReview API", "abstract": "We introduce backdrop, a flexible and simple-to-implement method, intuitively described as dropout acting only along the backpropagation pipeline. Backdrop is implemented via one or more masking layers which are inserted at specific points along the network. Each backdrop masking layer acts as the identity in the forward pass, but randomly masks parts of the backward gradient propagation.  Intuitively, inserting a backdrop layer after any  convolutional layer leads to stochastic  gradients corresponding to features of that scale.  Therefore, backdrop is well suited for problems in which the data have a multi-scale, hierarchical structure. Backdrop can also be applied to problems with non-decomposable loss functions where standard SGD methods are not well suited. We perform a number of experiments and demonstrate that backdrop leads to significant improvements in generalization.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=B1l8iiA9tQ", "arxiv_id": "1806.01337", "paper_pdf": "papers/B1l8iiA9tQ.pdf", "paper_pdf_sha256": "676b66edd712e4abfd38240ffa69a1a3674dde4cd57a14761171f7d60851f3c5", "paper_pdf_bytes": 3031581, "paper_pdf_source": "openreview", "code_url": "https://github.com/golkar/backdrop", "code_repository": "golkar/backdrop", "code_commit": "5e42191ec8a260d657a1dd29cebdae1870e215f4", "code_archive": "repos/B1l8iiA9tQ.zip", "code_archive_sha256": "c52d56a242c92f30b2ce1d9a78ce1a646dea991bac32c86f6257d0f1586f4ba8", "code_archive_bytes": 6660056, "code_file_count": 4, "code_extensions": {".py": 2, ".ipynb": 2}, "github_disk_usage_kb": 7059, "github_languages": {"Jupyter Notebook": 4601766, "Python": 24685}, "github_archived": false, "github_pushed_at": "2018-06-06T13:40:10Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/backdrop-stochastic-backpropagation"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "LQ1qYDOk6o", "year": 2026, "status": "rejected", "title": "ReCalKV: Low-Rank KV Cache Compression via Head Reordering and Offline Calibration", "authors": ["Xianglong Yan", "Zhiteng Li", "Tianao Zhang", "Haotong Qin", "Linghe Kong", "Yulun Zhang", "Xiaokang Yang"], "authorids": ["~Xianglong_Yan2", "~Zhiteng_Li2", "~Tianao_Zhang1", "~Haotong_Qin1", "~Linghe_Kong1", "~Yulun_Zhang1", "~Xiaokang_Yang1"], "authors_source": "OpenReview API", "abstract": "Large language models (LLMs) have demonstrated remarkable performance, but their long-context reasoning remains constrained by the excessive memory required for the Key-Value (KV) cache. This makes KV cache compression a critical step toward efficient long-context inference. Recent methods have explored low-rank techniques to reduce the hidden size of the KV cache. However, they neglect the distinct roles and varying importance of Keys and Values, leading to significant performance drops under high compression. To address this, we propose ReCalKV, a post-training low-rank KV cache compression approach with tailored strategies for Keys and Values. For Keys, we propose Head-wise Similarity–aware Reordering (HSR), which clusters structurally similar heads into groups, enabling more accurate low-rank approximation via grouped SVD. For Values, we propose Offline Value Calibration (OVC), which efficiently calibrates the value projection matrix using calibration data without training, ensuring an accurate representation of contextual information. Extensive experiments show that ReCalKV consistently outperforms existing low-rank compression methods, achieving high compression ratios with minimal performance loss. We will release all the code and models.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=LQ1qYDOk6o", "arxiv_id": "2505.24357", "paper_pdf": "papers/LQ1qYDOk6o.pdf", "paper_pdf_sha256": "6436531a3dddfd3d77d201fafe88a60cd00460a6e90bc255e9c7b1954d2e1312", "paper_pdf_bytes": 489009, "paper_pdf_source": "openreview", "code_url": "https://github.com/XIANGLONGYAN/ReCalKV", "code_repository": "XIANGLONGYAN/ReCalKV", "code_commit": "ea548d5c6c648316f009770a2e54b80fa16c6b06", "code_archive": "repos/LQ1qYDOk6o.zip", "code_archive_sha256": "64dd664dd5c750627d57b988c413040b91dce4e8017754a54a9282b6efeb15ba", "code_archive_bytes": 319716, "code_file_count": 36, "code_extensions": {".py": 35, ".sh": 1}, "github_disk_usage_kb": 1250, "github_languages": {"Python": 1064261, "Shell": 3287}, "github_archived": false, "github_pushed_at": "2026-07-09T07:41:13Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/recalkv-low-rank-kv-cache-compression-via"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "DfTWrTwLzD", "year": 2025, "status": "rejected", "title": "Two Are Better than One: Context Window Extension with Multi-Grained Self-Injection", "authors": ["Wei Han", "Pan Zhou", "Soujanya Poria", "Shuicheng YAN"], "authorids": ["~Wei_Han4", "~Pan_Zhou3", "~Soujanya_Poria1", "~Shuicheng_YAN3"], "authors_source": "OpenReview API", "abstract": "Limited longtext window has been an inherent constraint for large language models (LLMs), which significantly restricts their application scenarios.  Continual pre-training on long-context data is the most straightforward approach to further extend an LLM's context window, but it is at the expense of huge data acquisition and computation cost. \nThere are many cost-efficient context window extension methods which do not require pretraining process emerges as appealing solutions, such as extrapolation, attention manipulation, context compression, etc.\nIn this paper, we propose a novel approach named Shared-LLaMA. \nShared-LLaMA is composed of two short-context LLMs. \nOne of them works as compressor and the other works as decoder. \nThe decoder receives compressed multi-grained context information from \nthe compressor and performs context-aware modeling on the running text. \nInformation transfer between the compressor and decoder occurs only at the lowest layers to circumvent an entire forward pass and save the inference time. \nBoth LLMs are initialized from the same off-the-shelf checkpoint and thus can be directly trained without extra feature alignment stages.\nAdditionally, we propose a tree structure to store the multi-grained information and design a search algorithm to fast locate and retrieve related information from each level of that tree. \nWith these efficient design choices, Shared-LLaMA can greatly reduce memory consumption, and achieves apparent speed up over other advanced baselines (2$\\times$ over streaming, $3\\times$ over encoder-decoder architectures).\nIn our evaluation on long-context modeling and understanding tasks, Shared-LLaMA yields superior or comparable results to several strong baselines, indicating Shared-LLaMA achieves a good balance between efficiency and effectiveness.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=DfTWrTwLzD", "arxiv_id": "2410.19318", "paper_pdf": "papers/DfTWrTwLzD.pdf", "paper_pdf_sha256": "4bc2acdb1ec0ddc1f7eda01a70a2c7e9150b8f3c9641dcc88d38686ffac8ef1d", "paper_pdf_bytes": 553766, "paper_pdf_source": "openreview", "code_url": "https://github.com/Clement25/SharedLLM", "code_repository": "Clement25/SharedLLM", "code_commit": "79c3e15f8cead0b6877ffe1f2ff01ec8a1745bd7", "code_archive": "repos/DfTWrTwLzD.zip", "code_archive_sha256": "5e189fd627edca0a82d63e71eb41ff2f03b94d1f2d8424aa54ab4ef4842087b4", "code_archive_bytes": 91726, "code_file_count": 19, "code_extensions": {".py": 13, ".sh": 6}, "github_disk_usage_kb": 90, "github_languages": {"Python": 354136, "Shell": 2053}, "github_archived": false, "github_pushed_at": "2026-02-01T13:04:30Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/two-are-better-than-one-context-window"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "iTrd5xyHLP", "year": 2024, "status": "rejected", "title": "LLMatic: Neural Architecture Search via Large Language Models and Quality Diversity Optimization", "authors": ["Muhammad Umair Nasir", "Sam Earle", "Julian Togelius", "Steven James", "Christopher Wesley Cleghorn"], "authorids": ["~Muhammad_Umair_Nasir1", "~Sam_Earle1", "~Julian_Togelius1", "~Steven_James1", "~Christopher_Wesley_Cleghorn1"], "authors_source": "OpenReview API", "abstract": "Large Language Models (LLMs) have emerged as powerful tools capable of accomplishing a broad spectrum of tasks. Their abilities span numerous areas, and one area where they have made a significant impact is in the domain of code generation. Here, we propose to use the coding abilities of LLMs to introduce meaningful variations to code defining neural networks. Meanwhile, Quality-Diversity (QD) algorithms are known to discover diverse and robust solutions. By merging the code-generating abilities of LLMs with the diversity and\nrobustness of QD solutions, we introduce LLMatic, a Neural Architecture Search (NAS) algorithm. While LLMs struggle to conduct NAS directly through prompts, LLMatic uses a procedural approach, leveraging QD for prompts and network architecture to create diverse and high-performing networks. We test LLMatic on the CIFAR-10 and NAS-bench-201 benchmark, demonstrating that it can produce competitive networks while evaluating just 2, 000 candidates, even without prior knowledge of the benchmark domain or exposure to any previous top-performing models for the benchmark.", "decision": "Reject", "meta_review": "", "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=iTrd5xyHLP", "arxiv_id": "2306.01102", "paper_pdf": "papers/iTrd5xyHLP.pdf", "paper_pdf_sha256": "d32d664d3b3bdb8367d9f94612df7015977236f5bee4c7ab992b21757b6fecc5", "paper_pdf_bytes": 939503, "paper_pdf_source": "openreview", "code_url": "https://github.com/umair-nasir14/LLMatic", "code_repository": "umair-nasir14/LLMatic", "code_commit": "bd29c7e7d2348620083e21049f0c81a89c75ed07", "code_archive": "repos/iTrd5xyHLP.zip", "code_archive_sha256": "f54f3be8637d137734b54e865353ecbac0379d3cc603355ddf0a8ed30209baec", "code_archive_bytes": 46601, "code_file_count": 15, "code_extensions": {".py": 15}, "github_disk_usage_kb": 70, "github_languages": {"Python": 93398}, "github_archived": false, "github_pushed_at": "2024-08-14T10:42:00Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/llmatic-neural-architecture-search-via-large"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "lJX9okHBVMb", "year": 2023, "status": "rejected", "title": "Universal approximation and model compression for radial neural networks ", "authors": ["Iordan Ganev", "Twan van Laarhoven", "Robin Walters"], "authorids": ["~Iordan_Ganev1", "~Twan_van_Laarhoven1", "~Robin_Walters1"], "authors_source": "OpenReview API", "abstract": "We introduce a class of fully-connected neural networks whose activation functions, rather than being pointwise, rescale feature vectors by a function depending only on their norm. We call such networks radial neural networks, extending previous work on rotation equivariant networks that considers rescaling activations in less generality. We prove universal approximation theorems for radial neural networks, including in the more difficult cases of bounded widths and unbounded domains. Our proof techniques are novel, distinct from those in the pointwise case. Additionally, radial neural networks exhibit a rich group of orthogonal change-of-basis symmetries on the vector space of trainable parameters. Factoring out these symmetries leads to a practical lossless model compression algorithm. Optimization of the compressed model by gradient descent is equivalent to projected gradient descent for the full model.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=lJX9okHBVMb", "arxiv_id": "2107.02550", "paper_pdf": "papers/lJX9okHBVMb.pdf", "paper_pdf_sha256": "f124ec1a6de6efd6c33a97f95ce115f01c3f279203fe12d1dd9f6a0d5e70d44d", "paper_pdf_bytes": 813577, "paper_pdf_source": "openreview", "code_url": "https://github.com/ivganev/QR-decomposition-radial-NNs", "code_repository": "ivganev/QR-decomposition-radial-NNs", "code_commit": "37ff7009815f30651e45d8c1e130640b60cc0cd3", "code_archive": "repos/lJX9okHBVMb.zip", "code_archive_sha256": "25cda1d09ef290c6879d33166ce1e72270d0d66889906063883d57d3ce0ed9c8", "code_archive_bytes": 43238, "code_file_count": 3, "code_extensions": {".py": 3}, "github_disk_usage_kb": 138, "github_languages": {"Python": 24519}, "github_archived": false, "github_pushed_at": "2023-09-26T15:14:23Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/the-qr-decomposition-for-radial-neural"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "YgR1rRWETI", "year": 2022, "status": "rejected", "title": "Connectivity Matters: Neural Network Pruning Through the Lens of Effective Sparsity", "authors": ["Artem M Vysogorets", "Julia Kempe"], "authorids": ["~Artem_M_Vysogorets1", "~Julia_Kempe1"], "authors_source": "OpenReview API", "abstract": "Neural network pruning is a fruitful area of research with surging interest in high sparsity regimes. Benchmarking in this domain heavily relies on faithful representation of the sparsity of subnetworks, which has been traditionally computed as the fraction of removed connections (direct sparsity). This definition, however, fails to recognize unpruned parameters that detached from input or output layers of underlying subnetworks, potentially underestimating actual effective sparsity: the fraction of inactivated connections. While this effect might be negligible for moderately pruned networks (up to $10\\times-100\\times$ compression rates), we find that it plays an increasing role for thinner subnetworks, greatly distorting comparison between different pruning algorithms. For example, we show that effective compression of a randomly pruned LeNet-300-100 can be orders of magnitude larger than its direct counterpart, while no discrepancy is ever observed when using SynFlow for pruning (Tanaka et al., 2020). In this work, we adopt the lens of effective sparsity to reevaluate several recent pruning algorithms on common benchmark architectures (e.g., LeNet-300-100, VGG-19, ResNet-18) and discover that their absolute and relative performance changes dramatically in this new, and as we argue, more appropriate framework. To aim for effective, rather than direct, sparsity, we develop a low-cost extension to most pruning algorithms. Further, equipped with effective sparsity as a reference frame, we partially reconfirm that random pruning with appropriate sparsity allocation across layers performs as well or better than more sophisticated algorithms for pruning at initialization (Su et al., 2020). In response to this observation, using a simple analogy of pressure distribution in coupled cylinders from thermodynamics, we design novel layerwise sparsity quotas that outperform all existing baselines in the context of random pruning.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=YgR1rRWETI", "arxiv_id": "2107.02306", "paper_pdf": "papers/YgR1rRWETI.pdf", "paper_pdf_sha256": "10f57b347cb4a1cefab8153d608b1e66dad556eaf074acd2e6ad1b5b19bbc44c", "paper_pdf_bytes": 29691568, "paper_pdf_source": "openreview", "code_url": "https://github.com/avysogorets/effective-sparsity", "code_repository": "avysogorets/effective-sparsity", "code_commit": "7924cfe8a26106cbf65c3a760aa352e58765faf3", "code_archive": "repos/YgR1rRWETI.zip", "code_archive_sha256": "b70200f6aaba6cd0defc322ff34d5ef41b95371e5327d4588d1f496bef8ec7f8", "code_archive_bytes": 22066, "code_file_count": 10, "code_extensions": {".py": 10}, "github_disk_usage_kb": 185, "github_languages": {"Python": 106117}, "github_archived": false, "github_pushed_at": "2022-03-15T04:41:36Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/connectivity-matters-neural-network-pruning"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Mf4ZSXMZP7", "year": 2021, "status": "rejected", "title": "Improving Post Training Neural Quantization: Layer-wise Calibration and Integer Programming", "authors": ["Itay Hubara", "Yury Nahshan", "Yair Hanani", "Ron Banner", "Daniel Soudry"], "authorids": ["~Itay_Hubara1", "~Yury_Nahshan1", "~Yair_Hanani1", "~Ron_Banner1", "~Daniel_Soudry1"], "authors_source": "OpenReview API", "abstract": "Lately, post-training quantization methods have gained considerable attention, as they are simple to use, and require only a small unlabeled calibration set. This small dataset cannot be used to fine-tune the model without significant over-fitting. Instead, these methods only use the calibration set to set the activations' dynamic ranges. However, such methods always resulted in significant accuracy degradation, when used below 8-bits (except on small datasets). Here we aim to break the 8-bit barrier. To this end, we minimize the quantization errors of each layer separately by optimizing its parameters over the calibration set.  We empirically demonstrate that this approach is: (1) much less susceptible to over-fitting than the standard fine-tuning approaches, and can be used even on a very small calibration set; and (2) more powerful than previous methods, which only set the activations' dynamic ranges. Furthermore, we demonstrate how to optimally allocate the bit-widths for each layer, while constraining accuracy degradation or model compression by proposing a novel integer programming formulation. Finally, we suggest model global statistics tuning, to correct biases introduced during quantization. Together, these methods yield state-of-the-art results for both vision and text models. For instance, on ResNet50, we obtain less than 1\\% accuracy degradation --- with 4-bit weights and activations in all layers, but the smallest two. Our code is available at, https://github.com/papers-submission/CalibTIP", "decision": "Reject", "meta_review": null, "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=Mf4ZSXMZP7", "arxiv_id": "2006.10518", "paper_pdf": "papers/Mf4ZSXMZP7.pdf", "paper_pdf_sha256": "3a0213392a8fab771d33067efadf20d69cf76c30b8aa2f0c813919d410c96a4f", "paper_pdf_bytes": 549968, "paper_pdf_source": "openreview", "code_url": "https://github.com/itayhubara/CalibTIP", "code_repository": "itayhubara/CalibTIP", "code_commit": "8bffbe450fd48351b23bf873c59fb60ece5143d7", "code_archive": "repos/Mf4ZSXMZP7.zip", "code_archive_sha256": "b2c0006d171fb9ea2571b154c3fba69b413de0c5f4c7cf2cd1400ddc73e4a72e", "code_archive_bytes": 201212, "code_file_count": 66, "code_extensions": {".py": 51, ".sh": 15}, "github_disk_usage_kb": 214, "github_languages": {"Python": 498134, "Shell": 17724}, "github_archived": false, "github_pushed_at": "2021-06-10T13:20:51Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/improving-post-training-neural-quantization"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "BkeqATVYwr", "year": 2020, "status": "rejected", "title": "GRAPH NEIGHBORHOOD ATTENTIVE POOLING", "authors": ["Zekarias Tilahun Kefato", "Sarunas Girdzijauskas"], "authorids": ["zekarias@kth.se", "sarunasg@kth.se"], "authors_source": "OpenReview API", "abstract": "Network representation learning (NRL) is a powerful technique for learning low-dimensional vector representation of high-dimensional and sparse graphs. Most studies explore the structure and meta data associated with the graph using random walks and employ a unsupervised or semi-supervised learning schemes. Learning in these methods is context-free, because only a single representation per node is learned. Recently studies have argued on the sufficiency of a single representation and proposed a context-sensitive approach that proved to be highly effective in applications such as link prediction and ranking.\nHowever, most of these methods rely on additional textual features that require RNNs or CNNs to capture high-level features or rely on a community detection algorithm to identifying multiple contexts of a node.\nIn this study, without requiring additional features nor a community detection algorithm, we propose a novel context-sensitive algorithm called GAP that learns to attend on different part of a node’s neighborhood using attentive pooling networks. We show the efficacy of GAP using three real-world datasets on link prediction and node clustering tasks and compare it against 10 popular and state-of-the-art (SOTA) baselines. GAP consistently outperforms them and achieves up to ≈9% and ≈20% gain over the best performing methods on link prediction and clustering tasks, respectively.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=BkeqATVYwr", "arxiv_id": "2001.10394", "paper_pdf": "papers/BkeqATVYwr.pdf", "paper_pdf_sha256": "901b8289344e9ddc367fbcc556f54105f43fcee00b2f2c53e1e37fa172c83087", "paper_pdf_bytes": 364677, "paper_pdf_source": "openreview", "code_url": "https://github.com/zekarias-tilahun/GAP", "code_repository": "zekarias-tilahun/GAP", "code_commit": "0bb528135690de28517686453835babe17ff8f16", "code_archive": "repos/BkeqATVYwr.zip", "code_archive_sha256": "fdf6b492e0ea2ec5a66985d83ad2bd0c51c9f3de6fbbd05c35f0ef3a61aa9cc8", "code_archive_bytes": 247506, "code_file_count": 6, "code_extensions": {".py": 5, ".sh": 1}, "github_disk_usage_kb": 263, "github_languages": {"Python": 27027, "Shell": 236}, "github_archived": false, "github_pushed_at": "2021-03-10T09:58:08Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/graph-neighborhood-attentive-pooling-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "rk4Wf30qKQ", "year": 2019, "status": "rejected", "title": "Security Analysis of Deep Neural Networks Operating in the Presence of Cache Side-Channel Attacks", "authors": ["Sanghyun Hong", "Michael Davinroy", "Yigitcan Kaya", "Stuart Nevans Locke", "Ian Rackow", "Kevin Kulda", "Dana Dachman-Soled", "Tudor Dumitraș"], "authorids": ["shhong@cs.umd.edu", "mdavinr1@swarthmore.edu", "yigitcan@cs.umd.edu", "stnevans@mail.rit.edu", "ian.rackow@gmail.com", "kevin_kulda1@baylor.edu", "danadach@ece.umd.edu", "tdumitra@umiacs.umd.edu"], "authors_source": "OpenReview API", "abstract": "Recent work has introduced attacks that extract the architecture information of deep neural networks (DNN), as this knowledge enhances an adversary’s capability to conduct attacks on black-box networks. This paper presents the first in-depth security analysis of DNN fingerprinting attacks that exploit cache side-channels.  First, we define the threat model for these attacks:  our adversary does not need the ability to query the victim model; instead, she runs a co-located process on the host machine victim ’s deep learning  (DL) system is running and passively monitors the accesses of the target functions in the shared framework.  Second, we introduce DeepRecon, an attack that reconstructs the architecture of the victim network by using the internal information extracted via Flush+Reload, a cache side-channel technique. Once the attacker observes function invocations that map directly to architecture attributes of the victim network, the attacker can reconstruct the victim’s entire network architecture.  In our evaluation, we demonstrate that an attacker can accurately reconstruct two complex networks (VGG19 and ResNet50) having only observed one forward propagation. Based on the extracted architecture attributes, we also demonstrate that an attacker can build a meta-model that accurately fingerprints the architecture and family of the pre-trained model in a transfer learning setting. From this meta-model,  we evaluate the importance of the observed attributes in the fingerprinting process. Third, we propose and evaluate new framework-level defense techniques that obfuscate our attacker’s observations. Our empirical security analysis represents a step toward understanding the DNNs’ vulnerability to cache side-channel attacks.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=rk4Wf30qKQ", "arxiv_id": "1810.03487", "paper_pdf": "papers/rk4Wf30qKQ.pdf", "paper_pdf_sha256": "54dc3bb449bedc82a735dbc978719661b7f5c85c824f6e1e762d9dd4ae406218", "paper_pdf_bytes": 671133, "paper_pdf_source": "openreview", "code_url": "https://github.com/Sanghyun-Hong/DeepRecon", "code_repository": "Sanghyun-Hong/DeepRecon", "code_commit": "ad01b7034345fd6bdf463a4e09c73bf0049c06d0", "code_archive": "repos/rk4Wf30qKQ.zip", "code_archive_sha256": "eed534f55dbb407d22737b86bd2ca7acd06bfe0f88b6bd71c49d347c9ff56b57", "code_archive_bytes": 3543907, "code_file_count": 19, "code_extensions": {".py": 17, ".c": 1, ".sh": 1}, "github_disk_usage_kb": 7501, "github_languages": {"Python": 63674, "C": 7274, "Shell": 1264, "Makefile": 589}, "github_archived": false, "github_pushed_at": "2020-02-19T09:01:43Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/security-analysis-of-deep-neural-networks"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "bGVkAhR7Eu", "year": 2026, "status": "rejected", "title": "SAS-Bench: A Fine-Grained Benchmark for Evaluating Short Answer Scoring with Large Language Models", "authors": ["Peichao Lai", "Kexuan Zhang", "Yi Lin", "Linyihan Zhang", "Feiyang Ye", "Jinhao Yan", "Yanwei Xu", "Conghui He", "Yilei Wang", "Wentao Zhang", "Bin CUI"], "authorids": ["~Peichao_Lai1", "~Kexuan_Zhang1", "~Yi_Lin13", "~Linyihan_Zhang1", "~Feiyang_Ye3", "~Jinhao_Yan1", "~Yanwei_Xu1", "~Conghui_He2", "~Yilei_Wang3", "~Wentao_Zhang1", "~Bin_CUI2"], "authors_source": "OpenReview API", "abstract": "Short Answer Scoring (SAS) is a critical task in automated subjective answer grading, playing an essential role in education, standardized testing, and large-scale assessment systems. However, existing approaches often produce coarse-grained scores and lack detailed reasoning. Although large language models (LLMs) have demonstrated potential as zero-shot evaluators, they remain susceptible to bias, inconsistencies with human judgment, and limited transparency in scoring decisions. To overcome these limitations, we introduce SAS-Bench, a benchmark specifically designed for LLM-based SAS tasks. SAS-Bench provides fine-grained, step-wise scoring, expert-annotated error categories, and a diverse range of question types derived from real-world subject-specific exams. This benchmark facilitates detailed evaluation of model reasoning processes and explainability. We also release an open-source dataset containing 1,030 questions and 4,109 student responses, each annotated by domain experts. Furthermore, we conduct comprehensive experiments with various LLMs, identifying major challenges in scoring science-related questions and highlighting the effectiveness of few-shot prompting in improving scoring accuracy. Our work offers valuable insights into the development of more robust, fair, and educationally meaningful LLM-based evaluation systems.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=bGVkAhR7Eu", "arxiv_id": "2505.07247", "paper_pdf": "papers/bGVkAhR7Eu.pdf", "paper_pdf_sha256": "86b7a1795aa98286c9201640fc2682f91c1810dc6ce2d2a22ae6a586740d7b7a", "paper_pdf_bytes": 1735852, "paper_pdf_source": "openreview", "code_url": "https://github.com/PKU-DAIR/SAS-Bench", "code_repository": "PKU-DAIR/SAS-Bench", "code_commit": "4ef718b6433623d492d483a38d4045a4b32d6f45", "code_archive": "repos/bGVkAhR7Eu.zip", "code_archive_sha256": "7f5c5fa7b2065714bf5d4caffbbe145f0641dcd6707912a9d11d839f31d051fd", "code_archive_bytes": 1048715, "code_file_count": 32, "code_extensions": {".py": 32}, "github_disk_usage_kb": 1255, "github_languages": {"Python": 147605}, "github_archived": false, "github_pushed_at": "2025-05-15T11:12:40Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/sas-bench-a-fine-grained-benchmark-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Bff9RniI03", "year": 2025, "status": "rejected", "title": "Leveraging Skills from Unlabeled Prior Data for Efficient Online Exploration", "authors": ["Max Wilcoxson", "Qiyang Li", "Kevin Frans", "Sergey Levine"], "authorids": ["~Max_Wilcoxson1", "~Qiyang_Li1", "~Kevin_Frans1", "~Sergey_Levine1"], "authors_source": "OpenReview API", "abstract": "Unsupervised pretraining has been transformative in many supervised domains. However, applying such ideas to reinforcement learning (RL) presents a unique challenge in that fine-tuning does not involve mimicking task-specific data, but rather exploring and locating the solution through iterative self-improvement. In this work, we showcase how unlabeled prior trajectory data can be leveraged to learn efficient exploration strategies. The key insight is to use unlabelled trajectories twice, 1) to extract a set of low-level skills offline, and 2) as additional data for a high-level policy that composes these skills to explore. We utilize a simple strategy of learning an optimistic reward model from online samples, and relabeling past trajectories into high-level, task-relevant examples. We instantiate these insights as SUPE (Skills from Unlabeled Prior data for Exploration), and empirically show that SUPE reliably outperforms prior strategies, successfully solving a suite of long-horizon, sparse-reward tasks.", "decision": "Reject", "meta_review": "", "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=Bff9RniI03", "arxiv_id": "2410.18076", "paper_pdf": "papers/Bff9RniI03.pdf", "paper_pdf_sha256": "76e28b3ec521581c18dd9a7d09a3afa78044539d5f2018734d7ed3d05e5bf66f", "paper_pdf_bytes": 8531032, "paper_pdf_source": "openreview", "code_url": "https://github.com/rail-berkeley/SUPE", "code_repository": "rail-berkeley/SUPE", "code_commit": "b4d1dfbc1e40638ee4ab2d3cc06782f26950c5b5", "code_archive": "repos/Bff9RniI03.zip", "code_archive_sha256": "bff5af27b6d867e408f7af29b65874fd806b015e5ff0115710de7bb1c51dbb19", "code_archive_bytes": 108725, "code_file_count": 75, "code_extensions": {".py": 74, ".sh": 1}, "github_disk_usage_kb": 92, "github_languages": {"Python": 315447, "Shell": 1427}, "github_archived": false, "github_pushed_at": "2025-07-11T20:35:30Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/leveraging-skills-from-unlabeled-prior-data"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "3mY9aGiMn0", "year": 2024, "status": "rejected", "title": "Sparser, Better, Deeper, Stronger: Improving Sparse Training with Exact Orthogonal Initialization", "authors": ["Aleksandra Nowak", "Łukasz Gniecki", "Filip Szatkowski", "Jacek Tabor"], "authorids": ["~Aleksandra_Nowak1", "~Łukasz_Gniecki1", "~Filip_Szatkowski1", "~Jacek_Tabor1"], "authors_source": "OpenReview API", "abstract": "Sparse training aims to train sparse models from scratch, achieving remarkable results in recent years. A key design choice in sparse training is the sparse initialization, which determines the trainable sub-network through a binary mask. Existing methods mainly revolve around selecting the mask based on predefined dense weight initialization. However, such an approach may not efficiently leverage the mask's potential impact on training parameters and optimization. An alternative direction, inspired by research into dynamical isometry, is to introduce orthogonality in the sparse subnetwork. This helps prevent the gradient signal from vanishing or exploding, ultimately enhancing the reliability of the backpropagation process. In this work, we propose Exact Orthogonal Initialization (EOI), a novel sparse orthogonal initialization scheme based on composing random Givens rotations. Contrary to other existing approaches, our method provides exact (not approximated) orthogonality and enables the creation of layers with arbitrary densities. Through experiments on contemporary network architectures, we present the effectiveness of EOI and demonstrate that it consistently outperforms other commonly used sparse initialization techniques. Furthermore, to showcase the full potential of our method, we show that it enables the training of highly sparse 1000-layer MLP and CNN networks without any residual connections or normalization techniques.  Our research highlights the importance of weight initialization in sparse training, underscoring the vital part it plays alongside the sparse mask selection.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=3mY9aGiMn0", "arxiv_id": "2406.01755", "paper_pdf": "papers/3mY9aGiMn0.pdf", "paper_pdf_sha256": "6e84ed26ce495981d6a6ea5389650450d9ff16e1b925c6cfdeffb0d9f8d67811", "paper_pdf_bytes": 1003240, "paper_pdf_source": "openreview", "code_url": "https://github.com/woocash2/sparser-better-deeper-stronger", "code_repository": "woocash2/sparser-better-deeper-stronger", "code_commit": "4438360111aa0f84aa75e9750e608a6127405a32", "code_archive": "repos/3mY9aGiMn0.zip", "code_archive_sha256": "ef776ac92f4b3f38feadb3c3f5f9b63424a49be716cf89d214246723bc237870", "code_archive_bytes": 94557, "code_file_count": 53, "code_extensions": {".py": 52, ".sh": 1}, "github_disk_usage_kb": 74, "github_languages": {"Python": 300807, "Shell": 50}, "github_archived": false, "github_pushed_at": "2024-03-12T18:22:55Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/sparser-better-deeper-stronger-improving"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "0bLE93R9d0O", "year": 2023, "status": "rejected", "title": "Transformer Module Networks for Systematic Generalization in Visual Question Answering", "authors": ["Moyuru Yamada", "Vanessa D'Amario", "Kentaro Takemoto", "Xavier Boix", "Tomotake Sasaki"], "authorids": ["~Moyuru_Yamada1", "~Vanessa_D'Amario1", "~Kentaro_Takemoto1", "~Xavier_Boix1", "~Tomotake_Sasaki1"], "authors_source": "OpenReview API", "abstract": "Transformers achieve great performance on Visual Question Answering (VQA). However, their systematic generalization capabilities, i.e., handling novel combinations of known concepts, is unclear. We reveal that Neural Module Networks (NMNs), i.e., question-specific compositions of modules that tackle a sub-task, achieve better or similar systematic generalization performance than the conventional Transformers, even though NMNs' modules are CNN-based. In order to address this shortcoming of Transformers with respect to NMNs, in this paper we investigate whether and how modularity can bring benefits to Transformers. Namely, we introduce Transformer Module Network (TMN), a novel NMN based on compositions of Transformer modules. TMNs achieve state-of-the-art systematic generalization performance in three VQA datasets, improving more than 30% over standard Transformers for novel compositions of sub-tasks. We show that not only the module composition but also the module specialization for each sub-task are the key of such performance gain.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=0bLE93R9d0O", "arxiv_id": "2201.11316", "paper_pdf": "papers/0bLE93R9d0O.pdf", "paper_pdf_sha256": "f10035ba5ca0b24ae3d2ba6bb76e14cccbb9bb8264f74dd9c5c0c55ebb74cb4a", "paper_pdf_bytes": 771561, "paper_pdf_source": "openreview", "code_url": "https://github.com/FujitsuResearch/transformer_module_networks", "code_repository": "FujitsuResearch/transformer_module_networks", "code_commit": "8e91aa7d9be865ba89f850a2beda7b7d622d07d7", "code_archive": "repos/0bLE93R9d0O.zip", "code_archive_sha256": "757c83b1fe9589f2378631e40db83119c46120a978adecf4f7cf526d1e0dc305", "code_archive_bytes": 255519, "code_file_count": 61, "code_extensions": {".py": 61}, "github_disk_usage_kb": 147, "github_languages": {"Python": 486962}, "github_archived": false, "github_pushed_at": "2023-02-13T07:07:03Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/transformer-module-networks-for-systematic"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "KDAEc2nai83", "year": 2022, "status": "rejected", "title": "Human-Level Control without Server-Grade Hardware", "authors": ["Brett Daley", "Christopher Amato"], "authorids": ["~Brett_Daley1", "~Christopher_Amato1"], "authors_source": "OpenReview API", "abstract": "Deep Q-Network (DQN) marked a major milestone for reinforcement learning, demonstrating for the first time that human-level control policies could be learned directly from raw visual inputs via reward maximization. Even years after its introduction, DQN remains highly relevant to the research community since many of its innovations have been adopted by successor methods. Nevertheless, despite significant hardware advances in the interim, DQN's original Atari 2600 experiments remain extremely costly to replicate in full. This poses an immense barrier to researchers who cannot afford state-of-the-art hardware or lack access to large-scale cloud computing resources. To facilitate improved access to deep reinforcement learning research, we introduce a DQN implementation that leverages a novel concurrent and synchronized execution framework designed to maximally utilize a heterogeneous CPU-GPU desktop system. With just one NVIDIA GeForce GTX 1080 GPU, our implementation reduces the training time of a 200-million-frame Atari experiment from 25 hours to just 9 hours. The ideas introduced in our paper should be generalizable to a large number of off-policy deep reinforcement learning methods.", "decision": "Reject", "meta_review": null, "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=KDAEc2nai83", "arxiv_id": "2111.01264", "paper_pdf": "papers/KDAEc2nai83.pdf", "paper_pdf_sha256": "14e52cc1cf4f718fc9e8c2cf7710f017ef4f76e28b8519995ed101c1c881fdd5", "paper_pdf_bytes": 426996, "paper_pdf_source": "openreview", "code_url": "https://github.com/brett-daley/fast-dqn", "code_repository": "brett-daley/fast-dqn", "code_commit": "acf21e8bb193e52d73aa8e2d4e355957095bbd36", "code_archive": "repos/KDAEc2nai83.zip", "code_archive_sha256": "32695a08511e302aeba8ecdd1fbb5fd0b37dc6642d406c4c522280dbf4080ff5", "code_archive_bytes": 18663, "code_file_count": 12, "code_extensions": {".py": 12}, "github_disk_usage_kb": 204, "github_languages": {"Python": 30454}, "github_archived": false, "github_pushed_at": "2022-04-18T22:21:21Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/human-level-control-without-server-grade-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "o1O5nc48rn", "year": 2021, "status": "rejected", "title": "Optimal Transport Graph Neural Networks", "authors": ["Gary Bécigneul", "Octavian-Eugen Ganea", "Benson Chen", "Regina Barzilay", "Tommi S. Jaakkola"], "authorids": ["~Gary_Bécigneul1", "~Octavian-Eugen_Ganea1", "~Benson_Chen1", "~Regina_Barzilay1", "~Tommi_S._Jaakkola1"], "authors_source": "OpenReview API", "abstract": "Current graph neural network (GNN) architectures naively average or sum node embeddings into an aggregated graph representation---potentially losing structural or semantic information. We here introduce OT-GNN, a model that computes graph embeddings using parametric prototypes that highlight key facets of different graph aspects. Towards this goal, we are (to our knowledge) the first to successfully combine optimal transport with parametric graph models. Graph representations are obtained from Wasserstein distances between the set of GNN node embeddings and \"prototype\" point clouds as free parameters. We theoretically prove that, unlike traditional sum aggregation, our function class on point clouds satisfies a fundamental universal approximation theorem. Empirically, we address an inherent collapse optimization issue by proposing a noise contrastive regularizer to steer the model towards truly exploiting the optimal transport geometry. Finally, we consistently report better generalization performance on several molecular property prediction tasks, while exhibiting smoother graph representations.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=o1O5nc48rn", "arxiv_id": "2006.04804", "paper_pdf": "papers/o1O5nc48rn.pdf", "paper_pdf_sha256": "348e8543fb9192be1e97e74bba1e81ca30c114da4ebde5ac13864d189136d5ac", "paper_pdf_bytes": 2375559, "paper_pdf_source": "openreview", "code_url": "https://github.com/benatorc/OTGNN", "code_repository": "benatorc/OTGNN", "code_commit": "9282a8627ff3b28d40b538d75217cd87b4a45341", "code_archive": "repos/o1O5nc48rn.zip", "code_archive_sha256": "c308585f21e5e6793defb0545a9aaa2ed93d883bc84cb60103d9a50e5b40ce0a", "code_archive_bytes": 253339, "code_file_count": 18, "code_extensions": {".py": 18}, "github_disk_usage_kb": 228, "github_languages": {"Python": 77141}, "github_archived": false, "github_pushed_at": "2020-06-05T16:24:07Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/optimal-transport-graph-neural-networks"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "SJl1o2NFwS", "year": 2020, "status": "rejected", "title": "Understanding and Improving Transformer From a Multi-Particle Dynamic System Point of View", "authors": ["Yiping Lu", "Zhuohan Li", "Di He", "Zhiqing Sun", "Bin Dong", "Tao Qin", "Liwei Wang", "Tie-Yan Liu"], "authorids": ["yplu@stanford.edu", "zhuohan@berkeley.edu", "di_he@pku.edu.cn", "zhiqings@andrew.cmu.edu", "bindong@math.pku.edu.cn", "taoqin@microsoft.com", "wanglw@cis.pku.edu.cn", "tyliu@microsoft.com"], "authors_source": "OpenReview API", "abstract": "The Transformer architecture is widely used in natural language processing. Despite its success, the design principle of the Transformer remains elusive. In this paper, we provide a novel perspective towards understanding the architecture: we show that the Transformer can be mathematically interpreted as a numerical Ordinary Differential Equation (ODE) solver for a convection-diffusion equation in a multi-particle dynamic system. In particular, how words in a sentence are abstracted into contexts by passing through the layers of the Transformer can be interpreted as approximating multiple particles' movement in the space using the Lie-Trotter splitting scheme and the Euler's method. Given this ODE's perspective, the rich literature of numerical analysis can be brought to guide us in designing effective structures beyond the Transformer. As an example, we propose to replace the Lie-Trotter splitting scheme by the Strang-Marchuk splitting scheme, a scheme that is more commonly used and with much lower local truncation errors. The Strang-Marchuk splitting scheme suggests that the self-attention and position-wise feed-forward network (FFN) sub-layers should not be treated equally. Instead, in each layer, two position-wise FFN sub-layers should be used, and the self-attention sub-layer is placed in between. This leads to a brand new architecture. Such an FFN-attention-FFN layer is \"Macaron-like\", and thus we call the network with this new architecture the Macaron Net. Through extensive experiments, we show that the Macaron Net is superior to the Transformer on both supervised and unsupervised learning tasks. The reproducible code can be found on http://anonymized", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=SJl1o2NFwS", "arxiv_id": "1906.02762", "paper_pdf": "papers/SJl1o2NFwS.pdf", "paper_pdf_sha256": "0bf202dffcb42105416a9ebdb94c2ee3d6a0099dbdaec4d5d6f0ed044ea58f1e", "paper_pdf_bytes": 537504, "paper_pdf_source": "openreview", "code_url": "https://github.com/zhuohan123/macaron-net", "code_repository": "zhuohan123/macaron-net", "code_commit": "3a84e7a3323bd1a3a9a303194ba336d670a1fb2c", "code_archive": "repos/SJl1o2NFwS.zip", "code_archive_sha256": "5110fccecbc2708850cc1919a3a744bcefb941ccf6ef6e161c46e31b196fd9a9", "code_archive_bytes": 466657, "code_file_count": 252, "code_extensions": {".py": 229, ".sh": 15, ".cpp": 4, ".lua": 4}, "github_disk_usage_kb": 268, "github_languages": {"Python": 1322659, "Shell": 25736, "Lua": 8860, "C++": 7164}, "github_archived": false, "github_pushed_at": "2019-06-10T09:05:34Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/understanding-and-improving-transformer-from"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Hkl-di09FQ", "year": 2019, "status": "rejected", "title": "Decoupling feature extraction from policy learning: assessing benefits of state representation learning in goal based robotics", "authors": ["Antonin Raffin", "Ashley Hill", "René Traoré", "Timothée Lesort", "Natalia Díaz-Rodríguez", "David Filliat"], "authorids": ["antonin.raffin@ensta-paristech.fr", "ashley.hill@u-psud.fr", "krb.traore@protonmail.com", "timothee.lesort@ensta-paristech.fr", "diaz.rodriguez.natalia@gmail.com", "david.filliat@ensta-paristech.fr"], "authors_source": "OpenReview API", "abstract": "Scaling end-to-end reinforcement learning to control real robots from vision presents a series of challenges, in particular in terms of sample efficiency. Against end-to-end learning, state representation learning can help learn a compact, efficient and relevant representation of states that speeds up policy learning, reducing the number of samples needed, and that is easier to interpret. We evaluate several state representation learning methods on goal based robotics tasks and propose a new unsupervised model that stacks representations and combines strengths of several of these approaches. This method encodes all the relevant features, performs on par or better than end-to-end learning, and is robust to hyper-parameters change.", "decision": null, "meta_review": "Reject", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=Hkl-di09FQ", "arxiv_id": "1901.08651", "paper_pdf": "papers/Hkl-di09FQ.pdf", "paper_pdf_sha256": "a59b0a779281b5002b842dfea943c3405789bed729cf1d3e297f908f22c31e99", "paper_pdf_bytes": 40040956, "paper_pdf_source": "openreview", "code_url": "https://github.com/araffin/robotics-rl-srl", "code_repository": "araffin/robotics-rl-srl", "code_commit": "eae7c1ab310c79662f6e68c0d255e08641037ffa", "code_archive": "repos/Hkl-di09FQ.zip", "code_archive_sha256": "14a26443d3d5650d8fba5a8af4543a43c8a30f1df2e2aeb7e45a20518ef4cd93", "code_archive_bytes": 3891022, "code_file_count": 79, "code_extensions": {".py": 76, ".sh": 3}, "github_disk_usage_kb": 8154, "github_languages": {"Python": 471471, "Shell": 805}, "github_archived": true, "github_pushed_at": "2021-04-05T18:43:17Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/decoupling-feature-extraction-from-policy"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "2H6KhX1kJr", "year": 2025, "status": "rejected", "title": "Transformers and slot encoding for sample efficient physical world modelling", "authors": ["Francesco Petri", "Luigi Asprino", "Aldo Gangemi"], "authorids": ["~Francesco_Petri1", "~Luigi_Asprino1", "~Aldo_Gangemi1"], "authors_source": "OpenReview API", "abstract": "World modelling, i.e. building a representation of the rules that govern the world so as to predict its evolution, is an essential ability for any agent interacting with the physical world. Recent applications of the Transformer architecture to the problem of world modelling from video input show notable improvements in sample efficiency. However, existing approaches tend to work only at the image level thus disregarding that the environment is composed of objects interacting with each other. In this paper, we propose an architecture combining Transformers for world modelling with the slot-attention paradigm, an approach for learning representations of objects appearing in a scene. We describe the resulting neural architecture and report experimental results showing an improvement over the existing solutions in terms of sample efficiency and a reduction of the variation of the performance over the training examples.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=2H6KhX1kJr", "arxiv_id": "2405.20180", "paper_pdf": "papers/2H6KhX1kJr.pdf", "paper_pdf_sha256": "95ff9dc6e4249f8bf4051aa6ce747ee78fababf5328527b2922b754877ccd834", "paper_pdf_bytes": 792919, "paper_pdf_source": "openreview", "code_url": "https://github.com/torchipeppo/transformers-and-slot-encoding-for-wm", "code_repository": "torchipeppo/transformers-and-slot-encoding-for-wm", "code_commit": "7ea6e923a3c616610ee4af243f09f1b69e5aeb86", "code_archive": "repos/2H6KhX1kJr.zip", "code_archive_sha256": "746ea5726bd9b5d2293f214ac537ff9a3f32848968e941fa4cffe66edf739935", "code_archive_bytes": 133480, "code_file_count": 55, "code_extensions": {".py": 55}, "github_disk_usage_kb": 93, "github_languages": {"Python": 211262}, "github_archived": false, "github_pushed_at": "2024-05-31T13:50:46Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/transformers-and-slot-encoding-for-sample"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "vxmvbzw76R", "year": 2024, "status": "rejected", "title": "Split-and-Denoise: Protect large language model inference with local differential privacy", "authors": ["Peihua Mai", "Ran Yan", "Zhe Huang", "Youjia Yang", "Yan Pang"], "authorids": ["~Peihua_Mai1", "~Ran_Yan3", "~Zhe_Huang6", "~Youjia_Yang1", "~Yan_Pang1"], "authors_source": "OpenReview API", "abstract": "Large Language Models (LLMs) shows powerful capability in natural language understanding by capturing hidden semantics in vector space. This process enriches the value of the text embeddings for various downstream tasks, thereby fostering the Embedding-as-a-Service (EaaS) business model. However, the direct transmission of text to servers poses a largely unaddressed risk of privacy leakage. To mitigate this issue, we introduce Split-N-Denoise (SnD), an innovative framework that split the model to execute the token embedding layer on the client side at minimal computational cost. This allows the client to introduce noise prior to transmitting the embeddings to the server and subsequently receive and denoise the perturbed output embeddings for downstream tasks. Our approach is designed for the inference stage of LLMs and requires no modifications to the model parameters, while also being computationally efficient on the client side. Extensive experiments demonstrate SnD's effectiveness in optimizing the privacy-utility tradeoff across various LLM architectures and diverse downstream tasks. The results reveal an significant accuracy improvement under the same privacy budget compared to the baseline, offering clients a privacy-preserving solution for local privacy protection.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=vxmvbzw76R", "arxiv_id": "2310.09130", "paper_pdf": "papers/vxmvbzw76R.pdf", "paper_pdf_sha256": "8b6c49af013842fa8a35aba3ca20320987361b97db1fb54d13dc08f17a724ed4", "paper_pdf_bytes": 1450395, "paper_pdf_source": "openreview", "code_url": "https://github.com/NusIoraPrivacy/eaas-privacy", "code_repository": "NusIoraPrivacy/eaas-privacy", "code_commit": "156dc7b3db3892948b39da3c4a098e09328af019", "code_archive": "repos/vxmvbzw76R.zip", "code_archive_sha256": "85aef26d0510907c76ec53cbc333a2af49d1f55808f37d2b2212d2176196b26d", "code_archive_bytes": 88252, "code_file_count": 27, "code_extensions": {".py": 26, ".sh": 1}, "github_disk_usage_kb": 76, "github_languages": {"Python": 237361, "Shell": 1578}, "github_archived": false, "github_pushed_at": "2024-05-25T02:17:13Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/split-and-denoise-protect-large-language"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "MR8pqi9R7xP", "year": 2023, "status": "rejected", "title": "Watch What You Pretrain For: Targeted, Transferable Adversarial Examples on Self-Supervised Speech Recognition models", "authors": ["Raphael Olivier", "Hadi Abdullah", "Bhiksha Raj"], "authorids": ["~Raphael_Olivier1", "~Hadi_Abdullah1", "~Bhiksha_Raj1"], "authors_source": "OpenReview API", "abstract": "A targeted adversarial attack produces audio samples that can force an Automatic Speech Recognition (ASR) system to output attacker-chosen text. To exploit ASR models in real-world, black-box settings, an adversary can leverage the transferability property, i.e. that an adversarial sample produced for a proxy ASR can also fool a different remote ASR. However recent work has shown that transferability against large ASR models is very difficult. In this work, we show that modern ASR architectures, specifically ones based on Self-Supervised Learning, are in fact vulnerable to transferability. We successfully demonstrate this phenomenon by evaluating state-of-the-art self-supervised ASR models like Wav2Vec2, HuBERT, Data2Vec and WavLM. We show that with low-level additive noise achieving a 30dB Signal-Noise Ratio, we can achieve target transferability with up to 80\\% accuracy. Next, we  1) use an ablation study to show that Self-Supervised learning is the main cause of that phenomenon, and 2) we provide an explanation for this phenomenon. Through this we show that modern ASR architectures are uniquely vulnerable to adversarial security threats.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=MR8pqi9R7xP", "arxiv_id": "2209.13523", "paper_pdf": "papers/MR8pqi9R7xP.pdf", "paper_pdf_sha256": "204d3e68259bec3510af00883f2006374f6ce59afa63da697f68e792a078e28c", "paper_pdf_bytes": 444708, "paper_pdf_source": "openreview", "code_url": "https://github.com/RaphaelOlivier/asr_transferability", "code_repository": "RaphaelOlivier/asr_transferability", "code_commit": "458c193839ecbc818f8f5674f046e878f6ceba3e", "code_archive": "repos/MR8pqi9R7xP.zip", "code_archive_sha256": "7ef0385007c03d4bf771a828ca0598f22dfe63255a4f532222af76f6157af17f", "code_archive_bytes": 129337, "code_file_count": 8, "code_extensions": {".py": 8}, "github_disk_usage_kb": 149, "github_languages": {"Python": 53349}, "github_archived": false, "github_pushed_at": "2025-04-24T21:34:10Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/watch-what-you-pretrain-for-targeted"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Lwclw6u3Pcw", "year": 2022, "status": "rejected", "title": "Characterizing and Measuring the Similarity of Neural Networks with Persistent Homology ", "authors": ["David Pérez Fernández", "Asier Gutiérrez-Fandiño", "Jordi Armengol-Estapé", "Marta Villegas"], "authorids": ["~David_Pérez_Fernández1", "~Asier_Gutiérrez-Fandiño1", "~Jordi_Armengol-Estapé1", "~Marta_Villegas2"], "authors_source": "OpenReview API", "abstract": "Characterizing the structural properties of neural networks is crucial yet poorly understood, and there are no well-established similarity measures between networks. In this work, we observe that neural networks can be represented as abstract simplicial complex and analyzed using their topological 'fingerprints' via Persistent Homology (PH). We then describe a PH-based representation proposed for characterizing and measuring similarity of neural networks. We empirically show the effectiveness of this representation as a descriptor of different architectures in several datasets. This approach based on Topological Data Analysis is a step towards better understanding neural networks and a useful similarity measure.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=Lwclw6u3Pcw", "arxiv_id": "2101.07752", "paper_pdf": "papers/Lwclw6u3Pcw.pdf", "paper_pdf_sha256": "f34ac26de02e8fb877a036a626701bcaecbd59b3c391607ed2e439afe2b2635e", "paper_pdf_bytes": 309153, "paper_pdf_source": "openreview", "code_url": "https://github.com/asier-gutierrez/nn-similarity", "code_repository": "asier-gutierrez/nn-similarity", "code_commit": "5e88b34145ff93e8e3c97ffe460f861663bc524e", "code_archive": "repos/Lwclw6u3Pcw.zip", "code_archive_sha256": "de9bca3c2134ec460d28ac99986c1dbc95cfc8b9c0c07a330bd7db1f09cdeb74", "code_archive_bytes": 226013, "code_file_count": 21, "code_extensions": {".py": 21}, "github_disk_usage_kb": 218, "github_languages": {"Python": 104931}, "github_archived": false, "github_pushed_at": "2021-07-17T08:20:54Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/determining-structural-properties-of"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "J_pvI6ap5Mn", "year": 2021, "status": "rejected", "title": "Transfer Learning of Graph Neural Networks with Ego-graph Information Maximization", "authors": ["Qi Zhu", "Yidan Xu", "Haonan Wang", "Chao Zhang", "Jiawei Han", "Carl Yang"], "authorids": ["~Qi_Zhu7", "~Yidan_Xu1", "~Haonan_Wang1", "~Chao_Zhang9", "~Jiawei_Han1", "~Carl_Yang1"], "authors_source": "OpenReview API", "abstract": "Graph neural networks (GNNs) have been shown with superior performance in various applications, but training dedicated GNNs can be costly for large-scale graphs. Some recent work started to study the pre-training of GNNs. However, none of them provide theoretical insights into the design of their frameworks, or clear requirements and guarantees towards the transferability of GNNs. In this work, we establish a theoretically grounded and practically useful framework for the transfer learning of GNNs. Firstly, we propose a novel view towards the essential graph information and advocate the capturing of it as the goal of transferable GNN training, which motivates the design of EGI (ego-graph information maximization) to analytically achieve this goal. Secondly, we specify the requirement of structure-respecting node features as the GNN input, and conduct a rigorous analysis of GNN transferability based on the difference between the local graph Laplacians of the source and target graphs. Finally, we conduct controlled synthetic experiments to directly justify our theoretical conclusions. Extensive experiments on real-world networks towards role identification show consistent results in the rigorously analyzed setting of direct-transfering (freezing parameters), while those towards large-scale relation prediction show promising results in the more generalized and practical setting of transfering with fine-tuning.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=J_pvI6ap5Mn", "arxiv_id": "2009.05204", "paper_pdf": "papers/J_pvI6ap5Mn.pdf", "paper_pdf_sha256": "eeb409573db5281d3622578c2c221514c5fc724c07cdb94920304d075e3fdf63", "paper_pdf_bytes": 1218614, "paper_pdf_source": "openreview", "code_url": "https://github.com/GentleZhu/EGI", "code_repository": "GentleZhu/EGI", "code_commit": "f8b8ce8e987c63b97fb49a838fec600223533439", "code_archive": "repos/J_pvI6ap5Mn.zip", "code_archive_sha256": "955f048d301dbe1ea53e6ea9edd7286fcb7bb51fcd78e52f5e1db9e49ee29bac", "code_archive_bytes": 185519, "code_file_count": 10, "code_extensions": {".py": 10}, "github_disk_usage_kb": 246, "github_languages": {"Python": 121788}, "github_archived": false, "github_pushed_at": "2021-12-09T16:37:43Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/transfer-learning-of-graph-neural-networks"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "H1gz_nNYDS", "year": 2020, "status": "rejected", "title": "AutoSlim: Towards One-Shot Architecture Search for Channel Numbers", "authors": ["Jiahui Yu", "Thomas Huang"], "authorids": ["jyu79@illinois.edu", "t-huang1@illinois.edu"], "authors_source": "OpenReview API", "abstract": "\nWe study how to set the number of channels in a neural network to achieve better accuracy under constrained resources (e.g., FLOPs, latency, memory footprint or model size). A simple and one-shot approach, named AutoSlim, is presented. Instead of training many network samples and searching with reinforcement learning, we train a single slimmable network to approximate the network accuracy of different channel configurations. We then iteratively evaluate the trained slimmable model and greedily slim the layer with minimal accuracy drop. By this single pass, we can obtain the optimized channel configurations under different resource constraints. We present experiments with MobileNet v1, MobileNet v2, ResNet-50 and RL-searched MNasNet on ImageNet classification. We show significant improvements over their default channel configurations. We also achieve better accuracy than recent channel pruning methods and neural architecture search methods with 100X lower search cost.\n\nNotably, by setting optimized channel numbers, our AutoSlim-MobileNet-v2 at 305M FLOPs achieves 74.2% top-1 accuracy, 2.4% better than default MobileNet-v2 (301M FLOPs), and even 0.2% better than RL-searched MNasNet (317M FLOPs). Our AutoSlim-ResNet-50 at 570M FLOPs, without depthwise convolutions, achieves 1.3% better accuracy than MobileNet-v1 (569M FLOPs).\n", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=H1gz_nNYDS", "arxiv_id": "1903.11728", "paper_pdf": "papers/H1gz_nNYDS.pdf", "paper_pdf_sha256": "36b441e49cfd8acb0cdbb38f7ff8f010a9e3b9b0c51b819c726571f31d1a965b", "paper_pdf_bytes": 599945, "paper_pdf_source": "openreview", "code_url": "https://github.com/JiahuiYu/slimmable_networks", "code_repository": "JiahuiYu/slimmable_networks", "code_commit": "5dc14d0357ccfc596d706281acdc8a5b0b66c6d6", "code_archive": "repos/H1gz_nNYDS.zip", "code_archive_sha256": "4c507301cee871fa3074273131195b58b96f53cadc272667dbed40d07e121143", "code_archive_bytes": 49374, "code_file_count": 18, "code_extensions": {".py": 18}, "github_disk_usage_kb": 329, "github_languages": {"Python": 106555}, "github_archived": false, "github_pushed_at": "2023-03-09T08:37:20Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/network-slimming-by-slimmable-networks"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "wywgRd1MUQ", "year": 2026, "status": "rejected", "title": "InfiGUI-R1: Advancing Multimodal GUI Agents from Reactive Actors to Deliberative Reasoners", "authors": ["Yuhang Liu", "Pengxiang Li", "Congkai Xie", "Xueyu Hu", "Xiaotian Han", "Shengyu Zhang", "Hongxia Yang", "Fei Wu"], "authorids": ["~Yuhang_Liu9", "~Pengxiang_Li2", "~Congkai_Xie1", "~Xueyu_Hu1", "~Xiaotian_Han2", "~Shengyu_Zhang2", "~Hongxia_Yang2", "~Fei_Wu1"], "authors_source": "OpenReview API", "abstract": "Multimodal Large Language Models (MLLMs) have shown significant promise in powering Graphical User Interface (GUI) agents to automate complex digital tasks. However, the prevailing monolithic training paradigms often create a structural mismatch with the hierarchical nature of capabilities required for robust performance. Specifically, the efficacy of methods like Reinforcement Learning (RL) is critically predicated on the agent possessing a high-quality behavioral prior of key reasoning skills, such as spatial reasoning and goal decomposition, which are often absent. To resolve this impasse, we propose Actor2Reasoner, a novel two-stage hierarchical training paradigm grounded in the principle of Endow First, Internalize Later. The first stage, Cognitive Endowment, employs targeted supervised fine-tuning to instill these crucial thinking patterns, forging a Capable Actor. Subsequently, the second stage, Policy Internalization, utilizes RL to evolve this actor into a Deliberative Reasoner by internalizing the endowed abilities into a robust, context-aware decision-making policy. We instantiate our paradigm in InfiGUI-R1, an agent that achieves state-of-the-art performance on challenging benchmarks, including AndroidControl. Our work demonstrates that decoupling the endowment of foundational abilities from the internalization of policy provides a more effective and principled path toward developing sophisticated and resilient GUI agents.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=wywgRd1MUQ", "arxiv_id": "2504.14239", "paper_pdf": "papers/wywgRd1MUQ.pdf", "paper_pdf_sha256": "758780737e78b6c36c142526fc2960cbf5efbceee33925c9565c965e9ba46028", "paper_pdf_bytes": 525419, "paper_pdf_source": "openreview", "code_url": "https://github.com/InfiXAI/InfiGUI-R1", "code_repository": "InfiXAI/InfiGUI-R1", "code_commit": "a4fca17809a4395ba1fe08d481bb82c790ea7236", "code_archive": "repos/wywgRd1MUQ.zip", "code_archive_sha256": "8824c1100e804466b8ba3146a51cb4788dd217fff4dee0f3bda39c93ed49feeb", "code_archive_bytes": 1520968, "code_file_count": 3, "code_extensions": {".py": 3}, "github_disk_usage_kb": 1499, "github_languages": {"Python": 28852}, "github_archived": false, "github_pushed_at": "2025-12-04T13:36:09Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/infigui-r1-advancing-multimodal-gui-agents"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "eqVu9eaVAB", "year": 2025, "status": "rejected", "title": "What Matters in Hierarchical Search for Combinatorial Reasoning Problems?", "authors": ["Michał Zawalski", "Gracjan Góral", "Michał Tyrolski", "Emilia Wiśnios", "Franciszek Budrowski", "Łukasz Kuciński", "Piotr Miłoś"], "authorids": ["~Michał_Zawalski1", "~Gracjan_Góral1", "~Michał_Tyrolski1", "~Emilia_Wiśnios1", "~Franciszek_Budrowski1", "~Łukasz_Kuciński1", "~Piotr_Miłoś1"], "authors_source": "OpenReview API", "abstract": "Combinatorial reasoning problems, particularly the notorious NP-hard tasks, remain a significant challenge for AI research. A common approach to addressing them combines search with learned heuristics. Recent methods in this domain utilize hierarchical planning, executing strategies based on subgoals. Our goal is to advance research in this area and establish a solid conceptual and empirical foundation. Specifically, we identify the following key obstacles, whose presence favors the choice of hierarchical search methods: _hard-to-learn value functions_, _complex action spaces_, _presence of dead ends in the environment_, or _data collected from diverse sources_. Through in-depth empirical analysis, we establish that hierarchical search methods consistently outperform standard search methods across these dimensions, and we formulate insights for future research. On the practical side, we also propose a consistent evaluation methodology to enable meaningful comparisons between methods and to reassess the state-of-the-art algorithms.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=eqVu9eaVAB", "arxiv_id": "2406.03361", "paper_pdf": "papers/eqVu9eaVAB.pdf", "paper_pdf_sha256": "ce00deb1b508c3ddae286c259c1d1e872025438198475eaa0231468eb23d5bb6", "paper_pdf_bytes": 10711483, "paper_pdf_source": "openreview", "code_url": "https://github.com/subgoal-verification/what-matters-in-hierarchical-search", "code_repository": "subgoal-verification/what-matters-in-hierarchical-search", "code_commit": "ac0f72b23046d503ff40d75b43071cdf40fdfa31", "code_archive": "repos/eqVu9eaVAB.zip", "code_archive_sha256": "90aeff1d509c6d49fb0af18caf826877beab5fc77c2b37b360c7f9bf9781aa16", "code_archive_bytes": 72879, "code_file_count": 52, "code_extensions": {".py": 52}, "github_disk_usage_kb": 96, "github_languages": {"Python": 165544}, "github_archived": false, "github_pushed_at": "2024-05-18T14:48:05Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/what-matters-in-hierarchical-search-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "WfjJOEfAf7", "year": 2024, "status": "rejected", "title": "Information Flow in Self-Supervised Learning", "authors": ["Zhiquan Tan", "Jingqin Yang", "Weiran Huang", "Yang Yuan", "Yifan Zhang"], "authorids": ["~Zhiquan_Tan1", "~Jingqin_Yang2", "~Weiran_Huang1", "~Yang_Yuan4", "~Yifan_Zhang16"], "authors_source": "OpenReview API", "abstract": "In this paper, we provide a comprehensive toolbox for understanding and enhancing self-supervised learning (SSL) methods through the lens of matrix information theory. Specifically, by leveraging the principles of matrix mutual information and joint entropy, we offer a unified analysis for both contrastive and non-contrastive methods. Furthermore, we propose the matrix variational masked auto-encoder (M-MAE) method, grounded in matrix information theory, as an enhancement to masked image modeling. The empirical evaluations underscore the effectiveness of M-MAE compared with the state-of-the-art methods, including a 3.9% improvement in linear probing ViT-Base, and a 1% improvement in fine-tuning ViT-Large, both on ImageNet.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=WfjJOEfAf7", "arxiv_id": "2309.17281", "paper_pdf": "papers/WfjJOEfAf7.pdf", "paper_pdf_sha256": "127547cea9ad41185d9cce04fdfa35a815969bee45335bd68d65ea6f5a38efcd", "paper_pdf_bytes": 440808, "paper_pdf_source": "openreview", "code_url": "https://github.com/yifanzhang-pro/M-MAE", "code_repository": "yifanzhang-pro/M-MAE", "code_commit": "22dc5b248b3aa1788568d9b592a51ece9d28c847", "code_archive": "repos/WfjJOEfAf7.zip", "code_archive_sha256": "c8913c4f10624aa2949ad88305df541ae09bfa2f52095622bf0d69e0fbaf09a5", "code_archive_bytes": 46775, "code_file_count": 21, "code_extensions": {".py": 18, ".sh": 3}, "github_disk_usage_kb": 76, "github_languages": {"Python": 100780, "Shell": 1555}, "github_archived": false, "github_pushed_at": "2024-09-17T06:47:43Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/information-flow-in-self-supervised-learning"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "bTy4D3KHwWU", "year": 2023, "status": "rejected", "title": "Normalizing Flows for Interventional Density Estimation", "authors": ["Valentyn Melnychuk", "Dennis Frauen", "Stefan Feuerriegel"], "authorids": ["~Valentyn_Melnychuk1", "~Dennis_Frauen1", "~Stefan_Feuerriegel1"], "authors_source": "OpenReview API", "abstract": "Existing machine learning methods for causal inference usually estimate quantities expressed via the mean of potential outcomes (e.g., average treatment effect). However, such quantities do not capture the full information about the distribution of potential outcomes. In this work, we estimate the density of potential outcomes after interventions from observational data. For this, we propose a novel, fully-parametric deep learning method called Interventional Normalizing Flows. Specifically, we combine two normalizing flows, namely (i) a teacher flow for estimating nuisance parameters and (ii) a student flow for a parametric estimation of the density of potential outcomes. We further develop a tractable optimization objective via a one-step bias correction for an efficient and doubly robust estimation of the student flow parameters. As a result our Interventional Normalizing Flows offer a properly normalized density estimator.  Across various experiments, we demonstrate that our Interventional Normalizing Flows are expressive and highly effective, and scale well with both sample size and high-dimensional confounding. To the best of our knowledge, our Interventional Normalizing Flows are the first fully-parametric, deep learning method for density estimation of potential outcomes.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=bTy4D3KHwWU", "arxiv_id": "2209.06203", "paper_pdf": "papers/bTy4D3KHwWU.pdf", "paper_pdf_sha256": "e521635fe088dd215380b017d5a13ddfb5ab42b53ff106cfdaec150506ef58ed", "paper_pdf_bytes": 1068755, "paper_pdf_source": "openreview", "code_url": "https://github.com/Valentyn1997/INFs", "code_repository": "Valentyn1997/INFs", "code_commit": "b75d2e9f52a627145b5bd4a5b104337d04937e56", "code_archive": "repos/bTy4D3KHwWU.zip", "code_archive_sha256": "61410ad0fafe67d1454aa6670a71e162dc51a8d52ef116f45636b4ab7f1f46e2", "code_archive_bytes": 176658, "code_file_count": 22, "code_extensions": {".py": 22}, "github_disk_usage_kb": 149, "github_languages": {"Python": 125182}, "github_archived": false, "github_pushed_at": "2024-10-29T23:02:33Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/normalizing-flows-for-interventional-density"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ezbMFmQY7L", "year": 2022, "status": "rejected", "title": "C5T5: Controllable Generation of Organic Molecules with Transformers", "authors": ["Daniel Rothchild", "Alex Tamkin", "Julie Yu", "Ujval Misra", "Joseph E. Gonzalez"], "authorids": ["~Daniel_Rothchild1", "~Alex_Tamkin1", "~Julie_Yu2", "~Ujval_Misra1", "~Joseph_E._Gonzalez1"], "authors_source": "OpenReview API", "abstract": "Methods for designing organic materials with desired properties have high potential impact across fields such as medicine, renewable energy, petrochemical engineering, and agriculture. However, using generative models for this task is difficult because candidate compounds must satisfy many constraints, including synthetic accessibility, intellectual property attributes, ``chemical beauty'' (Bickerton et al., 2020), and other considerations that are intuitive to domain experts but can be challenging to quantify. We propose C5T5, a novel self-supervised pretraining method that works in tandem with domain experts by making zero-shot select-and-replace edits, altering organic substances towards desired property values. C5T5 operates on IUPAC names---a standardized molecular representation that intuitively encodes rich structural information for organic chemists but that has been largely ignored by the ML community. Our technique requires no edited molecule pairs to train and only a rough estimate of molecular properties, and it has the potential to model long-range dependencies and symmetric molecular structures more easily than graph-based methods. We demonstrate C5T5's effectiveness on four physical properties relevant for drug discovery, showing that it learns successful and chemically intuitive strategies for altering molecules towards desired property values.\n", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=ezbMFmQY7L", "arxiv_id": "2108.10307", "paper_pdf": "papers/ezbMFmQY7L.pdf", "paper_pdf_sha256": "8237b4ff69e1909247a9c368f6fc312120066812c7356f0928ab41331075ae9f", "paper_pdf_bytes": 1083477, "paper_pdf_source": "openreview", "code_url": "https://github.com/dhroth/c5t5", "code_repository": "dhroth/c5t5", "code_commit": "cba53cc881a87eb7d3451fd73e10df479e2ca057", "code_archive": "repos/ezbMFmQY7L.zip", "code_archive_sha256": "01d3595b9a56ec820573ac4a552f03e98a5653dc919fc2e39c3ac40a6fa979d1", "code_archive_bytes": 220637, "code_file_count": 16, "code_extensions": {".py": 13, ".sh": 3}, "github_disk_usage_kb": 223, "github_languages": {"Python": 60360, "Shell": 17555}, "github_archived": false, "github_pushed_at": "2021-12-17T17:39:31Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/c5t5-controllable-generation-of-organic"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "1s1T7xHc5l6", "year": 2021, "status": "rejected", "title": "FILTRA: Rethinking Steerable CNN by Filter Transform", "authors": ["Bo Li", "Qili Wang", "Gim Hee Lee"], "authorids": ["~Bo_Li16", "~Qili_Wang1", "~Gim_Hee_Lee1"], "authors_source": "OpenReview API", "abstract": "Steerable CNN imposes the prior knowledge of transformation invariance or equivariance in the network architecture to enhance the the network robustness on geometry transformation of data and reduce overfitting. Filter transform has been an intuitive and widely used technique to construct steerable CNN in the past decades. Recently, group representation theory is used to analyze steerable CNN and reveals the function space structure of a steerable kernel function. However, it is not yet clear on how this theory is related to the filter transform technique. In this paper, we show that kernel constructed by filter transform can also be interpreted in the group representation theory. Meanwhile, we show that filter transformed kernels can be used to convolve input/output features in different group representation. This interpretation help complete the puzzle of steerable CNN theory and provides a novel and simple approach to implement steerable convolution operators. Experiments are executed on multiple datasets to verify the feasibilty of the proposed approach.\n", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=1s1T7xHc5l6", "arxiv_id": "2105.11636", "paper_pdf": "papers/1s1T7xHc5l6.pdf", "paper_pdf_sha256": "b554d92887b7df121850744d01c96673ed4441497580a48c984accf8742363e9", "paper_pdf_bytes": 327026, "paper_pdf_source": "openreview", "code_url": "https://github.com/prclibo/filtra", "code_repository": "prclibo/filtra", "code_commit": "694e6f624f9221e0793b3132374731a7aebe0f18", "code_archive": "repos/1s1T7xHc5l6.zip", "code_archive_sha256": "96104740fe4c35dcff3b01090ba36c20d13a08cafaed87db086a97b62ae03bad", "code_archive_bytes": 87080, "code_file_count": 24, "code_extensions": {".py": 24}, "github_disk_usage_kb": 268, "github_languages": {"Python": 137175}, "github_archived": false, "github_pushed_at": "2022-03-01T14:51:45Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/filtra-rethinking-steerable-cnn-by-filter-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ByeL1R4FvS", "year": 2020, "status": "rejected", "title": "Unsupervised Data Augmentation for Consistency Training", "authors": ["Qizhe Xie", "Zihang Dai", "Eduard Hovy", "Minh-Thang Luong", "Quoc V. Le"], "authorids": ["qizhex@cs.cmu.edu", "dzihang@cs.cmu.edu", "hovy@cs.cmu.edu", "thangluong@google.com", "qvl@google.com"], "authors_source": "OpenReview API", "abstract": "Semi-supervised learning lately has shown much promise in improving deep learning models when labeled data is scarce. Common among recent approaches is the use of consistency training on a large amount of unlabeled data to constrain model predictions to be invariant to input noise. In this work, we present a new perspective on how to effectively noise unlabeled examples and argue that the quality of noising, specifically those produced by advanced data augmentation methods, plays a crucial role in semi-supervised learning. By substituting simple noising operations with advanced data augmentation methods, our method brings substantial improvements across six language and three vision tasks under the same consistency training framework. On the IMDb text classification dataset, with only 20 labeled examples, our method achieves an error rate of 4.20, outperforming the state-of-the-art model trained on 25,000 labeled examples. On a standard semi-supervised learning benchmark, CIFAR-10, our method outperforms all previous approaches and achieves an error rate of 2.7% with only 4,000 examples, nearly matching the performance of models trained on 50,000 labeled examples. Our method also combines well with transfer learning, e.g., when finetuning from BERT, and yields improvements in high-data regime, such as ImageNet, whether when there is only 10% labeled data or when a full labeled set with 1.3M extra unlabeled examples is used.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=ByeL1R4FvS", "arxiv_id": "1904.12848", "paper_pdf": "papers/ByeL1R4FvS.pdf", "paper_pdf_sha256": "9e11787b05769008c1480d2caca9641fce9c07ac5040a0a36642108943db4de9", "paper_pdf_bytes": 1395993, "paper_pdf_source": "openreview", "code_url": "https://github.com/google-research/uda", "code_repository": "google-research/uda", "code_commit": "960684e363251772a5938451d4d2bc0f1da9e24b", "code_archive": "repos/ByeL1R4FvS.zip", "code_archive_sha256": "00df1a73bf81e9dbdd95d5fe59b152272e3d391f5ee2e5e9e2eeaafa5ba29c49", "code_archive_bytes": 332834, "code_file_count": 44, "code_extensions": {".py": 30, ".sh": 14}, "github_disk_usage_kb": 342, "github_languages": {"Python": 234861, "Shell": 19001}, "github_archived": true, "github_pushed_at": "2021-08-28T07:16:56Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/unsupervised-data-augmentation-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "7cEMkTu7Lf", "year": 2026, "status": "rejected", "title": "Unlearning Isn't Deletion: Investigating Reversibility of Machine Unlearning in LLMs", "authors": ["Xiaoyu Xu", "Xiang Yue", "Yang Liu", "Qingqing Ye", "Huadi Zheng", "Peizhao Hu", "Minxin Du", "Haibo Hu"], "authorids": ["~Xiaoyu_Xu6", "~Xiang_Yue1", "~Yang_Liu3", "~Qingqing_Ye1", "~Huadi_Zheng1", "~Peizhao_Hu2", "~Minxin_Du1", "~Haibo_Hu2"], "authors_source": "OpenReview API", "abstract": "Unlearning in large language models (LLMs) aims to remove specified data, but its efficacy is typically assessed with task-level metrics like accuracy and perplexity. \nWe demonstrate that these metrics are often misleading, as models can appear to forget while their original behavior is easily restored through minimal fine-tuning. \nThis phenomenon of \\emph{reversibility} suggests that information is merely suppressed, not genuinely erased. To address this critical evaluation gap, we introduce a \\emph{representation-level analysis framework}. \nOur toolkit comprises PCA-based similarity and shift, centered kernel alignment (CKA), and Fisher information, complemented by a summary metric, the mean PCA distance, to measure representational drift. \nApplying this framework across six unlearning methods, three data domains, and two LLMs, we identify four distinct forgetting regimes based on their \\emph{reversibility} and \\emph{catastrophicity}. \nOur analysis reveals that achieving the ideal state--irreversible, non-catastrophic forgetting--is exceptionally challenging. \nBy probing the limits of unlearning, we identify a case of seemingly irreversible, targeted forgetting, offering new insights for designing more robust erasure algorithms. \nOur findings expose a fundamental gap in current evaluation practices and establish a representation-level foundation for trustworthy unlearning.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=7cEMkTu7Lf", "arxiv_id": "2505.16831", "paper_pdf": "papers/7cEMkTu7Lf.pdf", "paper_pdf_sha256": "d7c36abe236bdc5a53abbaaece6a7030abf69cfff38d757f86190275da20a19a", "paper_pdf_bytes": 6000515, "paper_pdf_source": "openreview", "code_url": "https://github.com/XiaoyuXU1/Representational_Analysis_Tools", "code_repository": "XiaoyuXU1/Representational_Analysis_Tools", "code_commit": "d0ddda6e794ed35b1c0e1c11a3d92e0787443da0", "code_archive": "repos/7cEMkTu7Lf.zip", "code_archive_sha256": "ac2b73617749ee2b5bbcc51e62a3af0cc7c8750c771f6a023ce1e549e68b5b7c", "code_archive_bytes": 933904, "code_file_count": 5, "code_extensions": {".py": 5}, "github_disk_usage_kb": 1533, "github_languages": {"Python": 26430}, "github_archived": false, "github_pushed_at": "2025-05-23T02:53:19Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/unlearning-isn-t-deletion-investigating"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "6Mdvq0bPyG", "year": 2025, "status": "rejected", "title": "EfficientQAT: Efficient Quantization-Aware Training for Large Language Models", "authors": ["Mengzhao Chen", "Wenqi Shao", "Peng Xu", "Jiahao Wang", "Peng Gao", "Kaipeng Zhang", "Ping Luo"], "authorids": ["~Mengzhao_Chen1", "~Wenqi_Shao2", "~Peng_Xu11", "~Jiahao_Wang1", "~Peng_Gao3", "~Kaipeng_Zhang1", "~Ping_Luo2"], "authors_source": "OpenReview API", "abstract": "Large language models (LLMs) are crucial in modern natural language processing and artificial intelligence. However, they face challenges in managing their significant memory requirements. Although quantization-aware training (QAT) offers a solution by reducing memory consumption through low-bit representations with minimal accuracy loss, it is impractical due to substantial training resources.  To address this, we propose Efficient Quantization-Aware Training (EfficientQAT), a more feasible QAT algorithm. EfficientQAT involves two consecutive phases: Block-wise training of all parameters (Block-AP) and end-to-end training of quantization parameters (E2E-QP).  To the best of our knowledge, Block-AP is the first method to enable direct training of all parameters in a block-wise manner, reducing accuracy loss in low-bit scenarios by enhancing the solution space during optimization. E2E-QP then trains only the quantization parameters (step sizes) end-to-end, further improving the performance of quantized models by considering interactions among all sub-modules. Extensive experiments demonstrate that EfficientQAT outperforms previous quantization methods across a range of models, including base LLMs, instruction-tuned LLMs, and multimodal LLMs, with scales from 7B to 70B parameters at various quantization bits. For instance, EfficientQAT obtains a 2-bit Llama-2-70B model on a single A100-80GB GPU in 41 hours, with less than 3 points accuracy degradation compared to the full precision (69.48 vs. 72.41).", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=6Mdvq0bPyG", "arxiv_id": "2407.11062", "paper_pdf": "papers/6Mdvq0bPyG.pdf", "paper_pdf_sha256": "1f67b01ee3663bc406d7a86a5c8a9eba91015f6840f98896da17d75c30a632ab", "paper_pdf_bytes": 470369, "paper_pdf_source": "openreview", "code_url": "https://github.com/OpenGVLab/EfficientQAT", "code_repository": "OpenGVLab/EfficientQAT", "code_commit": "39f37f3b6053681c9b1cd4c9dcaf692d8999459e", "code_archive": "repos/6Mdvq0bPyG.zip", "code_archive_sha256": "0ddb65323fff1409478fbbb077c703aa890f5cf11c39aef5130457d4dd37a0ed", "code_archive_bytes": 68414, "code_file_count": 43, "code_extensions": {".sh": 23, ".py": 20}, "github_disk_usage_kb": 99, "github_languages": {"Python": 158217}, "github_archived": false, "github_pushed_at": "2026-04-10T01:13:35Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/efficientqat-efficient-quantization-aware"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "hP4iZU8I3Y", "year": 2024, "status": "rejected", "title": "Understanding Inter-Session Intentions via Complex Logical Reasoning", "authors": ["Jiaxin Bai", "Chen Luo", "Zheng Li", "Qingyu Yin", "Yangqiu Song"], "authorids": ["~Jiaxin_Bai1", "~Chen_Luo3", "~Zheng_Li9", "~Qingyu_Yin2", "~Yangqiu_Song1"], "authors_source": "OpenReview API", "abstract": "Understanding user intentions is crucial for enhancing product recommendations, navigation suggestions, and query reformulations. However, user intentions can be intricate, involving multiple sessions and attribute requirements connected by logical operators such as And, Or, and Not. For example, a user may search for Nike or Adidas running shoes across various sessions, with a preference for the color purple. In another case, a user may have purchased a mattress in a previous session and is now seeking a corresponding bed frame without intending to buy another mattress. Prior research on session understanding has not sufficiently addressed how to make product or attribute recommendations for such complex intentions.\nIn this paper, we introduce the task of logical session query answering (LSQA), where sessions are treated as hyperedges of items. We formulate the problem of complex intention understanding as a task of answering logical queries on an aggregated hypergraph of sessions, items, and attributes. We also propose a new model, the Logical Session Graph Transformer (LSGT), which captures interactions among items across different sessions and their logical connections using a transformer structure.\nWe analyze the expressiveness of LSGT and prove the permutation invariance of the inputs for the logical operators. We evaluate LSGT on three datasets and demonstrate that it achieves state-of-the-art results.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=hP4iZU8I3Y", "arxiv_id": "2312.13866", "paper_pdf": "papers/hP4iZU8I3Y.pdf", "paper_pdf_sha256": "fa4928c34accc4011efe1c391e6785010ea7fa9a0182225d02b237fbb8be8838", "paper_pdf_bytes": 567896, "paper_pdf_source": "openreview", "code_url": "https://github.com/HKUST-KnowComp/SessionCQA", "code_repository": "HKUST-KnowComp/SessionCQA", "code_commit": "12241090b612f696e40c4c06583473ae18ea74a0", "code_archive": "repos/hP4iZU8I3Y.zip", "code_archive_sha256": "7ae3c5ed1405ded4ae98de99737d89c2297e837ab46b078d3607c9335fe273cd", "code_archive_bytes": 107006, "code_file_count": 52, "code_extensions": {".py": 28, ".sh": 24}, "github_disk_usage_kb": 78, "github_languages": {"Python": 509055, "Shell": 8435}, "github_archived": false, "github_pushed_at": "2024-07-26T08:42:16Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/understanding-inter-session-intentions-via"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "sKDtBKYOdIP", "year": 2023, "status": "rejected", "title": "Beam Tree Recursive Cells", "authors": ["Jishnu Ray Chowdhury", "Cornelia Caragea"], "authorids": ["~Jishnu_Ray_Chowdhury2", "~Cornelia_Caragea2"], "authors_source": "OpenReview API", "abstract": "Recursive Neural Networks (RvNNs) generalize Recurrent Neural Networks (RNNs) by allowing sequential composition in a more flexible order, typically, based on some tree structure. While initially user-annotated tree structures were used, in due time, several approaches were proposed to automatically induce tree-structures from raw text to guide the recursive compositions in RvNNs. In this paper, we present an approach called Beam Tree Recursive Cell (or BT-Cell) based on a simple yet overlooked backpropagation-friendly framework. BT-Cell applies beam search on easy-first parsing for simulating RvNNs with automatic structure-induction. Our results show that BT-Cell achieves near-perfect performance on several aspects of challenging structure-sensitive synthetic tasks like ListOps and also comparable performance in realistic data to other RvNN-based models. We further introduce and analyze several extensions of BT-Cell based on relaxations of the hard top-k operators in beam search. We evaluate the models in different out of distribution splits in both synthetic and realistic data. Additionally, we identify a previously unknown failure case for neural models in generalization to unseen number of arguments in ListOps. We will release our code.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=sKDtBKYOdIP", "arxiv_id": "2305.19999", "paper_pdf": "papers/sKDtBKYOdIP.pdf", "paper_pdf_sha256": "dd7127b71e71324ac0d186b152cc7dd0e4f45b8cf83e123a324a41d5823f2350", "paper_pdf_bytes": 508007, "paper_pdf_source": "openreview", "code_url": "https://github.com/JRC1995/BeamTreeRecursiveCells", "code_repository": "JRC1995/BeamTreeRecursiveCells", "code_commit": "30cbaae15635372249c9480c5553be5f654cf63f", "code_archive": "repos/sKDtBKYOdIP.zip", "code_archive_sha256": "38ecc30322d83a36c2585410d6bd6523e6c493bbdede088fc63ff792d37db9e0", "code_archive_bytes": 208896, "code_file_count": 136, "code_extensions": {".py": 135, ".cu": 1}, "github_disk_usage_kb": 154, "github_languages": {"Python": 639287, "Cuda": 5680}, "github_archived": false, "github_pushed_at": "2024-01-28T04:02:36Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/beam-tree-recursive-cells"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "7zFokR7k_86", "year": 2022, "status": "rejected", "title": "Learning Symbolic Rules for Reasoning in Quasi-Natural Language", "authors": ["Kaiyu Yang", "Jia Deng"], "authorids": ["~Kaiyu_Yang1", "~Jia_Deng1"], "authors_source": "OpenReview API", "abstract": "Symbolic reasoning, rule-based symbol manipulation, is a hallmark of human intelligence.  However, rule-based systems have had limited success competing with learning-based systems outside formalized domains such as automated theorem proving. We hypothesize that this is due to the manual construction of rules in past attempts. In this work, we ask how we can build a rule-based system that can reason with natural language input but without the manual construction of rules. We propose MetaQNL, a \"Quasi-Natural\" language that can express both formal logic and natural language sentences, and MetaInduce, a learning algorithm that induces MetaQNL rules from training data consisting of questions and answers, with or without intermediate reasoning steps. Our approach achieves state-of-the-art accuracy on multiple reasoning benchmarks; it learns compact models with much less data and produces not only answers but also checkable proofs.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=7zFokR7k_86", "arxiv_id": "2111.12038", "paper_pdf": "papers/7zFokR7k_86.pdf", "paper_pdf_sha256": "500460cedae7fa9db67a8506111a50c3fef5b9970a1f2e13fcd82c4929aaeab8", "paper_pdf_bytes": 345948, "paper_pdf_source": "openreview", "code_url": "https://github.com/princeton-vl/MetaQNL", "code_repository": "princeton-vl/MetaQNL", "code_commit": "dcff310641c8f999eb5354584f32660b98fd9ec9", "code_archive": "repos/7zFokR7k_86.zip", "code_archive_sha256": "7385df639bc5ccda91786cedbfe1e7f5f57cfb678ef1e68dbfd5a95ddac6ea75", "code_archive_bytes": 247531, "code_file_count": 69, "code_extensions": {".jl": 69}, "github_disk_usage_kb": 241, "github_languages": {"Julia": 256078}, "github_archived": false, "github_pushed_at": "2024-08-07T00:04:28Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/learning-symbolic-rules-for-reasoning-in-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "oj3bHNSq_2w", "year": 2021, "status": "rejected", "title": "Sample weighting as an explanation for mode collapse in generative adversarial networks", "authors": ["Aksel Wilhelm Wold Eide", "Eilif Solberg", "Ingebjørg Kåsen"], "authorids": ["~Aksel_Wilhelm_Wold_Eide1", "eilif.solberg@ffi.no", "ingebjorg.kasen@ffi.no"], "authors_source": "OpenReview API", "abstract": "Generative adversarial networks were introduced with a logistic MiniMax cost formulation, which normally fails to train due to saturation, and a Non-Saturating reformulation. While addressing the saturation problem, NS-GAN also inverts the generator's sample weighting, implicitly shifting emphasis from higher-scoring to lower-scoring samples when updating parameters. We present both theory and empirical results suggesting that this makes NS-GAN prone to mode dropping. We design MM-nsat, which preserves MM-GAN sample weighting while avoiding saturation by rescaling the MM-GAN minibatch gradient such that its magnitude approximates NS-GAN's gradient magnitude. MM-nsat has qualitatively different training dynamics, and on MNIST and CIFAR-10 it is stronger in terms of mode coverage, stability and FID. While the empirical results for MM-nsat are promising and favorable also in comparison with the LS-GAN and Hinge-GAN formulations, our main contribution is to show how and why NS-GAN's sample weighting causes mode dropping and training collapse.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=oj3bHNSq_2w", "arxiv_id": "2010.02035", "paper_pdf": "papers/oj3bHNSq_2w.pdf", "paper_pdf_sha256": "4c0d8b9973be3b23e1593f46aaf8642a8d678c4208ca12198915b8f746e48124", "paper_pdf_bytes": 20722498, "paper_pdf_source": "openreview", "code_url": "https://github.com/awweide/mm-nsat", "code_repository": "awweide/mm-nsat", "code_commit": "799b054e059f32eb0b3a3831f20c9648a2e0c282", "code_archive": "repos/oj3bHNSq_2w.zip", "code_archive_sha256": "15a2002e7bc2a5a4d27eaa7b2d048369e79ec2d15da07ea6ff587c8893fded0f", "code_archive_bytes": 272955, "code_file_count": 13, "code_extensions": {".py": 12, ".sh": 1}, "github_disk_usage_kb": 277, "github_languages": {"Python": 86592, "Shell": 7138}, "github_archived": false, "github_pushed_at": "2021-01-01T04:07:21Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/sample-weighting-as-an-explanation-for-mode-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "HJx4PAEYDH", "year": 2020, "status": "rejected", "title": "R-TRANSFORMER: RECURRENT NEURAL NETWORK ENHANCED TRANSFORMER", "authors": ["Zhiwei Wang", "Yao Ma", "Zitao Liu", "Jiliang Tang"], "authorids": ["wangzh65@msu.edu", "mayao4@msu.edu", "liuzitao@100tal.com", "tangjili@msu.edu"], "authors_source": "OpenReview API", "abstract": "Recurrent Neural Networks have long been the dominating choice for sequence modeling. However, it severely suffers from two issues: impotent in capturing very long-term dependencies and unable to parallelize the sequential computation procedure. Therefore, many non-recurrent sequence models that are built on convolution and attention operations have been proposed recently. Notably, models with multi-head attention such as Transformer have demonstrated extreme effectiveness in capturing long-term dependencies in a variety of sequence modeling tasks. Despite their success, however, these models lack necessary components to model local structures in sequences and heavily rely on position embeddings that have limited effects and require a considerable amount of design efforts. In this paper, we propose the R-Transformer which enjoys the advantages of both RNNs and the multi-head attention mechanism while avoids their respective drawbacks. The proposed model can effectively capture both local structures and global long-term dependencies in sequences without any use of position embeddings. We evaluate R-Transformer through extensive experiments with data from a wide range of domains and the empirical results show that R-Transformer outperforms the state-of-the-art methods by a large margin in most of the tasks.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=HJx4PAEYDH", "arxiv_id": "1907.05572", "paper_pdf": "papers/HJx4PAEYDH.pdf", "paper_pdf_sha256": "62e5856240e4211c528462d9f8152354ca0f7c6b44dc605c6625ffb0f1498c7e", "paper_pdf_bytes": 211103, "paper_pdf_source": "openreview", "code_url": "https://github.com/DSE-MSU/R-transformer", "code_repository": "DSE-MSU/R-transformer", "code_commit": "3103cdfd117c2d82d64fb5ca56a6399bd30684e4", "code_archive": "repos/HJx4PAEYDH.zip", "code_archive_sha256": "75a1d0283436607620008862a43f44a45dbca6be189a3676e371f07de0e1f9cb", "code_archive_bytes": 362600, "code_file_count": 13, "code_extensions": {".py": 13}, "github_disk_usage_kb": 347, "github_languages": {"Python": 42839}, "github_archived": false, "github_pushed_at": "2019-07-16T03:31:38Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/r-transformer-recurrent-neural-network"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "XUJcsLvpaQ", "year": 2025, "status": "rejected", "title": "G-Transformer for Conditional Average Potential Outcome Estimation over Time", "authors": ["Konstantin Hess", "Dennis Frauen", "Valentyn Melnychuk", "Stefan Feuerriegel"], "authorids": ["~Konstantin_Hess1", "~Dennis_Frauen1", "~Valentyn_Melnychuk1", "~Stefan_Feuerriegel1"], "authors_source": "OpenReview API", "abstract": "Estimating potential outcomes for treatments over time based on observational data is important for personalized decision-making in medicine. Yet, existing neural methods for this task either (1) do not perform proper adjustments for time-varying confounders, or (2) suffer from large estimation variance. In order to address both limitations, we introduce the G-transformer (GT). Our GT is a novel, neural end-to-end model which adjusts for time-varying confounders, and provides low-variance estimation of conditional average potential outcomes (CAPOs) over time. Specifically, our GT is the first neural model to perform regression-based iterative G-computation for CAPOs in the time-varying setting. We evaluate the effectiveness of our GT across various experiments. In sum, this work represents a significant step towards personalized decision-making from electronic health records.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=XUJcsLvpaQ", "arxiv_id": "2405.21012", "paper_pdf": "papers/XUJcsLvpaQ.pdf", "paper_pdf_sha256": "1596fd85f92d17081e408797830ecff564c3cc1521b128c5e23aebf9e519c7f0", "paper_pdf_bytes": 954115, "paper_pdf_source": "openreview", "code_url": "https://github.com/konstantinhess/G_transformer", "code_repository": "konstantinhess/G_transformer", "code_commit": "8452c89ef7da02f0793ca295d4e30455f5d77ee7", "code_archive": "repos/XUJcsLvpaQ.zip", "code_archive_sha256": "8a60b0bb112a220bd7b11837e07c575a6bc825744c6496ac93fa2f58514822a4", "code_archive_bytes": 136810, "code_file_count": 37, "code_extensions": {".py": 35, ".ipynb": 2}, "github_disk_usage_kb": 105, "github_languages": {"Python": 423591, "Jupyter Notebook": 33705}, "github_archived": false, "github_pushed_at": "2024-09-26T09:57:34Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/g-transformer-for-conditional-average"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "rFZtyj5kBz", "year": 2024, "status": "rejected", "title": "Certifiably Byzantine-Robust Federated Conformal Prediction", "authors": ["Mintong Kang", "Zhen Lin", "Jimeng Sun", "Cao Xiao", "Bo Li"], "authorids": ["~Mintong_Kang1", "~Zhen_Lin2", "~Jimeng_Sun3", "~Cao_Xiao2", "~Bo_Li19"], "authors_source": "OpenReview API", "abstract": "Conformal prediction has shown impressive capacity in constructing statistically rigorous prediction sets for machine learning models with exchangeable data samples. The burgeoning amount of large-scale data, coupled with the escalating privacy concerns related to local data sharing, has inspired recent innovations extending conformal prediction into federated environments with distributed data samples. However, this framework for distributed uncertainty quantification is susceptible to Byzantine failures. A minor subset of malicious clients can significantly compromise the practicality of coverage guarantees. To address this vulnerability, we introduce a novel algorithm Rob-FCP to execute robust federated conformal prediction, effectively countering malicious clients capable of reporting arbitrary statistics during the conformal calibration process. We theoretically provide the conformal coverage bound of Rob-FCP in the Byzantine setting and show that the coverage of Rob-FCP is asymptotically close to the desired coverage level under mild conditions in both IID and non-IID settings. We also propose a malicious client number estimator to tackle a more challenging setting where the number of malicious clients is unknown to the defender and theoretically show its effectiveness. We empirically demonstrate the robustness of Rob-FCP against diverse proportions of malicious clients under a variety of Byzantine attacks on five realistic benchmark and healthcare datasets.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=rFZtyj5kBz", "arxiv_id": "2406.01960", "paper_pdf": "papers/rFZtyj5kBz.pdf", "paper_pdf_sha256": "203e6269102add8a53b65741a4fc48517527c8ea7f67bcd84108c63cf7d32c4c", "paper_pdf_bytes": 1018447, "paper_pdf_source": "openreview", "code_url": "https://github.com/kangmintong/Rob-FCP", "code_repository": "kangmintong/Rob-FCP", "code_commit": "2475c51ed9c610eb81a0396d4283317108410d67", "code_archive": "repos/rFZtyj5kBz.zip", "code_archive_sha256": "3bae5ca9179b48e65d43e9960f51fe5f032210dd0ce03fbd0dab4f45e5a82c88", "code_archive_bytes": 90984, "code_file_count": 22, "code_extensions": {".py": 18, ".sh": 4}, "github_disk_usage_kb": 79, "github_languages": {"Python": 115303, "Shell": 10448}, "github_archived": false, "github_pushed_at": "2024-10-24T14:22:18Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/certifiably-byzantine-robust-federated"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "9Zx6tTcX0SE", "year": 2023, "status": "rejected", "title": "A Study of Biologically Plausible Neural Network: the Role and Interactions of Brain-Inspired Mechanisms in Continual Learning", "authors": ["Fahad Sarfraz", "Elahe Arani", "Bahram Zonooz"], "authorids": ["~Fahad_Sarfraz1", "~Elahe_Arani1", "~Bahram_Zonooz1"], "authors_source": "OpenReview API", "abstract": "Humans excel at continually acquiring, consolidating, and retaining information from an ever-changing environment, whereas artificial neural networks (ANNs) exhibit catastrophic forgetting. There are considerable differences in the complexity of synapses, the processing of information, and the learning mechanisms in biological neural networks and their artificial counterpart, which may explain the mismatch in performance. We consider a biologically plausible framework that constitutes separate populations of exclusively excitatory and inhibitory neurons which adhere to Dale's principle and the excitatory pyramidal neurons are augmented with dendritic-like structures for context-dependent processing of stimuli. We then conduct a comprehensive study on the role and interactions of different mechanisms inspired by the brain including sparse non-overlapping representations, Hebbian learning, synaptic consolidation, and replay of past activations that accompanied the learning event. Our study suggests that employing multiple complementary mechanisms in a biologically plausible architecture, similar to the brain, can be effective in enabling continual learning in ANNs.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=9Zx6tTcX0SE", "arxiv_id": "2304.06738", "paper_pdf": "papers/9Zx6tTcX0SE.pdf", "paper_pdf_sha256": "924b958ff149365651e83e4e5a9a67f1d0a2eb04b7582e700b13f2fca79d4507", "paper_pdf_bytes": 407538, "paper_pdf_source": "openreview", "code_url": "https://github.com/NeurAI-Lab/Bio-ANN", "code_repository": "NeurAI-Lab/Bio-ANN", "code_commit": "267d48d0651fafbff0acabee7afd4ccbacfc0e65", "code_archive": "repos/9Zx6tTcX0SE.zip", "code_archive_sha256": "779fef27fe0dc103b01964e7993fc9e8b48180f92bd9ca83e858e01e6d3672af", "code_archive_bytes": 227291, "code_file_count": 109, "code_extensions": {".py": 109}, "github_disk_usage_kb": 155, "github_languages": {"Python": 671006}, "github_archived": false, "github_pushed_at": "2023-10-11T14:06:17Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/a-study-of-biologically-plausible-neural"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Xg47v73CDaj", "year": 2022, "status": "rejected", "title": "Non-deep Networks", "authors": ["Ankit Goyal", "Alexey Bochkovskiy", "Jia Deng", "Vladlen Koltun"], "authorids": ["~Ankit_Goyal1", "alexeyab84@gmail.com", "~Jia_Deng1", "~Vladlen_Koltun1"], "authors_source": "OpenReview API", "abstract": "Depth is the hallmark of deep neural networks. But more depth means more sequential computation and higher latency. This begs the question -- is it possible to build high-performing ``non-deep\" neural networks? We show that it is. To do so, we use parallel subnetworks instead of stacking one layer after another. This helps effectively reduce depth while maintaining high performance. By utilizing parallel substructures, we show, for the first time, that a network with a depth of just 12 can achieve top-1 accuracy over 80% on ImageNet, 96% on CIFAR10, and 81% on CIFAR100. We also show that a network with a low-depth (12) backbone can achieve an AP of 48% on MS-COCO. We analyze the scaling rules for our design and show how to increase performance without changing the network's depth. Finally, we provide a proof of concept for how non-deep networks could be used to build low-latency recognition systems. We will open-source our code.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=Xg47v73CDaj", "arxiv_id": "2110.07641", "paper_pdf": "papers/Xg47v73CDaj.pdf", "paper_pdf_sha256": "3b7fe8bfe02cbf865c5c773cc27ee99295f592355283dc55950c4677c5e08e74", "paper_pdf_bytes": 312702, "paper_pdf_source": "openreview", "code_url": "https://github.com/imankgoyal/NonDeepNetworks", "code_repository": "imankgoyal/NonDeepNetworks", "code_commit": "777b286c2d6045625b1d37099241034985a1404e", "code_archive": "repos/Xg47v73CDaj.zip", "code_archive_sha256": "c2038fcd347ae8fa81d17811feb364adbaa272827d7d1e0be05985bbf7638574", "code_archive_bytes": 280351, "code_file_count": 116, "code_extensions": {".py": 111, ".sh": 5}, "github_disk_usage_kb": 258, "github_languages": {"Python": 548385, "Shell": 9236}, "github_archived": false, "github_pushed_at": "2022-10-14T16:58:14Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/non-deep-networks-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "TmkN9JmDJx1", "year": 2021, "status": "rejected", "title": "Thinking Like Transformers", "authors": ["Gail Weiss", "Yoav Goldberg", "Eran Yahav"], "authorids": ["~Gail_Weiss1", "~Yoav_Goldberg1", "~Eran_Yahav1"], "authors_source": "OpenReview API", "abstract": "What is the computational model behind a transformer? Where recurrent neural networks have direct parallels in finite state machines, allowing clear discussion and thought around architecture variants or trained models, transformers have no such familiar parallel. In this paper we aim to change that, proposing a computational model for the transformer-encoder in the form of a programming language. We map the basic components of a transformer-encoder – attention and feed-forward computation – into the simple primitives of select, aggregate, and zipmap, around which we form a programming language: the Restricted Access Sequence Process-ing Language (RASP). We show how RASP can be used to program solutions to tasks that could conceivably be learned by a transformer, augmenting it with tools we discover in our work. In particular, we provide RASP programs for histograms, sorting, and even logical inference similar to that of Clark et al. (2020). We further use our model to relate their difficulty in terms of the number of required layers and attention heads. Finally, we see how insights gained from our abstraction might be used to explain phenomena seen in recent works.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=TmkN9JmDJx1", "arxiv_id": "2106.06981", "paper_pdf": "papers/TmkN9JmDJx1.pdf", "paper_pdf_sha256": "01970681c267866a0ad467ee2249d20f959d6a5cef03176b5126c13ce9d50ab1", "paper_pdf_bytes": 910713, "paper_pdf_source": "openreview", "code_url": "https://github.com/tech-srl/RASP", "code_repository": "tech-srl/RASP", "code_commit": "61dde62577589b291db39d192de7d481ce97fdae", "code_archive": "repos/TmkN9JmDJx1.zip", "code_archive_sha256": "1a674b57063e6b08ac4d7de1a1c865b97fc74bd5dda685c2d7b60c441f986242", "code_archive_bytes": 238630, "code_file_count": 20, "code_extensions": {".py": 18, ".sh": 2}, "github_disk_usage_kb": 277, "github_languages": {"Python": 279124, "ANTLR": 3568, "Shell": 2028, "Vim Script": 1774}, "github_archived": false, "github_pushed_at": "2024-09-16T12:36:47Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/thinking-like-transformers-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "SJlPOCEKvH", "year": 2020, "status": "rejected", "title": "Compressing BERT: Studying the Effects of Weight Pruning on Transfer Learning", "authors": ["Mitchell A Gordon", "Kevin Duh", "Nicholas Andrews"], "authorids": ["mgordo37@jhu.edu", "kevinduh@cs.jhu.edu", "noa@jhu.edu"], "authors_source": "OpenReview API", "abstract": "Universal feature extractors, such as BERT for natural language processing and VGG for computer vision, have become effective methods for improving deep learning models without requiring more labeled data. A common paradigm is to pre-train a feature extractor on large amounts of data then fine-tune it as part of a deep learning model on some downstream task (i.e. transfer learning). While effective, feature extractors like BERT may be prohibitively large for some deployment scenarios. We explore weight pruning for BERT and ask: how does compression during pre-training affect transfer learning? We find that pruning affects transfer learning in three broad regimes. Low levels of pruning (30-40%) do not affect pre-training loss or transfer to downstream tasks at all. Medium levels of pruning increase the pre-training loss and prevent useful pre-training information from being transferred to downstream tasks. High levels of pruning additionally prevent models from fitting downstream datasets, leading to further degradation. Finally, we observe that fine-tuning BERT on a specific task does not improve its prunability. We conclude that BERT can be pruned once during pre-training rather than separately for each task without affecting performance.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=SJlPOCEKvH", "arxiv_id": "2002.08307", "paper_pdf": "papers/SJlPOCEKvH.pdf", "paper_pdf_sha256": "174bfe6a31d92af78db468a00480ad0ac4cbe674b0c09948e033f2a36ae7e14a", "paper_pdf_bytes": 4162391, "paper_pdf_source": "openreview", "code_url": "https://github.com/mitchellgordon95/bert-prune", "code_repository": "mitchellgordon95/bert-prune", "code_commit": "2754ae54d66a6d8d3fa33b3d64811256b21debfa", "code_archive": "repos/SJlPOCEKvH.zip", "code_archive_sha256": "e71072dd081c45c1aa373160738d250c5aa4bcd9873a5378e008d2d8ffd8f884", "code_archive_bytes": 198422, "code_file_count": 51, "code_extensions": {".py": 48, ".sh": 2, ".ipynb": 1}, "github_disk_usage_kb": 387, "github_languages": {"Python": 301882, "Jupyter Notebook": 66488, "Shell": 601}, "github_archived": false, "github_pushed_at": "2020-05-14T20:11:19Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/compressing-bert-studying-the-effects-of-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Asvt8BAaZT", "year": 2026, "status": "rejected", "title": "When Can You Get Away with Low Memory Adam?", "authors": ["Dayal Singh Kalra", "John Kirchenbauer", "Maissam Barkeshli", "Tom Goldstein"], "authorids": ["~Dayal_Singh_Kalra1", "~John_Kirchenbauer1", "~Maissam_Barkeshli1", "~Tom_Goldstein1"], "authors_source": "OpenReview API", "abstract": "Adam is the go-to optimizer for training modern machine learning models, but it requires additional memory to maintain the moving averages of the gradients and their squares. While various low-memory optimizers have been proposed that sometimes match the performance of Adam, their lack of reliability has left Adam as the default choice. In this work, we apply a simple layer-wise Signal-to-Noise Ratio (SNR) analysis to quantify when second-moment tensors can be effectively replaced by their means across different dimensions. Our SNR analysis reveals how architecture, training hyperparameters, and dataset properties impact compressibility along Adam's trajectory, naturally leading to \\emph{SlimAdam}, a memory-efficient Adam variant. \\emph{SlimAdam} compresses the second moments along dimensions with high SNR when feasible, and leaves when compression would be detrimental. Through experiments across a diverse set of architectures and training scenarios, we show that \\emph{SlimAdam} matches Adam's performance and stability while saving up to 98% of total second moments.", "decision": "Reject", "meta_review": "", "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=Asvt8BAaZT", "arxiv_id": "2503.01843", "paper_pdf": "papers/Asvt8BAaZT.pdf", "paper_pdf_sha256": "4bbbea433712864b39cd233b2054fc7ceee2e4d9e8faa5ba918307a4e66dbcb4", "paper_pdf_bytes": 3040815, "paper_pdf_source": "openreview", "code_url": "https://github.com/dayal-kalra/low-memory-adam", "code_repository": "dayal-kalra/low-memory-adam", "code_commit": "bb3652c2fd55eb3e975658ab844c14485f73f693", "code_archive": "repos/Asvt8BAaZT.zip", "code_archive_sha256": "d09081a6cc32722ebc12cde837431040c0a464c3669185e38f35cdf5ad4825ab", "code_archive_bytes": 1291781, "code_file_count": 9, "code_extensions": {".py": 9}, "github_disk_usage_kb": 1575, "github_languages": {"Python": 81530}, "github_archived": false, "github_pushed_at": "2025-03-02T22:27:52Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/when-can-you-get-away-with-low-memory-adam"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "iKgQOAtvsD", "year": 2025, "status": "rejected", "title": "Deciphering the Chaos: Enhancing Jailbreak Attacks via Adversarial Prompt Translation", "authors": ["Qizhang Li", "Xiaochen Yang", "Wangmeng Zuo", "Yiwen Guo"], "authorids": ["~Qizhang_Li1", "~Xiaochen_Yang2", "~Wangmeng_Zuo2", "~Yiwen_Guo1"], "authors_source": "OpenReview API", "abstract": "Automatic adversarial prompt generation provides remarkable success in jailbreaking safely-aligned large language models (LLMs). Existing gradient-based attacks, while demonstrating outstanding performance in jailbreaking white-box LLMs, often generate garbled adversarial prompts with chaotic appearance. These adversarial prompts are difficult to transfer to other LLMs, hindering their performance in attacking unknown victim models. In this paper, for the first time, we delve into the semantic meaning embedded in garbled adversarial prompts and propose a novel method that *\"translate\"* them into coherent, human-readable natural language adversarial prompts. In this way, we can effectively uncover the semantic information that triggers vulnerabilities in the model and unambiguously transfer it to the victim model, without overlooking the adversarial information hidden in the garbled text, to enhance jailbreak attacks. It also offers a new approach to discovering effective designs for jailbreak prompts, advancing the understanding of jailbreak attacks. Experimental results demonstrate that our method significantly improves the success rate of jailbreak attacks against various safety-aligned LLMs and outperforms state-of-the-arts by a large margin. With at most 10 queries, our method achieves an average attack success rate of 81.8% in attacking 7 commercial closed-source LLMs, including GPT and Claude-3 series, on HarmBench. Our method also achieves over 90% attack success rates against Llama-2-Chat models on AdvBench, despite their outstanding resistance to jailbreaks. Our code will be made publicly available.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=iKgQOAtvsD", "arxiv_id": "2410.11317", "paper_pdf": "papers/iKgQOAtvsD.pdf", "paper_pdf_sha256": "b2a1a96470b61dd6f4c81a13a89410bbd8c49ed44c84a17dcff57f2aa9825ea9", "paper_pdf_bytes": 4052233, "paper_pdf_source": "openreview", "code_url": "https://github.com/qizhangli/Adversarial-Prompt-Translator", "code_repository": "qizhangli/Adversarial-Prompt-Translator", "code_commit": "195111673b09c77609b7f3ef5b9cea4daecc7c74", "code_archive": "repos/iKgQOAtvsD.zip", "code_archive_sha256": "d0012f69438c0423f369fe00965840a9ece39dc95450825fe41e7bef19f9a7d0", "code_archive_bytes": 107289, "code_file_count": 9, "code_extensions": {".py": 7, ".sh": 2}, "github_disk_usage_kb": 107, "github_languages": {"Python": 20896, "Shell": 216}, "github_archived": false, "github_pushed_at": "2024-11-07T15:16:54Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/deciphering-the-chaos-enhancing-jailbreak"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "wmzFZ9lJrD", "year": 2024, "status": "rejected", "title": "Boolformer: Symbolic Regression of Logic Functions with Transformers", "authors": ["Stéphane d'Ascoli", "Samy Bengio", "Joshua M. Susskind", "Emmanuel Abbe"], "authorids": ["~Stéphane_d'Ascoli1", "~Samy_Bengio1", "~Joshua_M._Susskind1", "~Emmanuel_Abbe1"], "authors_source": "OpenReview API", "abstract": "In this work, we introduce the Boolformer, the first Transformer architecture trained to perform end-to-end symbolic regression of Boolean functions. First, we show that it can predict compact formulas for complex functions which were not seen during training, when provided a clean truth table. Then, we demonstrate its ability to find approximate expressions when provided incomplete and noisy observations. We compare it with classic machine learning approaches on a broad set of real-world binary classification datasets, demonstrating its potential as an interpretable alternative. Finally, we apply it to the widespread task of modelling the dynamics of gene regulatory networks. Using a recent benchmark, we show that Boolformer is competitive with state-of-the art genetic algorithms with a speedup of several orders of magnitude.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=wmzFZ9lJrD", "arxiv_id": "2309.12207", "paper_pdf": "papers/wmzFZ9lJrD.pdf", "paper_pdf_sha256": "22c4c2ae011a1c33d46ae119e2a72eb4dc7f9d54f5ca9949b6ee788b776f8bf7", "paper_pdf_bytes": 993583, "paper_pdf_source": "openreview", "code_url": "https://github.com/sdascoli/boolformer", "code_repository": "sdascoli/boolformer", "code_commit": "b6694ba5eda0abb6484176d309937ff72c370fb9", "code_archive": "repos/wmzFZ9lJrD.zip", "code_archive_sha256": "fd4d441b560e5a5a2b45de5a0eb9804f97ebde27750eaa0a5748cfc60bee1cde", "code_archive_bytes": 90431, "code_file_count": 27, "code_extensions": {".py": 26, ".ipynb": 1}, "github_disk_usage_kb": 85, "github_languages": {"Python": 282780, "Jupyter Notebook": 54272}, "github_archived": false, "github_pushed_at": "2024-03-25T09:54:02Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/boolformer-symbolic-regression-of-logic"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "wV09GfqYC-n", "year": 2023, "status": "rejected", "title": "Hierarchies of Reward Machines", "authors": ["Daniel Furelos-Blanco", "Mark Law", "Anders Jonsson", "Krysia Broda", "Alessandra Russo"], "authorids": ["~Daniel_Furelos-Blanco1", "~Mark_Law1", "~Anders_Jonsson1", "~Krysia_Broda1", "~Alessandra_Russo1"], "authors_source": "OpenReview API", "abstract": "Reward machines (RMs) are a recent formalism for representing the reward function of a reinforcement learning task through a finite-state machine whose edges encode landmarks of the task using high-level events. The structure of RMs enables the decomposition of a task into simpler and independently solvable subtasks that help tackle long-horizon and/or sparse reward tasks. We propose a formalism for further abstracting the subtask structure by endowing an RM with the ability to call other RMs, thus composing a hierarchy of RMs (HRM). We exploit HRMs by treating each call to an RM as an independently solvable subtask using the options framework, and describe a curriculum-based method to learn HRMs from traces observed by the agent. Our experiments reveal that exploiting a handcrafted HRM leads to faster convergence than with a flat HRM, and that learning an HRM remains feasible in cases where its equivalent flat representation is not.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=wV09GfqYC-n", "arxiv_id": "2205.15752", "paper_pdf": "papers/wV09GfqYC-n.pdf", "paper_pdf_sha256": "be58b64292249a699825a03d77d3210862712a9a5c32555336a5aa322ffe5717", "paper_pdf_bytes": 1794830, "paper_pdf_source": "openreview", "code_url": "https://github.com/ertsiger/hrm-learning", "code_repository": "ertsiger/hrm-learning", "code_commit": "23d0e25a367e1e8a559fa815d86696f0183ade43", "code_archive": "repos/wV09GfqYC-n.zip", "code_archive_sha256": "ef38c08fa0db18cad6f74d482dd09d0376eff55b4af29e7645cc41f5240f337a", "code_archive_bytes": 201897, "code_file_count": 84, "code_extensions": {".py": 77, ".sh": 6, ".rb": 1}, "github_disk_usage_kb": 158, "github_languages": {"Python": 579478, "Shell": 1255, "Ruby": 202}, "github_archived": false, "github_pushed_at": "2023-07-11T13:25:36Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/hierarchies-of-reward-machines"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "T_8wHvOkEi9", "year": 2022, "status": "rejected", "title": "Self-Organized Polynomial-time Coordination Graphs", "authors": ["Weijun Dong", "Qianlan Yang", "Zhizhou Ren", "Jianhao Wang", "Tonghan Wang", "Chongjie Zhang"], "authorids": ["~Weijun_Dong1", "~Qianlan_Yang1", "~Zhizhou_Ren1", "~Jianhao_Wang1", "~Tonghan_Wang1", "~Chongjie_Zhang1"], "authors_source": "OpenReview API", "abstract": "Coordination graph is a promising approach to model agent collaboration in multi-agent reinforcement learning. It factorizes a large multi-agent system into a suite of overlapping groups that represent the underlying coordination dependencies. One critical challenge in this paradigm is the complexity of computing maximum-value actions for a graph-based value factorization. It refers to the decentralized constraint optimization problem (DCOP), which and whose constant-ratio approximation are NP-hard problems. To bypass this fundamental hardness, this paper proposes a novel method, named Self-Organized Polynomial-time Coordination Graphs (SOP-CG), which uses structured graph classes to guarantee the optimality of the induced DCOPs with sufficient function expressiveness. We extend the graph topology to be state-dependent, formulate the graph selection as an imaginary agent, and finally derive an end-to-end learning paradigm from the unified Bellman optimality equation. In experiments, we show that our approach learns interpretable graph topologies, induces effective coordination, and improves performance across a variety of cooperative multi-agent tasks.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=T_8wHvOkEi9", "arxiv_id": "2112.03547", "paper_pdf": "papers/T_8wHvOkEi9.pdf", "paper_pdf_sha256": "b1ed82171cfecca4284a251ffc81c4070d62ecb3ec5061dc7ddd770b3af53c8b", "paper_pdf_bytes": 1481340, "paper_pdf_source": "openreview", "code_url": "https://github.com/yanQval/SOP-CG", "code_repository": "yanQval/SOP-CG", "code_commit": "8a719a11c0cf318ba5e39f803441cb99b79815b5", "code_archive": "repos/T_8wHvOkEi9.zip", "code_archive_sha256": "f006bc1fbe392a9bec534431bc32fbd17686478d2f5cca48d6a3b7a1c67809f3", "code_archive_bytes": 117928, "code_file_count": 64, "code_extensions": {".py": 58, ".sh": 4, ".cpp": 2}, "github_disk_usage_kb": 261, "github_languages": {"Python": 274940, "C++": 24804, "Shell": 1921, "Dockerfile": 1796}, "github_archived": false, "github_pushed_at": "2022-06-23T02:21:50Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/self-organized-polynomial-time-coordination-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "eyXknI5scWu", "year": 2021, "status": "rejected", "title": "Investigating and Simplifying Masking-based Saliency Methods for Model Interpretability", "authors": ["Jason Phang", "Jungkyu Park", "Krzysztof J. Geras"], "authorids": ["~Jason_Phang1", "~Jungkyu_Park1", "~Krzysztof_J._Geras1"], "authors_source": "OpenReview API", "abstract": "Saliency maps that identify the most informative regions of an image for a classifier are valuable for model interpretability. A common approach to creating saliency maps involves generating input masks that mask out portions of an image to maximally deteriorate classification performance, or mask in an image to preserve classification performance. Many variants of this approach have been proposed in the literature, such as counterfactual generation and optimizing over a Gumbel-Softmax distribution. Using a general formulation of masking-based saliency methods, we conduct an extensive evaluation study of a number of recently proposed variants to understand which elements of these methods meaningfully improve performance. Surprisingly, we find that a well-tuned, relatively simple formulation of a masking-based saliency model outperforms many more complex approaches. We find that the most important ingredients for high quality saliency map generation are (1) using both masked-in and masked-out objectives and (2) training the classifier alongside the masking model. Strikingly, we show that a masking model can be trained with as few as 10 examples per class and still generate saliency maps with only a 0.7-point increase in localization error.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=eyXknI5scWu", "arxiv_id": "2010.09750", "paper_pdf": "papers/eyXknI5scWu.pdf", "paper_pdf_sha256": "84babe42358de7a4757106491e86ae6c0cc7aa0ac6f52014b5395b88ca2f4e73", "paper_pdf_bytes": 2657504, "paper_pdf_source": "openreview", "code_url": "https://github.com/zphang/saliency_investigation", "code_repository": "zphang/saliency_investigation", "code_commit": "cdd03e92c4b8b26c43a60cb50f9d97ea2c7f9400", "code_archive": "repos/eyXknI5scWu.zip", "code_archive_sha256": "efa2002bc966d5830eaae746e5f096b162df4c28b3a9048e8774df33527093d4", "code_archive_bytes": 314668, "code_file_count": 23, "code_extensions": {".py": 23}, "github_disk_usage_kb": 300, "github_languages": {"Python": 194492}, "github_archived": false, "github_pushed_at": "2020-11-10T01:21:57Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/investigating-and-simplifying-masking-based-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "HJgS7p4FPH", "year": 2020, "status": "rejected", "title": "Accelerating Reinforcement Learning Through GPU Atari Emulation", "authors": ["Steven Dalton", "Michael Garland", "Iuri Frosio"], "authorids": ["sdalton@nvidia.com", "mgarland@nvidia.com", "ifrosio@nvidia.com"], "authors_source": "OpenReview API", "abstract": "We introduce CuLE (CUDA Learning Environment), a CUDA port of the Atari Learning Environment (ALE) which is used for the development of deep reinforcement algorithms.  CuLE overcomes many limitations of existing CPU-based emulators and scales naturally to multiple GPUs. It leverages GPU parallelization to run thousands of games simultaneously and it renders frames directly on the GPU, to avoid the bottleneck arising from the limited CPU-GPU communication bandwidth. CuLE generates up to 155M frames per hour on a single GPU, a finding previously achieved only through a cluster of CPUs. Beyond highlighting the differences between CPU and GPU emulators in the context of reinforcement learning, we show how to leverage the high throughput of CuLE by effective batching of the training data, and show accelerated convergence for A2C+V-trace. CuLE is available at [hidden URL].", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=HJgS7p4FPH", "arxiv_id": "1907.08467", "paper_pdf": "papers/HJgS7p4FPH.pdf", "paper_pdf_sha256": "8a0b050e67f648b9da67b126db5cfe3c63e762ef40fcae0a2d1baa403f306acd", "paper_pdf_bytes": 2961951, "paper_pdf_source": "openreview", "code_url": "https://github.com/NVlabs/cule", "code_repository": "NVlabs/cule", "code_commit": "dd0382b99ded6be23cd3c3e79e37938e7c873de0", "code_archive": "repos/HJgS7p4FPH.zip", "code_archive_sha256": "e4d47e43e37b366ca7ad4068342ec3da4f8e0d3511e70385ea3211edc3435490", "code_archive_bytes": 321653, "code_file_count": 155, "code_extensions": {".hpp": 117, ".py": 31, ".cpp": 6, ".cu": 1}, "github_disk_usage_kb": 418, "github_languages": {"C++": 564093, "Python": 32210, "Cuda": 8165, "Dockerfile": 1177}, "github_archived": false, "github_pushed_at": "2022-11-21T13:38:59Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/gpu-accelerated-atari-emulation-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "K7UfQFegK5", "year": 2026, "status": "rejected", "title": "FullFront: Benchmarking MLLMs Across the Full Front-End Engineering Workflow", "authors": ["Haoyu Sun", "Huichen Will Wang", "Jiawei Gu", "Linjie Li", "Yu Cheng"], "authorids": ["~Haoyu_Sun2", "~Huichen_Will_Wang1", "~Jiawei_Gu4", "~Linjie_Li1", "~Yu_Cheng1"], "authors_source": "OpenReview API", "abstract": "Front-end engineering involves a complex workflow where engineers conceptualize designs, translate them into code, and iteratively refine the implementation. While recent benchmarks primarily focus on converting visual designs to code, we present FullFront, a benchmark designed to evaluate Multimodal Large Language Models (MLLMs) \\textbf{across the full front-end development pipeline}. \nFullFront assesses three fundamental tasks that map directly to the front-end engineering pipeline: Webpage Design (conceptualization phase), Webpage Perception QA (comprehension of visual organization and elements), and Webpage Code Generation (implementation phase).\nUnlike existing benchmarks that use either scraped websites with bloated code or oversimplified LLM-generated HTML, FullFront employs a novel, two-stage process to transform real-world webpages into clean, standardized HTML while maintaining diverse visual designs and avoiding copyright issues.\nExtensive testing of state-of-the-art MLLMs reveals significant limitations in page perception, code generation (particularly for image handling and layout), and interaction implementation. Our results quantitatively demonstrate performance disparities across models and tasks, and highlight a substantial gap between current MLLM capabilities and human expert performance in front-end engineering.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=K7UfQFegK5", "arxiv_id": "2505.17399", "paper_pdf": "papers/K7UfQFegK5.pdf", "paper_pdf_sha256": "a6f661cd662ed4bd071548625f672830ecf075c4a5f0dd7d753c3db0a9c11974", "paper_pdf_bytes": 11101829, "paper_pdf_source": "openreview", "code_url": "https://github.com/Mikivishy/FullFront", "code_repository": "Mikivishy/FullFront", "code_commit": "c27e96dfe2679c286ebdd9df1b3316f5404a56f3", "code_archive": "repos/K7UfQFegK5.zip", "code_archive_sha256": "0172dcd19445ec852a0b7a690df25f80dfdcb13430df3a86f2c64b6586cc68e4", "code_archive_bytes": 1672700, "code_file_count": 32, "code_extensions": {".py": 24, ".sh": 8}, "github_disk_usage_kb": 1588, "github_languages": {"Python": 403317, "Shell": 2108}, "github_archived": false, "github_pushed_at": "2025-05-16T08:44:11Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/fullfront-benchmarking-mllms-across-the-full"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "el2pNeLrRC", "year": 2025, "status": "rejected", "title": "ECGN: A CLUSTER-AWARE APPROACH TO GRAPH NEURAL NETWORKS FOR IMBALANCED CLASSIFICATION.", "authors": ["Bishal Thapaliya", "Anh Nguyen", "Yao Lu", "Tian Xie", "Igor Grudetskyi", "Fudong Lin", "Antonios Valkanas", "Jingyu Liu", "Deepayan Chakrabarti", "Bilel Fehri"], "authorids": ["~Bishal_Thapaliya2", "~Anh_Nguyen8", "~Yao_Lu21", "~Tian_Xie13", "~Igor_Grudetskyi1", "~Fudong_Lin1", "~Antonios_Valkanas1", "~Jingyu_Liu5", "~Deepayan_Chakrabarti1", "~Bilel_Fehri1"], "authors_source": "OpenReview API", "abstract": "Classifying nodes in a graph is a common problem. The ideal classifier must\nadapt to any imbalances in the class distribution. It must also use information in\nthe clustering structure of real-world graphs. Existing Graph Neural Networks\n(GNNs) have not addressed both problems together. We propose the Enhanced\nCluster-aware Graph Network (ECGN), a novel method that addresses these is-\nsues by integrating cluster-specific training with synthetic node generation. Unlike\ntraditional GNNs that apply the same node update process for all nodes, ECGN\nlearns different aggregations for different clusters. We also use the clusters to gen-\nerate new minority-class nodes in a way that helps clarify the inter-class decision\nboundary. By combining cluster-aware embeddings with a global integration step,\nECGN enhances the quality of the resulting node embeddings. Our method works\nwith any underlying GNN and any cluster generation technique. Experimental\nresults show that ECGN consistently outperforms its closest competitors by up to\n11% on some widely-studied benchmark datasets. The GitHub implementation\nfor implementation and replication is publicly available on https://github.com/anonymous753341/ECGN.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=el2pNeLrRC", "arxiv_id": "2410.11765", "paper_pdf": "papers/el2pNeLrRC.pdf", "paper_pdf_sha256": "756397cb129dcb9f34de13a8383a126e3685ab053625e0a0dc1ffe7b7ff4a2fe", "paper_pdf_bytes": 1043339, "paper_pdf_source": "openreview", "code_url": "https://github.com/anonymous753341/ECGN", "code_repository": "anonymous753341/ECGN", "code_commit": "50fb116876969c278e0a3f48fd80ab38bb4ad725", "code_archive": "repos/el2pNeLrRC.zip", "code_archive_sha256": "c4a1a85ae67e1f00296e43e694192f18562f73b40a4c48e82282028a94855123", "code_archive_bytes": 104324, "code_file_count": 22, "code_extensions": {".py": 22}, "github_disk_usage_kb": 109, "github_languages": {"Python": 303204}, "github_archived": false, "github_pushed_at": "2024-10-02T04:15:40Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/ecgn-a-cluster-aware-approach-to-graph-neural"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "EAkjVCtRO2", "year": 2024, "status": "rejected", "title": "Variational quantization for state space models", "authors": ["Etienne David", "Jean Bellot", "Sylvain Le Corff"], "authorids": ["~Etienne_David1", "~Jean_Bellot1", "~Sylvain_Le_Corff1"], "authors_source": "OpenReview API", "abstract": "Forecasting tasks using large datasets gathering thousands of heterogeneous time series is a crucial statistical problem in numerous sectors. The main challenge is to model a rich variety of time series, leverage any available external signals and provide sharp predictions with statistical guarantees. In this work, we propose a new forecasting model  that combines discrete state space hidden Markov models with recent neural network architectures and training procedures inspired by vector quantized variational autoencoders. \nWe introduce a variational discrete posterior distribution of the latent states given the observations and a two-stage training procedure to alternatively train the parameters of the latent states and of the emission distributions. By learning a collection of emission laws and temporarily activating them depending on the hidden process dynamics, the proposed method allows to explore large datasets and leverage available external signals. We assess the performance of the proposed method using several datasets and show that it outperforms other state-of-the-art solutions.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=EAkjVCtRO2", "arxiv_id": "2404.11117", "paper_pdf": "papers/EAkjVCtRO2.pdf", "paper_pdf_sha256": "b14e25183f0170b58bf35fac8a2fd5eec456e718a9d7fc6bddfcd7e0ca1f26c6", "paper_pdf_bytes": 6245826, "paper_pdf_source": "openreview", "code_url": "https://github.com/etidav/next", "code_repository": "etidav/next", "code_commit": "7dde74c1d8875a05da714bf57a83feb28f5046f2", "code_archive": "repos/EAkjVCtRO2.zip", "code_archive_sha256": "600317e1cf2098fb329910f197c44508b06905541f5cf6f12135a7826a5275ff", "code_archive_bytes": 34231, "code_file_count": 20, "code_extensions": {".py": 11, ".sh": 9}, "github_disk_usage_kb": 86, "github_languages": {"Python": 133961, "Shell": 9465, "Dockerfile": 1204, "Makefile": 668}, "github_archived": false, "github_pushed_at": "2023-10-02T08:09:53Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/variational-quantization-for-state-space"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "En7lGmzT_x", "year": 2023, "status": "rejected", "title": "Sharper Rates and Flexible Framework for Nonconvex SGD with Client and Data Sampling", "authors": ["Alexander Tyurin", "Lukang Sun", "Konstantin Pavlovich Burlachenko", "Peter Richtárik"], "authorids": ["~Alexander_Tyurin1", "~Lukang_Sun1", "~Konstantin_Pavlovich_Burlachenko1", "~Peter_Richtárik1"], "authors_source": "OpenReview API", "abstract": "We revisit the classical problem of finding an approximately stationary point of the average of $n$ smooth and possibly nonconvex functions. The optimal complexity of stochastic first-order methods in terms of the number of gradient evaluations of individual functions is $\\mathcal{O}\\left(n + n^{1/2}\\varepsilon^{-1}\\right)$, attained by the optimal SGD methods SPIDER (Cong Fang et al., 2018) and PAGE (Zhize Li et al., 2020), for example, where $\\varepsilon$ is the error tolerance. However, i) the big-$\\mathcal{O}$ notation hides crucial dependencies on the smoothness constants associated with the functions, and ii) the rates and theory in these methods assume simplistic sampling mechanisms that do not offer any flexibility. In this work we remedy the situation. First, we generalize the PAGE algorithm so that it can provably work with virtually any (unbiased) sampling mechanism. This is particularly useful in federated learning, as it allows us to construct and better understand the impact of various combinations of client and data sampling strategies. Second, our analysis is sharper as we make explicit use of certain novel inequalities  that capture the intricate interplay between the smoothness constants and the sampling procedure. Indeed, our analysis is better even for the simple sampling procedure analyzed in the PAGE paper. However, this already improved bound can be further sharpened by a different sampling scheme which we propose. In summary, we provide the most general and most accurate analysis of optimal SGD in the smooth nonconvex regime. Finally, our theoretical findings are supposed with carefully designed experiments.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=En7lGmzT_x", "arxiv_id": "2206.02275", "paper_pdf": "papers/En7lGmzT_x.pdf", "paper_pdf_sha256": "5da1064c606c2c605df7bc00e88b10e3577c93442d822829650d009dd7b02d04", "paper_pdf_bytes": 1050259, "paper_pdf_source": "openreview", "code_url": "https://github.com/mysteryresearcher/sampling-in-optimal-sgd", "code_repository": "mysteryresearcher/sampling-in-optimal-sgd", "code_commit": "bd3f2e68783df06e4d7847843fe6d2d048226872", "code_archive": "repos/En7lGmzT_x.zip", "code_archive_sha256": "0a46391d570304a9a0c3d0ff585dc64f0d37dcdd70ce20310a0101402b1b226a", "code_archive_bytes": 231302, "code_file_count": 53, "code_extensions": {".py": 45, ".ipynb": 5, ".sh": 3}, "github_disk_usage_kb": 160, "github_languages": {"Python": 736820, "Jupyter Notebook": 90910, "Shell": 816}, "github_archived": false, "github_pushed_at": "2023-01-29T07:23:58Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/sharper-rates-and-flexible-framework-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "O0g6uPDLW7", "year": 2022, "status": "rejected", "title": "On the Adversarial Robustness of Vision Transformers", "authors": ["Rulin Shao", "Zhouxing Shi", "Jinfeng Yi", "Pin-Yu Chen", "Cho-Jui Hsieh"], "authorids": ["~Rulin_Shao1", "~Zhouxing_Shi1", "~Jinfeng_Yi1", "~Pin-Yu_Chen1", "~Cho-Jui_Hsieh1"], "authors_source": "OpenReview API", "abstract": "Following the success in advancing natural language processing and understanding, transformers are expected to bring revolutionary changes to computer vision. This work provides the first and comprehensive study on the robustness of vision transformers (ViTs) against adversarial perturbations. Tested on various white-box and transfer attack settings, we find that ViTs possess better adversarial robustness when compared with convolutional neural networks (CNNs). This observation also holds for certified robustness. We summarize the following main observations contributing to the improved robustness of ViTs:\n   1) Features learned by ViTs contain less low-level information and are more generalizable, which contributes to superior robustness against adversarial perturbations. \n   2) Introducing convolutional or tokens-to-token blocks for learning low-level features in ViTs can improve classification accuracy but at the cost of adversarial robustness.\n   3) Increasing the proportion of transformers in the model structure (when the model consists of both transformer and CNN blocks) leads to better robustness. But for a pure transformer model, simply increasing the size or adding layers cannot guarantee a similar effect.\n   4) Pre-training on larger datasets does not significantly improve adversarial robustness though it is critical for training ViTs. \n   5) Adversarial training is also applicable to ViT for training robust models.\nFurthermore, feature visualization and frequency analysis are conducted for explanation. The results show that ViTs are less sensitive to high-frequency perturbations than CNNs and there is a high correlation between how well the model learns low-level features and its robustness against different frequency-based perturbations.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=O0g6uPDLW7", "arxiv_id": "2103.15670", "paper_pdf": "papers/O0g6uPDLW7.pdf", "paper_pdf_sha256": "908b4c9fe1888894ad87afc44bd7be46ff3d91db7295bf5d3bfe1c0cd5c0f249", "paper_pdf_bytes": 23487093, "paper_pdf_source": "openreview", "code_url": "https://github.com/RulinShao/on-the-adversarial-robustness-of-visual-transformer", "code_repository": "RulinShao/on-the-adversarial-robustness-of-visual-transformer", "code_commit": "1cc19595c32b0c80f8a1a72d36e041844b339433", "code_archive": "repos/O0g6uPDLW7.zip", "code_archive_sha256": "cf9d74202d224ae957625c6d483287154b9f0fd65bc69baee6b9bff8551893d5", "code_archive_bytes": 269547, "code_file_count": 16, "code_extensions": {".py": 16}, "github_disk_usage_kb": 287, "github_languages": {"Python": 70597}, "github_archived": false, "github_pushed_at": "2021-11-18T18:31:19Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/2103-15670"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "szUsQ3NcQwV", "year": 2021, "status": "rejected", "title": "Randomized Entity-wise Factorization for Multi-Agent Reinforcement Learning", "authors": ["Shariq Iqbal", "Christian Schroeder de Witt", "Bei Peng", "Wendelin Boehmer", "Shimon Whiteson", "Fei Sha"], "authorids": ["~Shariq_Iqbal1", "~Christian_Schroeder_de_Witt1", "~Bei_Peng2", "~Wendelin_Boehmer1", "~Shimon_Whiteson1", "~Fei_Sha3"], "authors_source": "OpenReview API", "abstract": "Real world multi-agent tasks often involve varying types and quantities of agents and non-agent entities; however, agents within these tasks rarely need to consider all others at all times in order to act effectively. Factored value function approaches have historically leveraged such independences to improve learning efficiency, but these approaches typically rely on domain knowledge to select fixed subsets of state features to include in each factor. We propose to utilize value function factoring with random subsets of entities in each factor as an auxiliary objective in order to disentangle value predictions from irrelevant entities. This factoring approach is instantiated through a simple attention mechanism masking procedure. We hypothesize that such an approach helps agents learn more effectively in multi-agent settings by discovering common trajectories across episodes within sub-groups of agents/entities. Our approach, Randomized Entity-wise Factorization for Imagined Learning (REFIL), outperforms all strong baselines by a significant margin in challenging StarCraft micromanagement tasks.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=szUsQ3NcQwV", "arxiv_id": "2006.04222", "paper_pdf": "papers/szUsQ3NcQwV.pdf", "paper_pdf_sha256": "bf980830398d7625fd872560b1f0a2b47f308ca45091a21c963b899aae93baae", "paper_pdf_bytes": 3270420, "paper_pdf_source": "openreview", "code_url": "https://github.com/shariqiqbal2810/REFIL", "code_repository": "shariqiqbal2810/REFIL", "code_commit": "ffe23a1c62fd3f7304e700e69b30dcb0c29586cc", "code_archive": "repos/szUsQ3NcQwV.zip", "code_archive_sha256": "1f3565013c6f59005f205d0a955539d1b3883e65ef47f3cf682c5caa3e6b5820", "code_archive_bytes": 454521, "code_file_count": 43, "code_extensions": {".py": 39, ".sh": 4}, "github_disk_usage_kb": 321, "github_languages": {"Python": 264316, "Shell": 1958, "Dockerfile": 1657}, "github_archived": false, "github_pushed_at": "2021-05-22T22:46:58Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/ai-qmix-attention-and-imagination-for-dynamic"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "SkxHRySFvr", "year": 2020, "status": "rejected", "title": "LEARNING TO IMPUTE: A GENERAL FRAMEWORK FOR SEMI-SUPERVISED LEARNING", "authors": ["Wei-Hong Li", "Chuan-Sheng Foo", "Hakan Bilen"], "authorids": ["w.h.li@ed.ac.uk", "foo_chuan_sheng@i2r.a-star.edu.sg", "hbilen@ed.ac.uk"], "authors_source": "OpenReview API", "abstract": "Recent semi-supervised learning methods have shown to achieve comparable results to their supervised counterparts while using only a small portion of labels in image classification tasks thanks to their regularization strategies. In this paper, we take a more direct approach for semi-supervised learning and propose learning to impute the labels of unlabeled samples such that a network achieves better generalization when it is trained on these labels. We pose the problem in a learning-to-learn formulation which can easily be incorporated to the state-of-the-art semi-supervised techniques and boost their performance especially when the labels are limited. We demonstrate that our method is applicable to both classification and regression problems including image classification and facial landmark detection tasks.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=SkxHRySFvr", "arxiv_id": "1912.10364", "paper_pdf": "papers/SkxHRySFvr.pdf", "paper_pdf_sha256": "e79dffe4270c5ddd6cf2be625fa4e1a775f8b630eacd8b4277e376c0c0a441e8", "paper_pdf_bytes": 2760561, "paper_pdf_source": "openreview", "code_url": "https://github.com/VICO-UoE/L2I", "code_repository": "VICO-UoE/L2I", "code_commit": "0fda16b1d1eb353f9c34e3e9e1af3623d6f66575", "code_archive": "repos/SkxHRySFvr.zip", "code_archive_sha256": "11821666bb246f12dd54125c3e7ea45fb010a26c5e330e3ee86cfc08e0df80a5", "code_archive_bytes": 473092, "code_file_count": 23, "code_extensions": {".py": 23}, "github_disk_usage_kb": 429, "github_languages": {"Python": 222227}, "github_archived": false, "github_pushed_at": "2021-12-18T12:17:45Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/learning-to-impute-a-general-framework-for-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ZKkeA1G935", "year": 2026, "status": "rejected", "title": "Can LLMs Alleviate Catastrophic Forgetting in Graph Continual Learning? A Systematic Study", "authors": ["Ziyang Cheng", "Zhixun Li", "Yuhan Li", "Yixin Song", "Kangyi Zhao", "Dawei Cheng", "Jia Li", "Hong Cheng", "Jeffrey Xu Yu"], "authorids": ["~Ziyang_Cheng1", "~Zhixun_Li1", "~Yuhan_Li3", "~Yixin_Song2", "~Kangyi_Zhao1", "~Dawei_Cheng1", "~Jia_Li4", "~Hong_Cheng1", "~Jeffrey_Xu_Yu1"], "authors_source": "OpenReview API", "abstract": "Nowadays, real-world data, including graph-structure data, often arrives in a streaming manner, which means that learning systems need to continuously acquire new knowledge without forgetting previously learned information. Although substantial existing works attempt to address catastrophic forgetting in graph machine learning, they are all based on training from scratch with streaming data. With the rise of pretrained models, an increasing number of studies have leveraged their strong generalization ability for continual learning. Therefore, in this work, we attempt to answer whether large language models (LLMs) can mitigate catastrophic forgetting in graph continual learning}. We first evaluate the performance of LLMs and graph foundation models in graph continual learning scenarios, and found that with minimal modifications, they can easily achieve state-of-the-art results. Moreover, we found that certain current settings for graph continual learning tasks have significant flaws; it is possible to achieve zero forgetting with simple manipulations. Finally, based on extensive experiments, we propose a simple-yet-effective method, Simple Grpah Continual Learning (SimGCL), that surpasses the previous state-of-the-art baselines by around 20% under the rehearsal-free constraint.", "decision": "Reject", "meta_review": "", "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=ZKkeA1G935", "arxiv_id": "2505.18697", "paper_pdf": "papers/ZKkeA1G935.pdf", "paper_pdf_sha256": "8670c1708b38039295810b101422fea8dd543baa06a0e020499b9fa9f53cf406", "paper_pdf_bytes": 2437076, "paper_pdf_source": "openreview", "code_url": "https://github.com/ZhixunLEE/LLM4GCL", "code_repository": "ZhixunLEE/LLM4GCL", "code_commit": "e4b286560287289a184b20c34153af2716c89cec", "code_archive": "repos/ZKkeA1G935.zip", "code_archive_sha256": "d44e3251a95490b92fca6427690a532a10e8dae8dc552ae511fde8e510dcc750", "code_archive_bytes": 101569, "code_file_count": 43, "code_extensions": {".py": 37, ".sh": 6}, "github_disk_usage_kb": 1636, "github_languages": {"Python": 307475, "Shell": 6098}, "github_archived": false, "github_pushed_at": "2025-09-25T15:01:18Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/can-llms-alleviate-catastrophic-forgetting-in"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "kWtP5ZOErR", "year": 2025, "status": "rejected", "title": "EvoPress: Towards Optimal Dynamic Model Compression via Evolutionary Search", "authors": ["Oliver Sieberling", "Denis Kuznedelev", "Eldar Kurtic", "Dan Alistarh"], "authorids": ["~Oliver_Sieberling1", "~Denis_Kuznedelev1", "~Eldar_Kurtic1", "~Dan_Alistarh7"], "authors_source": "OpenReview API", "abstract": "The high computational costs of large language models (LLMs) have led to\na flurry of research on LLM compression, via methods such as quantization,\nsparsification, or structured pruning. A new frontier in this area is given by\ndynamic, non-uniform compression methods, which adjust the compression\nlevels (e.g., sparsity) per-block or even per-layer in order to minimize accuracy\nloss, while guaranteeing a global compression threshold. Yet, current methods\nrely on heuristics for identifying the “importance” of a given layer towards the\nloss, based on assumptions such as error monotonicity, i.e. that the end-to-end\nmodel compression error is proportional to the sum of layer-wise errors. In this\npaper, we revisit this area, and propose a new and general approach for dynamic\ncompression that is provably optimal in a given input range. We begin from\nthe motivating observation that, in general, error monotonicity does not hold for\nLLMs: compressed models with lower sum of per-layer errors can perform worse\nthan models with higher error sums. To address this, we propose a new general\nevolutionary framework for dynamic LLM compression called EvoPress, which\nhas provable convergence, low sample and evaluation complexity. We show that\nthese theoretical guarantees lead to highly competitive practical performance\nfor dynamic compression of Llama, Mistral and Phi models: via EvoPress, we\nset new state-of-the-art results for structural pruning (block/layer dropping),\nunstructured sparsity, as well as quantization with dynamic bitwidths.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=kWtP5ZOErR", "arxiv_id": "2410.14649", "paper_pdf": "papers/kWtP5ZOErR.pdf", "paper_pdf_sha256": "81aa15da647f999fd8367754a2bb94839c10307f22754b3d6174fc66ac3430ae", "paper_pdf_bytes": 847447, "paper_pdf_source": "openreview", "code_url": "https://github.com/IST-DASLab/EvoPress", "code_repository": "IST-DASLab/EvoPress", "code_commit": "a66a6974e91d27ede9043f07f885709732a9a9b1", "code_archive": "repos/kWtP5ZOErR.zip", "code_archive_sha256": "88be341b7385e379e4a59869fc8a2c6eea60bc7b9d30455dae7974555b0a41c3", "code_archive_bytes": 122483, "code_file_count": 35, "code_extensions": {".py": 28, ".sh": 7}, "github_disk_usage_kb": 112, "github_languages": {"Python": 217370, "Shell": 8478}, "github_archived": false, "github_pushed_at": "2026-06-14T15:13:15Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/evopress-towards-optimal-dynamic-model"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "wE1I9IGqeH", "year": 2024, "status": "rejected", "title": "Continual Learning in Open-vocabulary Classification with Complementary Memory Systems", "authors": ["Zhen Zhu", "Weijie Lyu", "Yao Xiao", "Derek Hoiem"], "authorids": ["~Zhen_Zhu1", "~Weijie_Lyu2", "~Yao_Xiao3", "~Derek_Hoiem1"], "authors_source": "OpenReview API", "abstract": "We introduce a method for flexible and efficient continual learning in open-vocabulary image classification, drawing inspiration from the complementary learning systems observed in human cognition. Specifically, we propose to combine predictions from a CLIP zero-shot model and the exemplar-based model, using the zero-shot estimated probability that a sample's class is within the exemplar classes. We also propose a \"tree probe\" method, an adaption of lazy learning principles, which enables fast learning from new examples with competitive accuracy to batch-trained linear models. We test in data incremental, class incremental, and task incremental settings, as well as ability to perform flexible inference on varying subsets of zero-shot and learned categories. Our proposed method achieves a good balance of learning speed, target task effectiveness, and zero-shot effectiveness.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=wE1I9IGqeH", "arxiv_id": "2307.01430", "paper_pdf": "papers/wE1I9IGqeH.pdf", "paper_pdf_sha256": "9f8c58e67298d91c4e3c4914ad261dc56b86965b538cd0fb7985c8bb0c16c8ec", "paper_pdf_bytes": 8557305, "paper_pdf_source": "openreview", "code_url": "https://github.com/jessemelpolio/TreeProbe", "code_repository": "jessemelpolio/TreeProbe", "code_commit": "cbb0b770a0b3c683f6a795f2b42d3934f7c8272a", "code_archive": "repos/wE1I9IGqeH.zip", "code_archive_sha256": "8fc471e892e369320fa4f93ef40ac456e842cd79099bc5cc330e224766f402c8", "code_archive_bytes": 104645, "code_file_count": 42, "code_extensions": {".py": 38, ".sh": 4}, "github_disk_usage_kb": 86, "github_languages": {"Python": 322128, "Shell": 16653}, "github_archived": false, "github_pushed_at": "2024-10-04T03:23:28Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/continual-learning-in-open-vocabulary"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "8Vxuz_PJNus", "year": 2023, "status": "rejected", "title": "Federated Learning of Large Models at the Edge via Principal Sub-Model Training", "authors": ["Yue Niu", "Saurav Prakash", "Souvik Kundu", "Sunwoo Lee", "Salman Avestimehr"], "authorids": ["~Yue_Niu1", "~Saurav_Prakash1", "~Souvik_Kundu2", "~Sunwoo_Lee1", "~Salman_Avestimehr1"], "authors_source": "OpenReview API", "abstract": "Limited compute, memory, and communication capabilities of edge users create a significant bottleneck for federated learning (FL) of large models. Current literature typically tackles the challenge with a heterogeneous client setting or allows training to be offloaded to the server. However, the former requires a fraction of clients to train near-full models, which may not be achievable at the edge; while the latter can compromise privacy with sharing of intermediate representations or labels. In this work, we consider a realistic, but much less explored, cross-device FL setting in which no client has the capacity to train a full large model nor is willing to share any intermediate representations with the server. To this end, we present Principal Sub-Model (PriSM) training methodology, which leverages models’ low-rank structure and kernel orthogonality to train sub-models in the orthogonal kernel space. More specifically, by applying singular value decomposition to original kernels in the server model, PriSM first obtains a set of principal orthogonal kernels with importance weighed by their singular values. Thereafter, PriSM utilizes a novel sampling strategy that selects different subsets of the principal kernels independently to create sub-models for clients with reduced computation and communication requirements. Importantly, a kernel with a large singular value is assigned with a high sampling probability. Thus, each sub-model is a low-rank approximation of the full large model, and all clients together achieve nearly full coverage of the principal kernels. To further improve memory efficiency, PriSM exploits low-rank structure in intermediate representations and allows each sub-model to learn only a subset of them while still preserving training performance. Our extensive evaluations on multiple datasets in various resource-constrained settings demonstrate that PriSM can yield an improved performance of up to $10\\%$ compared to existing alternatives, when training sub-models with only $20\\%$ principal kernels ($\\sim 5\\%$ of the full server model.).", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=8Vxuz_PJNus", "arxiv_id": "2208.13141", "paper_pdf": "papers/8Vxuz_PJNus.pdf", "paper_pdf_sha256": "7c4b4f9183dde522018b107ef32e78d424088a9eec962a1b9c7f599605497518", "paper_pdf_bytes": 1267211, "paper_pdf_source": "openreview", "code_url": "https://github.com/yuehniu/modeldecomp-fl", "code_repository": "yuehniu/modeldecomp-fl", "code_commit": "04e213968afd2dc7ae4097f8bc1d33b8bdb66ee2", "code_archive": "repos/8Vxuz_PJNus.zip", "code_archive_sha256": "ebf1b08048f776b240b511aea827134034d79f3b48a76fbc99fb5d40f1486417", "code_archive_bytes": 102796, "code_file_count": 30, "code_extensions": {".py": 29, ".sh": 1}, "github_disk_usage_kb": 167, "github_languages": {"Python": 240421, "Shell": 1127}, "github_archived": false, "github_pushed_at": "2023-10-05T18:27:16Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/federated-learning-of-large-models-at-the"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "cWlMII1LwTZ", "year": 2022, "status": "rejected", "title": "Task-aware Privacy Preservation for Multi-dimensional Data", "authors": ["Jiangnan Cheng", "Ao Tang", "Sandeep P. Chinchali"], "authorids": ["~Jiangnan_Cheng1", "~Ao_Tang1", "~Sandeep_P._Chinchali1"], "authors_source": "OpenReview API", "abstract": "Local differential privacy (LDP), a state-of-the-art technique for privacy preservation, has been successfully deployed in a few real-world applications. In the future, LDP can be adopted to anonymize richer user data attributes that will be input to more sophisticated machine learning (ML) tasks. However, today's LDP approaches are largely task-agnostic and often lead to sub-optimal performance - they will simply inject noise to all data attributes according to a given privacy budget, regardless of what features are most relevant for an ultimate task. In this paper, we address how to significantly improve the ultimate task performance for multi-dimensional user data by considering a task-aware privacy preservation problem. The key idea is to use an encoder-decoder framework to learn (and anonymize) a task-relevant latent representation of user data, which gives an analytical near-optimal solution for a linear setting with mean-squared error (MSE) task loss. We also provide an approximate solution through a learning algorithm for general nonlinear cases. Extensive experiments demonstrate that our task-aware approach significantly improves ultimate task accuracy compared to a standard benchmark LDP approach while guaranteeing the same level of privacy.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=cWlMII1LwTZ", "arxiv_id": "2110.02329", "paper_pdf": "papers/cWlMII1LwTZ.pdf", "paper_pdf_sha256": "b7c924d00150c871c86583ab5dd541450db469282f002fc757f9cb56526e5172", "paper_pdf_bytes": 510504, "paper_pdf_source": "openreview", "code_url": "https://github.com/chengjiangnan/task_aware_privacy", "code_repository": "chengjiangnan/task_aware_privacy", "code_commit": "2500da5b665945075c285ff575d519e256f66601", "code_archive": "repos/cWlMII1LwTZ.zip", "code_archive_sha256": "e655a53da9c6962cd52eea195c8060ae65ccb2c53f75149e70cd65db42546149", "code_archive_bytes": 351852, "code_file_count": 64, "code_extensions": {".sh": 39, ".py": 25}, "github_disk_usage_kb": 315, "github_languages": {"Python": 101625, "Shell": 10896}, "github_archived": false, "github_pushed_at": "2022-06-13T19:46:17Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/task-aware-privacy-preservation-for-multi"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "PuG6vCSbrV9", "year": 2021, "status": "rejected", "title": "Density estimation on low-dimensional manifolds: an inflation-deflation approach", "authors": ["Christian Horvat"], "authorids": ["~Christian_Horvat1"], "authors_source": "OpenReview API", "abstract": "Normalizing Flows (NFs) are universal density estimators based on Neuronal Networks. However, this universality is limited: the density's support needs to be diffeomorphic to a Euclidean space. In this paper, we propose a novel method to overcome this limitation without sacrificing the universality. The proposed method inflates the data manifold by adding noise in the normal space, trains an NF on this inflated manifold and, finally, deflates the learned density. Our main result provides sufficient conditions on the manifold and the specific choice of noise under which the corresponding estimator is exact. Our method has the same computational complexity as NFs, and does not require to compute an inverse flow. We also show that, if the embedding dimension is much larger than the manifold dimension, noise in the normal space can be well approximated by some Gaussian noise. This allows using our method for approximating arbitrary densities on non-flat manifolds provided that the manifold dimension is known. ", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=PuG6vCSbrV9", "arxiv_id": "2105.12152", "paper_pdf": "papers/PuG6vCSbrV9.pdf", "paper_pdf_sha256": "7a115a765945b97bdbbf087ea67c9bd19b97b1864968728e32feb7cc8beb074a", "paper_pdf_bytes": 2716141, "paper_pdf_source": "openreview", "code_url": "https://github.com/chrvt/Inflation-Deflation", "code_repository": "chrvt/Inflation-Deflation", "code_commit": "e19dc603e711df32821f4380acefc88ff64645a4", "code_archive": "repos/PuG6vCSbrV9.zip", "code_archive_sha256": "58ca7ace542656368cb74ed32ff54e9f3d158d91a314adb237b4500ff58fba94", "code_archive_bytes": 233213, "code_file_count": 36, "code_extensions": {".py": 32, ".sh": 4}, "github_disk_usage_kb": 332, "github_languages": {"Python": 284452, "Shell": 4811}, "github_archived": false, "github_pushed_at": "2022-02-02T14:22:08Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/density-estimation-an-inflation-deflation"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "B1xq264YvH", "year": 2020, "status": "rejected", "title": "Encoder-Agnostic Adaptation for Conditional Language Generation", "authors": ["Zachary M. Ziegler", "Luke Melas-Kyriazi", "Sebastian Gehrmann", "Alexander M. Rush"], "authorids": ["zziegler@g.harvard.edu", "lmelaskyriazi@college.harvard.edu", "gehrmann@seas.harvard.edu", "srush@seas.harvard.edu"], "authors_source": "OpenReview API", "abstract": "Large pretrained language models have changed the way researchers approach discriminative natural language understanding tasks, leading to the dominance of approaches that adapt a pretrained model for arbitrary downstream tasks. However, it is an open question how to use similar techniques for language generation. Early results in the encoder-agnostic setting have been mostly negative. In this work, we explore methods for adapting a pretrained language model to arbitrary conditional input. We observe that pretrained transformer models are sensitive to large parameter changes during tuning. Therefore, we propose an adaptation that directly injects arbitrary conditioning into self attention, an approach we call pseudo self attention. Through experiments on four diverse conditional text generation tasks, we show that this encoder-agnostic technique outperforms strong baselines, produces coherent generations, and is data-efficient.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=B1xq264YvH", "arxiv_id": "1908.06938", "paper_pdf": "papers/B1xq264YvH.pdf", "paper_pdf_sha256": "a2146da14ac4b2efe1b88056684a596e0ce69d19cd6fee4a291be2d955a36989", "paper_pdf_bytes": 371128, "paper_pdf_source": "openreview", "code_url": "https://github.com/harvardnlp/encoder-agnostic-adaptation", "code_repository": "harvardnlp/encoder-agnostic-adaptation", "code_commit": "5eff09874f25ac256f07daa0d3b9e7c03705086f", "code_archive": "repos/B1xq264YvH.zip", "code_archive_sha256": "832a1b98d5476b86b686f762ed17b41ce1e02fde3b6acd47b96a28b7b7ad65f5", "code_archive_bytes": 492951, "code_file_count": 100, "code_extensions": {".py": 96, ".sh": 3, ".ipynb": 1}, "github_disk_usage_kb": 449, "github_languages": {"Python": 656290, "Shell": 23927}, "github_archived": false, "github_pushed_at": "2024-07-25T10:15:13Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/encoder-agnostic-adaptation-for-conditional"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "OqVbYxDBHV", "year": 2026, "status": "rejected", "title": "Where do Large Vision-Language Models Look at when Answering Questions?", "authors": ["Xiaoying Xing", "Chia-Wen Kuo", "Li Fuxin", "Yulei Niu", "Fan Chen", "Ming Li", "Lianghui Wang", "Ying Wu", "Longyin Wen", "Sijie Zhu"], "authorids": ["~Xiaoying_Xing1", "~Chia-Wen_Kuo1", "~Li_Fuxin1", "~Yulei_Niu1", "~Fan_Chen5", "~Ming_Li19", "~Lianghui_Wang1", "~Ying_Wu7", "~Longyin_Wen1", "~Sijie_Zhu1"], "authors_source": "OpenReview API", "abstract": "Large Vision-Language Models (LVLMs) have shown promising performance in vision-language understanding and reasoning tasks. However, whether they truly understand the input image remain underexplored. A fundamental question arises: to what extent do LVLMs rely on visual input, and which image regions contribute to their responses? It is non-trivial to interpret the free-form generation of LVLMs due to their complicated visual architecture (e.g., multiple encoders and multi-resolution) and variable-length outputs. In this paper, we extend existing heatmap visualization methods for classification tasks to support LVLMs for open-ended visual question answering. We propose a method to select visually relevant tokens that reflect the relevance between generated responses and the input image.\nFurthermore, we conduct a comprehensive analysis of state-of-the-art LVLMs on benchmarks designed to require visual information to answer. Our findings offer several insights into LVLM behavior, including the relationship between focus region and answer correctness, differences in visual attention across architectures, and the impact of LLM scale on visual understanding. The code and data will be\nreleased", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=OqVbYxDBHV", "arxiv_id": "2503.13891", "paper_pdf": "papers/OqVbYxDBHV.pdf", "paper_pdf_sha256": "06f074da3f6a536df15a25854814555101a02218f65bb7015396f1aea73b64c8", "paper_pdf_bytes": 6078471, "paper_pdf_source": "openreview", "code_url": "https://github.com/bytedance/LVLM_Interpretation", "code_repository": "bytedance/LVLM_Interpretation", "code_commit": "378412f64c253aa497bd95e983395aada8940294", "code_archive": "repos/OqVbYxDBHV.zip", "code_archive_sha256": "87cf283127af82913ec33dcbadf9cd66e892bd63b34ed40f1c1715a2153e6254", "code_archive_bytes": 1778635, "code_file_count": 143, "code_extensions": {".py": 142, ".sh": 1}, "github_disk_usage_kb": 1647, "github_languages": {"Python": 1287958, "Shell": 399}, "github_archived": false, "github_pushed_at": "2026-01-07T03:10:03Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/where-do-large-vision-language-models-look-at"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "xreOs2yjqf", "year": 2025, "status": "rejected", "title": "EvalAlign: Supervised Fine-Tuning Multimodal LLMs with Human-Aligned Data for Evaluating Text-to-Image Models", "authors": ["Zhiyu Tan", "xiaomeng yang", "Luozheng Qin", "Mengping Yang", "Cheng Zhang", "Hao Li"], "authorids": ["~Zhiyu_Tan1", "~xiaomeng_yang3", "~Luozheng_Qin1", "~Mengping_Yang2", "~Cheng_Zhang8", "~Hao_Li16"], "authors_source": "OpenReview API", "abstract": "The recent advancements in text-to-image generative models have been remarkable. Yet, the field suffers from a lack of evaluation metrics that accurately reflect the performance of these models, particularly lacking fine-grained metrics that can guide the optimization of the models. In this paper, we propose EvalAlign, a metric characterized by its accuracy, stability, and fine granularity. Our approach leverages the capabilities of Multimodal Large Language Models (MLLMs) pre-trained on extensive data. We develop evaluation protocols that focus on two key dimensions: image faithfulness and text-image alignment. Each protocol comprises a set of detailed, fine-grained instructions linked to specific scoring options, enabling precise manual scoring of the generated images. We supervised fine-tune (SFT) the MLLM to align with human evaluative judgments, resulting in a robust evaluation model. Our evaluation across 24 text-to-image generation models demonstrate that EvalAlign not only provides superior metric stability but also aligns more closely with human preferences than existing metrics, confirming its effectiveness and utility in model assessment. We will make the code, data, and pre-trained models publicly available.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=xreOs2yjqf", "arxiv_id": "2406.16562", "paper_pdf": "papers/xreOs2yjqf.pdf", "paper_pdf_sha256": "5fa17083c06ce24c09c194cb91eb80dda63d337722abe3cf19d2f3d5aa00dcb3", "paper_pdf_bytes": 1150031, "paper_pdf_source": "openreview", "code_url": "https://github.com/SAIS-FUXI/EvalAlign", "code_repository": "SAIS-FUXI/EvalAlign", "code_commit": "d3fa7b49bd87e673d2a941d60c99b1d2c564205d", "code_archive": "repos/xreOs2yjqf.zip", "code_archive_sha256": "99f6aa425066ac533e10334e5e2ece4bc8dbe12e04c7dddb63541729c7df0b4b", "code_archive_bytes": 64055, "code_file_count": 25, "code_extensions": {".py": 23, ".sh": 2}, "github_disk_usage_kb": 112, "github_languages": {"Python": 102457, "Shell": 383}, "github_archived": false, "github_pushed_at": "2024-10-23T08:19:35Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/evalalign-evaluating-text-to-image-models"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "V0CUOBWUHa", "year": 2024, "status": "rejected", "title": "Scaling Sentence Embeddings with Large Language Models", "authors": ["Ting Jiang", "Shaohan Huang", "Zhongzhi Luan", "deqing wang", "Fuzhen Zhuang"], "authorids": ["~Ting_Jiang1", "~Shaohan_Huang1", "~Zhongzhi_Luan1", "~deqing_wang2", "~Fuzhen_Zhuang1"], "authors_source": "OpenReview API", "abstract": "Large language models (LLMs) have recently garnered significant interest. With in-context learning, LLMs achieve impressive results in various natural language tasks. However, the application of LLMs to sentence embeddings remains an area of ongoing research. In this work, we propose an in-context learning-based method aimed at improving sentence embeddings performance. Our approach involves adapting the previous prompt-based representation method for autoregressive models, constructing a demonstration set that enables LLMs to perform in-context learning, and scaling up the LLMs to different model sizes. Through extensive experiments, in-context learning enables LLMs to generate high-quality sentence embeddings without any fine-tuning. It helps LLMs achieve performance comparable to current contrastive learning methods. By scaling model size, we find scaling to more than tens of billion parameters harms the performance on semantic textual similarity (STS) tasks. However, the largest model outperforms other counterparts and achieves the new state-of-the-art result on transfer tasks. We also fine-tune LLMs with current contrastive learning approach, and the 2.7B OPT model, incorporating our prompt-based method, surpasses the performance of 4.8B ST5, achieving the new state-of-the-art results on STS tasks.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=V0CUOBWUHa", "arxiv_id": "2307.16645", "paper_pdf": "papers/V0CUOBWUHa.pdf", "paper_pdf_sha256": "965ec0863b409226598ca0354c7e8c3ad8439d9e69274de4208f82a2f0f9120b", "paper_pdf_bytes": 1168695, "paper_pdf_source": "openreview", "code_url": "https://github.com/kongds/scaling_sentemb", "code_repository": "kongds/scaling_sentemb", "code_commit": "8567aa083c1b3c77586670f91e7f78eb80694ad3", "code_archive": "repos/V0CUOBWUHa.zip", "code_archive_sha256": "b0cf75f8e6e7d9c590e82f080a72b27dafc8f682bd34333ee3b2cdcdeeaa156a", "code_archive_bytes": 84059, "code_file_count": 32, "code_extensions": {".py": 27, ".sh": 5}, "github_disk_usage_kb": 87, "github_languages": {"Python": 206324, "Shell": 7237}, "github_archived": false, "github_pushed_at": "2024-03-22T13:51:46Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/scaling-sentence-embeddings-with-large"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "LOTGOB5_Xh2", "year": 2023, "status": "rejected", "title": "Architecture-Agnostic Masked Image Modeling -- From ViT back to CNN", "authors": ["Siyuan Li", "Di Wu", "Fang Wu", "Zelin Zang", "Lei Shang", "Baigui Sun", "Xuansong Xie", "Stan Z. Li"], "authorids": ["~Siyuan_Li6", "~Di_Wu10", "~Fang_Wu1", "~Zelin_Zang2", "~Lei_Shang1", "~Baigui_Sun1", "~Xuansong_Xie1", "~Stan_Z._Li2"], "authors_source": "OpenReview API", "abstract": "Masked image modeling (MIM), an emerging self-supervised pre-training method, has shown impressive success across numerous downstream vision tasks with Vision transformers (ViTs). Its underlying idea is simple: a portion of the input image is randomly masked out and then reconstructed via the pre-text task. However, the working principle behind MIM is not well explained, and previous studies insist that MIM primarily works for the Transformer family but is incompatible with CNNs. In this paper, we first study interactions among patches to understand what knowledge is learned and how it is acquired via the MIM task. We observe that MIM essentially teaches the model to learn better middle-order interactions among patches and extract more generalized features. Based on this fact, we propose an Architecture-Agnostic Masked Image Modeling framework (A$^2$MIM), which is compatible with both Transformers and CNNs in a unified way. Extensive experiments on popular benchmarks show that our A$^2$MIM learns better representations without explicit design and endows the backbone model with the stronger capability to transfer to various downstream tasks for both Transformers and CNNs.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=LOTGOB5_Xh2", "arxiv_id": "2205.13943", "paper_pdf": "papers/LOTGOB5_Xh2.pdf", "paper_pdf_sha256": "c3b5019a459977720bc268a22455172fdd689c75ba31f8c46542193771e858e8", "paper_pdf_bytes": 8002029, "paper_pdf_source": "openreview", "code_url": "https://github.com/Westlake-AI/A2MIM", "code_repository": "Westlake-AI/A2MIM", "code_commit": "ccca6dc21995b8e39e5571bd189c1ebce93bf49d", "code_archive": "repos/LOTGOB5_Xh2.zip", "code_archive_sha256": "4313e080a4198f6abbe73fe55622798b89ad3a1c5627c503709a872fa8af07a3", "code_archive_bytes": 283762, "code_file_count": 224, "code_extensions": {".py": 210, ".sh": 14}, "github_disk_usage_kb": 178, "github_languages": {"Python": 516747, "Shell": 9223}, "github_archived": false, "github_pushed_at": "2024-08-15T22:42:24Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/architecture-agnostic-masked-image-modeling"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "i8d2kdxii1L", "year": 2022, "status": "rejected", "title": "$p$-Laplacian Based Graph Neural Networks", "authors": ["Guoji Fu", "Peilin Zhao", "Yatao Bian"], "authorids": ["~Guoji_Fu1", "~Peilin_Zhao2", "~Yatao_Bian1"], "authors_source": "OpenReview API", "abstract": "Graph neural networks (GNNs) have demonstrated superior performance for semi-supervised node classification on graphs, as a result of their ability to exploit node features and topological information simultaneously. However, most GNNs implicitly assume that the labels of nodes and their neighbors in a graph are the same or consistent, which does not hold in heterophilic graphs, where the labels of linked nodes are likely to differ. Hence, when the topology is non-informative for label prediction, ordinary GNNs may work significantly worse than simply applying multi-layer perceptrons (MLPs) on each node. To tackle the above problem, we propose a new $p$-Laplacian based GNN model, termed as $^p$GNN, whose message passing mechanism is derived from a discrete regularization framework and could be theoretically explained as an approximation of a polynomial graph filter defined on the spectral domain of $p$-Laplacians. The spectral analysis shows that the new message passing mechanism works simultaneously as low-pass and high-pass filters, thus making $^p$GNNs effective on both homophilic and heterophilic graphs. Empirical studies on real-world and synthetic datasets validate our findings and demonstrate that $^p$GNNs significantly outperform several state-of-the-art GNN architectures on heterophilic benchmarks while achieving competitive performance on homophilic benchmarks. Moreover, $^p$GNNs can adaptively learn aggregation weights and are robust to noisy edges.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=i8d2kdxii1L", "arxiv_id": "2111.07337", "paper_pdf": "papers/i8d2kdxii1L.pdf", "paper_pdf_sha256": "866c51c45e6d19e8fe905c4ecba62df507dfa3f6765f311cc38cf2b8ff2bfe2e", "paper_pdf_bytes": 3067889, "paper_pdf_source": "openreview", "code_url": "https://github.com/guoji-fu/pGNNs", "code_repository": "guoji-fu/pGNNs", "code_commit": "f9a0e2bbf553a81748bb3ed42e953f71af6028a2", "code_archive": "repos/i8d2kdxii1L.zip", "code_archive_sha256": "92c63acb901c3361f18be276a27abcdb550b942e7afbfbff6b1c48d29ad24ef3", "code_archive_bytes": 318773, "code_file_count": 9, "code_extensions": {".py": 7, ".sh": 2}, "github_disk_usage_kb": 331, "github_languages": {"Python": 41316, "Shell": 3020}, "github_archived": false, "github_pushed_at": "2025-08-26T04:15:37Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/p-laplacian-based-graph-neural-networks-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "0naHZ3gZSzo", "year": 2021, "status": "rejected", "title": "Optimizing Large-Scale Hyperparameters via Automated Learning Algorithm", "authors": ["Bin Gu", "Guodong Liu", "Yanfu Zhang", "Xiang Geng", "Heng Huang"], "authorids": ["~Bin_Gu1", "~Guodong_Liu2", "~Yanfu_Zhang1", "~Xiang_Geng1", "~Heng_Huang1"], "authors_source": "OpenReview API", "abstract": "Modern machine learning algorithms usually involve tuning multiple (from one to  thousands) hyperparameters which play a pivotal role  in terms of model  generalizability. Globally choosing appropriate values of hyperparameters is extremely  computationally challenging. Black-box optimization and gradient-based algorithms are  two dominant approaches to hyperparameter optimization while  they have totally  distinct advantages. How to design a new hyperparameter optimization technique inheriting all benefits  from both approaches is still an open problem. To address this challenging problem, in this paper, we  propose a new  hyperparameter optimization  method with zeroth-order hyper-gradients (HOZOG). Specifically, we first exactly formulate  hyperparameter optimization  as  an $\\mathcal{A}$-based constrained optimization problem, where $\\mathcal{A}$ is  a black-box optimization algorithm (such as deep neural network). Then, we use the average zeroth-order hyper-gradients  to update  hyperparameters. We provide the feasibility analysis of using  HOZOG to achieve hyperparameter optimization.  The experimental  results on three representative  hyperparameter (the size is from 1 to 1250) optimization tasks  demonstrate the benefits of HOZOG  in terms of  \\textit{simplicity, scalability, flexibility, effectiveness and efficiency} compared with the  state-of-the-art hyperparameter optimization methods.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=0naHZ3gZSzo", "arxiv_id": "2102.09026", "paper_pdf": "papers/0naHZ3gZSzo.pdf", "paper_pdf_sha256": "3db76b844f4d1092b9015f00dab01254223ced930aefa697222d43f3b1a019dd", "paper_pdf_bytes": 684065, "paper_pdf_source": "openreview", "code_url": "https://github.com/jsgubin/HOZOG", "code_repository": "jsgubin/HOZOG", "code_commit": "cae6ac386d1b43c70d269e47e10ba4a4ac7aed4a", "code_archive": "repos/0naHZ3gZSzo.zip", "code_archive_sha256": "b4aafcc5502a92ff03e97863c59fe0d1fa0e6c95203e46edc9ac96484aefec54", "code_archive_bytes": 278996, "code_file_count": 16, "code_extensions": {".py": 16}, "github_disk_usage_kb": 336, "github_languages": {"Python": 158062}, "github_archived": false, "github_pushed_at": "2021-04-12T07:09:54Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/optimizing-large-scale-hyperparameters-via-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "r1l-5pEtDr", "year": 2020, "status": "rejected", "title": "AdaX: Adaptive Gradient Descent with Exponential Long Term Memory", "authors": ["Wenjie Li", "Zhaoyang Zhang", "Xinjiang Wang", "Ping Luo"], "authorids": ["li3549@purdue.edu", "zhaoyangzhang@link.cuhk.edu.hk", "swanxinjiang@gmail.com", "pluo.lhi@gmail.com"], "authors_source": "OpenReview API", "abstract": "Adaptive optimization algorithms such as RMSProp and Adam have fast convergence and smooth learning process. Despite their successes, they are proven to have non-convergence issue even in convex optimization problems as well as weak performance compared with the first order gradient methods such as stochastic gradient descent (SGD). Several other algorithms, for example AMSGrad and AdaShift, have been proposed to alleviate these issues but only minor effect has been observed. This paper further analyzes the performance of such algorithms in a non-convex setting by extending their non-convergence issue into a simple non-convex case and show that Adam's design of update steps would possibly lead the algorithm to local minimums. To address the above problems, we propose a novel adaptive gradient descent algorithm, named AdaX, which accumulates the long-term past gradient information exponentially. We prove the convergence of AdaX in both convex and non-convex settings. Extensive experiments show that AdaX outperforms Adam in various tasks of computer vision and natural language processing and can catch up with SGD.\n", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=r1l-5pEtDr", "arxiv_id": "2004.09740", "paper_pdf": "papers/r1l-5pEtDr.pdf", "paper_pdf_sha256": "b17f9d2ab9934c085fc7a5295b48455bc398798bfc28245f300bbc215ef005f1", "paper_pdf_bytes": 747209, "paper_pdf_source": "openreview", "code_url": "https://github.com/switchablenorms/AdaX", "code_repository": "switchablenorms/AdaX", "code_commit": "e18f35f3d6ab99ad862f81d6ddf4d7dbc5f2f63d", "code_archive": "repos/r1l-5pEtDr.zip", "code_archive_sha256": "ebfc2d641b498b2184d82b523c9c929bb07c3f8491c6ebdd7009a9541abc8d16", "code_archive_bytes": 539620, "code_file_count": 12, "code_extensions": {".py": 12}, "github_disk_usage_kb": 529, "github_languages": {"Python": 81479}, "github_archived": false, "github_pushed_at": "2020-05-08T21:11:56Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/adax-adaptive-gradient-descent-with-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "M7TJaVg4sE", "year": 2026, "status": "rejected", "title": "TraSCE: Trajectory Steering for Concept Erasure", "authors": ["Anubhav Jain", "Takashi Shibuya", "Yuya Kobayashi", "Yuhta Takida", "Nasir Memon", "Julian Togelius", "Yuki Mitsufuji"], "authorids": ["~Anubhav_Jain1", "~Takashi_Shibuya1", "~Yuya_Kobayashi1", "~Yuhta_Takida1", "~Nasir_Memon1", "~Julian_Togelius1", "~Yuki_Mitsufuji1"], "authors_source": "OpenReview API", "abstract": "Recent advancements in text-to-image diffusion models have brought them to the public spotlight, becoming widely accessible and embraced by everyday users. However, these models have been shown to generate harmful content, such as not-safe-for-work (NSFW) images. While approaches have been proposed to erase such abstract concepts from the models, jail-breaking techniques have succeeded in bypassing such safety measures. In this paper, we propose TraSCE, an approach to guide the diffusion trajectory away from generating harmful content. Our approach is based on negative prompting, but as we show, the widely used negative prompting strategy is not a complete solution for concept erasure and can easily be bypassed in a simple corner case. To address this issue, we introduce two techniques. We first borrow the idea of concept negation from compositional generation and propose to use it instead of the conventional negative prompting. Second, we introduce a localized loss-based guidance that enhances the modified negative prompting technique by steering the diffusion trajectory. We demonstrate that our proposed method achieves state-of-the-art results on various benchmarks in erasing harmful content, artistic styles, and objects, including on benchmarks introduced by red teams without impacting unrelated concepts. Our proposed approach does not require any training, weight modifications, or training data (either image or prompt), making it easier for model owners to erase new concepts.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=M7TJaVg4sE", "arxiv_id": "2412.07658", "paper_pdf": "papers/M7TJaVg4sE.pdf", "paper_pdf_sha256": "da5252c05c4cbfdd9ea12e28a4b5e94b6a6eda160cd31b48314a1cf14c19576a", "paper_pdf_bytes": 15277066, "paper_pdf_source": "openreview", "code_url": "https://github.com/anubhav1997/TraSCE", "code_repository": "anubhav1997/TraSCE", "code_commit": "8e3813b2aea10311cbed4bb85e25be345dd3db35", "code_archive": "repos/M7TJaVg4sE.zip", "code_archive_sha256": "db108808a6143fb164e7e45de8835d6572d8309d627575a14be40f717dee534c", "code_archive_bytes": 1713193, "code_file_count": 7, "code_extensions": {".py": 7}, "github_disk_usage_kb": 1686, "github_languages": {"Python": 31762}, "github_archived": false, "github_pushed_at": "2025-02-01T16:48:21Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/trasce-trajectory-steering-for-concept"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "dTQmayPKMs", "year": 2025, "status": "rejected", "title": "Understanding Impact of Human Feedback via Influence Functions", "authors": ["Taywon Min", "Haeone Lee", "Hanho Ryu", "Yongchan Kwon", "Kimin Lee"], "authorids": ["~Taywon_Min1", "~Haeone_Lee1", "~Hanho_Ryu1", "~Yongchan_Kwon1", "~Kimin_Lee1"], "authors_source": "OpenReview API", "abstract": "In Reinforcement Learning from Human Feedback (RLHF), it is crucial to learn suitable reward models from human feedback to align large language models (LLMs) with human intentions. However, human feedback can often be noisy, inconsistent, or biased, especially when evaluating complex responses. Such feedback can lead to misaligned reward signals, potentially causing unintended side effects during the RLHF process. To address these challenges, we explore the use of influence functions to measure the impact of human feedback on the performance of reward models. We propose a compute-efficient approximation method that enables the application of influence functions to LLM-based reward models and large-scale preference datasets. In our experiments, we demonstrate two key applications of influence functions: (1) detecting common forms of labeler bias in human feedback datasets and (2) guiding labelers to refine their strategies to align more closely with expert feedback. By quantifying the impact of human feedback on reward models, we believe that influence functions can enhance feedback interpretability and contribute to scalable oversight in RLHF, helping labelers provide more accurate and consistent feedback.", "decision": "Reject", "meta_review": "", "num_reviews": 6, "openreview_url": "https://openreview.net/forum?id=dTQmayPKMs", "arxiv_id": "2501.05790", "paper_pdf": "papers/dTQmayPKMs.pdf", "paper_pdf_sha256": "4b5c1c488949e36a6503070217999598594d67257d064a6ceac2d1403b9800f5", "paper_pdf_bytes": 1440438, "paper_pdf_source": "openreview", "code_url": "https://github.com/mintaywon/IF_RLHF", "code_repository": "mintaywon/IF_RLHF", "code_commit": "e72cf5674faa4c19521a1856ee3558a564758062", "code_archive": "repos/dTQmayPKMs.zip", "code_archive_sha256": "0276d9f179d82f7bcc5fc4b97e4648393ee8f57c4dea60a8cbd8380aa1e28ffe", "code_archive_bytes": 108680, "code_file_count": 15, "code_extensions": {".py": 13, ".ipynb": 2}, "github_disk_usage_kb": 113, "github_languages": {"Python": 68845, "Jupyter Notebook": 14157}, "github_archived": false, "github_pushed_at": "2025-02-02T10:49:49Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/understanding-impact-of-human-feedback-via"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "7VVGO0kuuY", "year": 2024, "status": "rejected", "title": "Learning Causal Dynamics Models in Object-Oriented Environments", "authors": ["Zhongwei Yu", "Jingqing Ruan", "Dengpeng Xing"], "authorids": ["~Zhongwei_Yu1", "~Jingqing_Ruan1", "~Dengpeng_Xing1"], "authors_source": "OpenReview API", "abstract": "Causal Dynamics Models (CDMs) have demonstrated significant potential in addressing various challenges in reinforcement learning. Recent studies have incorporated causal discovery to capture the causal dependencies among environmental variables in the learning of CDMs. However, the learning of CDMs is still confined to small-scale environments due to computational complexity and sample efficiency constraints. This paper aims to extend CDMs to large-scale object-oriented environments, which consist of a multitude of objects classified into different categories. We introduce the Object-Oriented CDM (OOCDM) that shares causalities and parameters among objects belonging to the same class. Furthermore, we propose a learning method for OOCDM that enables it to adapt to a varying number of objects. Experimental results from large-scale tasks indicate that OOCDM outperforms existing CDMs in terms of causal discovery, prediction accuracy, generalization, and computational efficiency.", "decision": "Reject", "meta_review": "", "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=7VVGO0kuuY", "arxiv_id": "2405.12615", "paper_pdf": "papers/7VVGO0kuuY.pdf", "paper_pdf_sha256": "b946e7fc09051970674252bbbc7f488d24b01f49d67e958b1bf378b9c6c68b28", "paper_pdf_bytes": 1852025, "paper_pdf_source": "openreview", "code_url": "https://github.com/zwyu-ai/oocdm", "code_repository": "zwyu-ai/oocdm", "code_commit": "9bc24567986aa6b13be874d1a6599bef4fca02e8", "code_archive": "repos/7VVGO0kuuY.zip", "code_archive_sha256": "35697fa9de106b50f2c6a22fd8477f31ab0efc036148a77f3eb806744d5f8dac", "code_archive_bytes": 106512, "code_file_count": 63, "code_extensions": {".py": 63}, "github_disk_usage_kb": 90, "github_languages": {"Python": 341368}, "github_archived": false, "github_pushed_at": "2024-06-14T09:22:28Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/learning-causal-dynamics-models-in-object"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "mBkUeW8rpD6", "year": 2023, "status": "rejected", "title": "DiscoBAX - Discovery of optimal intervention sets in genomic experiment design", "authors": ["Clare Lyle", "Arash Mehrjou", "Pascal Notin", "Andrew Jesson", "Stefan Bauer", "Yarin Gal", "Patrick Schwab"], "authorids": ["~Clare_Lyle1", "~Arash_Mehrjou1", "~Pascal_Notin1", "~Andrew_Jesson1", "~Stefan_Bauer1", "~Yarin_Gal1", "~Patrick_Schwab1"], "authors_source": "OpenReview API", "abstract": "The discovery of novel therapeutics to cure genetic pathologies relies on the identification of the different genes involved in the underlying disease mechanism. With billions of potential hypotheses to test, an exhaustive  exploration of the entire space of potential interventions is impossible in practice. Sample-efficient methods based on active learning or bayesian optimization bear the promise of identifying interesting targets using the least experiments possible. However, genomic perturbation experiments typically rely on proxy outcomes measured in biological model systems that may not completely correlate with the outcome of interventions in humans. In practical experiment design, one aims to find a set of interventions which maximally move a target phenotype via a diverse set of mechanisms in order to reduce the risk of failure in future stages of trials. To that end, we introduce DiscoBAX — a sample-efficient algorithm for the discovery of genetic interventions that maximize the movement of a phenotype in a direction of interest while covering a diverse set of underlying mechanisms. We provide theoretical guarantees on the optimality of the approach under standard assumptions, conduct extensive experiments in synthetic and real-world settings relevant to genomic discovery and demonstrate that DiscoBAX outperforms state-of-the-art active learning and Bayesian optimization methods in this task. Better methods for selecting effective and diverse perturbations in biological systems could enable researchers to potentially discover novel therapeutics for a range of genetically-driven diseases.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=mBkUeW8rpD6", "arxiv_id": "2312.04064", "paper_pdf": "papers/mBkUeW8rpD6.pdf", "paper_pdf_sha256": "7003ec1b68d1f9e6729494fcefdac099451c36ac4ead1f5c27b8f8fd96ef4eaf", "paper_pdf_bytes": 804941, "paper_pdf_source": "openreview", "code_url": "https://github.com/amehrjou/DiscoBAX", "code_repository": "amehrjou/DiscoBAX", "code_commit": "84f01283bc7f6ab5f66b5ea2a63632b401cc0402", "code_archive": "repos/mBkUeW8rpD6.zip", "code_archive_sha256": "3f1b7708f0472388438757b64a9ba00e1844bffddcb6bfca6194bed954d628ca", "code_archive_bytes": 187751, "code_file_count": 37, "code_extensions": {".py": 35, ".ipynb": 1, ".sh": 1}, "github_disk_usage_kb": 179, "github_languages": {"Python": 209151, "Jupyter Notebook": 138840, "Shell": 2424}, "github_archived": false, "github_pushed_at": "2024-01-21T15:54:40Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/discobax-discovery-of-optimal-intervention"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "wNsNT56zDkG", "year": 2022, "status": "rejected", "title": "Adversarial Rademacher Complexity of Deep Neural Networks", "authors": ["Jiancong Xiao", "Yanbo Fan", "Ruoyu Sun", "Zhi-Quan Luo"], "authorids": ["~Jiancong_Xiao1", "~Yanbo_Fan1", "~Ruoyu_Sun1", "~Zhi-Quan_Luo1"], "authors_source": "OpenReview API", "abstract": "Deep neural networks are vulnerable to adversarial attacks. Adversarial training is one of the most effective algorithms to increase the model's robustness. However, the trained models cannot generalize well to the adversarial examples on the test set. In this paper, we study the generalization of adversarial training through the lens of adversarial Rademacher complexity. Current analysis of adversarial Rademacher complexity is up to two-layer neural networks. In adversarial settings, one major difficulty of generalizing these results to deep neural networks is that we cannot peel off the layer as the classical analysis for standard training. We provide a method to overcome this issue and provide upper bounds of adversarial Rademacher complexity of deep neural networks. Similar to the existing bounds of standard Rademacher complexity of neural nets, our bound also includes the product of weight norms. We provide experiments to show that the adversarially trained weight norms are larger than the standard trained weight norms, thus providing an explanation for the bad generalization performance of adversarial training.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=wNsNT56zDkG", "arxiv_id": "2211.14966", "paper_pdf": "papers/wNsNT56zDkG.pdf", "paper_pdf_sha256": "95eefe9f85dce2cef27ad47a78b2098d075eff59d4b07b9b28cfee2ff7bb8d31", "paper_pdf_bytes": 851245, "paper_pdf_source": "openreview", "code_url": "https://github.com/JiancongXiao/Adversarial-Rademacher-Complexity", "code_repository": "JiancongXiao/Adversarial-Rademacher-Complexity", "code_commit": "adc891cdc79f13014cf994db3837a66cbb06392c", "code_archive": "repos/wNsNT56zDkG.zip", "code_archive_sha256": "e307414deb5544a4b266d775639e8b5a973544b38c5b14cc3d151e8982da4f10", "code_archive_bytes": 350410, "code_file_count": 11, "code_extensions": {".py": 11}, "github_disk_usage_kb": 336, "github_languages": {"Python": 55343}, "github_archived": false, "github_pushed_at": "2023-11-27T19:48:26Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/adversarial-rademacher-complexity-of-deep-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "gp5Uzbl-9C-", "year": 2021, "status": "rejected", "title": "Systematic Evaluation of Causal Discovery in Visual Model Based Reinforcement Learning", "authors": ["Nan Rosemary Ke", "Aniket Rajiv Didolkar", "Sarthak Mittal", "Anirudh Goyal", "Guillaume Lajoie", "Stefan Bauer", "Danilo Jimenez Rezende", "Michael Curtis Mozer", "Yoshua Bengio", "Christopher Pal"], "authorids": ["~Nan_Rosemary_Ke1", "~Aniket_Rajiv_Didolkar1", "~Sarthak_Mittal1", "~Anirudh_Goyal1", "~Guillaume_Lajoie1", "~Stefan_Bauer1", "~Danilo_Jimenez_Rezende2", "~Michael_Curtis_Mozer1", "~Yoshua_Bengio1", "~Christopher_Pal1"], "authors_source": "OpenReview API", "abstract": "Inducing causal relationships from observations is a classic problem in machine learning. Most work in causality starts from the premise that the causal variables themselves have known semantics or are observed. However, for AI agents such as robots trying to make sense of their environment, the only observables are low-level variables like pixels in images. To generalize well, an agent must induce high-level variables, particularly those which are causal or are affected by causal variables. A central goal for AI and causality is thus the joint discovery of abstract representations and causal structure. In this work, we systematically evaluate the agent's ability to learn underlying causal structure. We note that existing environments for studying causal induction are poorly suited for this  objective because they have complicated task-specific causal graphs with many confounding factors. Hence, to facilitate research in learning the representation of high-level variables as well as causal structure among these  variables, we present a suite of RL environments created to systematically probe the ability of methods to identify variables as well as causal structure among those variables. We evaluate various representation learning algorithms from literature and found that  explicitly incorporating structure and modularity in the model can help causal induction in model-based reinforcement learning.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=gp5Uzbl-9C-", "arxiv_id": "2107.00848", "paper_pdf": "papers/gp5Uzbl-9C-.pdf", "paper_pdf_sha256": "51d03eb06c9c227a225cb4f475d548f0e9644ef8e08612ed469e9c2959742564", "paper_pdf_bytes": 285872, "paper_pdf_source": "openreview", "code_url": "https://github.com/dido1998/CausalMBRL", "code_repository": "dido1998/CausalMBRL", "code_commit": "2c595988b06b3d568fdde53213029e97841acb03", "code_archive": "repos/gp5Uzbl-9C-.zip", "code_archive_sha256": "9b0b6d9b9fa79474cc28219afd8560d48dd7c4d3c031f9b53205ee70fc5a8713", "code_archive_bytes": 160266, "code_file_count": 85, "code_extensions": {".py": 53, ".sh": 32}, "github_disk_usage_kb": 367, "github_languages": {"Python": 527951, "Shell": 58570}, "github_archived": false, "github_pushed_at": "2021-07-15T17:13:11Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/systematic-evaluation-of-causal-discovery-in-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "r1g6MCEtwr", "year": 2020, "status": "rejected", "title": "Zero-Shot Out-of-Distribution Detection with Feature Correlations", "authors": ["Chandramouli S Sastry", "Sageev Oore"], "authorids": ["chandramouli.sastry@gmail.com", "osageev@gmail.com"], "authors_source": "OpenReview API", "abstract": "When presented with Out-of-Distribution (OOD) examples, deep neural networks yield confident, incorrect predictions. Detecting OOD examples is challenging, and the potential risks are high. In this paper, we propose to detect OOD examples by identifying inconsistencies between activity patterns and class predicted. We find that characterizing activity patterns by feature correlations and identifying anomalies in pairwise feature correlation values can yield high OOD detection rates. We identify anomalies in the pairwise feature correlations by simply comparing each pairwise correlation value with its respective range observed over the training data. Unlike many approaches, this can be used with any pre-trained softmax classifier and does not require access to OOD data for fine-tuning hyperparameters, nor does it require OOD access for inferring parameters. The method is applicable across a variety of architectures and vision datasets and generally performs better than or equal to state-of-the-art OOD detection methods, including those that do assume access to OOD examples.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=r1g6MCEtwr", "arxiv_id": null, "paper_pdf": "papers/r1g6MCEtwr.pdf", "paper_pdf_sha256": "64dbf01f5c4eb61ff9cb704e81da311f35b296e2ea4d588c85dcc8846d8430b1", "paper_pdf_bytes": 625825, "paper_pdf_source": "openreview", "code_url": "https://github.com/zeroshot-ood/ood-detection", "code_repository": "zeroshot-ood/ood-detection", "code_commit": "a7c08eae66b72109daf5b7f3e46754f5c583b64a", "code_archive": "repos/r1g6MCEtwr.zip", "code_archive_sha256": "42fd17cabc42fc84f758f0431dbad6b97bdda892eccb2d082c1409718305206e", "code_archive_bytes": 593891, "code_file_count": 8, "code_extensions": {".ipynb": 6, ".py": 2}, "github_disk_usage_kb": 585, "github_languages": {"Jupyter Notebook": 146932, "Python": 13512}, "github_archived": false, "github_pushed_at": "2020-01-19T23:38:04Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/zero-shot-out-of-distribution-detection-with"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "z1yPH2Rska", "year": 2026, "status": "rejected", "title": "Diagonal Batching Unlocks Parallelism in Recurrent Memory Transformers for Long Contexts", "authors": ["Danil Sivtsov", "Ivan Rodkin", "Gleb Kuzmin", "Yuri Kuratov", "Ivan Oseledets"], "authorids": ["~Danil_Sivtsov1", "~Ivan_Rodkin2", "~Gleb_Kuzmin1", "~Yuri_Kuratov2", "~Ivan_Oseledets1"], "authors_source": "OpenReview API", "abstract": "Long-context inference with Transformers is constrained by quadratic attention\nand linear memory growth. Many linear-time alternatives require pretraining from\nscratch, whereas Recurrent Memory Transformers (RMTs) convert pretrained \nmodels into segment-recurrent variants via finetuning without modifying the original\nmodel architecture. However, their sequential memory updates underutilize GPUs.\nWe show that RMT-style architectures with layer-level memory (PRMTs) \n(e.g., ARMT) can be among the most latency-efficient linear approaches when scheduled\nproperly. We introduce Diagonal Batching, a compute-reordering scheme that\npreserves exact recurrence while exposing inter-step parallelism by executing \n\"diagonals\" concurrently with grouped layers. On LLaMA (1B/3B/8B) up to 131,072\ntokens on A100/H100, Diagonal Batching achieves up to 3.3× lower latency than\nfull-attention inference and 1.8× over a sequential ARMT baseline, with no \ncustom CUDA kernels. With the right scheduling, PRMTs achieve linear scaling with\ncontext length and stand out as competitive, scalable architectures among linear\nrecurrent models.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=z1yPH2Rska", "arxiv_id": "2506.05229", "paper_pdf": "papers/z1yPH2Rska.pdf", "paper_pdf_sha256": "6fcb10e2ca0f5614dd017ab1ee764c42b648819267bb3fd2ae4e15c96d5e3752", "paper_pdf_bytes": 983470, "paper_pdf_source": "openreview", "code_url": "https://github.com/svtdanny/diagonal-batching", "code_repository": "svtdanny/diagonal-batching", "code_commit": "641ab9fc2c027e8b2614d76368e81cd2ee4bd1f8", "code_archive": "repos/z1yPH2Rska.zip", "code_archive_sha256": "f36de8bd1624bffc6cc5c85dcad131d5f8af721cf935866fdb42d7c7a616cded", "code_archive_bytes": 994323, "code_file_count": 30, "code_extensions": {".py": 20, ".ipynb": 10}, "github_disk_usage_kb": 1703, "github_languages": {"Jupyter Notebook": 1157285, "Python": 184303}, "github_archived": false, "github_pushed_at": "2025-09-02T19:39:37Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/diagonal-batching-unlocks-parallelism-in"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "WL4BmXG7Pl", "year": 2025, "status": "rejected", "title": "Crafting Heavy-Tails in Weight Matrix Spectrum without Gradient Noise", "authors": ["Vignesh Kothapalli", "Tianyu Pang", "Shenyang Deng", "Zongmin Liu", "Yaoqing Yang"], "authorids": ["~Vignesh_Kothapalli1", "~Tianyu_Pang2", "~Shenyang_Deng2", "~Zongmin_Liu1", "~Yaoqing_Yang1"], "authors_source": "OpenReview API", "abstract": "Training strategies for modern deep neural networks (NNs) tend to induce a heavy-\ntailed (HT) empirical spectral density (ESD) in the layer weights. While previous\nefforts have shown that the HT phenomenon correlates with good generalization\nin large NNs, a theoretical explanation of its occurrence is still lacking. Especially,\nunderstanding the conditions which lead to this phenomenon can shed light on the\ninterplay between generalization and weight spectra. Our work aims to bridge this\ngap by presenting a simple, rich setting to model the emergence of HT ESD. In\nparticular, we present a theory-informed analysis for 'crafting' heavy tails in the\nESD of two-layer NNs without any gradient noise. This is the first work to analyze a noise-free setting and incorporate optimizer (GD/Adam) dependent (large)\nlearning rates into the HT ESD analysis. Our results highlight the role of learning\nrates on the Bulk+Spike and HT shape of the ESDs in the early phase of training,\nwhich can facilitate generalization in the two-layer NN. These observations shed\nlight on the behavior of large-scale NNs, albeit in a much simpler setting. Last\nbut not least, we present a novel perspective on the ESD evolution dynamics by\nanalyzing the singular vectors of weight matrices and optimizer updates", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=WL4BmXG7Pl", "arxiv_id": "2406.04657", "paper_pdf": "papers/WL4BmXG7Pl.pdf", "paper_pdf_sha256": "52a01629a41d6a38a5fb40587fb51c30df93b1e880de7e186038dbf19afe3d1d", "paper_pdf_bytes": 6872743, "paper_pdf_source": "openreview", "code_url": "https://github.com/kvignesh1420/single-index-ht", "code_repository": "kvignesh1420/single-index-ht", "code_commit": "901ee634e80f2151d2b4f92c08acf3649fb2a871", "code_archive": "repos/WL4BmXG7Pl.zip", "code_archive_sha256": "1d64731b3b52d4f013d62b8f481d4c5f0e4fa1519f083d7d5765cb1a79c4933c", "code_archive_bytes": 25948, "code_file_count": 9, "code_extensions": {".py": 9}, "github_disk_usage_kb": 122, "github_languages": {"Python": 65585}, "github_archived": false, "github_pushed_at": "2026-03-23T20:08:59Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/crafting-heavy-tails-in-weight-matrix"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "CupHThqQl3", "year": 2024, "status": "rejected", "title": "It's About Time: Temporal References in Emergent Communication", "authors": ["Olaf Lipinski", "Adam Sobey", "Federico Cerutti", "Timothy J. Norman"], "authorids": ["~Olaf_Lipinski1", "~Adam_Sobey1", "~Federico_Cerutti1", "~Timothy_J._Norman1"], "authors_source": "OpenReview API", "abstract": "As humans, we use linguistic elements referencing time, such as “before” or “tomorrow”, to easily share past experiences and future predictions. While temporal aspects of the language have been considered in computational linguistics, no such exploration has been done within the field of emergent communication. We research this gap, providing the first reported temporal vocabulary within emergent communication literature. Our experimental analysis shows that a different agent architecture is sufficient for the natural emergence of temporal references, and that no additional losses are necessary. Our readily transferable architectural insights provide the basis for the incorporation of temporal referencing into other emergent communication environments.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=CupHThqQl3", "arxiv_id": "2310.06555", "paper_pdf": "papers/CupHThqQl3.pdf", "paper_pdf_sha256": "aee571dc23005cd4dfc81ac8e696e69737334e8af03a998c5c104858f5c85fa0", "paper_pdf_bytes": 1851232, "paper_pdf_source": "openreview", "code_url": "https://github.com/olipinski/TRG", "code_repository": "olipinski/TRG", "code_commit": "39cd5d55d5e35be3f204f447ce7d399f56a53bcd", "code_archive": "repos/CupHThqQl3.zip", "code_archive_sha256": "8f22ac4df9cdeae26c03427036a3c701082ec38c2b6e1ab7eebbfe12c30b0fef", "code_archive_bytes": 102281, "code_file_count": 12, "code_extensions": {".py": 9, ".ipynb": 2, ".sh": 1}, "github_disk_usage_kb": 93, "github_languages": {"Jupyter Notebook": 243580, "Python": 107779, "Shell": 1388}, "github_archived": false, "github_pushed_at": "2024-05-21T09:34:12Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/on-temporal-references-in-emergent"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "EJka_dVXEcr", "year": 2023, "status": "rejected", "title": "TabDDPM: Modelling Tabular Data with Diffusion Models", "authors": ["Akim Kotelnikov", "Dmitry Baranchuk", "Ivan Rubachev", "Artem Babenko"], "authorids": ["~Akim_Kotelnikov1", "~Dmitry_Baranchuk2", "~Ivan_Rubachev1", "~Artem_Babenko1"], "authors_source": "OpenReview API", "abstract": "Denoising diffusion probabilistic models are currently becoming the leading paradigm of generative modeling for many important data modalities. Being the most prevalent in the computer vision community, diffusion models have also recently gained some attention for other domains, including speech, NLP, and graph-like data. In this work, we investigate if the framework of diffusion models can be advantageous for general tabular problems, where datapoints are typically represented by vectors of heterogeneous features. The inherent heterogeneity of tabular data makes it quite challenging for accurate modeling, since the individual features can be of completely different nature, i.e., some of them can be continuous and some of them can be discrete. To address such data types, we introduce TabDDPM --- a diffusion model that can be universally applied to any tabular dataset and handles any types of features. We extensively evaluate TabDDPM on a wide set of benchmarks and demonstrate its superiority over existing GAN/VAE alternatives, which is consistent with the advantage of diffusion models in other fields. Additionally, we show that TabDDPM can be successfully used in privacy-oriented setups, where the original datapoints cannot be shared.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=EJka_dVXEcr", "arxiv_id": "2209.15421", "paper_pdf": "papers/EJka_dVXEcr.pdf", "paper_pdf_sha256": "3ee9d8d61be221ec9a504830143bb43e473548a25ac954725f7028c132375764", "paper_pdf_bytes": 512026, "paper_pdf_source": "openreview", "code_url": "https://github.com/yandex-research/tab-ddpm", "code_repository": "yandex-research/tab-ddpm", "code_commit": "b476257dd460b778ba09eb97f7a51d6490fa17f8", "code_archive": "repos/EJka_dVXEcr.zip", "code_archive_sha256": "2508743d42ec590f5eee24f83bd7fa7ce5183d6adc5280594e226b9fb124f8b4", "code_archive_bytes": 436897, "code_file_count": 73, "code_extensions": {".py": 72, ".ipynb": 1}, "github_disk_usage_kb": 187, "github_languages": {"Python": 504054, "Jupyter Notebook": 9799, "Makefile": 6617}, "github_archived": false, "github_pushed_at": "2024-07-13T04:02:14Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/tabddpm-modelling-tabular-data-with-diffusion"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "lgOylcEZQgr", "year": 2022, "status": "rejected", "title": "Online Unsupervised Learning of Visual Representations and Categories", "authors": ["Mengye Ren", "Tyler R. Scott", "Michael Louis Iuzzolino", "Michael Curtis Mozer", "Richard Zemel"], "authorids": ["~Mengye_Ren1", "~Tyler_R._Scott1", "~Michael_Louis_Iuzzolino1", "~Michael_Curtis_Mozer1", "~Richard_Zemel1"], "authors_source": "OpenReview API", "abstract": "Real world learning scenarios involve a nonstationary distribution of classes with sequential dependencies among the samples, in contrast to the standard machine learning formulation of drawing samples independently from a fixed, typically uniform distribution. Furthermore, real world interactions demand learning on-the-fly from few or no class labels. In this work, we propose an unsupervised model that simultaneously performs online visual representation learning and few-shot learning of new categories without relying on any class labels. Our model is a prototype-based memory network with a control component that determines when to form a new class prototype. We formulate it as an online Gaussian mixture model, where components are created online with only a single new example, and assignments do not have to be balanced, which permits an approximation to natural imbalanced distributions from uncurated raw data. Learning includes a contrastive loss that encourages different views of the same image to be assigned to the same prototype. The result is a mechanism that forms categorical representations of objects in nonstationary environments. Experiments show that our method can learn from an online stream of visual input data and is significantly better at category recognition compared to state-of-the-art self-supervised learning methods.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=lgOylcEZQgr", "arxiv_id": "2109.05675", "paper_pdf": "papers/lgOylcEZQgr.pdf", "paper_pdf_sha256": "7b223dc096107b56655e324fbf60babe884d0e02c855e23a32dd6da05433b6b2", "paper_pdf_bytes": 8423587, "paper_pdf_source": "openreview", "code_url": "https://github.com/renmengye/online-unsup-proto-net", "code_repository": "renmengye/online-unsup-proto-net", "code_commit": "67a14718fe5e00015d58694dcaa21a45bf5faf62", "code_archive": "repos/lgOylcEZQgr.zip", "code_archive_sha256": "58c459b4e90c5fd3b227d1a4aa6e09729f58eb76f2165d0ef5893c19bab740ba", "code_archive_bytes": 489318, "code_file_count": 227, "code_extensions": {".py": 210, ".sh": 17}, "github_disk_usage_kb": 350, "github_languages": {"Python": 1203470, "Shell": 9432, "Roff": 4498, "Dockerfile": 2835, "Makefile": 136}, "github_archived": false, "github_pushed_at": "2022-10-01T05:01:13Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/online-unsupervised-learning-of-visual"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "cQzf26aA3vM", "year": 2021, "status": "rejected", "title": "Design-Bench: Benchmarks for Data-Driven Offline Model-Based Optimization", "authors": ["Brandon Trabucco", "Aviral Kumar", "Xinyang Geng", "Sergey Levine"], "authorids": ["~Brandon_Trabucco1", "~Aviral_Kumar2", "~Xinyang_Geng1", "~Sergey_Levine1"], "authors_source": "OpenReview API", "abstract": "Black-box model-based optimization (MBO) problems, where the goal is to find a design input that maximizes an unknown objective function, are ubiquitous in a wide range of domains, such as the design of drugs, aircraft, and robot morphology. Typically, such problems are solved by actively querying the black-box objective on design proposals and using the resulting feedback to improve the proposed designs. However, when the true objective function is expensive or dangerous to evaluate in the real world, we might instead prefer a method that can optimize this function using only previously collected data, for example from a set of previously conducted experiments. This data-driven offline MBO set- ting presents a number of unique challenges, but a number of recent works have demonstrated that viable offline MBO methods can be developed even for high- dimensional problems, using high-capacity deep neural network function approximators. Unfortunately, the lack of standardized evaluation tasks in this emerg- ing new field has made tracking progress and comparing recent methods difficult. To address this problem, we present Design-Bench, a benchmark suite of offline MBO tasks with a unified evaluation protocol and reference implementations of recent methods. Our benchmark suite includes diverse and realistic tasks derived from real-world problems in biology, material science, and robotics that present distinct challenges for offline MBO methods. Our benchmarks, together with the reference implementations, are available at sites.google.com/view/design-bench. We hope that our benchmark can serve as a meaningful metric for the progress of offline MBO methods and guide future algorithmic development.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=cQzf26aA3vM", "arxiv_id": "2202.08450", "paper_pdf": "papers/cQzf26aA3vM.pdf", "paper_pdf_sha256": "450b763881dc0b9489a9b3578a45b0af980f0ac902b5b7b6a43c719174f8a8ef", "paper_pdf_bytes": 2179947, "paper_pdf_source": "openreview", "code_url": "https://github.com/rail-berkeley/design-bench", "code_repository": "rail-berkeley/design-bench", "code_commit": "95a495359eebbe2280c3ddbfcbdf80e611744fbf", "code_archive": "repos/cQzf26aA3vM.zip", "code_archive_sha256": "83b225ff667f89d1801bdc37695e341f361100776b9c535d0c3543a208c40100", "code_archive_bytes": 175757, "code_file_count": 69, "code_extensions": {".py": 69}, "github_disk_usage_kb": 447, "github_languages": {"Python": 687234}, "github_archived": false, "github_pushed_at": "2022-02-16T07:00:27Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/design-bench-benchmarks-for-data-driven-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "r1gzdhEKvH", "year": 2020, "status": "rejected", "title": "Neural Linear Bandits: Overcoming Catastrophic Forgetting through Likelihood Matching", "authors": ["Tom Zahavy", "Shie Mannor"], "authorids": ["tomzahavy@gmail.com", "shiemannor@gmail.com"], "authors_source": "OpenReview API", "abstract": "We study neural-linear bandits for solving problems where both exploration and representation learning play an important role. Neural-linear bandits leverage the representation power of deep neural networks and combine it with efficient exploration mechanisms, designed for linear contextual bandits, on top of the last hidden layer. Since the representation is being optimized during learning, information regarding exploration with \"old\" features is lost. Here, we propose the first limited memory neural-linear bandit that is resilient to this catastrophic forgetting phenomenon. We perform simulations on a variety of real-world problems, including regression, classification, and sentiment analysis, and observe that our algorithm achieves superior performance and shows resilience to catastrophic forgetting. ", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=r1gzdhEKvH", "arxiv_id": null, "paper_pdf": "papers/r1gzdhEKvH.pdf", "paper_pdf_sha256": "dfaafd1711639999570f040a73517687ef67cefe118bdcf26cce5ece56efbe91", "paper_pdf_bytes": 1775445, "paper_pdf_source": "openreview", "code_url": "https://github.com/anonymousneurips/neurips", "code_repository": "anonymousneurips/neurips", "code_commit": "3637ee7e10c9c83a6c89dacb4d4bfcac72c7175a", "code_archive": "repos/r1gzdhEKvH.zip", "code_archive_sha256": "506d7f77e522fe4ca94718cbd0cbc1b45608513f347d3f2ad473b14be3c7d649", "code_archive_bytes": 516554, "code_file_count": 23, "code_extensions": {".py": 23}, "github_disk_usage_kb": 600, "github_languages": {"Python": 190942}, "github_archived": false, "github_pushed_at": "2019-05-20T07:24:00Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/neural-linear-bandits-overcoming-catastrophic"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "kaW0pt8pLR", "year": 2026, "status": "rejected", "title": "Labelling Data with Unknown References", "authors": ["Adrian de Wynter"], "authorids": ["~Adrian_de_Wynter1"], "authors_source": "OpenReview API", "abstract": "An evaluator is trustworthy when there exists some agreed-upon way to measure its performance as a labeller. The two ways to establish trustworthiness are either by testing it, or by assuming the evaluator 'knows' somehow the way to label the corpus. However, if labelled references (e.g., a development set) are unavailable, neither of these approaches work: the former requires data, and the latter is an assumption, not evidence. To address this, we introduce an algorithm (the 'No-Data Algorithm') by which to establish trust in an evaluator without any existing references. Our algorithm works by successively posing challenges to said evaluator. We show that this is sufficient to establish trustworthiness w.h.p., in such a way that when the evaluator actually knows the way to label the corpus, the No-Data Algorithm accepts its output; and, conversely, flags untrustworthy evaluators when these are unable to prove it. We present formal proofs of correctness, empirical tests, and applications to LLMs-as-judges on low-resource languages.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=kaW0pt8pLR", "arxiv_id": "2506.03083", "paper_pdf": "papers/kaW0pt8pLR.pdf", "paper_pdf_sha256": "b7006a2fbab7b6305f77b358456382272f8d1fc221ef25ed89921b649332a92d", "paper_pdf_bytes": 591154, "paper_pdf_source": "openreview", "code_url": "https://github.com/adewynter/no_data_algorithm", "code_repository": "adewynter/no_data_algorithm", "code_commit": "541e58cc1fc39ac9307930ab6f5a49beaa7c4b12", "code_archive": "repos/kaW0pt8pLR.zip", "code_archive_sha256": "97f0f0d9bbad4f9ce924d309be5cb8d46031374362fe27b68291cbd6c642e0b0", "code_archive_bytes": 1764104, "code_file_count": 7, "code_extensions": {".py": 4, ".ipynb": 3}, "github_disk_usage_kb": 1808, "github_languages": {"Jupyter Notebook": 243131, "Python": 32965}, "github_archived": false, "github_pushed_at": "2026-05-05T03:44:49Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/labelling-data-with-unknown-references"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "sMWkTWh2JF", "year": 2025, "status": "rejected", "title": "ReLIC: A Recipe for 64k Steps of In-Context Reinforcement Learning for Embodied AI", "authors": ["Ahmad Elawady", "Gunjan Chhablani", "Ram Ramrakhya", "Karmesh Yadav", "Dhruv Batra", "Zsolt Kira", "Andrew Szot"], "authorids": ["~Ahmad_Elawady1", "~Gunjan_Chhablani1", "~Ram_Ramrakhya2", "~Karmesh_Yadav1", "~Dhruv_Batra1", "~Zsolt_Kira1", "~Andrew_Szot1"], "authors_source": "OpenReview API", "abstract": "Intelligent embodied agents need to quickly adapt to new scenarios by integrating long histories of experience into decision-making. For instance, a robot in an unfamiliar house initially wouldn't know the locations of objects needed for tasks and might perform inefficiently. However, as it gathers more experience, it should learn the layout of its environment and remember where objects are, allowing it to complete new tasks more efficiently. To enable such rapid adaptation to new tasks, we present ReLIC, a new approach for in-context reinforcement learning (RL) for embodied agents. With ReLIC, agents are capable of adapting to new environments using 64,000 steps of in-context experience with full attention while being trained through self-generated experience via RL. We achieve this by proposing a novel policy update scheme for on-policy RL called \"partial updates\" as well as a Sink-KV mechanism that enables effective utilization of a long observation history for embodied agents. Our method outperforms a variety of meta-RL baselines in adapting to unseen houses in an embodied multi-object navigation task. In addition, we find that ReLIC is capable of few-shot imitation learning despite never being trained with expert demonstrations. We also provide a comprehensive analysis of ReLIC, highlighting that the combination of large-scale RL training, the proposed partial updates scheme, and the Sink-KV are essential for effective in-context learning.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=sMWkTWh2JF", "arxiv_id": "2410.02751", "paper_pdf": "papers/sMWkTWh2JF.pdf", "paper_pdf_sha256": "09de2695265e86d91aa57c9322faf8a22bde8912bd09dfca93e10739c4fc1414", "paper_pdf_bytes": 3613664, "paper_pdf_source": "openreview", "code_url": "https://github.com/aielawady/relic", "code_repository": "aielawady/relic", "code_commit": "1c3b7a13b2a2c5873223de2680b6cd5ed6a8fc96", "code_archive": "repos/sMWkTWh2JF.zip", "code_archive_sha256": "b35d95bb1976d23422c72c678a37d6636856ee5a5680751e57fc86f513f88214", "code_archive_bytes": 155858, "code_file_count": 41, "code_extensions": {".py": 39, ".sh": 2}, "github_disk_usage_kb": 123, "github_languages": {"Python": 536329, "Shell": 3350}, "github_archived": false, "github_pushed_at": "2024-09-07T21:56:12Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/relic-a-recipe-for-64k-steps-of-in-context"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "8w6FzR68DS", "year": 2024, "status": "rejected", "title": "PriViT: Vision Transformers for Fast Private Inference", "authors": ["Naren Dhyani", "Jianqiao Cambridge Mo", "Minsu Cho", "Ameya Joshi", "Siddharth Garg", "Brandon Reagen", "Chinmay Hegde"], "authorids": ["~Naren_Dhyani1", "~Jianqiao_Cambridge_Mo1", "~Minsu_Cho2", "~Ameya_Joshi2", "~Siddharth_Garg1", "~Brandon_Reagen1", "~Chinmay_Hegde1"], "authors_source": "OpenReview API", "abstract": "The Vision Transformer (ViT) architecture has emerged as the backbone of choice for state-of-the-art deep models for computer vision applications. However, ViTs are ill-suited for private inference using secure multi-party computation (MPC) protocols, due to the large number of non-polynomial operations (self-attention, feed-forward rectifiers, layer normalization). We propose PriViT, a gradient-based algorithm to selectively Taylorize nonlinearities in ViTs while maintaining their prediction accuracy. Our algorithm is conceptually simple, easy to implement, and achieves improved performance over existing approaches for designing MPC-friendly transformer architectures in terms of achieving the Pareto frontier in latency-accuracy. We confirm these improvements via experiments on several standard image classification tasks.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=8w6FzR68DS", "arxiv_id": "2310.04604", "paper_pdf": "papers/8w6FzR68DS.pdf", "paper_pdf_sha256": "5a132caef3a0e0475bc4d157c4529b71fb4b56a5e898a86a2c8184082e44ae7a", "paper_pdf_bytes": 686634, "paper_pdf_source": "openreview", "code_url": "https://github.com/NYU-DICE-Lab/privit", "code_repository": "NYU-DICE-Lab/privit", "code_commit": "6e4ef9071761ea25bb496ac989dc5c25256a1c71", "code_archive": "repos/8w6FzR68DS.zip", "code_archive_sha256": "75f261a43e2c687fade1dcd0336dc3a19488c823228b13851a6e94e1d3ac8245", "code_archive_bytes": 99188, "code_file_count": 18, "code_extensions": {".py": 18}, "github_disk_usage_kb": 94, "github_languages": {"Python": 142319, "Shell": 2217}, "github_archived": false, "github_pushed_at": "2024-09-10T16:06:04Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/privit-vision-transformers-for-fast-private"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "TtxOsdYU92d", "year": 2023, "status": "rejected", "title": "Generalized Sum Pooling for Metric Learning", "authors": ["Yeti Z. Gürbüz", "Ozan Sener", "A. Aydın Alatan"], "authorids": ["~Yeti_Z._Gürbüz1", "~Ozan_Sener1", "~A._Aydın_Alatan1"], "authors_source": "OpenReview API", "abstract": "A common architectural choice for deep metric learning is a convolutional neural network followed by global average pooling (GAP). Albeit simple, GAP is a highly effective way to aggregate information. One possible explanation for the effectiveness of GAP is considering each feature vector as representing a different semantic entity and GAP as a convex combination of them. Following this perspective, we generalize GAP and propose a learnable generalized sum pooling method (GSP). GSP improves GAP with two distinct abilities: i) the ability to choose a subset of semantic entities, effectively learning to ignore nuisance information, and ii) learning the weights corresponding to the importance of each entity. Formally, we propose an entropy-smoothed optimal transport problem and show that it is a strict generalization of GAP, \\ie a specific realization of the problem gives back GAP. We show that this optimization problem enjoys analytical gradients enabling us to use it as a direct learnable replacement for GAP. We further propose a zero-shot loss to ease the learning of GSP. We show the effectiveness of our method with extensive evaluations on 4 popular metric learning benchmarks. Code is available at: GSP-DML Framework", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=TtxOsdYU92d", "arxiv_id": "2308.09228", "paper_pdf": "papers/TtxOsdYU92d.pdf", "paper_pdf_sha256": "710ea533c1d9a337fc2326f9dea6804c7011cebe671029b572119f63131654f5", "paper_pdf_bytes": 4033299, "paper_pdf_source": "openreview", "code_url": "https://github.com/yetigurbuz/generalized-sum-pooling", "code_repository": "yetigurbuz/generalized-sum-pooling", "code_commit": "b765b8890f82bc01e55ed32e7432a0fd71a1d4a6", "code_archive": "repos/TtxOsdYU92d.zip", "code_archive_sha256": "b0892871dd4a27aa7557a04d84ecfcdbcfb1785091e4d09f9ee9dee2c5715fe4", "code_archive_bytes": 169913, "code_file_count": 93, "code_extensions": {".py": 92, ".sh": 1}, "github_disk_usage_kb": 191, "github_languages": {"Python": 556723, "Shell": 721}, "github_archived": false, "github_pushed_at": "2023-08-12T15:51:04Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/generalized-sum-pooling-for-metric-learning"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ZDYhm_o8MX", "year": 2022, "status": "rejected", "title": "Neural Manifold Clustering and Embedding", "authors": ["ZENGYI LI", "Yubei Chen", "Yann LeCun", "Friedrich Sommer"], "authorids": ["~ZENGYI_LI1", "~Yubei_Chen1", "~Yann_LeCun1", "~Friedrich_Sommer1"], "authors_source": "OpenReview API", "abstract": "Given a union of non-linear manifolds, non-linear subspace clustering or manifold clustering aims to cluster data points based on manifold structures and also learn to parameterize each manifold as a linear subspace in a feature space. Deep neural networks have the potential to achieve this goal under highly non-linear settings given their large capacity and flexibility. We argue that achieving manifold clustering with neural networks requires two essential ingredients: a domain-specific constraint that ensures the identification of the manifolds, and a learning algorithm for embedding each manifold to a linear subspace in the feature space. This work shows that many constraints can be implemented by data augmentation. For subspace feature learning, Maximum Coding Rate Reduction (MCR$^2$) objective can be used. Putting them together yields Neural Manifold Clustering and Embedding (NMCE), a novel method for general purpose manifold clustering, which significantly outperforms autoencoder-based deep subspace clustering and achieve state-of-the-art performance on several important benchmarks. Further, on more challenging natural image datasets, NMCE can also outperform other algorithms specifically designed for clustering. Qualitatively, we demonstrate that NMCE learns a meaningful and interpretable feature space. As the formulation of NMCE is closely related to several important Self-supervised learning (SSL) methods, we believe this work can help us build a deep understanding on SSL representation learning.", "decision": "Reject", "meta_review": null, "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=ZDYhm_o8MX", "arxiv_id": "2201.10000", "paper_pdf": "papers/ZDYhm_o8MX.pdf", "paper_pdf_sha256": "acfe723950465e0b5422dbfb67a423935a02243501be2ff6db97457b30ded9ae", "paper_pdf_bytes": 4647410, "paper_pdf_source": "openreview", "code_url": "https://github.com/zengyi-li/NMCE-release", "code_repository": "zengyi-li/NMCE-release", "code_commit": "b60a05eedcaa115b0efeb45686bfae865c0398de", "code_archive": "repos/ZDYhm_o8MX.zip", "code_archive_sha256": "f80ae0ffa96aba9ffe8b69c2e568223a2bbb52f491a8ede503c93308fd05553d", "code_archive_bytes": 368323, "code_file_count": 19, "code_extensions": {".py": 18, ".ipynb": 1}, "github_disk_usage_kb": 352, "github_languages": {"Python": 123762, "Jupyter Notebook": 66360}, "github_archived": false, "github_pushed_at": "2022-03-11T01:52:42Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/neural-manifold-clustering-and-embedding-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "eYgI3cTPTq9", "year": 2021, "status": "rejected", "title": "How to Avoid Being Eaten by a Grue: Structured Exploration Strategies for Textual Worlds", "authors": ["Prithviraj Ammanabrolu", "Ethan Tien", "Matthew Hausknecht", "Mark Riedl"], "authorids": ["~Prithviraj_Ammanabrolu1", "~Ethan_Tien1", "~Matthew_Hausknecht1", "~Mark_Riedl1"], "authors_source": "OpenReview API", "abstract": "Text-based games are long puzzles or quests, characterized by a sequence of sparse and potentially deceptive rewards. They provide an ideal platform to develop agents that perceive and act upon the world using a combinatorially sized natural language state-action space. Standard Reinforcement Learning agents are poorly equipped to effectively explore such spaces and often struggle to overcome bottlenecks---states that agents are unable to pass through simply because they do not see the right action sequence enough times to be sufficiently reinforced. We introduce Q*BERT, an agent that learns to build a knowledge graph of the world by answering questions, which leads to greater sample efficiency. To overcome bottlenecks, we further introduce MC!Q*BERT an agent that uses an knowledge-graph-based intrinsic motivation to detect bottlenecks and a novel exploration strategy to efficiently learn a chain of policy modules to overcome them. We present an ablation study and results demonstrating how our method outperforms the current state-of-the-art on nine text games, including the popular game, Zork, where, for the first time, a learning agent gets past the bottleneck where the player is eaten by a Grue.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=eYgI3cTPTq9", "arxiv_id": "2006.07409", "paper_pdf": "papers/eYgI3cTPTq9.pdf", "paper_pdf_sha256": "22cbfa152665c70a38184813aa508ec3b6f55e956997a95c2391ceee2e7c9b64", "paper_pdf_bytes": 3161716, "paper_pdf_source": "openreview", "code_url": "https://github.com/rajammanabrolu/Q-BERT", "code_repository": "rajammanabrolu/Q-BERT", "code_commit": "9ca2df39d9ab5968f7a5034381483cfcef122a03", "code_archive": "repos/eYgI3cTPTq9.zip", "code_archive_sha256": "de8dd37eaf5ee9acc6a63cb6283c529153cbbfd3644ac43dc267c629429dea8e", "code_archive_bytes": 484483, "code_file_count": 26, "code_extensions": {".py": 26}, "github_disk_usage_kb": 481, "github_languages": {"Python": 333890}, "github_archived": false, "github_pushed_at": "2023-08-09T22:03:56Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/how-to-avoid-being-eaten-by-a-grue-structured"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "BJlEEaEFDS", "year": 2020, "status": "rejected", "title": "Towards an Adversarially Robust Normalization Approach", "authors": ["Muhammad Awais", "Fahad Shamshad", "Sung-Ho Bae"], "authorids": ["awais@khu.ac.kr", "fahad.shamshad@itu.edu.pk", "shbae@khu.ac.kr"], "authors_source": "OpenReview API", "abstract": "Batch Normalization (BatchNorm) has shown to be effective for improving and accelerating the training of deep neural networks. However, recently it has been shown that it is also vulnerable to adversarial perturbations. In this work, we aim to investigate the cause of adversarial vulnerability of the BatchNorm. We hypothesize that the use of different normalization statistics during training and inference (mini-batch statistics for training and moving average of these values at inference) is the main cause of this adversarial vulnerability in the BatchNorm layer. We empirically proved this by experiments on various neural network architectures and datasets. Furthermore, we introduce Robust Normalization (RobustNorm) and experimentally show that it is not only resilient to adversarial perturbation but also inherit the benefits of BatchNorm.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=BJlEEaEFDS", "arxiv_id": "2006.11007", "paper_pdf": "papers/BJlEEaEFDS.pdf", "paper_pdf_sha256": "31a33ee05aae850fda5e8fa6aaed89ea747a04cbfd682d7e7b9cb8038b29095c", "paper_pdf_bytes": 966068, "paper_pdf_source": "openreview", "code_url": "https://github.com/awaisrauf/RobustNorm", "code_repository": "awaisrauf/RobustNorm", "code_commit": "d14d68cb8990be3c50b387e08ec72dd20a01dba2", "code_archive": "repos/BJlEEaEFDS.zip", "code_archive_sha256": "36d6149294885a8be0a93e3d7f5b5d936a5ced6225d29dd93240db589baed0fd", "code_archive_bytes": 656239, "code_file_count": 29, "code_extensions": {".py": 28, ".sh": 1}, "github_disk_usage_kb": 633, "github_languages": {"Python": 121993, "Shell": 355}, "github_archived": false, "github_pushed_at": "2020-10-06T05:31:45Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/towards-an-adversarially-robust-normalization-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "14vBE8iAEx", "year": 2026, "status": "rejected", "title": "Beyond Reward Hacking: Causal Rewards for Large Language Model Alignment", "authors": ["Chaoqi Wang", "Zhuokai Zhao", "Yibo Jiang", "Zhaorun Chen", "Chen Zhu", "Yuxin Chen", "Jiayi Liu", "Lizhu Zhang", "Xiangjun Fan", "Hao Ma", "Sinong Wang"], "authorids": ["~Chaoqi_Wang1", "~Zhuokai_Zhao1", "~Yibo_Jiang2", "~Zhaorun_Chen1", "~Chen_Zhu2", "~Yuxin_Chen1", "~Jiayi_Liu1", "~Lizhu_Zhang2", "~Xiangjun_Fan1", "~Hao_Ma1", "~Sinong_Wang1"], "authors_source": "OpenReview API", "abstract": "Recent advances in large language models (LLMs) have demonstrated significant progress in performing complex tasks. While Reinforcement Learning from Human Feedback (RLHF) has been effective in aligning LLMs with human preferences, it is susceptible to spurious correlations in reward modeling. Consequently, it often introduces biases—such as length bias, sycophancy, conceptual bias, and discrimination—that hinder the model’s ability to capture true causal relationships. To address this, we propose a novel causal reward modeling approach that integrates causality to mitigate these spurious correlations. Our method enforces counterfactual invariance, ensuring reward predictions remain consistent when irrelevant variables are altered. Through experiments on both synthetic and real-world datasets, we show that our approach mitigates various types of spurious correlations effectively, resulting in more reliable and fair alignment of LLMs with human preferences. As a drop-in enhancement to the existing RLHF workflow, our causal reward modeling provides a practical way to improve the trustworthiness and fairness of LLM finetuning.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=14vBE8iAEx", "arxiv_id": "2501.09620", "paper_pdf": "papers/14vBE8iAEx.pdf", "paper_pdf_sha256": "b7584a51aa6bbddfb257383c20c40dbcc3c05db84687be2f9ed3cb76022a71c8", "paper_pdf_bytes": 425357, "paper_pdf_source": "openreview", "code_url": "https://github.com/tatsu-lab/alpaca_farm", "code_repository": "tatsu-lab/alpaca_farm", "code_commit": "30717ddae735365de756ee2085191b491a71788d", "code_archive": "repos/14vBE8iAEx.zip", "code_archive_sha256": "3c56a0b5a4a8d6965665e4914e578041c188ab617a6b6473d43496e2e02bc779", "code_archive_bytes": 2205761, "code_file_count": 59, "code_extensions": {".py": 51, ".sh": 6, ".js": 1, ".ipynb": 1}, "github_disk_usage_kb": 1895, "github_languages": {"Python": 261484, "JavaScript": 69}, "github_archived": false, "github_pushed_at": "2024-07-01T18:53:35Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/beyond-reward-hacking-causal-rewards-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "2Oh2EOcFSO", "year": 2025, "status": "rejected", "title": "Can a Bayesian oracle prevent harm from an agent?", "authors": ["Yoshua Bengio", "Michael K. Cohen", "Nikolay Malkin", "Matt MacDermott", "Damiano Fornasiere", "Pietro Greiner", "Younesse Kaddar"], "authorids": ["~Yoshua_Bengio1", "~Michael_K._Cohen1", "~Nikolay_Malkin1", "~Matt_MacDermott1", "~Damiano_Fornasiere1", "~Pietro_Greiner1", "~Younesse_Kaddar1"], "authors_source": "OpenReview API", "abstract": "Is there a way to design powerful AI systems based on machine learning methods that would satisfy probabilistic safety guarantees? With the long-term goal of obtaining a probabilistic guarantee that would apply in every context, we consider estimating a context-dependent bound on the probability of violating a given safety specification.  Such a risk evaluation would need to be performed at run-time to provide a guardrail against dangerous actions of an AI. Noting that different plausible hypotheses about the world could produce very different outcomes, and because we do not know which one is right, we derive bounds on the safety violation probability predicted  under the true but unknown hypothesis. Such bounds could be used to reject potentially dangerous actions. Our main results involve searching for cautious but plausible hypotheses, obtained by a maximization that involves Bayesian posteriors over hypotheses. We consider two forms of this result, in the i.i.d. case and in the non-i.i.d. case, and conclude with open problems towards turning such theoretical results into practical AI guardrails.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=2Oh2EOcFSO", "arxiv_id": "2408.05284", "paper_pdf": "papers/2Oh2EOcFSO.pdf", "paper_pdf_sha256": "52a00b3b306bcd160fdc4e40a990c855137743b90f1fd91704818d0062a5639a", "paper_pdf_bytes": 424788, "paper_pdf_source": "openreview", "code_url": "https://github.com/saifh-github/conservative-bayesian-public", "code_repository": "saifh-github/conservative-bayesian-public", "code_commit": "7ab10e0654b8675536e369f5064e58c18cadb582", "code_archive": "repos/2Oh2EOcFSO.zip", "code_archive_sha256": "d403e6dfe2036c31e40d8a5ce32a4aa8a752e10a49f55b35a93e454a50e888a3", "code_archive_bytes": 40511, "code_file_count": 15, "code_extensions": {".py": 13, ".sh": 2}, "github_disk_usage_kb": 129, "github_languages": {"Python": 112121, "Shell": 1231}, "github_archived": false, "github_pushed_at": "2025-05-30T12:23:33Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/can-a-bayesian-oracle-prevent-harm-from-an"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "o0C2v4xTdS", "year": 2024, "status": "rejected", "title": "CoarsenConf: Equivariant Coarsening with Aggregated Attention for Molecular Conformer Generation", "authors": ["Danny Reidenbach", "Aditi S. Krishnapriyan"], "authorids": ["~Danny_Reidenbach1", "~Aditi_S._Krishnapriyan1"], "authors_source": "OpenReview API", "abstract": "Molecular conformer generation (MCG) is an important task in cheminformatics and drug discovery. The ability to efficiently generate low-energy 3D structures can avoid expensive quantum mechanical simulations, leading to accelerated virtual screenings and enhanced structural exploration. Several generative models have been developed for MCG, but many struggle to consistently produce high-quality conformers. To address these issues, we introduce CoarsenConf, which coarse-grains molecular graphs based on torsional angles and integrates them into an SE(3)-equivariant hierarchical variational autoencoder. Through equivariant coarse-graining, we aggregate the fine-grained atomic coordinates of subgraphs connected via rotatable bonds, creating a variable-length coarse-grained latent representation. Our model uses a novel aggregated attention mechanism to restore fine-grained coordinates from the coarse-grained latent representation, enabling efficient generation of accurate conformers. Furthermore, we evaluate the chemical and biochemical quality of our generated conformers on multiple downstream applications, including property prediction and oracle-based protein docking. Overall, CoarsenConf generates more accurate conformer ensembles compared to prior generative models.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=o0C2v4xTdS", "arxiv_id": "2306.14852", "paper_pdf": "papers/o0C2v4xTdS.pdf", "paper_pdf_sha256": "80351d2da14573ec35251b8a919633546f798c071d37dc031488a118deac1862", "paper_pdf_bytes": 1982418, "paper_pdf_source": "openreview", "code_url": "https://github.com/ASK-Berkeley/CoarsenConf", "code_repository": "ASK-Berkeley/CoarsenConf", "code_commit": "1751189fbfc9874515f8bdd396b586a3ac2ff53b", "code_archive": "repos/o0C2v4xTdS.zip", "code_archive_sha256": "7260925cb09875ca2ee4a1b96415498c49cf1a3ddf7d4f2390c3f45c5b66fdf0", "code_archive_bytes": 120594, "code_file_count": 20, "code_extensions": {".py": 20}, "github_disk_usage_kb": 99, "github_languages": {"Python": 486162}, "github_archived": false, "github_pushed_at": "2024-10-24T06:59:35Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/coarsenconf-equivariant-coarsening-with"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "WFBksaezAs", "year": 2023, "status": "rejected", "title": "Multi-Prompt Alignment for Multi-source Unsupervised Domain Adaptation", "authors": ["Haoran Chen", "Zuxuan Wu", "Yu-Gang Jiang"], "authorids": ["~Haoran_Chen4", "~Zuxuan_Wu1", "~Yu-Gang_Jiang1"], "authors_source": "OpenReview API", "abstract": "Most existing methods for multi-source unsupervised domain adaptation (UDA) rely on a common feature encoder to extract domain-invariant features. However, learning such an encoder involves updating the parameters of the entire network, which makes the optimization computationally expensive, particularly when coupled with min-max objectives. Inspired by recent advances in prompt learning that adapts high-capacity deep models for downstream tasks in a computationally economic way, we introduce Multi-Prompt Alignment (MPA), a simple yet efficient two-stage framework for multi-source UDA. Given a source and target domain pair, MPA first trains an individual prompt to minimize the domain gap through a contrastive loss, while tuning only a small set of parameters. Then, MPA derives a low-dimensional latent space through an auto-encoding process that maximizes the agreement of multiple learned prompts. The resulting embedding further facilitates generalization to unseen domains. Extensive experiments show that our method achieves state-of-the-art results on popular benchmark datasets while requiring substantially fewer tunable parameters. To the best of our knowledge, we are the first to apply prompt learning to the multi-source UDA problem and our method achieves the highest reported average accuracy of 54.1% on DomainNet, the most challenging UDA dataset to date, with only 15.9M parameters trained. More importantly, we demonstrate that the learned embedding space can be easily adapted to novel unseen domains.", "decision": "Reject", "meta_review": null, "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=WFBksaezAs", "arxiv_id": "2209.15210", "paper_pdf": "papers/WFBksaezAs.pdf", "paper_pdf_sha256": "c5cf38a8094aabbb98d1acd0a4f00b7e3b9da1fe36a08d263262524ef7d4591b", "paper_pdf_bytes": 1341029, "paper_pdf_source": "openreview", "code_url": "https://github.com/HaoranChen/Multi-Prompt-Alignment-for-MSUDA", "code_repository": "HaoranChen/Multi-Prompt-Alignment-for-MSUDA", "code_commit": "a80a7eb7dec9a7d18a90db70c7c3677d95a608e2", "code_archive": "repos/WFBksaezAs.zip", "code_archive_sha256": "cbfaf48bee95e9aca9be1866bb8d3c09d80955b2e44c7ec60d4a0c4e6827ca51", "code_archive_bytes": 182450, "code_file_count": 9, "code_extensions": {".py": 9}, "github_disk_usage_kb": 193, "github_languages": {"Python": 56463}, "github_archived": false, "github_pushed_at": "2025-01-26T14:29:03Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/multi-prompt-alignment-for-multi-source"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "PtuQ8bk9xF5", "year": 2022, "status": "rejected", "title": "Learning to Act with Affordance-Aware Multimodal Neural SLAM", "authors": ["Zhiwei Jia", "Kaixiang Lin", "Yizhou Zhao", "Qiaozi Gao", "Govind Thattai", "Gaurav S. Sukhatme"], "authorids": ["~Zhiwei_Jia1", "~Kaixiang_Lin1", "~Yizhou_Zhao1", "~Qiaozi_Gao1", "~Govind_Thattai1", "~Gaurav_S._Sukhatme1"], "authors_source": "OpenReview API", "abstract": "Recent years have witnessed an emerging paradigm shift toward embodied artificial intelligence, in which an agent must learn to solve challenging tasks by interacting with its environment. There are several challenges in solving embodied multimodal tasks, including long-horizon planning, vision-and-language grounding, and efficient exploration. We focus on a critical bottleneck, namely the performance of planning and navigation. To tackle this challenge, we propose a Neural SLAM approach that, for the first time, utilizes several modalities for exploration, predicts an affordance-aware semantic map, and plans over it at the same time. This significantly improves exploration efficiency, leads to robust long-horizon planning, and enables effective vision-and-language grounding. With the proposed Affordance-aware Multimodal Neural SLAM (AMSLAM) approach, we obtain more than 40% improvement over prior published work on the ALFRED benchmark and set a new state-of-the-art generalization performance at a success rate of 23.48% on the test unseen scenes.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=PtuQ8bk9xF5", "arxiv_id": "2201.09862", "paper_pdf": "papers/PtuQ8bk9xF5.pdf", "paper_pdf_sha256": "03966d3ae031a91e1ebbb81d190f98b28a8ed1d7ee7207d0904ed62cc0c3cff4", "paper_pdf_bytes": 3882110, "paper_pdf_source": "openreview", "code_url": "https://github.com/amazon-science/multimodal-neuralslam", "code_repository": "amazon-science/multimodal-neuralslam", "code_commit": "530558fdfa31c6e048fc3e7b253f681f6786b04d", "code_archive": "repos/PtuQ8bk9xF5.zip", "code_archive_sha256": "13c5ccbc0c43adc841eb6a20339c1aeae5030e61be3cd0117b4e5c99e9077800", "code_archive_bytes": 529752, "code_file_count": 91, "code_extensions": {".py": 68, ".c": 11, ".h": 11, ".sh": 1}, "github_disk_usage_kb": 367, "github_languages": {"Python": 598816, "C": 568481, "Yacc": 38288, "Lex": 5270, "Makefile": 1465, "Shell": 81}, "github_archived": false, "github_pushed_at": "2021-09-18T23:00:50Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/learning-to-act-with-affordance-aware-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ot9bYHvuULl", "year": 2021, "status": "rejected", "title": "Augmented Sliced Wasserstein Distances", "authors": ["Xiongjie Chen", "Yongxin Yang", "Yunpeng Li"], "authorids": ["~Xiongjie_Chen1", "~Yongxin_Yang1", "~Yunpeng_Li1"], "authors_source": "OpenReview API", "abstract": "While theoretically appealing, the application of the Wasserstein distance to large-scale machine learning problems has been hampered by its prohibitive computational cost. The sliced Wasserstein distance and its variants improve the computational efficiency through random projection, yet they suffer from low projection efficiency because the majority of projections result in trivially small values. In this work, we propose a new family of distance metrics, called augmented sliced Wasserstein distances (ASWDs), constructed by first mapping samples to higher-dimensional hypersurfaces parameterized by neural networks. It is derived from a key observation that (random) linear projections of samples residing on these hypersurfaces would translate to much more flexible nonlinear projections in the original sample space, so they can capture complex structures of the data distribution. We show that the hypersurfaces can be optimized by gradient ascent efficiently. We provide the condition under which the ASWD is a valid metric and show that this can be obtained by an injective neural network architecture. Numerical results demonstrate that the ASWD significantly outperforms other Wasserstein variants for both synthetic and real-world problems.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=ot9bYHvuULl", "arxiv_id": "2006.08812", "paper_pdf": "papers/ot9bYHvuULl.pdf", "paper_pdf_sha256": "c70413f8c6074aa320d42b715a68a9de03eb56b6ca45aa6e5687ccc1f194ab49", "paper_pdf_bytes": 1617521, "paper_pdf_source": "openreview", "code_url": "https://github.com/xiongjiechen/ASWD", "code_repository": "xiongjiechen/ASWD", "code_commit": "86ebb8303c7baeda56a903f9598928b70986e07b", "code_archive": "repos/ot9bYHvuULl.zip", "code_archive_sha256": "b5fb4974069f46995bffb6885513d0f7b7d09c2757baf064f7366b3b1e5085dc", "code_archive_bytes": 492580, "code_file_count": 8, "code_extensions": {".py": 7, ".ipynb": 1}, "github_disk_usage_kb": 492, "github_languages": {"Python": 65200, "Jupyter Notebook": 35383}, "github_archived": false, "github_pushed_at": "2025-06-14T12:47:37Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/augmented-sliced-wasserstein-distances"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "BJgctpEKwr", "year": 2020, "status": "rejected", "title": "RPGAN: random paths as a latent space for GAN interpretability", "authors": ["Andrey Voynov", "Artem Babenko"], "authorids": ["an.voynov@gmail.com", "artem.babenko@phystech.edu"], "authors_source": "OpenReview API", "abstract": "In this paper, we introduce Random Path Generative Adversarial Network (RPGAN) --- an alternative scheme of GANs that can serve as a tool for generative model analysis. While the latent space of a typical GAN consists of input vectors, randomly sampled from the standard Gaussian distribution, the latent space of RPGAN consists of random paths in a generator network. As we show, this design allows to associate different layers of the generator with different regions of the latent space, providing their natural interpretability. With experiments on standard benchmarks, we demonstrate that RPGAN reveals several interesting insights about roles that different layers play in the image generation process. Aside from interpretability, the RPGAN model also provides competitive generation quality and allows efficient incremental learning on new data.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=BJgctpEKwr", "arxiv_id": null, "paper_pdf": "papers/BJgctpEKwr.pdf", "paper_pdf_sha256": "f14fb66da947161c58af4aca74fb9eb40b7029241227b72edb6b326d3c6acbdd", "paper_pdf_bytes": 23078746, "paper_pdf_source": "openreview", "code_url": "https://github.com/rpgan-ICLR2020/RPGAN", "code_repository": "rpgan-ICLR2020/RPGAN", "code_commit": "62b4e4be24d012be0191e8290f9b1b249e0e1b7f", "code_archive": "repos/BJgctpEKwr.zip", "code_archive_sha256": "1d29eac546de74a4cb5a369d5f2f0a297968de59bf2c4b5201d07de220e64847", "code_archive_bytes": 662765, "code_file_count": 27, "code_extensions": {".py": 27}, "github_disk_usage_kb": 640, "github_languages": {"Python": 86801}, "github_archived": false, "github_pushed_at": "2023-07-06T21:42:22Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/rpgan-random-paths-as-a-latent-space-for-gan"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "BFb4ACHayj", "year": 2026, "status": "rejected", "title": "SafeAgentBench: A Benchmark for Safe Task Planning of Embodied LLM Agents", "authors": ["Sheng Yin", "Xianghe Pang", "Yuanzhuo Ding", "Menglan Chen", "Yutong Bi", "Yichen Xiong", "Wenhao Huang", "Zhen Xiang", "Jing Shao", "Siheng Chen"], "authorids": ["~Sheng_Yin1", "~Xianghe_Pang1", "~Yuanzhuo_Ding1", "~Menglan_Chen1", "~Yutong_Bi1", "~Yichen_Xiong1", "~Wenhao_Huang4", "~Zhen_Xiang1", "~Jing_Shao3", "~Siheng_Chen1"], "authors_source": "OpenReview API", "abstract": "With the integration of large language models (LLMs), embodied agents have strong capabilities to understand and plan complicated natural language instructions. However, a foreseeable issue is that those embodied agents can also flawlessly execute some hazardous tasks, potentially causing damages in the real world. Existing benchmarks predominantly overlook critical safety risks, focusing solely on planning performance, while a few evaluate LLMs' safety awareness only on non-interactive image-text data. To address this gap, we present \\textbf{SafeAgentBench}—the first comprehensive benchmark for safety-aware task planning of embodied LLM agents in interactive simulation environments, covering both explicit and implicit hazards. SafeAgentBench includes: (1) an executable, diverse, and high-quality dataset of 750 tasks, rigorously curated to cover 10 potential hazards and 3 task types; (2) SafeAgentEnv, a universal embodied environment with a low-level controller, supporting multi-agent execution with 17 high-level actions for 9 state-of-the-art baselines; and (3) reliable evaluation methods from both execution and semantic perspectives. Experimental results show that, although agents based on different design frameworks exhibit substantial differences in task success rates, their overall safety awareness remains weak. The most safety-conscious baseline achieves only a 10\\% rejection rate for detailed hazardous tasks. Moreover, simply replacing the LLM driving the agent does not lead to notable improvements in safety awareness. Dataset and codes are available and shown in the reproducibility statement.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=BFb4ACHayj", "arxiv_id": "2412.13178", "paper_pdf": "papers/BFb4ACHayj.pdf", "paper_pdf_sha256": "b6168f71cc6cfeaa70c4e10d1527d38f31f96f484b3b5112db470c4737339807", "paper_pdf_bytes": 5445029, "paper_pdf_source": "openreview", "code_url": "https://github.com/shengyin1224/SafeAgentBench", "code_repository": "shengyin1224/SafeAgentBench", "code_commit": "38ca3ab27eb8a5f5034a50bdcc5cbab23ce8f089", "code_archive": "repos/BFb4ACHayj.zip", "code_archive_sha256": "18955464ca05f1500164776fab83da05dceb5b7c2e1a116fb27c1601b2b10553", "code_archive_bytes": 1950191, "code_file_count": 7, "code_extensions": {".py": 7}, "github_disk_usage_kb": 1901, "github_languages": {"Python": 146079}, "github_archived": false, "github_pushed_at": "2025-02-25T15:07:23Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/safeagentbench-a-benchmark-for-safe-task"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "mkjKqeBXkt", "year": 2024, "status": "rejected", "title": "KITS: Inductive Spatio-Temporal Kriging with Increment Training Strategy", "authors": ["Qianxiong Xu", "Cheng Long", "Ziyue Li", "Sijie Ruan", "Rui Zhao", "zhishuai Li"], "authorids": ["~Qianxiong_Xu1", "~Cheng_Long1", "~Ziyue_Li2", "~Sijie_Ruan1", "~Rui_Zhao6", "~zhishuai_Li1"], "authors_source": "OpenReview API", "abstract": "Sensors are commonly deployed to perceive the environment. However, due to the high cost, sensors are usually sparsely deployed. Kriging is the tailored task to infer the unobserved nodes (without sensors) using the observed source nodes (with sensors). The essence of kriging task is transferability. Recently, several inductive spatio-temporal kriging methods have been proposed based on graph neural networks, being trained based on a graph built on top of observed nodes via pretext tasks such as masking nodes out and reconstructing them. However, the graph in training is inevitably much sparser than the graph in inference that includes all the observed and unobserved nodes. The learned pattern cannot be well generalized for inference, denoted as graph gap. To address this issue, we first present a novel Increment training strategy: instead of masking nodes (and reconstructing them), we add virtual nodes into the training graph so as to mitigate the graph gap issue naturally. Nevertheless, the empty-shell virtual nodes without labels could have bad-learned features and lack supervision signals. To solve these issues, we pair each virtual node with its most similar observed node and fuse their features together; to enhance the supervision signal, we construct reliable pseudo labels for virtual nodes. As a result, the learned pattern of virtual nodes could be safely transferred to real unobserved nodes for reliable kriging. We name our new Kriging model with Increment Training Strategy as KITS. Extensive experiments demonstrate that KITS consistently outperforms existing kriging methods by large margins, e.g., the improvement over MAE score could be as high as 18.33\\%.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=mkjKqeBXkt", "arxiv_id": "2311.02565", "paper_pdf": "papers/mkjKqeBXkt.pdf", "paper_pdf_sha256": "7fa60569c66cb07b2d6ed312576d392cd66927fdeb1807040b4963bf42c9ef65", "paper_pdf_bytes": 2011634, "paper_pdf_source": "openreview", "code_url": "https://github.com/Sam1224/KITS", "code_repository": "Sam1224/KITS", "code_commit": "957e20179758ea97cce17cf40bc08c668287724f", "code_archive": "repos/mkjKqeBXkt.zip", "code_archive_sha256": "9528d3af291ea73b18911ef131b1de91394ac0c340769068304952cfe9f167d4", "code_archive_bytes": 131277, "code_file_count": 36, "code_extensions": {".py": 36}, "github_disk_usage_kb": 106, "github_languages": {"Python": 149801}, "github_archived": false, "github_pushed_at": "2024-12-17T06:26:17Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/kits-inductive-spatio-temporal-kriging-with"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Su84ELBdm5U", "year": 2023, "status": "rejected", "title": "How does overparametrization affect performance on minority groups?", "authors": ["Subha Maity", "Saptarshi Roy", "Songkai Xue", "Mikhail Yurochkin", "Yuekai Sun"], "authorids": ["~Subha_Maity1", "~Saptarshi_Roy1", "~Songkai_Xue1", "~Mikhail_Yurochkin1", "~Yuekai_Sun1"], "authors_source": "OpenReview API", "abstract": "The benefits of overparameterization for the overall performance of modern machine learning (ML) models are well known. However, the effect of overparameterization at a more granular level of data subgroups is less understood. Recent empirical studies demonstrate encouraging results: (i) when groups are not known, overparameterized models trained with empirical risk minimization (ERM) perform better on minority groups; (ii) when groups are known, ERM on data subsampled to equalize group sizes yields state-of-the-art worst-group-accuracy in the overparameterized regime. In this paper, we complement these empirical studies with a theoretical investigation of the risk of overparameterized random feature models on minority groups. In a setting in which the regression functions for the majority and minority groups are different, we show that overparameterization always improves minority group performance.", "decision": "Reject", "meta_review": null, "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=Su84ELBdm5U", "arxiv_id": "2206.03515", "paper_pdf": "papers/Su84ELBdm5U.pdf", "paper_pdf_sha256": "d178cad140d851044caf7f1a50149a186a7eb0b634bde42ffd99b53f620518a0", "paper_pdf_bytes": 584861, "paper_pdf_source": "openreview", "code_url": "https://github.com/smaityumich/overparameterization", "code_repository": "smaityumich/overparameterization", "code_commit": "e66a8bcf32982848ca2a354fde02172811ed2aa6", "code_archive": "repos/Su84ELBdm5U.zip", "code_archive_sha256": "9505a35970f636ac1cc3d651ea26f55eb6e58d091a505343be7c8f1b74b9aaa6", "code_archive_bytes": 339545, "code_file_count": 11, "code_extensions": {".py": 10, ".ipynb": 1}, "github_disk_usage_kb": 196, "github_languages": {"Python": 62925, "Jupyter Notebook": 72}, "github_archived": false, "github_pushed_at": "2022-06-07T17:02:46Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/how-does-overparametrization-affect"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "XhMa8XPHxpw", "year": 2022, "status": "rejected", "title": "Low-Precision Stochastic Gradient Langevin Dynamics", "authors": ["Ruqi Zhang", "Andrew Gordon Wilson", "Christopher De Sa"], "authorids": ["~Ruqi_Zhang1", "~Andrew_Gordon_Wilson1", "~Christopher_De_Sa2"], "authors_source": "OpenReview API", "abstract": "Low-precision optimization is widely used to accelerate large-scale deep learning. Despite providing better uncertainty estimation and generalization, sampling methods remain mostly unexplored in this space. In this paper, we provide the first study of low-precision Stochastic Gradient Langevin Dynamics (SGLD), arguing that it is particularly suited to low-bit arithmetic due to its intrinsic ability to handle system noise. We prove the convergence of low-precision SGLD on strongly log-concave distributions, showing that with full-precision gradient accumulators, SGLD is more robust to quantization error than SGD; however, with low-precision gradient accumulators, SGLD can diverge arbitrarily far from the target distribution with small stepsizes. To remedy this issue, we develop a new quantization function that preserves the correct variance in each update step. We demonstrate that the resulting low-precision SGLD algorithm is comparable to full-precision SGLD and outperforms low-precision SGD on deep learning tasks.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=XhMa8XPHxpw", "arxiv_id": "2206.09909", "paper_pdf": "papers/XhMa8XPHxpw.pdf", "paper_pdf_sha256": "211a4a0d344a2ffb4ac7d8b66493f8a51c3a4e44d80ce91d147906f55baa5608", "paper_pdf_bytes": 392781, "paper_pdf_source": "openreview", "code_url": "https://github.com/ruqizhang/low-precision-sgld", "code_repository": "ruqizhang/low-precision-sgld", "code_commit": "5f9192e10b2437f0de27c72e8b582bf8621dae26", "code_archive": "repos/XhMa8XPHxpw.zip", "code_archive_sha256": "afaf0c089e0cf96f3891dff22fd6dfbf91f3d81d07f53c3c10fb3ba0018ce41a", "code_archive_bytes": 425131, "code_file_count": 14, "code_extensions": {".py": 9, ".sh": 5}, "github_disk_usage_kb": 413, "github_languages": {"Python": 50121, "Shell": 2895}, "github_archived": false, "github_pushed_at": "2022-06-22T03:13:37Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/low-precision-stochastic-gradient-langevin-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "WPO0vDYLXem", "year": 2021, "status": "rejected", "title": "Hyperparameter Transfer Across Developer Adjustments", "authors": ["Danny Stoll", "Jörg K.H. Franke", "Diane Wagner", "Simon Selg", "Frank Hutter"], "authorids": ["~Danny_Stoll1", "~Jörg_K.H._Franke1", "wagnerd@cs.uni-freiburg.de", "selgs@cs.uni-freiburg.de", "~Frank_Hutter1"], "authors_source": "OpenReview API", "abstract": "After developer adjustments to a machine learning (ML) algorithm, how can the results of an old hyperparameter optimization (HPO) automatically be used to speedup a new HPO? This question poses a challenging problem, as developer adjustments can change which hyperparameter settings perform well, or even the hyperparameter search space itself. While many approaches exist that leverage knowledge obtained on previous tasks, so far, knowledge from previous development steps remains entirely untapped. In this work, we remedy this situation and propose a new research framework: hyperparameter transfer across adjustments (HT-AA). To lay a solid foundation for this research framework, we provide four simple HT-AA baseline algorithms and eight benchmarks\nchanging various aspects of ML algorithms, their hyperparameter search spaces, and the neural architectures used. The best baseline, on average and depending on the budgets for the old and new HPO, reaches a given performance 1.2-3.6x faster than a prominent HPO algorithm without transfer. As HPO is a crucial step in ML development but requires extensive computational resources, this speedup would lead to faster development cycles, lower costs, and reduced environmental impacts. To make these benefits available to ML developers off-the-shelf and to facilitate future research on HT-AA, we provide python packages for our baselines and benchmarks.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=WPO0vDYLXem", "arxiv_id": "2010.13117", "paper_pdf": "papers/WPO0vDYLXem.pdf", "paper_pdf_sha256": "fb57b6b9691760ca88b708e830c616478ac05a17b1a7187f1d3a560da7030da6", "paper_pdf_bytes": 4031177, "paper_pdf_source": "openreview", "code_url": "https://github.com/hp-transfer/htaa_experiments", "code_repository": "hp-transfer/htaa_experiments", "code_commit": "7b5c3b9b53d4e623c25bb44fd703884f314c3bf6", "code_archive": "repos/WPO0vDYLXem.zip", "code_archive_sha256": "05e6a3ff2ba7da11b9f7fa64710e06346c1eda60f36df79b95fcc414ca127c7d", "code_archive_bytes": 86641, "code_file_count": 7, "code_extensions": {".py": 7}, "github_disk_usage_kb": 505, "github_languages": {"Python": 43464}, "github_archived": false, "github_pushed_at": "2021-10-04T01:12:07Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/hyperparameter-transfer-across-developer-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "1dMsKsfwJ1", "year": 2026, "status": "rejected", "title": "Dataset Distillation via Committee Voting", "authors": ["Jiacheng Cui", "Zhaoyi Li", "Xiaochen Ma", "Xinyue Bi", "Yaxin Luo", "Zhiqiang Shen"], "authorids": ["~Jiacheng_Cui1", "~Zhaoyi_Li6", "~Xiaochen_Ma1", "~Xinyue_Bi1", "~Yaxin_Luo1", "~Zhiqiang_Shen1"], "authors_source": "OpenReview API", "abstract": "Dataset distillation aims to synthesize a smaller, representative dataset that preserves the essential properties of the original data, enabling efficient model training with reduced computational resources. Prior work has primarily focused on improving the alignment or matching process between original and synthetic data, or on enhancing the efficiency of distilling large datasets. In this work, we introduce $\\bf C$ommittee $\\bf V$otingfor $\\bf D$ataset $\\bf D$istillation (CV-DD), a novel and orthogonal approach that leverages the collective wisdom of multiple models or experts to create high-quality distilled datasets. We start by showing how to establish a strong baseline that already achieves state-of-the-art accuracy through leveraging recent advancements and thoughtful adjustments in model design and optimization processes. By integrating distributions and predictions from a committee of models while generating accurate soft labels, our method captures a wider spectrum of data features, reduces model-specific biases and mitigates distributional shifts between synthetic data and original data. This voting-based strategy not only promotes diversity and robustness within the distilled dataset but also significantly reduces overfitting, resulting in improved performance on post-eval tasks. Extensive experiments across various datasets and IPCs (images per class) demonstrate that Committee Voting leads to more reliable and adaptable distilled data compared to single/multi-model distillation methods, demonstrating its potential for efficient and accurate dataset distillation.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=1dMsKsfwJ1", "arxiv_id": "2501.07575", "paper_pdf": "papers/1dMsKsfwJ1.pdf", "paper_pdf_sha256": "a9624299834fcd360a87b90fad5332c350c2798cc3528dc652d95a6f0fda9319", "paper_pdf_bytes": 5268947, "paper_pdf_source": "openreview", "code_url": "https://github.com/Jiacheng8/CV-DD", "code_repository": "Jiacheng8/CV-DD", "code_commit": "24097e9c7058954b90eb442daa7cc6ba4b774830", "code_archive": "repos/1dMsKsfwJ1.zip", "code_archive_sha256": "82a6ef59ca3d400274a346f14fa856fe52735b1c86ebf96a18ed6e617f985ae3", "code_archive_bytes": 1493776, "code_file_count": 174, "code_extensions": {".sh": 159, ".py": 15}, "github_disk_usage_kb": 1913, "github_languages": {"Shell": 148806, "Python": 118403}, "github_archived": false, "github_pushed_at": "2025-07-28T04:55:09Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/dataset-distillation-via-committee-voting"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "PRKFRzOEq8", "year": 2025, "status": "rejected", "title": "Estimating the conformal prediction threshold from noisy labels", "authors": ["Coby Penso", "Jacob Goldberger", "Ethan Fetaya"], "authorids": ["~Coby_Penso1", "~Jacob_Goldberger1", "~Ethan_Fetaya1"], "authors_source": "OpenReview API", "abstract": "Conformal Prediction (CP) is a method to control prediction uncertainty by producing a small prediction set,   ensuring a predetermined probability that the true class lies within this set.   This is commonly done by defining a score, based on the model predictions, and setting a threshold on this score using a validation set. In this study, we address the problem of CP calibration when we only have access to a validation set with noisy labels. We show how we can estimate the noise-free conformal threshold based on the noisy labeled data.   Our solution is flexible and can accommodate various modeling assumptions regarding the label contamination process, without needing any information about the underlying data distribution or the internal mechanisms of the machine learning classifier.    We develop a coverage guarantee for uniform noise that is effective even in tasks with a large number of classes. We dub our approach Noise-Aware Conformal Prediction (NACP) and show on several natural and medical image classification datasets, including ImageNet, that it significantly outperforms current noisy label methods and achieves results comparable to those obtained with a clean validation set.", "decision": "Reject", "meta_review": "", "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=PRKFRzOEq8", "arxiv_id": "2501.12749", "paper_pdf": "papers/PRKFRzOEq8.pdf", "paper_pdf_sha256": "b58b40f9f7798d53831eeba9ae9ecc252056d1f76d574a8c6a7af29291622ef7", "paper_pdf_bytes": 437942, "paper_pdf_source": "openreview", "code_url": "https://github.com/cobypenso/Noise-Aware-Conformal-Prediction", "code_repository": "cobypenso/Noise-Aware-Conformal-Prediction", "code_commit": "64ce4d978cc6b734054839cc5b856de06570fd4d", "code_archive": "repos/PRKFRzOEq8.zip", "code_archive_sha256": "1c9cc9b891ef2d8644d507e0b5bdeabdf4dd79c930a6eb3209b08665ffc96c72", "code_archive_bytes": 140595, "code_file_count": 7, "code_extensions": {".py": 7}, "github_disk_usage_kb": 133, "github_languages": {"Python": 36273}, "github_archived": false, "github_pushed_at": "2024-05-21T09:45:50Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/estimating-the-conformal-prediction-threshold"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "F0XXA9OG13", "year": 2024, "status": "rejected", "title": "MediTab: Scaling Medical Tabular Data Predictors via Data Consolidation, Enrichment, and Refinement", "authors": ["Zifeng Wang", "Chufan Gao", "Cao Xiao", "Jimeng Sun"], "authorids": ["~Zifeng_Wang3", "~Chufan_Gao1", "~Cao_Xiao2", "~Jimeng_Sun3"], "authors_source": "OpenReview API", "abstract": "Tabular data prediction has been employed in medical applications such as patient health risk prediction. However, existing methods usually revolve around the algorithm design while overlooking the significance of data engineering. Medical tabular datasets frequently exhibit significant heterogeneity across different sources, with limited sample sizes per source. As such, previous predictors are often trained on manually curated small datasets that struggle to generalize across different tabular datasets during inference.\n\nThis paper proposes to scale medical tabular data predictors (MediTab) to various tabular inputs with varying features. The method uses a data engine that leverages large language models (LLMs) to consolidate tabular samples to overcome the barrier across tables with distinct schema. It also aligns out-domain data with the target task using a \"learn, annotate, and refinement'' pipeline. The expanded training data then enables the pre-trained MediTab to infer for arbitrary tabular input in the domain without fine-tuning, resulting in significant improvements over supervised baselines: it reaches an average ranking of 1.57 and 1.00 on 7 patient outcome prediction datasets and 3 trial outcome prediction datasets, respectively. In addition, MediTab exhibits impressive zero-shot performances: it outperforms supervised XGBoost models by 8.9% and 17.2% on average in two prediction tasks, respectively.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=F0XXA9OG13", "arxiv_id": "2305.12081", "paper_pdf": "papers/F0XXA9OG13.pdf", "paper_pdf_sha256": "59152de65b6ad2f9d76f5ec54b658093b75fde93a179fc7f5f99d932e34e5ffc", "paper_pdf_bytes": 1110037, "paper_pdf_source": "openreview", "code_url": "https://github.com/RyanWangZf/MediTab", "code_repository": "RyanWangZf/MediTab", "code_commit": "e50db96e97efa9ec594e9bcb99cbf3b9140ca166", "code_archive": "repos/F0XXA9OG13.zip", "code_archive_sha256": "7be0da25b967d2f2d45911a5bd64ff197f5cdfdb21ec1684ab6df6dd42e06232", "code_archive_bytes": 125975, "code_file_count": 21, "code_extensions": {".py": 13, ".ipynb": 8}, "github_disk_usage_kb": 109, "github_languages": {"Jupyter Notebook": 412132, "Python": 119504}, "github_archived": false, "github_pushed_at": "2024-05-08T04:59:19Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/anypredict-foundation-model-for-tabular"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "PHtzmXK8am", "year": 2023, "status": "rejected", "title": "TAN without a burn: Scaling laws of DP-SGD", "authors": ["Tom Sander", "Pierre Stock", "Alexandre Sablayrolles"], "authorids": ["~Tom_Sander1", "~Pierre_Stock1", "~Alexandre_Sablayrolles1"], "authors_source": "OpenReview API", "abstract": "Differentially Private methods for training Deep Neural Networks (DNNs) have progressed recently, in particular with the use of massive batches and aggregated data augmentations for a large number of steps. These techniques require much more compute than their non-private counterparts, shifting the traditional privacy-accuracy trade-off to a privacy-accuracy-compute trade-off and making hyper-parameter search virtually impossible for realistic scenarios. In this work, we decouple privacy analysis and experimental behavior of noisy training to explore the trade-off with minimal computational requirements. We first use the tools of Rényi Differential Privacy (RDP) to show that the privacy budget, when not overcharged, only depends on the total amount of noise (TAN) injected throughout training. We then derive scaling laws for training models with DP-SGD to optimize hyper-parameters with more than a $100\\times$ reduction in computational budget. We apply the proposed method on CIFAR-10 and ImageNet and, in particular, strongly improve the state-of-the-art on ImageNet with a $+9$ points gain in accuracy for a privacy budget $\\varepsilon=8$.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=PHtzmXK8am", "arxiv_id": "2210.03403", "paper_pdf": "papers/PHtzmXK8am.pdf", "paper_pdf_sha256": "e430534137b455b33f6428275508339aac8c3d1aaa9acdb983569e6cdf841ea4", "paper_pdf_bytes": 1390439, "paper_pdf_source": "openreview", "code_url": "https://github.com/facebookresearch/tan", "code_repository": "facebookresearch/tan", "code_commit": "9d8be82b62bdffaabb4a01f0f2938b6e026018bc", "code_archive": "repos/PHtzmXK8am.zip", "code_archive_sha256": "0f46d203914d5a044b4af8d9b55b00d958b70110c5304289d48a73bbaa683122", "code_archive_bytes": 210808, "code_file_count": 20, "code_extensions": {".py": 20}, "github_disk_usage_kb": 199, "github_languages": {"Python": 205589}, "github_archived": true, "github_pushed_at": "2025-01-07T19:37:20Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/tan-without-a-burn-scaling-laws-of-dp-sgd"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "EXe93Md8RqS", "year": 2022, "status": "rejected", "title": "Data Quality Matters For Adversarial Training: An Empirical Study", "authors": ["Chengyu Dong", "Liyuan Liu", "Jingbo Shang"], "authorids": ["~Chengyu_Dong1", "~Liyuan_Liu3", "~Jingbo_Shang2"], "authors_source": "OpenReview API", "abstract": "Multiple intriguing problems are hovering in adversarial training, including robust overfitting, robustness overestimation, and robustness-accuracy trade-off. These problems pose great challenges to both reliable evaluation and practical deployment. Here, we empirically show that these problems share one common cause --- low-quality samples in the dataset. Specifically, we first propose a strategy to measure the data quality based on the learning behaviors of the data during adversarial training and find that low-quality data may not be useful and even detrimental to the adversarial robustness. We then design controlled experiments to investigate the interconnections between data quality and problems in adversarial training. We find that when low-quality data is removed, robust overfitting and robustness overestimation can be largely alleviated; and robustness-accuracy trade-off becomes less significant. These observations not only verify our intuition about data quality but may also open new opportunities to advance adversarial training. ", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=EXe93Md8RqS", "arxiv_id": "2102.07437", "paper_pdf": "papers/EXe93Md8RqS.pdf", "paper_pdf_sha256": "c780d3f082d0422127e2eb0a56ce626327639cf9593c3d302b90c6ecd192ec02", "paper_pdf_bytes": 1478760, "paper_pdf_source": "openreview", "code_url": "https://github.com/shwinshaker/RobustDataProfiling", "code_repository": "shwinshaker/RobustDataProfiling", "code_commit": "7f29c2ea5b304771bd6d1d123fbb89633317ee5f", "code_archive": "repos/EXe93Md8RqS.zip", "code_archive_sha256": "0c9f3e216a54ec532ec0378aeba9c868d2dc1834e5dcbae24563bd55c1590c78", "code_archive_bytes": 398973, "code_file_count": 33, "code_extensions": {".py": 32, ".sh": 1}, "github_disk_usage_kb": 418, "github_languages": {"Python": 241372, "Shell": 1383}, "github_archived": false, "github_pushed_at": "2021-02-24T06:32:14Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/data-profiling-for-adversarial-training-on"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "tv8n52XbO4p", "year": 2021, "status": "rejected", "title": "Learning to Generate Noise for Multi-Attack Robustness", "authors": ["Divyam Madaan", "Jinwoo Shin", "Sung Ju Hwang"], "authorids": ["~Divyam_Madaan1", "~Jinwoo_Shin1", "~Sung_Ju_Hwang1"], "authors_source": "OpenReview API", "abstract": "Adversarial learning has emerged as one of the successful techniques to circumvent the susceptibility of existing methods against adversarial perturbations. However, the majority of existing defense methods are tailored to defend against a single category of adversarial perturbation (e.g. $\\ell_\\infty$-attack). In safety-critical applications, this makes these methods extraneous as the attacker can adopt diverse adversaries to deceive the system. Moreover, training on multiple perturbations simultaneously significantly increases the computational overhead during training. To address these challenges, we propose a novel meta-learning framework that explicitly learns to generate noise to improve the model's robustness against multiple types of attacks. Its key component is Meta Noise Generator (MNG) that outputs optimal noise to stochastically perturb a given sample, such that it helps lower the error on diverse adversarial perturbations. By utilizing samples generated by MNG, we train a model by enforcing the label consistency across multiple perturbations. We validate the robustness of models trained by our scheme on various datasets and against a wide variety of perturbations, demonstrating that it significantly outperforms the baselines across multiple perturbations with a marginal computational cost.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=tv8n52XbO4p", "arxiv_id": "2006.12135", "paper_pdf": "papers/tv8n52XbO4p.pdf", "paper_pdf_sha256": "ab8ac8c648509a90a97e4e250c851b7e6c8713ca47486becd709203daab88988", "paper_pdf_bytes": 2433764, "paper_pdf_source": "openreview", "code_url": "https://github.com/divyam3897/MNG_AC", "code_repository": "divyam3897/MNG_AC", "code_commit": "40c25a0c162c32cd858803f6bf1790180c727d06", "code_archive": "repos/tv8n52XbO4p.zip", "code_archive_sha256": "4c713f7a7e95aa3a4f4c476fc6526d71fb1f8387c800767cc7965c38e4e603a3", "code_archive_bytes": 537200, "code_file_count": 11, "code_extensions": {".py": 11}, "github_disk_usage_kb": 519, "github_languages": {"Python": 110764}, "github_archived": false, "github_pushed_at": "2021-06-25T05:27:43Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/learning-to-generate-noise-for-robustness"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Jri6onxoMY", "year": 2026, "status": "rejected", "title": "Causal Cartographer: From Mapping to Reasoning Over Counterfactual Worlds", "authors": ["Gael Gendron", "Joze M. Rozanec", "Michael J. Witbrock", "Gillian Dobbie"], "authorids": ["~Gael_Gendron1", "~Joze_M._Rozanec1", "~Michael_J._Witbrock1", "~Gillian_Dobbie1"], "authors_source": "OpenReview API", "abstract": "Causal world models are systems that can answer counterfactual questions about an environment of interest, i.e., predict how it would have evolved if an arbitrary subset of events had been realized differently. The ability to answer such questions is crucial for models to reliably understand the world. However, this task currently eludes large language models (LLMs), which do not have demonstrated causal reasoning capabilities beyond the memorization of existing causal relationships. Furthermore, evaluating counterfactuals in real-world applications is challenging since only the factual world is observed, limiting evaluation to synthetic datasets. We address these problems by proposing the Causal Cartographer, a twofold system composed of two agents: the first extracts causal relationships from data and builds a vast repository of causal knowledge, while the second uses them as constraints to perform reliable step-by-step causal inference. We evaluate our approach on real-world counterfactuals obtained by matching data from diverse news sources. We show that our approach can extract accurate causal knowledge and enhance the robustness of LLMs for causal reasoning tasks. In particular, the proposed causal conditioning mitigates the impact of spurious correlations and greatly reduces inference costs (by up to 70\\%) compared to chain-of-thought reasoning.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=Jri6onxoMY", "arxiv_id": "2505.14396", "paper_pdf": "papers/Jri6onxoMY.pdf", "paper_pdf_sha256": "77dcce5a6be1344f330b5921b371ba5836cc6ac0dc1a1972118e019857cf6d4b", "paper_pdf_bytes": 1051050, "paper_pdf_source": "openreview", "code_url": "https://github.com/ggendro/causal-cartographer", "code_repository": "ggendro/causal-cartographer", "code_commit": "a69b23e6f4a4477a08e65e13813662d9a053b882", "code_archive": "repos/Jri6onxoMY.zip", "code_archive_sha256": "32c3fdf9204cb02746d258c5a6d3e45fddb2bb3fbf75617b66a9d56ad8f4f56d", "code_archive_bytes": 526372, "code_file_count": 45, "code_extensions": {".py": 44, ".js": 1}, "github_disk_usage_kb": 1943, "github_languages": {"Python": 480960, "JavaScript": 59235, "CSS": 17196, "HTML": 5759}, "github_archived": false, "github_pushed_at": "2025-05-17T23:02:56Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/causal-cartographer-from-mapping-to-reasoning"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "rapXZIfwbX", "year": 2025, "status": "rejected", "title": "The Persian Rug: solving toy models of superposition using large-scale symmetries", "authors": ["Aditya Cowsik", "Kfir Dolev", "Alex Infanger"], "authorids": ["~Aditya_Cowsik1", "~Kfir_Dolev1", "~Alex_Infanger1"], "authors_source": "OpenReview API", "abstract": "We present a complete mechanistic description of the algorithm learned by a minimal non-linear sparse data autoencoder in the limit of large input dimension. The model, originally presented in \\cite{elhage2022superposition}, compresses sparse data vectors through a linear layer and decompresses using another linear layer followed by a ReLU activation. We notice that when the data is permutation symmetric (no input feature is privileged) large models reliably learn an algorithm that is sensitive to individual weights only through their large-scale statistics. For these models, the loss function becomes analytically tractable. Using this understanding, we give explicit upper bounds on the loss, which show that the model is near-optimal among recently proposed architectures. In particular, changes to the elementwise activation function or the addition of gating can at best improve its performance by a constant factor. Finally, we forward-engineer a model with the requisite symmetries and show that its loss precisely matches that of the trained models. Unlike the trained model weights, the minimal randomness in the artificial weights results in miraculous fractal structures resembling a Persian rug, to which the algorithm is oblivious. Our work contributes to neural network interpretability by introducing techniques for understanding the structure of autoencoders.", "decision": "Reject", "meta_review": "", "num_reviews": 6, "openreview_url": "https://openreview.net/forum?id=rapXZIfwbX", "arxiv_id": "2410.12101", "paper_pdf": "papers/rapXZIfwbX.pdf", "paper_pdf_sha256": "75d369cb48f910cb471f7dc5f9adce8da4e3c89d5bf5cc5dad8b9cd733dd86d6", "paper_pdf_bytes": 5818731, "paper_pdf_source": "openreview", "code_url": "https://github.com/KfirD/PersianRug", "code_repository": "KfirD/PersianRug", "code_commit": "e178d0bbe5acd1da0a8641a14ca4c3cf27daafe9", "code_archive": "repos/rapXZIfwbX.zip", "code_archive_sha256": "e3ec13a02e1930461323f05038665b8975b5974e1fe18b0f08d8765d72d63e18", "code_archive_bytes": 126195, "code_file_count": 15, "code_extensions": {".py": 14, ".ipynb": 1}, "github_disk_usage_kb": 134, "github_languages": {"Jupyter Notebook": 140702, "Python": 116782}, "github_archived": false, "github_pushed_at": "2024-12-07T22:13:55Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/the-persian-rug-solving-toy-models-of"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "DZyhUXpEee", "year": 2024, "status": "rejected", "title": "SpaFL: Communication-Efficient Federated Learning with Sparse Models and Low Computational Overhead", "authors": ["Minsu Kim", "Walid Saad", "Merouane Abdelkader DEBBAH", "Choong Seon Hong"], "authorids": ["~Minsu_Kim6", "~Walid_Saad1", "~Merouane_Abdelkader_DEBBAH1", "~Choong_Seon_Hong1"], "authors_source": "OpenReview API", "abstract": "The large communication and computation overhead of federated learning (FL) is one of the main challenges facing its practical deployment over resource-constrained clients and systems. In this work, SpaFL: a communication-efficient FL framework is proposed to optimize both personalized model parameters and sparse model structures  with low computational overhead. In SpaFL, a trainable threshold is defined for each neuron/filter to prune its connected parameters. Both model parameters and thresholds are jointly optimized to enable the automatic sparsification of the models while recovering prematurely pruned parameters during training. To reduce communication costs, only thresholds are communicated between a server and clients instead of parameters, thereby enabling the clients to learn how to prune. Further, global thresholds are used to update model parameters by extracting aggregated parameter importance. The convergence of SpaFL is analyzed, and the results provide new insights into the tradeoff between computation overhead and learning performance. Experimental results show that SpaFL improves accuracy while requiring much less communication and computing resources compared to both dense and sparse personalized baselines.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=DZyhUXpEee", "arxiv_id": "2406.00431", "paper_pdf": "papers/DZyhUXpEee.pdf", "paper_pdf_sha256": "f956c908d682235d1bd2da9f795a959f9f617cad646fe186abc9de5de0523249", "paper_pdf_bytes": 1906004, "paper_pdf_source": "openreview", "code_url": "https://github.com/news-vt/SpaFL_NeruIPS_2024", "code_repository": "news-vt/SpaFL_NeruIPS_2024", "code_commit": "d3639ab3c6ecae3a3d09cf897b264fc44551a56c", "code_archive": "repos/DZyhUXpEee.zip", "code_archive_sha256": "ace2a4816fba57d6b6bad61cf84fae411e323aef7f9ee1db8bc8908ff3e17cff", "code_archive_bytes": 171180, "code_file_count": 34, "code_extensions": {".py": 34}, "github_disk_usage_kb": 111, "github_languages": {"Python": 135364}, "github_archived": false, "github_pushed_at": "2025-01-15T16:22:30Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/spafl-communication-efficient-federated"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "1EVPT82ttr", "year": 2023, "status": "rejected", "title": "Learning Unified Representations for Multi-Resolution Face Recognition", "authors": ["Hulingxiao He", "Wu Yuan", "Yidian Huang", "Shilong Zhao", "Wen Yuan", "HanQing Li"], "authorids": ["~Hulingxiao_He1", "~Wu_Yuan1", "~Yidian_Huang1", "~Shilong_Zhao1", "~Wen_Yuan1", "~HanQing_Li2"], "authors_source": "OpenReview API", "abstract": "In this work, we propose Branch-to-Trunk network (BTNet), a novel representation learning method for multi-resolution face recognition. It consists of a trunk network (TNet), namely a unified encoder, and multiple branch networks (BNets), namely resolution adapters. As per the input, a resolution-specific BNet is used and the output are implanted as feature maps in the feature pyramid of TNet, at a layer with the same resolution. The discriminability of tiny faces is significantly improved, as the interpolation error introduced by rescaling, especially up-sampling, is mitigated on the inputs. With branch distillation and backward-compatible training, BTNet transfers discriminative high-resolution information to multiple branches while guaranteeing representation compatibility. Our experiments demonstrate strong performance on face recognition benchmarks, both for multi-resolution face verification and face identification, with much less computation amount and parameter storage. We establish new state-of-the-art on the challenging QMUL-SurvFace 1: N face identification task.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=1EVPT82ttr", "arxiv_id": "2310.09563", "paper_pdf": "papers/1EVPT82ttr.pdf", "paper_pdf_sha256": "c1a004a0d2c5dfd3583544a22c5b4fca0471a724cfba96248f63e4170a2702b6", "paper_pdf_bytes": 10915076, "paper_pdf_source": "openreview", "code_url": "https://github.com/hehulingxiao/BTNet", "code_repository": "hehulingxiao/BTNet", "code_commit": "df06e02fa4cff2fafec52078ea6cf858f36d5f76", "code_archive": "repos/1EVPT82ttr.zip", "code_archive_sha256": "a7110c532d9cb54ff2970962fcf0504aac4a2b6e69bd3544cf29b417af98ad3f", "code_archive_bytes": 223984, "code_file_count": 34, "code_extensions": {".py": 34}, "github_disk_usage_kb": 200, "github_languages": {"Python": 171003}, "github_archived": false, "github_pushed_at": "2023-10-03T07:27:40Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/learning-unified-representations-for-multi"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "DvcMMKmDJ3q", "year": 2022, "status": "rejected", "title": "Generating Symbolic Reasoning Problems with Transformer GANs", "authors": ["Jens U. Kreber", "Christopher Hahn"], "authorids": ["~Jens_U._Kreber1", "~Christopher_Hahn1"], "authors_source": "OpenReview API", "abstract": "Constructing training data for symbolic reasoning domains is challenging: Existing instances are typically hand-crafted and too few to be trained on directly and synthetically generated instances are often hard to evaluate in terms of their meaningfulness. We study the capabilities of GANs and Wasserstein GANs equipped with Transformer encoders to generate sensible and challenging training data for symbolic reasoning domains. We conduct experiments on two problem domains where Transformers have been successfully applied recently: symbolic mathematics and temporal specifications in verification. Even without autoregression, our GAN models produce syntactically correct instances and we show that these can be used as meaningful substitutes for real training data when training a classifier. Using a GAN setting also allows us to alter the target distribution: We show that by adding a classifier uncertainty part to the generator objective, we obtain a dataset that is even harder to solve for a classifier than our original dataset.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=DvcMMKmDJ3q", "arxiv_id": "2110.10054", "paper_pdf": "papers/DvcMMKmDJ3q.pdf", "paper_pdf_sha256": "6ee02cd175a5f6a096895f3beec1316e3857e4e87d2e0bb4fdf90defb5412e5f", "paper_pdf_bytes": 452208, "paper_pdf_source": "openreview", "code_url": "https://github.com/reactive-systems/TGAN-SR", "code_repository": "reactive-systems/TGAN-SR", "code_commit": "43a5a93d0864ca77a693328cc07c8088be1f8f50", "code_archive": "repos/DvcMMKmDJ3q.zip", "code_archive_sha256": "005ca68f7a2e718ae90e57edf8830508737776f59854a7496c6af2db954c6ca3", "code_archive_bytes": 437605, "code_file_count": 22, "code_extensions": {".py": 22}, "github_disk_usage_kb": 424, "github_languages": {"Python": 160421}, "github_archived": false, "github_pushed_at": "2021-10-20T07:53:51Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/generating-symbolic-reasoning-problems-with-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "i7aMbliTkHs", "year": 2021, "status": "rejected", "title": "TAM: Temporal Adaptive Module for Video Recognition", "authors": ["Zhaoyang Liu", "Limin Wang", "Wayne Wu", "Chen Qian", "Tong Lu"], "authorids": ["~Zhaoyang_Liu1", "~Limin_Wang1", "~Wayne_Wu1", "~Chen_Qian1", "~Tong_Lu1"], "authors_source": "OpenReview API", "abstract": "Temporal modeling is crucial for capturing spatiotemporal structure in videos for action recognition. Video data is with extremely complex dynamics along its temporal dimension due to various factors such as camera motion, speed variation, and different activities. To effectively capture this diverse motion pattern, this paper presents a new temporal adaptive module ({\\bf TAM}) to generate video-specific kernels based on its own feature maps. TAM proposes a unique two-level adaptive modeling scheme by decoupling dynamic kernels into a location sensitive importance map and a location invariant aggregation weight. The importance map is learned in a local temporal window to capture short term information, while the aggregation weight is generated from a global view with a focus on long-term structure. TAM is a principled module and could be integrated into 2D CNNs to yield a powerful video architecture (TANet) with a very small extra computational cost. The extensive experiments on Kinetics-400 and Something-Something datasets, demonstrate that the TAM outperforms other temporal modeling methods consistently owing to its temporal adaptive modeling strategy.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=i7aMbliTkHs", "arxiv_id": "2005.06803", "paper_pdf": "papers/i7aMbliTkHs.pdf", "paper_pdf_sha256": "7963a986053be91e75a695b6e553b50af05c9cacccc7da913b14a538b691c893", "paper_pdf_bytes": 35096691, "paper_pdf_source": "openreview", "code_url": "https://github.com/liu-zhy/temporal-adaptive-module", "code_repository": "liu-zhy/temporal-adaptive-module", "code_commit": "5fcb46fe671ead2f7958d8ffc3321a0e274d6491", "code_archive": "repos/i7aMbliTkHs.zip", "code_archive_sha256": "9c6c33114bc2975422e2903f8f23f8a9692b4a7e057e5029baba4a935f6e7704", "code_archive_bytes": 582026, "code_file_count": 27, "code_extensions": {".py": 21, ".sh": 6}, "github_disk_usage_kb": 564, "github_languages": {"Python": 173536, "Shell": 1666}, "github_archived": false, "github_pushed_at": "2022-08-18T09:33:50Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/tam-temporal-adaptive-module-for-video"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "SylR6n4tPS", "year": 2020, "status": "rejected", "title": "Learning to Generate Grounded Visual Captions without Localization Supervision", "authors": ["Chih-Yao Ma", "Yannis Kalantidis", "Ghassan AlRegib", "Peter Vajda", "Marcus Rohrbach", "Zsolt Kira"], "authorids": ["cyma@gatech.edu", "ykalant@image.ntua.gr", "vajdap@fb.com", "alregib@gatech.edu", "maroffm@gmail.com", "zkira@gatech.edu"], "authors_source": "OpenReview API", "abstract": "When automatically generating a sentence description for an image or video, it often remains unclear how well the generated caption is grounded, or if the model hallucinates based on priors in the dataset and/or the language model. The most common way of relating image regions with words in caption models is through an attention mechanism over the regions that are used as input to predict the next word. The model must therefore learn to predict the attentional weights without knowing the word it should localize. This is difficult to train without grounding supervision since recurrent models can propagate past information and there is no explicit signal to force the captioning model to properly ground the individual decoded words. In this work, we help the model to achieve this via a novel cyclical training regimen that forces the model to localize each word in the image after the sentence decoder generates it, and then reconstruct the sentence from the localized image region(s) to match the ground-truth. Our proposed framework only requires learning one extra fully-connected layer (the localizer), a layer that can be removed at test time. We show that our model significantly improves grounding accuracy without relying on grounding supervision or introducing extra computation during inference for both image and video captioning tasks.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=SylR6n4tPS", "arxiv_id": null, "paper_pdf": "papers/SylR6n4tPS.pdf", "paper_pdf_sha256": "4bab70ed0033a01af38d3698ee7ec5b1d9bb991b837eb37ce8e03f2a7d3650ad", "paper_pdf_bytes": 6798464, "paper_pdf_source": "openreview", "code_url": "https://github.com/chihyaoma/cyclical-visual-captioning", "code_repository": "chihyaoma/cyclical-visual-captioning", "code_commit": "337eee17a8789e58d7bf687ad9ecf27c8da44f14", "code_archive": "repos/SylR6n4tPS.zip", "code_archive_sha256": "f8f5a036cb7ded7d2e2ab40c2533cb5023ac497531b0706e172063bc1a34e869", "code_archive_bytes": 566261, "code_file_count": 17, "code_extensions": {".py": 16, ".sh": 1}, "github_disk_usage_kb": 945, "github_languages": {"Python": 161584, "Shell": 2140}, "github_archived": false, "github_pushed_at": "2020-07-29T22:22:45Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/learning-to-generate-grounded-visual-captions"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "xQIvTUL1Tf", "year": 2026, "status": "rejected", "title": "Efficient Reasoning with Hidden Thinking", "authors": ["Xuan Shen", "Yizhou Wang", "Yufa Zhou", "Xiangxi Shi", "Pu Zhao", "Yanzhi Wang", "Jiuxiang Gu"], "authorids": ["~Xuan_Shen1", "~Yizhou_Wang3", "~Yufa_Zhou1", "~Xiangxi_Shi1", "~Pu_Zhao1", "~Yanzhi_Wang3", "~Jiuxiang_Gu2"], "authors_source": "OpenReview API", "abstract": "Chain-of-Thought (CoT) reasoning has become a powerful framework for improving complex problem-solving capabilities in Multimodal Large Language Models (MLLMs).\nHowever, the verbose nature of textual reasoning introduces significant inefficiencies.\nIn this work, we propose **Heima** (as hidden llama), an effective CoT compression framework that condenses lengthy CoTs into a small set of abstract thinking tokens, preserving essential reasoning while removing redundancy.\nWe then conduct a theoretical analysis from an information-theoretic perspective, quantifying the information gap induced by compression, showing that reasoning capability is preserved when non-trivial mutual information is retained.\nTo further explore and quantify this information gap, we design the adaptive interpreter that maps thinking tokens back to variable-length textual sequences, thereby reconstructing the reasoning process.\nExperiments across diverse reasoning benchmarks demonstrate that Heima improves reasoning efficiency, while maintaining or even achieving better zero-shot accuracy.\nMoreover, the interpreter reconstructs coherent reasoning progresses from compressed thinking tokens, revealing that the information gap is minimal and validating the effectiveness of the proposed framework.\nThis work paves the way for scalable latent reasoning models and advances our understanding of efficient reasoning processes in large models.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=xQIvTUL1Tf", "arxiv_id": "2501.19201", "paper_pdf": "papers/xQIvTUL1Tf.pdf", "paper_pdf_sha256": "6bbcc422e3b2577e8b180d9a4c45961676a2d72b31f323fa3c694b5e06aae769", "paper_pdf_bytes": 773553, "paper_pdf_source": "openreview", "code_url": "https://github.com/shawnricecake/Heima", "code_repository": "shawnricecake/Heima", "code_commit": "cebd6ef3ef862f62c601ced191b64b45d2bf3de2", "code_archive": "repos/xQIvTUL1Tf.zip", "code_archive_sha256": "3d241265227604042b3b93df008ef1564b24bbb219b954a535b4f1a1e701c15f", "code_archive_bytes": 2304672, "code_file_count": 375, "code_extensions": {".py": 365, ".sh": 10}, "github_disk_usage_kb": 2054, "github_languages": {"Python": 3038372, "Shell": 7653}, "github_archived": false, "github_pushed_at": "2026-05-20T03:51:41Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/efficient-reasoning-with-hidden-thinking"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "a6XE2GJHjk", "year": 2025, "status": "rejected", "title": "TabGraphs: A Benchmark and Strong Baselines for Learning on Graphs with Tabular Node Features", "authors": ["Gleb Bazhenov", "Oleg Platonov", "Liudmila Prokhorenkova"], "authorids": ["~Gleb_Bazhenov1", "~Oleg_Platonov1", "~Liudmila_Prokhorenkova1"], "authors_source": "OpenReview API", "abstract": "Tabular machine learning is an important field for industry and science. In this field, table rows are typically treated as independent data samples, but additional information about the relations between these samples is sometimes available and can be used to improve predictive performance. Such information can be naturally modeled with a graph, hence tabular machine learning may benefit from graph machine learning methods. However, graph machine learning models are typically evaluated on datasets with homogeneous, most often text-based node features, which are very different from heterogeneous mixtures of numerical and categorical features present in tabular datasets. Thus, there is a critical difference between the data used in tabular and graph machine learning studies, which does not allow one to understand how successfully graph models can be transferred to tabular data. To bridge this gap, we propose a new benchmark of diverse graphs with heterogeneous tabular node features and realistic prediction tasks. We use this benchmark to evaluate a vast set of models, including simple methods previously overlooked in the literature. Our experiments show that graph neural networks indeed can often bring gains in predictive performance for tabular data, but standard tabular models can also be adapted to work with graph data by using simple graph-based feature augmentation, which sometimes enables them to compete with and even outperform graph neural models. Based on our empirical study, we provide insights for researchers and practitioners in both tabular and graph machine learning fields.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=a6XE2GJHjk", "arxiv_id": "2409.14500", "paper_pdf": "papers/a6XE2GJHjk.pdf", "paper_pdf_sha256": "f8bfd453f82e37ab43ab222eba058d94559eefb3431694ee4a87107fa47f8cba", "paper_pdf_bytes": 413075, "paper_pdf_source": "openreview", "code_url": "https://github.com/yandex-research/tabgraphs", "code_repository": "yandex-research/tabgraphs", "code_commit": "0ddeca9d715c3e920d98b2b1dbe2f24c19f23b02", "code_archive": "repos/a6XE2GJHjk.zip", "code_archive_sha256": "33b838186071d1db1e275ce57a7f4c9e4976f262b1a6416a1f91513b08c8e172", "code_archive_bytes": 238810, "code_file_count": 61, "code_extensions": {".py": 56, ".ipynb": 4, ".sh": 1}, "github_disk_usage_kb": 143, "github_languages": {"Python": 426420, "Jupyter Notebook": 52844, "Shell": 3874, "Makefile": 422}, "github_archived": true, "github_pushed_at": "2025-10-29T17:41:55Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/tabgraphs-a-benchmark-and-strong-baselines"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "UbOzNf6hGq", "year": 2024, "status": "rejected", "title": "FiLM: Fill-in Language Models for Any-Order Generation", "authors": ["Tianxiao Shen", "Hao Peng", "Ruoqi Shen", "Yao Fu", "Zaid Harchaoui", "Yejin Choi"], "authorids": ["~Tianxiao_Shen1", "~Hao_Peng4", "~Ruoqi_Shen1", "~Yao_Fu3", "~Zaid_Harchaoui1", "~Yejin_Choi1"], "authors_source": "OpenReview API", "abstract": "Language models have become the backbone of today's AI systems. However, their predominant left-to-right generation limits the use of bidirectional context, which is essential for tasks that involve filling text in the middle. We propose the Fill-in Language Model (FiLM), a new language modeling approach that allows for flexible generation at any position without adhering to a specific generation order. Its training extends the masked language modeling objective by adopting varying mask probabilities sampled from the Beta distribution to enhance the generative capabilities of FiLM. During inference, FiLM can seamlessly insert missing phrases, sentences, or paragraphs, ensuring that the outputs are fluent and are coherent with the surrounding context. In both automatic and human evaluations, FiLM outperforms existing infilling methods that rely on left-to-right language models trained on rearranged text segments. FiLM is easy to implement and can be either trained from scratch or fine-tuned from a left-to-right language model. Notably, as the model size grows, FiLM's perplexity approaches that of strong left-to-right language models of similar sizes, indicating FiLM's scalability and potential as a large language model.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=UbOzNf6hGq", "arxiv_id": "2310.09930", "paper_pdf": "papers/UbOzNf6hGq.pdf", "paper_pdf_sha256": "d5ade475df72e444da11545c8d4588600ae7cbcb16de5f98ca2a078ca52b34f3", "paper_pdf_bytes": 999602, "paper_pdf_source": "openreview", "code_url": "https://github.com/shentianxiao/FiLM", "code_repository": "shentianxiao/FiLM", "code_commit": "ca22c9fe51786b6b5544deb7b418969afb26eaf3", "code_archive": "repos/UbOzNf6hGq.zip", "code_archive_sha256": "af53b38123381ccc6b9d71c3f707b47b636881804cf8050c3087f24ea833d883", "code_archive_bytes": 121259, "code_file_count": 21, "code_extensions": {".py": 19, ".sh": 2}, "github_disk_usage_kb": 115, "github_languages": {"Python": 31378, "Shell": 1197}, "github_archived": false, "github_pushed_at": "2023-10-18T15:36:31Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/film-fill-in-language-models-for-any-order"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "MPoQtFC588n", "year": 2022, "status": "rejected", "title": "RMNet: Equivalently Removing Residual Connection from Networks", "authors": ["Fanxu Meng", "Hao Cheng", "Jia-Xin Zhuang", "Ke Li", "Xing Sun"], "authorids": ["~Fanxu_Meng1", "~Hao_Cheng5", "~Jia-Xin_Zhuang1", "~Ke_Li4", "~Xing_Sun1"], "authors_source": "OpenReview API", "abstract": "Although residual connection enables training very deep neural networks, it is not friendly for online inference due to its multi-branch topology. This encourages many researchers to work on designing DNNs without residual connections at inference. For example, RepVGG re-parameterizes multi-branch topology to a VGG-like (single-branch) model when deploying, showing great performance when the network is relatively shallow. However, RepVGG can not transform ResNet to VGG equivalently because re-parameterizing methods can only be applied to linear Blocks and the non-linear layers (ReLU) have to be put outside of the residual connection which results in limited representation ability, especially for deeper networks. In this paper, we aim to remedy this problem and propose to remove the residual connection in a vanilla ResNet equivalently by a reserving and merging (RM) operation on ResBlock. Specifically, RM operation allows input feature maps to pass through the block while reserving their information and merges all the information at the end of each block, which can remove residual connection without changing original output. RMNet basically has two advantages: 1) it achieves a better accuracy-speed trade-off compared with ResNet and RepVGG; 2) its implementation makes it naturally friendly for high ratio network pruning. Extensive experiments are performed to verify the effectiveness of RMNet. We believe the ideology of RMNet can inspire many insights on model design for the community in the future.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=MPoQtFC588n", "arxiv_id": "2111.00687", "paper_pdf": "papers/MPoQtFC588n.pdf", "paper_pdf_sha256": "c725adf4fc788b93ed506bd79bf8be38386b750621e8f11d9d2744623a801738", "paper_pdf_bytes": 684926, "paper_pdf_source": "openreview", "code_url": "https://github.com/fxmeng/RMNet", "code_repository": "fxmeng/RMNet", "code_commit": "4754b8e8a0c8b0f6e4e0c8c77f6cde47712c8feb", "code_archive": "repos/MPoQtFC588n.zip", "code_archive_sha256": "61b4b441fb55c495307169fe01965a6d6baa11f1f4ce248bfe47a2386e01af45", "code_archive_bytes": 436106, "code_file_count": 21, "code_extensions": {".py": 21}, "github_disk_usage_kb": 444, "github_languages": {"Python": 144239}, "github_archived": false, "github_pushed_at": "2023-06-17T02:23:58Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/rmnet-equivalently-removing-residual-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "SVsLxTfHa1", "year": 2021, "status": "rejected", "title": "Towards Multi-Sense Cross-Lingual Alignment of Contextual Embeddings", "authors": ["Linlin Liu", "Thien Hai Nguyen", "Shafiq Joty", "Lidong Bing", "Luo Si"], "authorids": ["~Linlin_Liu2", "~Thien_Hai_Nguyen1", "~Shafiq_Joty1", "~Lidong_Bing2", "~Luo_Si2"], "authors_source": "OpenReview API", "abstract": "Cross-lingual word embeddings (CLWE) have been proven useful in many cross-lingual tasks. However, most existing approaches to learn CLWE including the ones with contextual embeddings are sense agnostic. In this work, we propose a novel framework to align contextual embeddings at the sense level by leveraging cross-lingual signal from bilingual dictionaries only. We operationalize our framework by first proposing a novel sense-aware cross entropy loss to model word senses explicitly. The monolingual ELMo and BERT models pretrained with our sense-aware cross entropy loss demonstrate significant performance improvement for word sense disambiguation tasks. We then propose a sense alignment objective on top of the sense-aware cross entropy loss for cross-lingual model pretraining, and pretrain cross-lingual models for several language pairs (English to German/Spanish/Japanese/Chinese). Compared with the best baseline results, our cross-lingual models achieve 0.52%, 2.09% and 1.29% average performance improvements on zero-shot cross-lingual NER, sentiment classification and XNLI tasks, respectively. We will release our code.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=SVsLxTfHa1", "arxiv_id": "2103.06459", "paper_pdf": "papers/SVsLxTfHa1.pdf", "paper_pdf_sha256": "0fa02abb9517834a7bd8b17220e2f60f46d31fdbf3e039975bf575ff48777cd0", "paper_pdf_bytes": 2103083, "paper_pdf_source": "openreview", "code_url": "https://github.com/ntunlp/multisense_embedding_alignment", "code_repository": "ntunlp/multisense_embedding_alignment", "code_commit": "6cf157ec4eefafcdb1d3403fa1cc0dd2b5169833", "code_archive": "repos/SVsLxTfHa1.zip", "code_archive_sha256": "709922b27a817d0ba19c6de936ce5532fe5761744242e402b469d9e9b537e22c", "code_archive_bytes": 620056, "code_file_count": 39, "code_extensions": {".py": 32, ".sh": 7}, "github_disk_usage_kb": 589, "github_languages": {"Python": 421997, "Shell": 3341}, "github_archived": false, "github_pushed_at": "2022-10-14T13:19:51Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/towards-multi-sense-cross-lingual-alignment-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "rkgFXR4KPr", "year": 2020, "status": "rejected", "title": "A Simple Recurrent Unit with Reduced Tensor Product Representations", "authors": ["Shuai Tang", "Paul Smolensky", "Virginia R. de Sa"], "authorids": ["shuaitang93@ucsd.edu", "paul.smolensky@gmail.com", "desa@ucsd.edu"], "authors_source": "OpenReview API", "abstract": "Widely used recurrent units, including Long-short Term Memory (LSTM) and Gated Recurrent Unit (GRU), perform well on natural language tasks, but their ability to learn structured representations is still questionable. Exploiting reduced Tensor Product Representations (TPRs) --- distributed representations of symbolic structure in which vector-embedded symbols are bound to vector-embedded structural positions --- we propose the TPRU, a simple recurrent unit that, at each time step, explicitly executes structural-role binding and unbinding operations to incorporate structural information into learning. The gradient analysis of our proposed TPRU is conducted to support our model design, and its performance on multiple datasets shows the effectiveness of it. Furthermore, observations on linguistically grounded study demonstrate the interpretability of our TPRU.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=rkgFXR4KPr", "arxiv_id": "1810.12456", "paper_pdf": "papers/rkgFXR4KPr.pdf", "paper_pdf_sha256": "b2f265f3892a64708fa0c891f98f63a88c6a338080527d6b22dc9bd17d86e9bc", "paper_pdf_bytes": 637961, "paper_pdf_source": "openreview", "code_url": "https://github.com/shuaitang/TPRU", "code_repository": "shuaitang/TPRU", "code_commit": "bb5db2cb81cb6b248ec19211a15bc908dc4bef98", "code_archive": "repos/rkgFXR4KPr.zip", "code_archive_sha256": "51c5f960c6623a624443c5b96d845afbeffc697e845b4799c3cfda07a6d6c2d8", "code_archive_bytes": 51736, "code_file_count": 35, "code_extensions": {".py": 30, ".sh": 5}, "github_disk_usage_kb": 992, "github_languages": {"Python": 103450, "Shell": 1506}, "github_archived": false, "github_pushed_at": "2019-05-29T00:52:42Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/learning-distributed-representations-of-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "FQ8KX1pxeo", "year": 2026, "status": "rejected", "title": "Progressive Binarization with Semi-Structured Pruning for LLMs", "authors": ["Xianglong Yan", "Tianao Zhang", "Zhiteng Li", "Haotong Qin", "Yulun Zhang"], "authorids": ["~Xianglong_Yan2", "~Tianao_Zhang1", "~Zhiteng_Li2", "~Haotong_Qin1", "~Yulun_Zhang1"], "authors_source": "OpenReview API", "abstract": "Large language models (LLMs) have achieved remarkable progress in natural language processing, but their high computational and memory costs hinder deployment on resource-constrained devices. Binarization represents the most extreme form of quantization, yet binarized models still contain redundancy that can be further removed. Pruning provides a natural way to eliminate such redundancy, but naïve combination with binarization often results in severe performance degradation. In this paper, we propose Progressive Binarization with Semi-Structured Pruning (PBS$^2$P), a novel post-training framework that seamlessly integrates binarization and semi-structured pruning. We first propose Stepwise semi-structured Pruning with Binarization Optimization (SPBO), which progressively introduces sparsity while optimizing binarization parameters to jointly reduce pruning and quantization error, yielding more stable and accurate compression. Additionally, we propose a Coarse-to-Fine Search (CFS) that first allocates pruning ratios and then refines element selection, further enhancing overall performance. Extensive experiments across multiple LLM families show that PBS$^2$P consistently outperforms state-of-the-art (SOTA) binary post-training quantization methods in both perplexity and downstream accuracy. We will release all the code and models.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=FQ8KX1pxeo", "arxiv_id": "2502.01705", "paper_pdf": "papers/FQ8KX1pxeo.pdf", "paper_pdf_sha256": "9193806b34f46f6cecdf33d1ecf48f889ab500f6ec26cde374f9593752fef38a", "paper_pdf_bytes": 494632, "paper_pdf_source": "openreview", "code_url": "https://github.com/XIANGLONGYAN/PBS2P", "code_repository": "XIANGLONGYAN/PBS2P", "code_commit": "923548007fe2d40e146b47f4a05c9d9526b496d9", "code_archive": "repos/FQ8KX1pxeo.zip", "code_archive_sha256": "680299f689c80f1ce625ea3a2709009bdefb199668c32b09fb8c22aae4ed527c", "code_archive_bytes": 516334, "code_file_count": 89, "code_extensions": {".py": 88, ".sh": 1}, "github_disk_usage_kb": 2073, "github_languages": {"Python": 534298, "Shell": 624}, "github_archived": false, "github_pushed_at": "2026-07-11T12:52:02Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/progressive-binarization-with-semi-structured"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "xAM9VaXZnY", "year": 2025, "status": "rejected", "title": "What Can We Learn from State Space Models for Machine Learning on Graphs?", "authors": ["Yinan Huang", "Siqi Miao", "Pan Li"], "authorids": ["~Yinan_Huang1", "~Siqi_Miao1", "~Pan_Li2"], "authors_source": "OpenReview API", "abstract": "Machine learning on graphs has recently found extensive applications across domains. However, the commonly used Message Passing Neural Networks (MPNNs) suffer from limited expressive power and struggle to capture long-range dependencies. Graph transformers offer a strong alternative due to their global attention mechanism, but they come with great computational overheads, especially for large graphs. In recent years, State Space Models (SSMs) have emerged as a compelling approach to replace full attention in transformers to model sequential data. It blends the strengths of RNNs and CNNs, offering a) efficient computation, b) the ability to capture long-range dependencies, and c) good generalization across sequences of various lengths. However, extending SSMs to graph-structured data presents unique challenges due to the lack of canonical node ordering in graphs. In this work, we propose Graph State Space Convolution (GSSC) as a principled extension of SSMs to graph-structured data. By leveraging global permutation-equivariant set aggregation and factorizable graph kernels that rely on relative node distances as the convolution kernels, GSSC preserves all three advantages of SSMs. We demonstrate the provably stronger expressiveness of GSSC than MPNNs in counting graph substructures and show its effectiveness across 11 real-world, widely used benchmark datasets. GSSC achieves the best results on 6 out of 11 datasets with all significant improvements compared to the state-of-the-art baselines and second-best results on the other 5 datasets. Our findings highlight the potential of GSSC as a powerful and scalable model for graph machine learning. Anonymous code\nis available at https://anonymous.4open.science/r/GSSC-5ED8.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=xAM9VaXZnY", "arxiv_id": "2406.05815", "paper_pdf": "papers/xAM9VaXZnY.pdf", "paper_pdf_sha256": "afe402430c5f59e297a1a6ce55c6c5d557e2e9e1e976a107cebb7b93a7db4395", "paper_pdf_bytes": 2570169, "paper_pdf_source": "openreview", "code_url": "https://github.com/Graph-COM/GSSC", "code_repository": "Graph-COM/GSSC", "code_commit": "ebb0ce4c52b6904ff95514b68f4a72aa48dfeb0f", "code_archive": "repos/xAM9VaXZnY.zip", "code_archive_sha256": "aa0cd27bdfcc8c56bfd63cacfd2e128bdbea7f031148d1a5476ee379765cffc1", "code_archive_bytes": 189536, "code_file_count": 101, "code_extensions": {".py": 101}, "github_disk_usage_kb": 149, "github_languages": {"Python": 539696}, "github_archived": false, "github_pushed_at": "2024-06-11T15:24:35Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/what-can-we-learn-from-state-space-models-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "XbLffB0T2z", "year": 2024, "status": "rejected", "title": "Transferable Availability Poisoning Attacks", "authors": ["Yiyong Liu", "Michael Backes", "Xiao Zhang"], "authorids": ["~Yiyong_Liu1", "~Michael_Backes3", "~Xiao_Zhang2"], "authors_source": "OpenReview API", "abstract": "We study availability data poisoning attacks, where an adversary aims to degrade the overall test accuracy of a machine learning model by crafting imperceptible perturbations to its training data. Existing strategies can achieve the attack goal but usually assume that the victim employs the same training method as what the adversary uses to mount the attack. In this paper, we argue that this assumption is strong, since the victim may choose any learning algorithm to train the model as long as it achieves some targeted performance on clean data. In addition, we observe a large decrease in the effectiveness of prior poisoning attacks, when the victim uses a different learning paradigm to train the model, and marked differences in frequency-level characteristics between perturbations generated with different learners and attack methods. To enhance the attack transferability, we propose _Transferable Poisoning_, which generates high-frequency poisoning perturbations by alternately leveraging the gradient information with two specific algorithms respectively selected from supervised and unsupervised contrastive learning paradigms. Through extensive experiments on benchmark image datasets, we show that transferable poisoning can produce poisoned samples with significantly improved transferability, which not only applies to the two learners used to devise the attack but also works for learning algorithms and even paradigms beyond.", "decision": "Reject", "meta_review": "", "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=XbLffB0T2z", "arxiv_id": "2310.05141", "paper_pdf": "papers/XbLffB0T2z.pdf", "paper_pdf_sha256": "2f0bebb474ddba5509c11e0a75967c123478e6d32be1f4d14e90bd3721b6f7c0", "paper_pdf_bytes": 1121599, "paper_pdf_source": "openreview", "code_url": "https://github.com/TrustMLRG/TransPoison", "code_repository": "TrustMLRG/TransPoison", "code_commit": "2d08a9c0b10ff5da332b3c0b800c10c6dca8c518", "code_archive": "repos/XbLffB0T2z.zip", "code_archive_sha256": "b41d22510b8378efa95f5caa5ca257f1217b01d1bf27c59a61a65159eb2d4584", "code_archive_bytes": 138601, "code_file_count": 59, "code_extensions": {".py": 59}, "github_disk_usage_kb": 116, "github_languages": {"Python": 310559}, "github_archived": false, "github_pushed_at": "2023-10-11T09:29:25Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/transferable-availability-poisoning-attacks"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Z-CqSH6J_VK", "year": 2023, "status": "rejected", "title": "Differentiable and transportable structure learning", "authors": ["Jeroen Berrevoets", "Nabeel Seedat", "Fergus Imrie", "Mihaela van der Schaar"], "authorids": ["~Jeroen_Berrevoets1", "~Nabeel_Seedat1", "~Fergus_Imrie1", "~Mihaela_van_der_Schaar2"], "authors_source": "OpenReview API", "abstract": "Directed acyclic graphs (DAGs) encode a lot of information about a particular distribution in its structure. However, compute required to infer these structures is typically super-exponential in the number of variables, as inference requires a sweep of a combinatorially large space of potential structures. That is, until recent advances made it possible to search this space using a differentiable metric, drastically reducing search time. While this technique– named NOTEARS –is widely considered a seminal work in DAG-discovery, it concedes an important property in favour of differentiability: transportability. To be transportable, the structures discovered on one dataset must apply to another dataset from the same domain. In our paper, we introduce D-Struct which recovers transportability in the discovered structures through a novel architecture and loss function, while remaining completely differentiable. Because D-Struct remains differentiable, our method can be easily adopted in existing differentiable architectures, as was previously done with NOTEARS. In our experiments, we empirically validate D-Struct with respect to edge accuracy and structural Hamming distance in a variety of settings.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=Z-CqSH6J_VK", "arxiv_id": "2206.06354", "paper_pdf": "papers/Z-CqSH6J_VK.pdf", "paper_pdf_sha256": "f7faf8e2820d7794311bb16928bd4314bc338569e11ff3668b0aaa6573169e7f", "paper_pdf_bytes": 634533, "paper_pdf_source": "openreview", "code_url": "https://github.com/jeroenbe/d-struct", "code_repository": "jeroenbe/d-struct", "code_commit": "4fda078cb7015fd8fe2a9ad366ea93b36664ac5b", "code_archive": "repos/Z-CqSH6J_VK.zip", "code_archive_sha256": "e8e069deee3dfd27af0ea592bfce2bfbf6af69b53e588b8f398912ddac05e401", "code_archive_bytes": 125889, "code_file_count": 17, "code_extensions": {".sh": 9, ".py": 8}, "github_disk_usage_kb": 214, "github_languages": {"Python": 66972, "Shell": 18511}, "github_archived": false, "github_pushed_at": "2023-07-26T08:37:50Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/differentiable-and-transportable-structure"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "qDx6DXD3Fzt", "year": 2022, "status": "rejected", "title": "Provably Robust Detection of Out-of-distribution Data (almost) for free", "authors": ["Alexander Meinke", "Julian Bitterwolf", "Matthias Hein"], "authorids": ["~Alexander_Meinke1", "~Julian_Bitterwolf1", "~Matthias_Hein2"], "authors_source": "OpenReview API", "abstract": "The application of machine learning in safety-critical systems requires a reliable assessment of uncertainy. However, deep neural networks are known to produce highly overconfident predictions on out-of-distribution (OOD) data. Even if trained to be non-confident on OOD data one can still adversarially manipulate OOD data so that the classifier again assigns high confidence to the manipulated samples. In this paper we propose a novel method that combines a certifiable OOD detector with a standard classifier from first principles into an OOD aware classifier. This way we achieve the best of two worlds: certifiably adversarially robust OOD detection, even for OOD samples close to the in-distribution, without loss in either prediction accuracy or detection performance for non-manipulated OOD data. Moreover, due to the particular construction our classifier provably avoids the asymptotic overconfidence problem of standard neural networks.", "decision": "Reject", "meta_review": null, "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=qDx6DXD3Fzt", "arxiv_id": "2106.04260", "paper_pdf": "papers/qDx6DXD3Fzt.pdf", "paper_pdf_sha256": "a6ff5fa80bfc8dadf75d47b4a4f9f12c0464a901d5346415947a0b2e0b0a7b96", "paper_pdf_bytes": 547405, "paper_pdf_source": "openreview", "code_url": "https://github.com/AlexMeinke/Provable-OOD-Detection", "code_repository": "AlexMeinke/Provable-OOD-Detection", "code_commit": "9cef8d284811b938a6182835c4fd16590b1c2699", "code_archive": "repos/qDx6DXD3Fzt.zip", "code_archive_sha256": "3b3ff895cbf1dd4a24c0c7428cb7247089303265d8b670bd22e759819cab605e", "code_archive_bytes": 483105, "code_file_count": 32, "code_extensions": {".py": 32}, "github_disk_usage_kb": 483, "github_languages": {"Python": 302833}, "github_archived": false, "github_pushed_at": "2022-10-12T18:23:17Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/provably-robust-detection-of-out-of"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "fgX9O5q0BT", "year": 2021, "status": "rejected", "title": "On Noise Injection in Generative Adversarial Networks", "authors": ["Ruili Feng", "Deli Zhao", "Zheng-Jun Zha"], "authorids": ["~Ruili_Feng1", "~Deli_Zhao1", "~Zheng-Jun_Zha2"], "authors_source": "OpenReview API", "abstract": "Noise injection is an effective way of circumventing overfitting  and enhancing generalization in machine learning, the rationale of which has been validated in deep learning as well.  Recently, noise injection exhibits surprising performance when\n  generating high-fidelity images in Generative Adversarial Networks (GANs). Despite its successful applications in GANs, the mechanism of its validity is still unclear. In this paper, we propose a geometric framework to theoretically analyze the role of noise injection in GANs. Based on Riemannian geometry, we successfully model the noise injection framework as fuzzy equivalence on geodesic normal coordinates. Guided by our theories, we find that existing methods are incomplete and a new strategy for noise injection is devised. Experiments on image generation and GAN inversion demonstrate the superiority of our method.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=fgX9O5q0BT", "arxiv_id": "2006.05891", "paper_pdf": "papers/fgX9O5q0BT.pdf", "paper_pdf_sha256": "0eb1fce18b442aa63f72e38354ef8b77c5cf04c2ca1e8f05c33d7152724b9328", "paper_pdf_bytes": 4516587, "paper_pdf_source": "openreview", "code_url": "https://github.com/NVlabs/stylegan2", "code_repository": "NVlabs/stylegan2", "code_commit": "bf0fe0baba9fc7039eae0cac575c1778be1ce3e3", "code_archive": "repos/fgX9O5q0BT.zip", "code_archive_sha256": "2594614039946bb63b347c8d4e1a50609833062ccccc1df78d2183385895b82b", "code_archive_bytes": 599614, "code_file_count": 41, "code_extensions": {".py": 38, ".cu": 3}, "github_disk_usage_kb": 594, "github_languages": {"Python": 369008, "Cuda": 23649, "Dockerfile": 362}, "github_archived": false, "github_pushed_at": "2024-05-18T15:55:42Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/on-noise-injection-in-generative-adversarial"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "B1xZD1rtPr", "year": 2020, "status": "rejected", "title": "The Dual Information Bottleneck", "authors": ["Zoe Piran", "Naftali Tishby"], "authorids": ["zoe.piran@mail.huji.ac.il", "tishby@cs.huji.ac.il"], "authors_source": "OpenReview API", "abstract": "The Information-Bottleneck (IB) framework suggests a general characterization of optimal representations in learning, and deep learning in particular. It is based on the optimal trade off between the representation complexity and accuracy, both of which are quantified by mutual information. The problem is solved by alternating projections between the encoder and decoder of the representation, which can be performed locally at each representation level. The framework, however, has practical drawbacks, in that mutual information is notoriously difficult to handle at high dimension, and only has closed form solutions in special cases. Further, because it aims to extract representations which are minimal sufficient statistics of the data with respect to the desired label, it does not necessarily optimize the actual prediction of unseen labels. Here we present a  formal dual problem to the IB which has several interesting properties. By switching the order in the KL-divergence between the representation decoder and data, the optimal decoder becomes the geometric rather than the arithmetic mean of the input points. While providing a good approximation to the original IB, it also preserves the form of exponential families, and optimizes the mutual information on the predicted label rather than the desired one. We also analyze the critical points of the dualIB and discuss their importance for the quality of this approach.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=B1xZD1rtPr", "arxiv_id": "2006.04641", "paper_pdf": "papers/B1xZD1rtPr.pdf", "paper_pdf_sha256": "0036af0f6acc1d04c71bb1750bbdacf0b050b30baaf0f2e13b450e4aeab132f3", "paper_pdf_bytes": 2153987, "paper_pdf_source": "openreview", "code_url": "https://github.com/ravidziv/dual_IB", "code_repository": "ravidziv/dual_IB", "code_commit": "a3c8ae96a0581c467c00241b0fa7da7b7d50ddaa", "code_archive": "repos/B1xZD1rtPr.zip", "code_archive_sha256": "ee5316ab1067489f9f70b0ca73e89a3b5eb851070d69a04a54c955d41117ae1e", "code_archive_bytes": 81260, "code_file_count": 8, "code_extensions": {".py": 8}, "github_disk_usage_kb": 1291, "github_languages": {"Python": 46844}, "github_archived": false, "github_pushed_at": "2024-10-22T14:31:25Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/the-dual-information-bottleneck-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "PTTmPHS7OE", "year": 2026, "status": "rejected", "title": "LocationReasoner: Evaluating LLMs on Real-World Site Selection Reasoning", "authors": ["Miho Koda", "Yu Zheng", "Ruixian Ma", "Mingyang Sun", "Devesh Pansare", "Fabio Duarte", "Paolo Santi"], "authorids": ["~Miho_Koda1", "~Yu_Zheng7", "~Ruixian_Ma1", "~Mingyang_Sun7", "~Devesh_Pansare1", "~Fabio_Duarte1", "~Paolo_Santi2"], "authors_source": "OpenReview API", "abstract": "Recent advances in large language models (LLMs), particularly those enhanced through reinforced post-training, have demonstrated impressive reasoning capabilities, as exemplified by models such as OpenAI o1 and DeepSeek-R1. However, these capabilities are predominantly benchmarked on domains like mathematical problem solving and code generation, leaving open the question of whether such reasoning skills generalize to complex real-world scenarios. In this paper, we introduce LocationReasoner, a benchmark designed to evaluate LLMs’ reasoning abilities in the context of real-world site selection, where models must identify feasible locations by reasoning over diverse and complicated spatial, environmental, and logistic constraints. The benchmark covers carefully crafted queries of varying\ndifficulty levels and is supported by a sandbox environment with in-house tools for constraint-based location search. Automated verification further guarantees the scalability of the benchmark, enabling the addition of arbitrary number of queries. Extensive evaluations on real-world site selection data from Boston, New York, and Tampa reveal that state-of-the-art reasoning models offer limited improvement over their non-reasoning predecessors in real-world contexts, with even the latest OpenAI o4 model failing on 30% of site selection tasks. Moreover, agentic strategies such as ReAct and Reflexion often suffer from over-reasoning, leading to worse outcomes than direct prompting. With key limitations of LLMs in holistic and non-linear reasoning highlighted, we release LocationReasoner to foster the development of LLMs and agents capable of robust, grounded reasoning in real-world decision-making tasks. Codes and data for our benchmark are available at https://anonymous.4open.science/r/LocationReasoner-DC5D.", "decision": "Reject", "meta_review": "", "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=PTTmPHS7OE", "arxiv_id": "2506.13841", "paper_pdf": "papers/PTTmPHS7OE.pdf", "paper_pdf_sha256": "17f13fb35126f8d12d59a61c92c3214b22689d4d6a49b9b810203f8bb9bd49ee", "paper_pdf_bytes": 341084, "paper_pdf_source": "openreview", "code_url": "https://github.com/miho-koda/LocationReasoner", "code_repository": "miho-koda/LocationReasoner", "code_commit": "54c643e3cad533325238530840e3146556e59b24", "code_archive": "repos/PTTmPHS7OE.zip", "code_archive_sha256": "9e1d4e87f3760a955f533ddd5d09b86d311e6f9e225563fe6ca5ef1fef33672b", "code_archive_bytes": 249029, "code_file_count": 89, "code_extensions": {".py": 87, ".ipynb": 2}, "github_disk_usage_kb": 2087, "github_languages": {"Python": 851288, "Jupyter Notebook": 44029}, "github_archived": false, "github_pushed_at": "2026-07-13T02:00:18Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/locationreasoner-evaluating-llms-on-real"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "YGflij9S6x", "year": 2025, "status": "rejected", "title": "Continual Learning: Less Forgetting, More OOD Generalization via Adaptive Contrastive Replay", "authors": ["Hossein Rezaei", "Mohammad Khalooei", "Mohammad Sabokrou"], "authorids": ["~Hossein_Rezaei1", "~Mohammad_Khalooei2", "~Mohammad_Sabokrou1"], "authors_source": "OpenReview API", "abstract": "Machine learning models often suffer from catastrophic forgetting of previously learned knowledge when learning new classes. Various methods have been proposed to mitigate this issue. However, rehearsal-based learning, which retains samples from previous classes, typically achieves good performance but tends to memorize specific instances, struggling with Out-of-Distribution (OOD) generalization. This often leads to high forgetting rates and poor generalization. Surprisingly, the OOD generalization capabilities of these methods have been largely unexplored. In this paper, we highlight this issue and propose a simple yet effective strategy inspired by contrastive learning and data-centric principles to address it.\nWe introduce Adaptive Contrastive Replay (ACR), a method that employs dual optimization to simultaneously train both the encoder and the classifier. ACR adaptively populates the replay buffer with misclassified samples while ensuring a balanced representation of classes and tasks. By refining the decision boundary in this way, ACR achieves a balance between stability and plasticity. Our method significantly outperforms previous approaches in terms of OOD generalization, achieving an improvement of 13.41\\% on Split CIFAR-100, 9.91\\% on Split Mini-ImageNet, and 5.98\\% on Split Tiny-ImageNet.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=YGflij9S6x", "arxiv_id": "2410.07110", "paper_pdf": "papers/YGflij9S6x.pdf", "paper_pdf_sha256": "c3b34cf1533b20adf3802b894db7af1344f5ce1155f76d525b861728ce7947c6", "paper_pdf_bytes": 3559766, "paper_pdf_source": "openreview", "code_url": "https://github.com/hossein-rezaei624/ACR", "code_repository": "hossein-rezaei624/ACR", "code_commit": "339b4734065e203e9ae1da289573de6f515806cf", "code_archive": "repos/YGflij9S6x.zip", "code_archive_sha256": "d515281c32aabf365268c115e5964a5332432b6359013e53f325d86037e3fe83", "code_archive_bytes": 151142, "code_file_count": 83, "code_extensions": {".py": 83}, "github_disk_usage_kb": 151, "github_languages": {"Python": 472469}, "github_archived": false, "github_pushed_at": "2024-10-02T10:03:46Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/continual-learning-less-forgetting-more-ood"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "T16M4SzH1v", "year": 2024, "status": "rejected", "title": "Distributional Bellman Operators over Mean Embeddings", "authors": ["Li Kevin Wenliang", "Gregoire Deletang", "Matthew Aitchison", "Marcus Hutter", "Anian Ruoss", "Arthur Gretton", "Mark Rowland"], "authorids": ["~Li_Kevin_Wenliang1", "~Gregoire_Deletang1", "~Matthew_Aitchison1", "~Marcus_Hutter1", "~Anian_Ruoss1", "~Arthur_Gretton1", "~Mark_Rowland1"], "authors_source": "OpenReview API", "abstract": "We propose a novel algorithmic framework for distributional reinforcement learning, based on learning finite-dimensional mean embeddings  of return distributions. We derive several new algorithms for dynamic programming and temporal-difference learning based on this framework, provide asymptotic convergence theory, and examine the empirical performance of the algorithms on a suite of tabular tasks.   Further, we show that this approach can be straightforwardly combined with deep reinforcement learning, and obtain a new deep reinforcement learning agent that improves over baseline distributional approaches on the Arcade Learning Environment.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=T16M4SzH1v", "arxiv_id": "2312.07358", "paper_pdf": "papers/T16M4SzH1v.pdf", "paper_pdf_sha256": "ed76f24e59312599fa3c165dbc2fba6fd740cd02bb030c587768a7a851026ace", "paper_pdf_bytes": 1010551, "paper_pdf_source": "openreview", "code_url": "https://github.com/google-deepmind/sketch_dqn", "code_repository": "google-deepmind/sketch_dqn", "code_commit": "b8cf8133fb9e33f3557f901c4136c9372997f0f0", "code_archive": "repos/T16M4SzH1v.zip", "code_archive_sha256": "1e1ab23ef8528fd981852c77ad3a9b3e281788158b2d74ea4964a41a9e0c3abe", "code_archive_bytes": 72195, "code_file_count": 20, "code_extensions": {".py": 19, ".sh": 1}, "github_disk_usage_kb": 117, "github_languages": {"Python": 202590, "Shell": 1871, "Dockerfile": 1748}, "github_archived": false, "github_pushed_at": "2026-04-13T23:44:04Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/distributional-bellman-operators-over-mean"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "jCpTofV7iY_", "year": 2023, "status": "rejected", "title": "Pre-trained Language Models can be Fully Zero-Shot Learners", "authors": ["Xuandong Zhao", "Siqi Ouyang", "Zhiguo Yu", "Ming Wu", "Lei Li"], "authorids": ["~Xuandong_Zhao1", "~Siqi_Ouyang2", "~Zhiguo_Yu2", "~Ming_Wu4", "~Lei_Li11"], "authors_source": "OpenReview API", "abstract": "How can we extend a pre-trained model to many language understanding tasks, without labeled or additional unlabeled data? Pre-trained language models (PLMs) have been effective for a wide range of NLP tasks. However, existing approaches either require fine-tuning on downstream labeled datasets or manually constructing proper prompts. In this paper, we propose nonparametric prompting PLM (NPPrompt) for fully zero-shot language understanding. Unlike previous methods, NPPrompt uses only pre-trained language models and does not require any labeled data or additional raw corpus for further fine-tuning, nor does it rely on humans to construct a comprehensive set of prompt label words. We evaluate NPPrompt against previous major few-shot and zero-shot learning methods on diverse NLP tasks: including text classification, text entailment, similar text retrieval, and paraphrasing. Experimental results demonstrate that our NPPrompt outperforms the previous best fully zero-shot method by big margins, with absolute gains of 12.8% in accuracy on text classification and 18.9% on the GLUE benchmark.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=jCpTofV7iY_", "arxiv_id": "2212.06950", "paper_pdf": "papers/jCpTofV7iY_.pdf", "paper_pdf_sha256": "771c9cd1dea75b28c1d41fbed92df94bd5920d1652e8b9b12911f41893873204", "paper_pdf_bytes": 434752, "paper_pdf_source": "openreview", "code_url": "https://github.com/XuandongZhao/NPPrompt", "code_repository": "XuandongZhao/NPPrompt", "code_commit": "2c30c7ba47bbbccd8e3e627e7b06b16c5229cf52", "code_archive": "repos/jCpTofV7iY_.zip", "code_archive_sha256": "d55f1e35c947d1fa2c3ed6f3b59bf282ec57106d180443f46450a4bcba3996f2", "code_archive_bytes": 266871, "code_file_count": 68, "code_extensions": {".py": 66, ".sh": 2}, "github_disk_usage_kb": 216, "github_languages": {"Python": 555249, "Shell": 2161}, "github_archived": false, "github_pushed_at": "2024-02-02T18:33:05Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/pre-trained-language-models-can-be-fully-zero"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Fj1Tpym9KxH", "year": 2022, "status": "rejected", "title": "A Closer Look at Smoothness in Domain Adversarial Training", "authors": ["Harsh Rangwani", "Sumukh K Aithal", "Arihant Jain", "Venkatesh Babu Radhakrishnan"], "authorids": ["~Harsh_Rangwani1", "~Sumukh_K_Aithal1", "arihantjain@iisc.ac.in", "~Venkatesh_Babu_Radhakrishnan2"], "authors_source": "OpenReview API", "abstract": "Domain adversarial training has been ubiquitous for achieving invariant representations and is used widely for various domain adaptation tasks. In recent times methods converging to smooth optima have shown improved generalization for supervised learning tasks like classification.  In this work, we analyze the effect of smoothness enhancing formulations on domain adversarial training, the objective of which is a combination of classification and adversarial terms. In contrast to classification loss, our analysis shows that \\textit{converging to smooth minima w.r.t. adversarial loss leads to sub-optimal generalization on the target domain}. Based on the analysis, we introduce the Smooth Domain Adversarial training (SDAT) procedure, which effectively enhances the performance of existing domain adversarial methods for both classification and object detection tasks. Our smoothness analysis also provides insight into the extensive usage of SGD over Adam in domain adversarial training.   ", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=Fj1Tpym9KxH", "arxiv_id": "2206.08213", "paper_pdf": "papers/Fj1Tpym9KxH.pdf", "paper_pdf_sha256": "5f2fca3ef01ecf3c34dbe7f76aea417c827bd988feba63940c8e4d2de81717b9", "paper_pdf_bytes": 3183844, "paper_pdf_source": "openreview", "code_url": "https://github.com/val-iisc/SDAT", "code_repository": "val-iisc/SDAT", "code_commit": "33d7429d37522972edda608cb802a686e4b9e794", "code_archive": "repos/Fj1Tpym9KxH.zip", "code_archive_sha256": "27f4c120735b5d9a03c5a947f5e1e2b9c57ff65b672f72ac3768e65cf4a78811", "code_archive_bytes": 541112, "code_file_count": 39, "code_extensions": {".py": 37, ".sh": 2}, "github_disk_usage_kb": 491, "github_languages": {"Python": 84834}, "github_archived": false, "github_pushed_at": "2024-04-11T22:11:25Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/a-closer-look-at-smoothness-in-domain-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "aYbCpFNnHdh", "year": 2021, "status": "rejected", "title": "Visual Question Answering From Another Perspective: CLEVR Mental Rotation Tests", "authors": ["Christopher Beckham", "Martin Weiss", "Florian Golemo", "Sina Honari", "Derek Nowrouzezahrai", "Christopher Pal"], "authorids": ["~Christopher_Beckham1", "~Martin_Weiss4", "~Florian_Golemo1", "~Sina_Honari1", "~Derek_Nowrouzezahrai1", "~Christopher_Pal1"], "authors_source": "OpenReview API", "abstract": "Different types of \\emph{mental rotation tests} have been used extensively in psychology to understand human visual reasoning and perception. Understanding what an object or visual scene would look like from another viewpoint is a challenging problem that is made even harder if it must be performed from a single image. 3D computer vision has a long history of examining related problems. However, often what one is most interested in is the answer to a relatively simple question posed in another visual frame of reference -- as opposed to creating a full 3D reconstruction. \nMental rotations tests can also manifest as consequential questions in the real world such as: does the pedestrian that I see, see the car that I am driving?\nWe explore a controlled setting whereby questions are posed about the properties of a scene if the scene were observed from another viewpoint. To do this we have created a new version of the CLEVR VQA problem setup and dataset that we call CLEVR Mental Rotation Tests or CLEVR-MRT, where the goal is to answer questions about the original CLEVR viewpoint given a single image obtained from a different viewpoint of the same scene. Using CLEVR Mental Rotation Tests we examine standard state of the art methods, show how they fall short, then explore novel neural architectures that involve inferring representations encoded as feature volumes describing a scene. Our new methods use rigid transformations of feature volumes conditioned on the viewpoint camera. We examine the efficacy of different model variants through performing a rigorous ablation study. Furthermore, we examine the use of contrastive learning to infer a volumetric encoder in a self-supervised manner and find that this approach yields the best results of our study using CLEVR-MRT.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=aYbCpFNnHdh", "arxiv_id": "2212.01639", "paper_pdf": "papers/aYbCpFNnHdh.pdf", "paper_pdf_sha256": "854dad77542a35aba62feb94b5f0507e981ba37c198c618690f1dcb583188815", "paper_pdf_bytes": 3838909, "paper_pdf_source": "openreview", "code_url": "https://github.com/christopher-beckham/clevr-mrt", "code_repository": "christopher-beckham/clevr-mrt", "code_commit": "3eb04b7fb8348b0620ae569bc0f99dc400c8d51a", "code_archive": "repos/aYbCpFNnHdh.zip", "code_archive_sha256": "368b2e9cac3cbce77014a34c0417117f3b4caf5954d2ca003dcb4e0af0347e81", "code_archive_bytes": 197126, "code_file_count": 79, "code_extensions": {".py": 76, ".sh": 3}, "github_disk_usage_kb": 596, "github_languages": {"Python": 476242, "Shell": 905, "Dockerfile": 698}, "github_archived": false, "github_pushed_at": "2023-04-11T16:30:15Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/visual-question-answering-from-another-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "SylpBgrKPH", "year": 2020, "status": "rejected", "title": "MissDeepCausal: causal inference from incomplete data using deep latent variable models", "authors": ["Julie Josse", "Imke Mayer", "Jean-Philippe Vert"], "authorids": ["julie.josse@polytechnique.edu", "imke.mayer@polytechnique.edu", "jpvert@google.com"], "authors_source": "OpenReview API", "abstract": "Inferring causal effects of a treatment, intervention or policy from observational data is central to many applications. However, state-of-the-art methods for causal inference seldom consider the possibility that covariates have missing values, which is ubiquitous in many real-world analyses.  Missing data greatly complicate causal inference procedures as they require an adapted unconfoundedness hypothesis which can be difficult to justify in practice. We circumvent this issue by considering latent confounders whose distribution is learned through variational autoencoders adapted to missing values. They can be used either as a pre-processing step prior to causal inference but we also suggest to embed them in a multiple imputation strategy to take into account the variability due to missing values.  Numerical experiments demonstrate the effectiveness of the proposed methodology especially for non-linear models compared to competitors.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=SylpBgrKPH", "arxiv_id": "2002.10837", "paper_pdf": "papers/SylpBgrKPH.pdf", "paper_pdf_sha256": "c6ef826c8bd03fb13249cd1a147351c35c1ab098c10bd274128fc548a705ee6b", "paper_pdf_bytes": 504496, "paper_pdf_source": "openreview", "code_url": "https://github.com/imkemayer/MissDeepCausal", "code_repository": "imkemayer/MissDeepCausal", "code_commit": "6faecaa47234c457af54d492aa669a8f8b3b6b04", "code_archive": "repos/SylpBgrKPH.zip", "code_archive_sha256": "c2976250071e00561b1dd871d5a09e0aae1b322195325e6d99a57351ba6e5906", "code_archive_bytes": 1241821, "code_file_count": 20, "code_extensions": {".py": 20}, "github_disk_usage_kb": 1292, "github_languages": {"Python": 192281}, "github_archived": false, "github_pushed_at": "2022-12-08T07:30:46Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/missdeepcausal-causal-inference-from-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "MYmHznX3aB", "year": 2026, "status": "rejected", "title": "Information Bargaining: Bilateral Commitment in Bayesian Persuasion", "authors": ["Yue Lin", "Shuhui Zhu", "William A Cunningham", "Wenhao Li", "Pascal Poupart", "Hongyuan Zha", "Baoxiang Wang"], "authorids": ["~Yue_Lin2", "~Shuhui_Zhu1", "~William_A_Cunningham1", "~Wenhao_Li2", "~Pascal_Poupart2", "~Hongyuan_Zha1", "~Baoxiang_Wang1"], "authors_source": "OpenReview API", "abstract": "Bayesian persuasion studies how an informed sender can influence a receiver’s actions through committed signaling schemes. While effective in one-shot settings, extending Bayesian persuasion to real-world long-term interactions becomes NP-hard.\nA separate empirical mismatch is also evident, where people tend to be much more truthful in practice than the signaling scheme predicted by Bayesian persuasion equilibria.\nTo address these issues, we first prove that long-term Bayesian persuasion can be decomposed into a bargaining stage and a realization stage while preserving optimality and equilibria. This decomposition disentangles the previously conflated informational and first-mover advantages of the sender. Based on the results, we establish a unified, realistic, and fairness-oriented framework called \\emph{information bargaining}. \nVariants of bargaining-game settings and their associated solution concepts, such as the Nash cooperative bargaining model and the Nash bargaining solution, can be applied directly to persuasion games to yield more realistic models.\nTo validate our framework, we identify and employ capable reasoning LLMs that can solve persuasion game equilibria effectively. In long-term persuasion task variants and in the corresponding bargaining game variants, these capable LLMs demonstrate that they reach the same equilibrium.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=MYmHznX3aB", "arxiv_id": "2506.05876", "paper_pdf": "papers/MYmHznX3aB.pdf", "paper_pdf_sha256": "4eb9c8c62cf3a9f3b104a4ab9acc11030cc6e97afd2c1cd69fd06759ab94c6c9", "paper_pdf_bytes": 477451, "paper_pdf_source": "openreview", "code_url": "https://github.com/YueLin301/InformationBargaining", "code_repository": "YueLin301/InformationBargaining", "code_commit": "d5fa03c60d02d2e7c4572cc21dd8c1bff14f4be8", "code_archive": "repos/MYmHznX3aB.zip", "code_archive_sha256": "f2af77294af1f13a090b88433fcdd3146569d45027d7508a92dbd5170ac11b59", "code_archive_bytes": 4480690, "code_file_count": 9, "code_extensions": {".py": 9}, "github_disk_usage_kb": 2100, "github_languages": {"Python": 78794}, "github_archived": false, "github_pushed_at": "2025-06-06T07:37:29Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/information-bargaining-bilateral-commitment"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "BdPbmgJ2jo", "year": 2025, "status": "rejected", "title": "High-dimensional Asymptotics of VAEs: Threshold of Posterior Collapse and Dataset-Size Dependence of Rate-Distortion Curve", "authors": ["Yuma Ichikawa", "Koji Hukushima"], "authorids": ["~Yuma_Ichikawa1", "~Koji_Hukushima1"], "authors_source": "OpenReview API", "abstract": "In variational autoencoders (VAEs), the variational posterior often aligns closely with the prior, known as posterior collapse, which leads to poor representation learning quality. An adjustable hyperparameter beta has been introduced in VAE to address this issue. This study sharply evaluates the conditions under which the posterior collapse occurs with respect to beta and dataset size by analyzing a minimal VAE in a high-dimensional limit. Additionally, this setting enables the evaluation of the rate-distortion curve in the VAE. This result shows that, unlike typical regularization parameters, VAEs face \"inevitable posterior collapse\" beyond a certain beta threshold, regardless of dataset size. The dataset-size dependence of the derived rate-distortion curve also suggests that relatively large datasets are required to achieve a rate-distortion curve with high rates. These results robustly explain generalization behavior across various real datasets with highly non-linear VAEs.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=BdPbmgJ2jo", "arxiv_id": "2309.07663", "paper_pdf": "papers/BdPbmgJ2jo.pdf", "paper_pdf_sha256": "e2274e498095f338820ba10d7ccbd5f6f0fd4aaeb6ba087658ab73ed254ae50d", "paper_pdf_bytes": 1124052, "paper_pdf_source": "openreview", "code_url": "https://github.com/Yuma-Ichikawa/vae-replica", "code_repository": "Yuma-Ichikawa/vae-replica", "code_commit": "9f6f430e2a6399e689984077921706fc02fa2a4e", "code_archive": "repos/BdPbmgJ2jo.zip", "code_archive_sha256": "a05463930303dec4c91ea8bc842b972884be4f990d7b9172235b0ecdee4ce12b", "code_archive_bytes": 59985, "code_file_count": 5, "code_extensions": {".py": 4, ".ipynb": 1}, "github_disk_usage_kb": 152, "github_languages": {"Jupyter Notebook": 211800, "Python": 8665}, "github_archived": false, "github_pushed_at": "2025-03-20T02:03:18Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/dataset-size-dependence-of-rate-distortion"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "AfiM6F2YPY", "year": 2024, "status": "rejected", "title": "Applying language models to algebraic topology: generating simplicial cycles using multi-labeling in Wu's formula", "authors": ["Kirill Brilliantov", "Fedor Pavutnitskiy", "Dmitry Pasechnyuk", "German Magai"], "authorids": ["~Kirill_Brilliantov1", "~Fedor_Pavutnitskiy1", "~Dmitry_Pasechnyuk1", "~German_Magai1"], "authors_source": "OpenReview API", "abstract": "Computing homotopy groups of spheres has long been a fundamental objective in algebraic topology. Various theoretical and algorithmic approaches have been developed to tackle this problem. In this paper we take a step towards the goal of comprehending the group-theoretic structure of the generators of these homotopy groups by leveraging the power of machine learning. Specifically, in the simplicial group setting of Wu's formula, we reformulate the problem of generating simplicial cycles as a problem of sampling from the intersection of algorithmic datasets related to Dyck languages. We present and evaluate language modelling approaches that employ multi-label information for input sequences, along with the necessary group-theoretic toolkit and non-neural baselines.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=AfiM6F2YPY", "arxiv_id": "2306.16951", "paper_pdf": "papers/AfiM6F2YPY.pdf", "paper_pdf_sha256": "24d2320351b78649e01beba2a5acb2c31220c1a93bf8b1f08254e2360be1ec2c", "paper_pdf_bytes": 785192, "paper_pdf_source": "openreview", "code_url": "https://github.com/ml-in-algebraic-topology/gen-simplicial-cycles", "code_repository": "ml-in-algebraic-topology/gen-simplicial-cycles", "code_commit": "e9149c42de7c586704be1f593d36193561622911", "code_archive": "repos/AfiM6F2YPY.zip", "code_archive_sha256": "a863891145e16f041efbba923623dde580b43523c5b7b4ecfd500f91b84d15b3", "code_archive_bytes": 90952, "code_file_count": 5, "code_extensions": {".py": 4, ".ipynb": 1}, "github_disk_usage_kb": 120, "github_languages": {"Jupyter Notebook": 40765, "Python": 19388}, "github_archived": false, "github_pushed_at": "2024-05-28T09:23:52Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/applying-language-models-to-algebraic"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ZBMpG7fWwOP", "year": 2023, "status": "rejected", "title": "Game Theoretic Mixed Experts for Combinational Adversarial Machine Learning", "authors": ["Ethan Rathbun", "Kaleel Mahmood", "Sohaib Ahmad", "Caiwen Ding", "Marten van Dijk"], "authorids": ["~Ethan_Rathbun1", "~Kaleel_Mahmood1", "sohaib.ahmad@uconn.edu", "~Caiwen_Ding1", "~Marten_van_Dijk1"], "authors_source": "OpenReview API", "abstract": "Recent advances in adversarial machine learning have shown that defenses considered to be robust are actually susceptible to adversarial attacks which are specifically tailored to target their weaknesses. These defenses include Barrage of Random Transforms (BaRT), Friendly Adversarial Training (FAT), Trash is Treasure (TiT) and ensemble models made up of Vision Transformers (ViTs), Big Transfer models and Spiking Neural Networks (SNNs). It remains an open question, however, as to whether the adversarial examples designed to target one defense will be similarly misclassified by another defense. In this paper, we provide the first adversarial defense transferability study, as well as a game theoretic framework for ensemble adversarial attacks and defenses. Our framework is called Game theoretic Mixed Experts (GaME) and is designed to find the Mixed-Nash strategy for an attacker that can employ compositional adversarial attacks. We show that this framework creates an ensemble of defenses with greater robustness than a combinational defense with a uniform or random probability distribution. Overall, our framework and analyses advance the field of adversarial machine learning by yielding new insights into compositional attack and defense formulations.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=ZBMpG7fWwOP", "arxiv_id": "2211.14669", "paper_pdf": "papers/ZBMpG7fWwOP.pdf", "paper_pdf_sha256": "9b699b192900249e4300f19a18b3a74f4249bb6682af18d300aadef44fc7e8c2", "paper_pdf_bytes": 323014, "paper_pdf_source": "openreview", "code_url": "https://github.com/EthanRath/Game-Theoretic-Mixed-Experts", "code_repository": "EthanRath/Game-Theoretic-Mixed-Experts", "code_commit": "561d3e7549dff1f4d494f6622dc558403d57e3ce", "code_archive": "repos/ZBMpG7fWwOP.zip", "code_archive_sha256": "92b00fe0bf6109649fc1ce5db0e36177848063f8a5e8f324e6727c94c6b6f273", "code_archive_bytes": 247566, "code_file_count": 36, "code_extensions": {".py": 36}, "github_disk_usage_kb": 221, "github_languages": {"Python": 636880}, "github_archived": false, "github_pushed_at": "2024-03-24T23:51:39Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/game-theoretic-mixed-experts-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "1oEvY1a67c1", "year": 2022, "status": "rejected", "title": "If your data distribution shifts, use self-learning", "authors": ["Evgenia Rusak", "Steffen Schneider", "George Pachitariu", "Luisa Eck", "Peter Vincent Gehler", "Oliver Bringmann", "Wieland Brendel", "Matthias Bethge"], "authorids": ["~Evgenia_Rusak1", "~Steffen_Schneider1", "~George_Pachitariu2", "~Luisa_Eck1", "~Peter_Vincent_Gehler1", "~Oliver_Bringmann1", "~Wieland_Brendel1", "~Matthias_Bethge1"], "authors_source": "OpenReview API", "abstract": "In this paper, we demonstrate that self-learning techniques like entropy minimization or pseudo-labeling are simple, yet effective techniques for increasing test performance under domain shifts. Our results show that self-learning consistently increases performance under distribution shifts, irrespective of the model architecture, the pre-training technique or the type of distribution shift. At the same time, self-learning is simple to use in practice because it does not require knowledge or access to the original training data or scheme, is robust to hyperparameter choices, is straight-forward to implement and requires only a few training epochs. This makes self-learning techniques highly attractive for any practitioner who applies machine learning algorithms in the real world. We present state-of-the art adaptation results on CIFAR10-C (8.5% error),  ImageNet-C (22.0% mCE), ImageNet-R (17.4% error) and ImageNet-A (14.8% error), theoretically study the dynamics of self-supervised adaptation methods and propose a new classification dataset (ImageNet-D) which is challenging even with adaptation.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=1oEvY1a67c1", "arxiv_id": "2104.12928", "paper_pdf": "papers/1oEvY1a67c1.pdf", "paper_pdf_sha256": "784e52fd47b843a3bef9af8daff038edf7e39e9df1e1379911ddb9bdb81e42d3", "paper_pdf_bytes": 1652815, "paper_pdf_source": "openreview", "code_url": "https://github.com/bethgelab/robustness", "code_repository": "bethgelab/robustness", "code_commit": "0ef82b178e3526f63d8017dc870072d611d883e9", "code_archive": "repos/1oEvY1a67c1.zip", "code_archive_sha256": "19cdd6297102dc21c81b704b16b852d4a76b2be403853f862bdd47661a5408d1", "code_archive_bytes": 478310, "code_file_count": 39, "code_extensions": {".py": 33, ".sh": 5, ".ipynb": 1}, "github_disk_usage_kb": 512, "github_languages": {"Python": 110845, "Shell": 93}, "github_archived": false, "github_pushed_at": "2023-07-05T23:11:00Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/adapting-imagenet-scale-models-to-complex"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "86t2GlfzFo", "year": 2021, "status": "rejected", "title": "Deep Curvature Suite", "authors": ["Diego Granziol", "Xingchen Wan", "Timur Garipov"], "authorids": ["~Diego_Granziol1", "~Xingchen_Wan1", "~Timur_Garipov1"], "authors_source": "OpenReview API", "abstract": "The curvature of the loss, provides rich information on the geometry underlying neural networks, with applications in second order optimisation and Bayesian deep learning. However, accessing curvature information is still a daunting engineering challenge, inaccessible to most practitioners. We hence provide a software package the \\textbf{Deep Curvature Suite}, which allows easy curvature evaluation for large modern neural networks. Beyond the calculation of a highly accurate moment matched approximation of the Hessian spectrum using Lanczos, our package provides: extensive \\emph{loss surface visualisation}, the calculation of the \\emph{Hessian variance} and \\emph{stochastic second order optimisers}. We further address and disprove many common misconceptions in the literature about the Lanczos algorithm, namely that it learns eigenvalues from the top down. We prove using high dimensional concentration inequalities that for specific matrices a single random vector is sufficient for accurate spectral estimation, informing our spectral visualisation method. We showcase our package practical utility on a series of examples based on realistic modern neural networks such as the VGG-$16$ and Preactivated ResNets on the CIFAR-$10$/$100$ datasets. We further detail $3$ specific potential use cases enabled by our software: research in stochastic second order optimisation for deep learning, learning rate scheduling using known optimality formulae for convex surfaces and empirical verification of deep learning theory based on comparing empirical and theoretically implied spectra.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=86t2GlfzFo", "arxiv_id": "1912.09656", "paper_pdf": "papers/86t2GlfzFo.pdf", "paper_pdf_sha256": "5905b75bc1164f7cf5d74000d5e0220be0e5c6ae1d75693661645c283d7017b6", "paper_pdf_bytes": 727277, "paper_pdf_source": "openreview", "code_url": "https://github.com/xingchenwan/MLRG_DeepCurvature", "code_repository": "xingchenwan/MLRG_DeepCurvature", "code_commit": "818ea8631e7aafc659035310c45270145fd174d3", "code_archive": "repos/86t2GlfzFo.zip", "code_archive_sha256": "0d1f3286d747db87256526fa6df002840a2007116d4206a7e8149630a938a45f", "code_archive_bytes": 343504, "code_file_count": 37, "code_extensions": {".py": 36, ".ipynb": 1}, "github_disk_usage_kb": 606, "github_languages": {"Jupyter Notebook": 375459, "Python": 199652}, "github_archived": false, "github_pushed_at": "2020-04-19T09:34:30Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/mlrg-deep-curvature"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "S1e0ZlHYDB", "year": 2020, "status": "rejected", "title": "Progressive Compressed Records: Taking a Byte Out of Deep Learning Data", "authors": ["Michael Kuchnik", "George Amvrosiadis", "Virginia Smith"], "authorids": ["mkuchnik@andrew.cmu.edu", "gamvrosi@cmu.edu", "smithv@cmu.edu"], "authors_source": "OpenReview API", "abstract": "Deep learning training accesses vast amounts of data at high velocity, posing challenges for datasets retrieved over commodity networks and storage devices. We introduce a way to dynamically reduce the overhead of fetching and transporting training data with a method we term Progressive Compressed Records (PCRs). PCRs deviate from previous formats by leveraging progressive compression to split each training example into multiple examples of increasingly higher fidelity, without adding to the total data size. Training examples of similar fidelity are grouped together, which reduces both the system overhead and data bandwidth needed to train a model. We show that models can be trained on aggressively compressed representations of the training data and still retain high accuracy, and that PCRs can enable a 2x speedup on average over baseline formats using JPEG compression. Our results hold across deep learning architectures for a wide range of datasets: ImageNet, HAM10000, Stanford Cars, and CelebA-HQ.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=S1e0ZlHYDB", "arxiv_id": "1911.00472", "paper_pdf": "papers/S1e0ZlHYDB.pdf", "paper_pdf_sha256": "fc3e6ee90a925f417ef46b5b71db7e857923498ed90acb811ba1da2a47cfd0e3", "paper_pdf_bytes": 3943379, "paper_pdf_source": "openreview", "code_url": "https://github.com/mkuchnik/PCR_Release", "code_repository": "mkuchnik/PCR_Release", "code_commit": "aa76741f5dc6dcb6d7676c62e2fd53a78f43f142", "code_archive": "repos/S1e0ZlHYDB.zip", "code_archive_sha256": "adbf803296d6dcbc860e2d92039251c6d99429ed10637ebbbb11d9fac14bc829", "code_archive_bytes": 1365478, "code_file_count": 42, "code_extensions": {".py": 27, ".sh": 8, ".cpp": 3, ".c": 3, ".h": 1}, "github_disk_usage_kb": 1311, "github_languages": {"Python": 249111, "C": 20669, "Shell": 9203, "C++": 8497, "CMake": 2487, "Dockerfile": 1308}, "github_archived": false, "github_pushed_at": "2021-08-20T00:57:32Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/progressive-compressed-records-taking-a-byte-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "8gdfWRilR7", "year": 2026, "status": "rejected", "title": "TimeSeriesGym: A Scalable Benchmark for (Time Series) Machine Learning Engineering Agents", "authors": ["Yifu Cai", "Xinyu Li", "Mononito Goswami", "Michał Wiliński", "Gus Welter", "Artur Dubrawski"], "authorids": ["~Yifu_Cai1", "~Xinyu_Li7", "~Mononito_Goswami1", "~Michał_Wiliński1", "~Gus_Welter1", "~Artur_Dubrawski2"], "authors_source": "OpenReview API", "abstract": "We introduce TimeSeriesGym, a scalable benchmarking framework for evaluating Artificial Intelligence (AI) agents on time series machine learning engineering challenges. Existing benchmarks lack scalability, focus narrowly on model building in well-defined settings, and evaluate only a limited set of research artifacts (e.g., CSV submission files). To make AI agent benchmarking more relevant to the practice of machine learning engineering, our framework scales along two critical dimensions. First, recognizing that effective ML engineering requires a range of diverse skills, TimeSeriesGym incorporates challenges from diverse sources spanning multiple domains and tasks. We design challenges to evaluate both isolated capabilities (including data handling, understanding research repositories, and code translation) and their combinations, and rather than addressing each challenge independently, we develop tools that support designing multiple challenges at scale. Second, we implement evaluation mechanisms for multiple research artifacts, including submission files, code, and models, using precise numeric measures and _optionally_ LLM-based qualitative assessments. This strategy complements objective evaluation with subjective assessment when appropriate. Although our initial focus is on time series applications, our framework can be readily extended to other data modalities, broadly enhancing the comprehensiveness and practical utility of agentic AI evaluation. We [open-source](https://anonymous.4open.science/r/TimeSeriesGym-9CF6/) our benchmarking framework to facilitate future research on the ML engineering capabilities of AI agents.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=8gdfWRilR7", "arxiv_id": "2505.13291", "paper_pdf": "papers/8gdfWRilR7.pdf", "paper_pdf_sha256": "589c81fe9b3697eea2f0f91a048098c1879d46cdd0e30c06e094c6e6d66f6538", "paper_pdf_bytes": 1142833, "paper_pdf_source": "openreview", "code_url": "https://github.com/moment-timeseries-foundation-model/TimeSeriesGym", "code_repository": "moment-timeseries-foundation-model/TimeSeriesGym", "code_commit": "853b4782f8ca18121c04ee2c2cf96e9b189b0c11", "code_archive": "repos/8gdfWRilR7.zip", "code_archive_sha256": "d124546b6872a0adf8f0098d59f34e98759d697b24c45ec510633ba2fd067583", "code_archive_bytes": 2390403, "code_file_count": 114, "code_extensions": {".py": 106, ".sh": 8}, "github_disk_usage_kb": 2125, "github_languages": {"Python": 397513, "Shell": 13091, "Dockerfile": 8552}, "github_archived": false, "github_pushed_at": "2025-11-30T22:28:51Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/timeseriesgym-a-scalable-benchmark-for-time"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "OyjMJjfhiw", "year": 2025, "status": "rejected", "title": "LLM Embeddings Improve Test-Time Adaptation to Tabular $Y|X$-Shifts", "authors": ["Yibo Zeng", "Jiashuo Liu", "Henry Lam", "Hongseok Namkoong"], "authorids": ["~Yibo_Zeng1", "~Jiashuo_Liu1", "~Henry_Lam1", "~Hongseok_Namkoong2"], "authors_source": "OpenReview API", "abstract": "For tabular datasets, the change in the relationship between the label and covariates ($Y|X$-shifts) is common due to missing variables. Since it is impossible to generalize to a completely new and unknown domain, we study models that are easy to adapt to the target domain even with few labeled examples. We focus on building more informative representations of tabular data that can mitigate $Y|X$-shifts, and propose to leverage the prior world knowledge in LLMs by serializing the tabular data to encode it. We find LLM embeddings alone provide inconsistent improvements in robustness, but models trained on them can be well adapted to the target domain even using 32 labeled observations. Our finding is based on a systematic study consisting of 7650 source-target pairs and benchmark against **261,000** model configurations trained by 20 algorithms. Our observation holds when ablating the size of accessible target data and different adaptation strategies.", "decision": "Reject", "meta_review": "", "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=OyjMJjfhiw", "arxiv_id": "2410.07395", "paper_pdf": "papers/OyjMJjfhiw.pdf", "paper_pdf_sha256": "729336df048bddee27c5ee4c405a83eb1ccf64f2e421c8d600e8c0f4b692d6d9", "paper_pdf_bytes": 1056479, "paper_pdf_source": "openreview", "code_url": "https://github.com/namkoong-lab/LLM-Tabular-Shifts", "code_repository": "namkoong-lab/LLM-Tabular-Shifts", "code_commit": "9bab7ea2c4f0dc63412492025d8e3a2ec72e1f96", "code_archive": "repos/OyjMJjfhiw.zip", "code_archive_sha256": "d90e360a582a0f2b858b6b57dffc7e13472c92ac82ea4ddadf517570948510f7", "code_archive_bytes": 243625, "code_file_count": 87, "code_extensions": {".py": 82, ".ipynb": 5}, "github_disk_usage_kb": 153, "github_languages": {"Python": 596460, "Jupyter Notebook": 127816}, "github_archived": false, "github_pushed_at": "2024-10-17T17:31:47Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/llm-embeddings-improve-test-time-adaptation"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "bCNYFOaWsy", "year": 2024, "status": "rejected", "title": "Class-Imbalanced Graph Learning without Class Rebalancing", "authors": ["Zhining Liu", "Zhichen Zeng", "Ruizhong Qiu", "Hyunsik Yoo", "David Zhou", "Zhe Xu", "Yada Zhu", "Kommy Weldemariam", "Jingrui He", "Hanghang Tong"], "authorids": ["~Zhining_Liu1", "~Zhichen_Zeng1", "~Ruizhong_Qiu1", "~Hyunsik_Yoo1", "~David_Zhou1", "~Zhe_Xu5", "~Yada_Zhu1", "~Kommy_Weldemariam1", "~Jingrui_He1", "~Hanghang_Tong3"], "authors_source": "OpenReview API", "abstract": "Class imbalance is prevalent in real-world node classification tasks and poses great challenges for graph machine-learning models. Most existing studies are rooted in a class-rebalancing (CR) perspective and aim to address class imbalance with class-wise reweighting or resampling. In this work, we approach the root cause of class-imbalance bias from an orthogonal topological paradigm. Specifically, we theoretically reveal and empirically observe two fundamental phenomena in the underlying graph topology that can greatly exacerbate the predictive bias stemming from class imbalance. In light of these findings, we devise a lightweight topological augmentation framework called TOBE to mitigate the class-imbalance bias without class rebalancing. Being orthogonal to CR, the proposed TOBE is a model-agnostic and efficient solution that can be seamlessly combined with and further boost existing CR techniques. Systematic experiments on real-world imbalanced graph learning tasks show that TOBE can deliver up to 46.27% performance gain and up to 72.74% bias reduction over existing techniques. Code is available at https://anonymous.4open.science/r/ToBE/.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=bCNYFOaWsy", "arxiv_id": "2308.14181", "paper_pdf": "papers/bCNYFOaWsy.pdf", "paper_pdf_sha256": "6da258c21c1f743445f8d5f34045a90a5f14900c11b21e4d0455b5446e200f3b", "paper_pdf_bytes": 2847168, "paper_pdf_source": "openreview", "code_url": "https://github.com/ZhiningLiu1998/BAT", "code_repository": "ZhiningLiu1998/BAT", "code_commit": "de58799f8da953abecf5eb32c1fdb05ff89c10f7", "code_archive": "repos/bCNYFOaWsy.zip", "code_archive_sha256": "aedf17fc52b2476dbd4a86e3bad3e4d8a67a24ba810ff1d05ede24fa91c22e49", "code_archive_bytes": 111796, "code_file_count": 44, "code_extensions": {".py": 42, ".ipynb": 2}, "github_disk_usage_kb": 125, "github_languages": {"Python": 292188, "Jupyter Notebook": 110469}, "github_archived": false, "github_pushed_at": "2024-11-27T22:27:59Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/topological-augmentation-for-class-imbalanced"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "AsOLzq1S-p", "year": 2023, "status": "rejected", "title": "Offline Policy Comparison with Confidence: Benchmarks and Baselines", "authors": ["Anurag Koul", "Mariano Phielipp", "Alan Fern"], "authorids": ["~Anurag_Koul1", "~Mariano_Phielipp2", "~Alan_Fern1"], "authors_source": "OpenReview API", "abstract": "Decision makers often wish to use offline historical data to compare sequential-action policies at various world states. Importantly, computational tools should produce confidence values for such offline policy comparison (OPC) to account for statistical variance and limited data coverage. Nevertheless, there is little work that directly evaluates the quality of confidence values for OPC. In this work, we address this issue by creating benchmarks for OPC with Confidence (OPCC), derived by adding sets of policy comparison queries to datasets from offline reinforcement learning. In addition, we present an empirical evaluation of the \"risk versus coverage\" trade-off for a class of model-based baselines. In particular, the baselines learn ensembles of dynamics models, which are used in various ways to produce simulations for answering queries with confidence values. While our results suggest advantages for certain baseline variations, there appears to be significant room for improvement in future work.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=AsOLzq1S-p", "arxiv_id": "2205.10739", "paper_pdf": "papers/AsOLzq1S-p.pdf", "paper_pdf_sha256": "391c3aff8599bff76e1c67358bb3004156480a33faf6bac48bc2a8936fdc98fb", "paper_pdf_bytes": 10351002, "paper_pdf_source": "openreview", "code_url": "https://github.com/koulanurag/opcc-baselines", "code_repository": "koulanurag/opcc-baselines", "code_commit": "4eef0ad4f9059883a9f38da9ff832960b12647f8", "code_archive": "repos/AsOLzq1S-p.zip", "code_archive_sha256": "7e09a10467431214f71b62b451f7c04421ef6eec6af399d6d15f8e0242521e5f", "code_archive_bytes": 41793, "code_file_count": 18, "code_extensions": {".py": 18}, "github_disk_usage_kb": 223, "github_languages": {"Python": 120522}, "github_archived": false, "github_pushed_at": "2023-10-20T04:39:24Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/offline-policy-comparison-with-confidence"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "vdKncX1WclT", "year": 2022, "status": "rejected", "title": "Red Alarm for Pre-trained Models: Universal Vulnerability to Neuron-Level Backdoor Attacks", "authors": ["Zhengyan Zhang", "Guangxuan Xiao", "Yongwei Li", "Tian Lv", "Fanchao Qi", "Zhiyuan Liu", "Yasheng Wang", "Xin Jiang", "Maosong Sun"], "authorids": ["~Zhengyan_Zhang1", "~Guangxuan_Xiao1", "~Yongwei_Li1", "~Tian_Lv1", "~Fanchao_Qi1", "~Zhiyuan_Liu1", "~Yasheng_Wang1", "~Xin_Jiang1", "~Maosong_Sun1"], "authors_source": "OpenReview API", "abstract": "The pre-training-then-fine-tuning paradigm has been widely used in deep learning. Due to the huge computation cost for pre-training, practitioners usually download pre-trained models from the Internet and fine-tune them on downstream datasets while the downloaded models may suffer backdoor attacks. Different from previous attacks aiming at a target task, we show that a backdoored pre-trained model can behave maliciously in various downstream tasks without foreknowing task information. Attackers can restrict the output representations of trigger-embedded samples to arbitrary predefined values through additional training, namely Neuron-level Backdoor Attack (NeuBA). Since fine-tuning has little effect on model parameters, the fine-tuned model will retain the backdoor functionality and predict a specific label for the samples embedded with the same trigger. To provoke multiple labels in a specific task, attackers can introduce several triggers with contrastive predefined values. In the experiments of both natural language processing (NLP) and computer vision (CV), we show that NeuBA can well control the predictions for trigger-embedded instances with different trigger designs. Our findings sound a red alarm for the wide use of pre-trained models. Finally, we apply several defense methods to NeuBA and find that model pruning is a promising technique to resist NeuBA by omitting backdoored neurons.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=vdKncX1WclT", "arxiv_id": "2101.06969", "paper_pdf": "papers/vdKncX1WclT.pdf", "paper_pdf_sha256": "98e80ec09447ed7b41cf9f1acdfd52ee48246c2c47cad4a31d8fcab8cd0153da", "paper_pdf_bytes": 449533, "paper_pdf_source": "openreview", "code_url": "https://github.com/thunlp/NeuBA", "code_repository": "thunlp/NeuBA", "code_commit": "df8d545e49ecfbdbb6b9d71238d0dd43527c97da", "code_archive": "repos/vdKncX1WclT.zip", "code_archive_sha256": "60bdd81c6ece5e1416615cbb32a58543f13ca3ab883e874a1f8f98ef5e299660", "code_archive_bytes": 66200, "code_file_count": 25, "code_extensions": {".sh": 13, ".py": 12}, "github_disk_usage_kb": 520, "github_languages": {"Python": 228617, "Shell": 14187}, "github_archived": false, "github_pushed_at": "2021-06-23T07:54:31Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/red-alarm-for-pre-trained-models-universal"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "jAJrc-kzVd0", "year": 2021, "status": "rejected", "title": "Revisiting Prioritized Experience Replay: A Value Perspective", "authors": ["Ang A. Li", "Zongqing Lu", "Chenglin Miao"], "authorids": ["amazingang@pku.edu.cn", "~Zongqing_Lu2", "chenglin.miao@pku.edu.cn"], "authors_source": "OpenReview API", "abstract": "Reinforcement learning (RL) agents need to learn from past experiences. Prioritized experience replay that weighs experiences by their surprise (the magnitude of the temporal-difference error) significantly improves the learning efficiency for RL algorithms. Intuitively, surprise quantifies the unexpectedness of an experience to the learning agent. But how surprise is related to the importance of experience is not well understood. To address this problem, we derive three value metrics to quantify the importance of experience, which consider the extra reward would be earned by accessing the experience. We theoretically show these value metrics are upper-bounded by surprise for Q-learning. Furthermore, we successfully extend our theoretical framework to maximum-entropy RL by deriving the lower and upper bounds of these value metrics for soft Q-learning, which is also related to surprise. Our framework links two important quantities in RL, i.e., surprise and value of experience, and provides a theoretical basis to estimate the value of experience by surprise. We empirically show that the upper bounds hold in practice, and experience replay using the upper bound as priority improves maximum-entropy RL in Atari games. ", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=jAJrc-kzVd0", "arxiv_id": "2102.03261", "paper_pdf": "papers/jAJrc-kzVd0.pdf", "paper_pdf_sha256": "413c9304337d973d56087c5b43d7beac63c8d7de19ae8131e9715991306f24c0", "paper_pdf_bytes": 1818566, "paper_pdf_source": "openreview", "code_url": "https://github.com/AmazingAng/VER", "code_repository": "AmazingAng/VER", "code_commit": "cdfcdc96defe117d6123940e6aa4174db5c6f6c6", "code_archive": "repos/jAJrc-kzVd0.zip", "code_archive_sha256": "4d62b0048232de39ad3f88023efa8a75bc10febf8cc59d84929ba0e0843d64a2", "code_archive_bytes": 619847, "code_file_count": 10, "code_extensions": {".ipynb": 5, ".py": 5}, "github_disk_usage_kb": 630, "github_languages": {"Jupyter Notebook": 988075, "Python": 49142}, "github_archived": false, "github_pushed_at": "2021-02-01T16:51:57Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/revisiting-prioritized-experience-replay-a-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "BygSZAVKvr", "year": 2020, "status": "rejected", "title": "Energy-Aware Neural Architecture Optimization with Fast Splitting Steepest Descent", "authors": ["Dilin Wang", "Meng Li", "Lemeng Wu", "Vikas Chandra", "Qiang Liu"], "authorids": ["dilin@cs.utexas.edu", "meng.li@fb.com", "lmwu@cs.utexas.edu", "vchandra@fb.com", "lqiang@cs.utexas.edu"], "authors_source": "OpenReview API", "abstract": "Designing energy-efficient networks is of critical importance for enabling state-of-the-art deep learning in mobile and edge settings where the computation and energy budgets are highly limited. Recently, Wu et al. (2019) framed the search of efficient neural architectures into a continuous splitting process: it iteratively splits existing neurons into multiple off-springs to achieve progressive loss minimization,  thus finding novel architectures by gradually growing the neural network. However, this method was not specifically tailored for designing energy-efficient networks, and is computationally expensive on large-scale benchmarks.  In this work, we substantially improve Wu et al. (2019) in two significant ways:  1) we incorporate the energy cost of splitting different neurons to better guide the splitting process, thereby discovering more energy-efficient network architectures; 2) we substantially speed up the splitting process of Wu et al. (2019), which requires expensive eigen-decomposition, by proposing a highly scalable Rayleigh-quotient stochastic gradient algorithm.  Our fast algorithm allows us to reduce the computational cost of splitting to the same level of typical back-propagation updates and enables efficient implementation on GPU. Extensive empirical results show that our method can train highly accurate and energy-efficient networks on challenging datasets such as ImageNet,  improving a variety of baselines,  including the pruning-based methods and expert-designed architectures.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=BygSZAVKvr", "arxiv_id": "1910.03103", "paper_pdf": "papers/BygSZAVKvr.pdf", "paper_pdf_sha256": "113e3daabcc3343d29558952932584d7a3a167bbce31e21184cf286cc17c9ced", "paper_pdf_bytes": 5004873, "paper_pdf_source": "openreview", "code_url": "https://github.com/dilinwang820/fast-energy-aware-splitting", "code_repository": "dilinwang820/fast-energy-aware-splitting", "code_commit": "73a1c9e5fec403cf86711aaa133ddec0f58ba36a", "code_archive": "repos/BygSZAVKvr.zip", "code_archive_sha256": "e4609dfe005e5f5cc137fbdf76a388f7655c4d6c0d3310dc88c3822f91dfb59c", "code_archive_bytes": 629947, "code_file_count": 34, "code_extensions": {".py": 31, ".sh": 3}, "github_disk_usage_kb": 1450, "github_languages": {"Python": 250014, "Shell": 576}, "github_archived": false, "github_pushed_at": "2020-02-06T21:12:34Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/energy-aware-neural-architecture-optimization"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "guXFowAqLQ", "year": 2026, "status": "rejected", "title": "Time series saliency maps: Explaining models across multiple domains", "authors": ["Christodoulos Kechris", "Jonathan Dan", "David Atienza"], "authorids": ["~Christodoulos_Kechris1", "~Jonathan_Dan1", "~David_Atienza1"], "authors_source": "OpenReview API", "abstract": "Traditional saliency map methods, popularized in computer vision, highlight individual points (pixels) of the input that contribute the most to the model's output. However, in time-series, they offer limited insights, as semantically meaningful features are often found in other domains. We introduce Cross-domain Integrated Gradients, a generalization of Integrated Gradients. Our method enables feature attributions in any domain that can be formulated as an invertible, differentiable transformation of the time domain. Crucially, our derivation extends the original Integrated Gradients into the complex domain, enabling frequency-based attributions. We provide the necessary theoretical guarantees, namely, path independence and completeness. Our approach reveals interpretable, problem-specific attributions that time-domain methods cannot capture in three real-world tasks: wearable sensor heart rate extraction, electroencephalography-based seizure detection, and zero-shot time-series forecasting. We release an open-source Tensorflow/PyTorch library to enable plug-and-play cross-domain explainability for time-series models. These results demonstrate the ability of cross-domain integrated gradients to provide semantically meaningful insights into time series models that are impossible to achieve with traditional time‑domain saliency.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=guXFowAqLQ", "arxiv_id": "2505.13100", "paper_pdf": "papers/guXFowAqLQ.pdf", "paper_pdf_sha256": "8860bc819b6cc14405adbb570d5da70781eda61792f43feb55348cd528d625e7", "paper_pdf_bytes": 3653808, "paper_pdf_source": "openreview", "code_url": "https://github.com/esl-epfl/cross-domain-saliency-maps", "code_repository": "esl-epfl/cross-domain-saliency-maps", "code_commit": "e4fee40c5a05601218a7268c9fb4ec27790dc760", "code_archive": "repos/guXFowAqLQ.zip", "code_archive_sha256": "68ca18eb6f14cb9bd95d895991c859403d1a09ce0bc8f927202c7bff2dfee63d", "code_archive_bytes": 2695630, "code_file_count": 20, "code_extensions": {".py": 16, ".ipynb": 4}, "github_disk_usage_kb": 2209, "github_languages": {"Python": 99091}, "github_archived": false, "github_pushed_at": "2026-05-04T23:49:34Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/time-series-saliency-maps-explaining-models"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "P98KMCf60l", "year": 2025, "status": "rejected", "title": "Theoretical Insights into Fine-Tuning Attention Mechanism: Generalization and Optimization", "authors": ["Xinhao Yao", "Hongjin Qian", "Xiaolin Hu", "Gengze Xu", "Yong Liu"], "authorids": ["~Xinhao_Yao1", "~Hongjin_Qian1", "~Xiaolin_Hu6", "~Gengze_Xu1", "~Yong_Liu7"], "authors_source": "OpenReview API", "abstract": "Large Language Models (LLMs), built on Transformer architectures, exhibit remarkable generalization across a wide range of tasks. However, fine-tuning these models for specific tasks remains resource-intensive due to their extensive parameterization.  \nIn this paper, we investigate two remarkable phenomena related to the attention mechanism during the fine-tuning of LLMs. The first phenomenon, termed “Unequal Importance of Attention Matrices,” highlights the impact of fine-tuning different weight matrices. It shows that optimizing the $\\mathbf{W}_v$ matrix yields significantly better performance than optimizing the $\\mathbf{W}_k$ matrix. Fine-tuning only the $\\mathbf{W}_q$ and $\\mathbf{W}_v$ matrices is computationally efficient while delivering results comparable to, or even better than fine-tuning all three matrices ($\\mathbf{W}_q$, $\\mathbf{W}_k$, and $\\mathbf{W}_v$). The second phenomenon, “Attention Matrices with Customized Learning Rate Leads to Better Convergence,” emphasizes the importance of assigning distinct learning rates to these matrices. Specifically, a higher learning rate for the $\\mathbf{W}_v$ matrix compared to $\\mathbf{W}_q$ and $\\mathbf{W}_k$ accelerates convergence and improves performance. Building on these insights, we propose a new strategy that improves fine-tuning efficiency in terms of both storage and time. Experimental results on benchmark datasets validate the effectiveness of this approach, supporting our theoretical findings. Our analysis lays the theoretical groundwork for configuring and improving lightweight algorithms in LLMs fine-tuning.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=P98KMCf60l", "arxiv_id": "2410.02247", "paper_pdf": "papers/P98KMCf60l.pdf", "paper_pdf_sha256": "416981fbcc92b73def6bed50eec45b587bb156210dfb9a271d5c51b83bd6b2ce", "paper_pdf_bytes": 598252, "paper_pdf_source": "openreview", "code_url": "https://github.com/chen123CtrlS/LightweightAtt", "code_repository": "chen123CtrlS/LightweightAtt", "code_commit": "6e6cfc3d981194abd565c298e372df9899a2fc5f", "code_archive": "repos/P98KMCf60l.zip", "code_archive_sha256": "c7b0b6e4e940130771147bafb71ef24abbf192f562bed60007394a4f37ffc985", "code_archive_bytes": 167302, "code_file_count": 13, "code_extensions": {".sh": 8, ".py": 5}, "github_disk_usage_kb": 157, "github_languages": {"Python": 43600, "Shell": 10627}, "github_archived": false, "github_pushed_at": "2024-10-01T08:28:30Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/theoretical-insights-into-fine-tuning"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "OMwD6pGYB4", "year": 2024, "status": "rejected", "title": "A Distributional Analogue to the Successor Representation", "authors": ["Jesse Farebrother", "Harley Wiltzer", "Arthur Gretton", "Yunhao Tang", "Andre Barreto", "Will Dabney", "Marc G Bellemare", "Mark Rowland"], "authorids": ["~Jesse_Farebrother1", "~Harley_Wiltzer1", "~Arthur_Gretton1", "~Yunhao_Tang1", "~Andre_Barreto1", "~Will_Dabney1", "~Marc_G_Bellemare1", "~Mark_Rowland1"], "authors_source": "OpenReview API", "abstract": "This paper contributes a new approach for distributional reinforcement learning which allows for a \nclean separation of transition structure and reward in the learning process. Analogous to how the successor representation (SR) describes the expected consequences from behaving according to a given policy, our distributional successor measure (SM) describes the distributional consequences of this behaviour.\nWe model the distributional SM as a distribution over distributions and provide theory connecting it with distributional and model-based reinforcement learning. \nExtending γ-models (Janner et al., 2020), we propose an algorithm that learns the distributional SM from samples by minimizing a two-level maximum mean discrepancy. Key to our method are a number of algorithmic techniques that are independently valuable in the context of learning generative models of state.\nAs an illustration of the practical usefulness of the distributional successor measure, we show that it\nenables zero-shot risk-sensitive policy evaluation in a way that was not previously possible.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=OMwD6pGYB4", "arxiv_id": "2402.08530", "paper_pdf": "papers/OMwD6pGYB4.pdf", "paper_pdf_sha256": "0845864f5f06f3199b4414634b20e102c72eb4597a185d96a25755b65f69cfdf", "paper_pdf_bytes": 1981327, "paper_pdf_source": "openreview", "code_url": "https://github.com/JesseFarebro/distributional-sr", "code_repository": "JesseFarebro/distributional-sr", "code_commit": "f379d0f65914459db069cf0089bdc54451f3e563", "code_archive": "repos/OMwD6pGYB4.zip", "code_archive_sha256": "7d433a5be9bd4a09a739222a86251ab8346b4fc199e0e82fcf93db80185526f8", "code_archive_bytes": 128944, "code_file_count": 21, "code_extensions": {".py": 21}, "github_disk_usage_kb": 135, "github_languages": {"Python": 91877}, "github_archived": false, "github_pushed_at": "2024-11-08T00:11:13Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/a-distributional-analogue-to-the-successor"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "lEFM4OTz62c", "year": 2023, "status": "rejected", "title": "Group-level Brain Decoding with Deep Learning", "authors": ["Richard Csaky", "Mats W.J. Van Es", "Oiwi Parker Jones", "Mark Woolrich"], "authorids": ["~Richard_Csaky1", "~Mats_W.J._Van_Es1", "~Oiwi_Parker_Jones1", "~Mark_Woolrich1"], "authors_source": "OpenReview API", "abstract": "Decoding experimental variables from brain imaging data is gaining popularity, with applications in brain-computer interfaces and the study of neural representations. Decoding is typically subject-specific and does not generalise well over subjects. Here, we propose a method that uses subject embedding, analogous to word embedding in Natural Language Processing, to learn and exploit the structure in between subject variability as part of a decoding model, our adaptation of the WaveNet architecture for classification. We apply this to magnetoencephalography data, where 15 subjects viewed 118 different images, with 30 examples per image; to classify images using the entire 1s window following image presentation. We show that the combination of deep learning and subject embedding is crucial to closing the performance gap between subject- and group-level decoding models. Importantly, group models outperform subject models on low-accuracy subjects (but impair high-accuracy subjects) and can be helpful for initialising subject models. The potential of such group modelling is even higher with bigger datasets. To better enable physiological interpretation at the group level we demonstrate the use of permutation feature importance developing insights into the spatio-temporal and spectral information encoded in the models. All code is available on GitHub.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=lEFM4OTz62c", "arxiv_id": "2205.14102", "paper_pdf": "papers/lEFM4OTz62c.pdf", "paper_pdf_sha256": "52f5b635365f22f821fcb5f343b0a412d581002e15fce5d7a37d0a6b64c2379e", "paper_pdf_bytes": 18413707, "paper_pdf_source": "openreview", "code_url": "https://github.com/ricsinaruto/MEG-group-decode", "code_repository": "ricsinaruto/MEG-group-decode", "code_commit": "3a4e7daae885b387fd4e19903e5d295eae9c0fea", "code_archive": "repos/lEFM4OTz62c.zip", "code_archive_sha256": "e21b28d53b527b5953c0d22ae747ee5db10937ba5b396a3589aeaa26a70f0fda", "code_archive_bytes": 219323, "code_file_count": 27, "code_extensions": {".py": 24, ".ipynb": 3}, "github_disk_usage_kb": 228, "github_languages": {"Jupyter Notebook": 255845, "Python": 92038}, "github_archived": false, "github_pushed_at": "2024-01-18T17:11:03Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/generalizing-brain-decoding-across-subjects"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "UgNQM-LcVpN", "year": 2022, "status": "rejected", "title": "A Modulation Layer to Increase Neural Network Robustness Against Data Quality Issues", "authors": ["Mohamed Abdelhack", "Jiaming Zhang", "Sandhya Tripathi", "Bradley A Fritz", "Michael Avidan", "Yixin Chen", "Christopher Ryan King"], "authorids": ["~Mohamed_Abdelhack1", "~Jiaming_Zhang2", "~Sandhya_Tripathi1", "~Bradley_A_Fritz1", "~Michael_Avidan1", "~Yixin_Chen1", "~Christopher_Ryan_King1"], "authors_source": "OpenReview API", "abstract": "Data quality is a common problem in machine learning, especially in high-stakes settings such as healthcare. Missing data affects accuracy, calibration, and feature attribution in complex patterns. Developers often train models on carefully curated datasets to minimize missing data bias; however, this reduces the usability of such models in production environments, such as real-time healthcare records. Making machine learning models robust to missing data is therefore crucial for practical application. While some classifiers naturally handle missing data, others, such as deep neural networks, are not designed for unknown values. We propose a novel neural network modification to mitigate the impacts of missing data. The approach is inspired by neuromodulation that is performed by biological neural networks. Our proposal replaces the fixed weights of a fully-connected layer with a function of an additional input (reliability score) at each input, mimicking the ability of cortex to up- and down-weight inputs based on the presence  of other data. The modulation function is jointly learned with the main task using a multi-layer perceptron. We tested our modulating fully connected layer on multiple classification, regression, and imputation problems, and it either improved performance or generated comparable performance to conventional neural network architectures concatenating reliability to the inputs. Models with modulating layers were more robust against degradation of data quality by introducing additional missingness at evaluation time. These results suggest that explicitly accounting for reduced information quality with a modulating fully connected layer can enable the deployment of artificial intelligence systems in real-time settings.\n", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=UgNQM-LcVpN", "arxiv_id": "2107.08574", "paper_pdf": "papers/UgNQM-LcVpN.pdf", "paper_pdf_sha256": "5d1a5b41b462af0110940f2f46b929f4fb3c12a5661d83707fc097259f2b4391", "paper_pdf_bytes": 455793, "paper_pdf_source": "openreview", "code_url": "https://github.com/mabdelhack/mfcl", "code_repository": "mabdelhack/mfcl", "code_commit": "7be8449b06d41251e2cb549fbf6ffc1338dc37d0", "code_archive": "repos/UgNQM-LcVpN.zip", "code_archive_sha256": "c58c64e31bb07b472e18594f2ff9dacc335f91d5fd6c78563610c9298f2fd480", "code_archive_bytes": 506029, "code_file_count": 19, "code_extensions": {".py": 19}, "github_disk_usage_kb": 547, "github_languages": {"Python": 139130}, "github_archived": false, "github_pushed_at": "2023-04-04T08:32:18Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/a-modulation-layer-to-increase-neural-network"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "yvuk0RsLoP7", "year": 2021, "status": "rejected", "title": "Improving Model Robustness with Latent Distribution Locally and Globally", "authors": ["Zhuang QIAN", "Shufei Zhang", "Kaizhu Huang", "Qiufeng Wang", "Rui Zhang", "Xinping Yi"], "authorids": ["~Zhuang_QIAN1", "~Shufei_Zhang1", "~Kaizhu_Huang1", "~Qiufeng_Wang2", "~Rui_Zhang10", "xinping.yi@liverpool.ac.uk"], "authors_source": "OpenReview API", "abstract": "We propose a novel adversarial training method which leverages both the local and global information to defend  adversarial attacks. Existing adversarial training methods usually generate adversarial perturbations  locally in a supervised manner and fail to consider the data manifold information in a global way. Consequently, the resulting  adversarial examples may corrupt the underlying data structure and are typically biased towards the decision boundary. In this work, we exploit both the local and global information of data manifold to generate adversarial examples in an unsupervised manner. Specifically, we design our novel framework  via an adversarial game between  a discriminator and a classifier: the discriminator is  learned to differentiate the latent distributions  of the natural data and the perturbed counterpart, while the classifier is trained to recognize accurately the perturbed examples as well as enforcing the invariance between the two latent distributions. We conduct a series of analysis on the model robustness and also verify the effectiveness of our proposed method empirically. Experimental results show that our method  substantially outperforms the recent state-of-the-art (i.e. Feature Scattering) in defending adversarial attacks  by a large accuracy margin  (e.g. $17.0\\%$ and $18.1\\%$ on SVHN dataset, $9.3\\%$ and $17.4\\%$ on CIFAR-10 dataset, $6.0\\%$ and $16.2\\%$ on CIFAR-100 dataset for defending PGD20 and CW20 attacks respectively).", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=yvuk0RsLoP7", "arxiv_id": "2107.04401", "paper_pdf": "papers/yvuk0RsLoP7.pdf", "paper_pdf_sha256": "1142db4cb99504ec3da5363fdf6fbdfd61aee9071718e48bb6695f7830a66479", "paper_pdf_bytes": 4605606, "paper_pdf_source": "openreview", "code_url": "https://github.com/LitterQ/ATLD-pytorch", "code_repository": "LitterQ/ATLD-pytorch", "code_commit": "36bc00c32ff92995ad89622285ec30018e7d3f40", "code_archive": "repos/yvuk0RsLoP7.zip", "code_archive_sha256": "985c549f5bf1eaecbe054c4be86653a03992868b042450ac2a6bf3700683fa9d", "code_archive_bytes": 469729, "code_file_count": 59, "code_extensions": {".py": 49, ".sh": 10}, "github_disk_usage_kb": 693, "github_languages": {"Python": 334224, "Shell": 3683}, "github_archived": false, "github_pushed_at": "2025-03-23T10:19:18Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/improving-model-robustness-with-latent-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "SJeoE0VKDS", "year": 2020, "status": "rejected", "title": "Novelty Search in representational space for sample efficient exploration", "authors": ["Ruo Yu Tao", "Vincent François-Lavet", "Joelle Pineau"], "authorids": ["ruo.tao@mail.mcgill.ca", "vincent.francois-lavet@mail.mcgill.ca", "jpineau@cs.mcgill.ca"], "authors_source": "OpenReview API", "abstract": "We present a new approach for efficient exploration which leverages a low-dimensional encoding of the environment learned with a combination of model-based and model-free objectives. Our approach uses intrinsic rewards that are based on a weighted distance of nearest neighbors in the low dimensional representational space to gauge novelty.\nWe then leverage these intrinsic rewards for sample-efficient exploration with planning routines in representational space.\nOne key element of our approach is that we perform more gradient steps in-between every environment step in order to ensure the model accuracy. We test our approach on a number of maze tasks, as well as a control problem and show that our exploration approach is more sample-efficient compared to strong baselines. ", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=SJeoE0VKDS", "arxiv_id": "2009.13579", "paper_pdf": "papers/SJeoE0VKDS.pdf", "paper_pdf_sha256": "832d9a85d4f22d440c38518189218d3f46fd7389d5c4c1960296be996dfe6b4d", "paper_pdf_bytes": 1756879, "paper_pdf_source": "openreview", "code_url": "https://github.com/taodav/nsrs", "code_repository": "taodav/nsrs", "code_commit": "7d9329134cd1f3305db8c8afe87504349875ef90", "code_archive": "repos/SJeoE0VKDS.zip", "code_archive_sha256": "aa7c2788c29e68cceb2331ec092fed4d9a392d3e7fd250a671f8c3b1cf1d3db1", "code_archive_bytes": 1560024, "code_file_count": 60, "code_extensions": {".py": 45, ".sh": 8, ".ipynb": 7}, "github_disk_usage_kb": 1731, "github_languages": {"Jupyter Notebook": 4485362, "Python": 433443, "Shell": 4228}, "github_archived": false, "github_pushed_at": "2024-07-16T14:51:56Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/novelty-search-in-representational-space-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "39CEoMXTHh", "year": 2026, "status": "rejected", "title": "$\\texttt{LLINBO}$: Trustworthy LLM-in-the-Loop Bayesian Optimization", "authors": ["Chih-Yu Chang", "Milad Azvar", "Chinedum Okwudire", "Raed Al Kontar"], "authorids": ["~Chih-Yu_Chang1", "~Milad_Azvar1", "~Chinedum_Okwudire1", "~Raed_Al_Kontar1"], "authors_source": "OpenReview API", "abstract": "Bayesian optimization (BO) is a sequential decision-making tool widely used for optimizing expensive black-box functions. Recently, Large Language Models (LLMs) have shown remarkable adaptability in low-data regimes, making them promising tools for black-box optimization by leveraging contextual knowledge to propose high-quality query points. However, relying solely on LLMs as optimization agents introduces risks due to their lack of explicit surrogate modeling and calibrated uncertainty, as well as their inherently opaque internal mechanisms. This structural opacity makes it difficult to characterize or control the exploration–exploitation trade-off, ultimately undermining theoretical tractability and reliability. To address this, we propose $\\texttt{LLINBO}$: LLM-in-the-Loop BO, a hybrid framework for BO that combines LLMs with statistical surrogate experts (e.g., Gaussian Processes ($\\mathcal{GP}$)). The core philosophy is to leverage contextual reasoning strengths of LLMs for early exploration, while relying on principled statistical models to guide efficient exploitation. Specifically, we introduce three mechanisms that enable this collaboration and establish their theoretical guarantees. We end the paper with a real-life proof-of-concept in the context of 3D printing.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=39CEoMXTHh", "arxiv_id": "2505.14756", "paper_pdf": "papers/39CEoMXTHh.pdf", "paper_pdf_sha256": "21b890b29c7771600da688fe419c66487b298a65e920a5cbfc01852952ca2c6f", "paper_pdf_bytes": 7566648, "paper_pdf_source": "openreview", "code_url": "https://github.com/UMDataScienceLab/LLM-in-the-Loop-BO", "code_repository": "UMDataScienceLab/LLM-in-the-Loop-BO", "code_commit": "105701dffee2999f31de1c0f42241d0e82b48e55", "code_archive": "repos/39CEoMXTHh.zip", "code_archive_sha256": "c04ddc8e9aa7decee1433f04e6f9294984c5ccee99951688537370a0d9844d92", "code_archive_bytes": 2232443, "code_file_count": 7, "code_extensions": {".py": 4, ".ipynb": 3}, "github_disk_usage_kb": 2226, "github_languages": {"Jupyter Notebook": 139972, "Python": 63501}, "github_archived": false, "github_pushed_at": "2025-05-20T09:06:11Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/texttt-llinbo-trustworthy-llm-in-the-loop"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "tGOOP7DGxs", "year": 2024, "status": "rejected", "title": "Graph Transformers for Large Graphs", "authors": ["Vijay Prakash Dwivedi", "Yozen Liu", "Anh Tuan Luu", "Xavier Bresson", "Neil Shah", "Tong Zhao"], "authorids": ["~Vijay_Prakash_Dwivedi1", "~Yozen_Liu1", "~Anh_Tuan_Luu2", "~Xavier_Bresson6", "~Neil_Shah2", "~Tong_Zhao3"], "authors_source": "OpenReview API", "abstract": "Transformers have recently emerged as powerful neural networks for graph learning, showcasing state-of-the-art performance on several graph property prediction tasks. However, these results have been limited to small-scale graphs, such as ligand molecules with fewer than a hundred atoms, where the computational feasibility of the global attention mechanism is possible. The next goal is to scale up these architectures to handle very large graphs on the scale of millions or even billions of nodes. With large-scale graphs, global attention learning is proven impractical due to its quadratic complexity w.r.t. the number of nodes. On the other hand, neighborhood sampling techniques become essential to manage large graph sizes, yet finding the optimal trade-off between speed and accuracy with sampling techniques remains challenging. This work advances representation learning on single large-scale graphs with a focus on identifying model characteristics and critical design constraints for developing scalable graph transformer (GT) architectures. We argue such GT requires layers that can adeptly learn both local and global graph representations while swiftly sampling the graph topology. As such, a key innovation of this work lies in the creation of a fast neighborhood sampling technique coupled with a local attention mechanism that encompasses a 4-hop reception field, but achieved through just 2-hop operations. This local node embedding is then integrated with a global node embedding, acquired via another self-attention layer with an approximate global codebook, before finally sent through a downstream layer for node predictions. The proposed GT framework, named LargeGT, overcomes previous computational bottlenecks and is validated on three large-scale node classification benchmarks. We report a 3× speedup and 16.8% performance gain on ogbn-products and snap-patents compared to their nearest baselines respectively, while we also scale LargeGT on ogbn-papers100M with a 5.9% improvement in performance.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=tGOOP7DGxs", "arxiv_id": "2312.11109", "paper_pdf": "papers/tGOOP7DGxs.pdf", "paper_pdf_sha256": "fbe21b0dd6e556753be60eb90bdb3dccbe2ec8b54bd9fba09e1ddeb27f2671a1", "paper_pdf_bytes": 708610, "paper_pdf_source": "openreview", "code_url": "https://github.com/snap-research/LargeGT", "code_repository": "snap-research/LargeGT", "code_commit": "532326966ef25f2f28c25d034ef06c1c3b9a141a", "code_archive": "repos/tGOOP7DGxs.zip", "code_archive_sha256": "90a2c848c0003f24cf7ce5a85444bc6d1819bb3d6dc0819a99b3a18ca6e92a45", "code_archive_bytes": 103115, "code_file_count": 9, "code_extensions": {".py": 8, ".sh": 1}, "github_disk_usage_kb": 135, "github_languages": {"Python": 55366, "Shell": 4914}, "github_archived": false, "github_pushed_at": "2024-04-26T22:28:56Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/graph-transformers-for-large-graphs"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "_DYi95e8CAe", "year": 2023, "status": "rejected", "title": "Multi-Task Structural Learning using Local Task Similarity induced Neuron Creation and Removal", "authors": ["Naresh Kumar Gurulingan", "Elahe Arani", "Bahram Zonooz"], "authorids": ["~Naresh_Kumar_Gurulingan1", "~Elahe_Arani1", "~Bahram_Zonooz1"], "authors_source": "OpenReview API", "abstract": "Multi-task learning has the potential to improve generalization by maximizing positive transfer between tasks while reducing task interference. Fully achieving this potential is hindered by manually designed architectures that remain static throughout training. In contrast, learning in the brain occurs through structural changes that are in tandem with changes in synaptic strength. Therefore, we propose Multi-Task Structural Learning (MTSL) which simultaneously learns the multi-task architecture and its parameters. MTSL begins with an identical single task network for each task and alternates between a task learning phase and a structural learning phase. In the task learning phase, each network specializes in the corresponding task. In each of the structural learning phases, starting from the earliest layer, locally similar task layers first transfer their knowledge to a newly created group layer after which they become redundant and are removed. Our experimental results show that MTSL achieves competitive generalization with various baselines and improves robustness to out-of-distribution data.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=_DYi95e8CAe", "arxiv_id": "2305.00441", "paper_pdf": "papers/_DYi95e8CAe.pdf", "paper_pdf_sha256": "fea0f81c436a996951924dce32dc69dc654ae7e90ff4c4eff83123ed0473778f", "paper_pdf_bytes": 528478, "paper_pdf_source": "openreview", "code_url": "https://github.com/NeurAI-Lab/MTSL", "code_repository": "NeurAI-Lab/MTSL", "code_commit": "28db1ac294052062dc59d48762f94724146f8e51", "code_archive": "repos/_DYi95e8CAe.zip", "code_archive_sha256": "91b8e4705c1229b69444fc5cc0a4a9f9ba11a5fbea8a2897fe25d88e6d358824", "code_archive_bytes": 248732, "code_file_count": 56, "code_extensions": {".py": 56}, "github_disk_usage_kb": 235, "github_languages": {"Python": 253044}, "github_archived": false, "github_pushed_at": "2023-06-21T14:44:24Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/multi-task-structural-learning-using-local"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "MRGFutr0p5e", "year": 2022, "status": "rejected", "title": "Graph Barlow Twins: A self-supervised representation learning framework for graphs", "authors": ["Piotr Bielak", "Tomasz Jan Kajdanowicz", "Nitesh Chawla"], "authorids": ["~Piotr_Bielak1", "~Tomasz_Jan_Kajdanowicz1", "~Nitesh_Chawla1"], "authors_source": "OpenReview API", "abstract": "The self-supervised learning (SSL) paradigm is an essential exploration area, which tries to eliminate the need for expensive data labeling. Despite the great success of SSL methods in computer vision and natural language processing, most of them employ contrastive learning objectives that require negative samples, which are hard to define. This becomes even more challenging in the case of graphs and is a bottleneck for achieving robust representations. To overcome such limitations, we propose a framework for self-supervised graph representation learning - Graph Barlow Twins, which utilizes a cross-correlation-based loss function instead of negative samples. Moreover, it does not rely on non-symmetric neural network architectures - in contrast to state-of-the-art self-supervised graph representation learning method BGRL. We show that our method achieves as competitive results as the best self-supervised methods and fully supervised ones while requiring fewer hyperparameters and substantially shorter computation time (ca. 30 times faster than BGRL).", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=MRGFutr0p5e", "arxiv_id": "2106.02466", "paper_pdf": "papers/MRGFutr0p5e.pdf", "paper_pdf_sha256": "82e54fe756c61cc415cd38fc3f2ee4ec7c36e4d1822f3079316083c69193f270", "paper_pdf_bytes": 361179, "paper_pdf_source": "openreview", "code_url": "https://github.com/pbielak/graph-barlow-twins", "code_repository": "pbielak/graph-barlow-twins", "code_commit": "ec62580aa89bf3f0d20c92e7549031deedc105ab", "code_archive": "repos/MRGFutr0p5e.zip", "code_archive_sha256": "f59e245b6fc3cffd6a4e083ab004d03aae5d71ceb80f94c61f72b061995a89ba", "code_archive_bytes": 233851, "code_file_count": 59, "code_extensions": {".py": 57, ".sh": 2}, "github_disk_usage_kb": 550, "github_languages": {"Python": 203808, "Dockerfile": 782, "Shell": 623}, "github_archived": false, "github_pushed_at": "2023-10-15T10:42:29Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/graph-barlow-twins-a-self-supervised"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "1ibNKMp8SKc", "year": 2021, "status": "rejected", "title": "On Disentangled Representations Learned From Correlated Data", "authors": ["Frederik Träuble", "Elliot Creager", "Niki Kilbertus", "Anirudh Goyal", "Francesco Locatello", "Bernhard Schölkopf", "Stefan Bauer"], "authorids": ["~Frederik_Träuble1", "~Elliot_Creager1", "~Niki_Kilbertus1", "~Anirudh_Goyal1", "~Francesco_Locatello1", "~Bernhard_Schölkopf1", "~Stefan_Bauer1"], "authors_source": "OpenReview API", "abstract": "Despite impressive progress in the last decade, it still remains an open challenge to build models that generalize well across multiple tasks and datasets. One path to achieve this is to learn meaningful and compact representations, in which different semantic aspects of data are structurally disentangled. The focus of disentanglement approaches has been on separating independent factors of variation despite the fact that real-world observations are often not structured into meaningful independent causal variables. In this work, we bridge the gap to real-world scenarios by analyzing the behavior of most prominent methods and disentanglement scores on correlated data in a large scale empirical study (including 4260 models). We show that systematically induced correlations in the dataset are being learned and reflected in the latent representations, while widely used disentanglement scores fall short of capturing these latent correlations. Finally, we demonstrate how to disentangle these latent correlations using weak supervision, even if we constrain this supervision to be causally plausible. Our results thus support the argument to learn independent mechanisms rather than independent factors of variations.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=1ibNKMp8SKc", "arxiv_id": "2006.07886", "paper_pdf": "papers/1ibNKMp8SKc.pdf", "paper_pdf_sha256": "dbeeab21fb6fec7ba99c09c04b8fb42ac3694b4fd1be7224fddf3bd0545d0809", "paper_pdf_bytes": 12556365, "paper_pdf_source": "openreview", "code_url": "https://github.com/ftraeuble/disentanglement_lib", "code_repository": "ftraeuble/disentanglement_lib", "code_commit": "39584ad5270a090723c8f12567c47d84f02444c9", "code_archive": "repos/1ibNKMp8SKc.zip", "code_archive_sha256": "6cce794571d5298cc269ce94d36e1c3f3508487542d8518f0b2bec8e2bfc5128", "code_archive_bytes": 912960, "code_file_count": 126, "code_extensions": {".py": 126}, "github_disk_usage_kb": 741, "github_languages": {"Python": 660604, "Shell": 6767}, "github_archived": false, "github_pushed_at": "2021-06-08T18:37:00Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/is-independence-all-you-need-on-the"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "rJlDoT4twr", "year": 2020, "status": "rejected", "title": "Unified Probabilistic Deep Continual Learning through Generative Replay and Open Set Recognition", "authors": ["Martin Mundt", "Sagnik Majumder", "Iuliia Pliushch", "Visvanathan Ramesh"], "authorids": ["mmundt@em.uni-frankfurt.de", "majumder@ccc.cs.uni-frankfurt.de", "pliushch@em.uni-frankfurt.de", "ramesh@fias.uni-frankfurt.de"], "authors_source": "OpenReview API", "abstract": "We introduce a unified probabilistic approach for deep continual learning based on variational Bayesian inference with open set recognition. Our model combines a joint probabilistic encoder with a generative model and a linear classifier that get shared across tasks. The open set recognition bounds the approximate posterior by fitting regions of high density on the basis of correctly classified data points and balances open set detection with recognition errors. Catastrophic forgetting is significantly alleviated through generative replay, where the open set recognition is used to sample from high density areas of the class specific posterior and reject statistical outliers. Our approach naturally allows for forward and backward transfer while maintaining past knowledge without the necessity of storing old data, regularization or inferring task labels. We demonstrate compelling results in the challenging scenario of incrementally expanding the single-head classifier for both class incremental visual and audio classification tasks, as well as incremental learning of datasets across modalities.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=rJlDoT4twr", "arxiv_id": "1905.12019", "paper_pdf": "papers/rJlDoT4twr.pdf", "paper_pdf_sha256": "1c1a0e20a0bcea8cc9981ce51d8f5ceeacc78494b649f204598fa3e66352818a", "paper_pdf_bytes": 8387536, "paper_pdf_source": "openreview", "code_url": "https://github.com/MrtnMndt/OpenVAE_ContinualLearning", "code_repository": "MrtnMndt/OpenVAE_ContinualLearning", "code_commit": "2cca59334c8993f96af15daf70d99ba89fea003f", "code_archive": "repos/rJlDoT4twr.zip", "code_archive_sha256": "e779449113aa7315d08be861bbc253c06c146fb2b3d1cac9de079c131ed7829c", "code_archive_bytes": 1293385, "code_file_count": 17, "code_extensions": {".py": 17}, "github_disk_usage_kb": 1945, "github_languages": {"Python": 236893}, "github_archived": false, "github_pushed_at": "2022-04-01T12:53:56Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/unified-probabilistic-deep-continual-learning"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "jimY7NF6aJ", "year": 2026, "status": "rejected", "title": "MUSEG: Reinforcing Video Temporal Understanding via Timestamp-Aware Multi-Segment Grounding", "authors": ["Fuwen Luo", "Shengfeng Lou", "Chi Chen", "Ziyue Wang", "Chenliang Li", "Weizhou Shen", "Jiyue Guo", "Peng Li", "Ming Yan", "Fei Huang", "Yang Liu"], "authorids": ["~Fuwen_Luo1", "~Shengfeng_Lou1", "~Chi_Chen1", "~Ziyue_Wang4", "~Chenliang_Li2", "~Weizhou_Shen1", "~Jiyue_Guo1", "~Peng_Li2", "~Ming_Yan2", "~Fei_Huang2", "~Yang_Liu19"], "authors_source": "OpenReview API", "abstract": "Video temporal understanding is crucial for multimodal large language models (MLLMs) to reason over events in videos. Despite recent advances in general video understanding, current MLLMs still struggle with fine-grained temporal reasoning. While reinforcement learning (RL) has been explored to address this issue recently, existing RL approaches remain limited in performance on time-sensitive tasks. In this work, we propose **MUSEG**, a novel RL-based method that enhances temporal understanding by introducing timestamp-aware multi-segment grounding. MUSEG enables MLLMs to align queries with multiple relevant video segments, promoting more comprehensive temporal reasoning. To facilitate effective learning, we design a customized RL training recipe with phased rewards that progressively guides the model toward temporally grounded reasoning. Extensive experiments on temporal grounding and time-sensitive video question answering (QA) tasks demonstrate that \\methodname significantly outperforms existing methods and generalizes well across diverse temporal understanding scenarios.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=jimY7NF6aJ", "arxiv_id": "2505.20715", "paper_pdf": "papers/jimY7NF6aJ.pdf", "paper_pdf_sha256": "0a1bcf0323282b28113d75da834ad10103ede1f83ceef716b544035e4bfb5e7c", "paper_pdf_bytes": 1219029, "paper_pdf_source": "openreview", "code_url": "https://github.com/THUNLP-MT/MUSEG", "code_repository": "THUNLP-MT/MUSEG", "code_commit": "889565e78b21e95f9ee2d2f6d250c3248b1587bb", "code_archive": "repos/jimY7NF6aJ.zip", "code_archive_sha256": "dc020eedecddfe0ff4144a41437edb96bd1c268d2655133d490878ef655f80ae", "code_archive_bytes": 2336496, "code_file_count": 14, "code_extensions": {".sh": 8, ".py": 6}, "github_disk_usage_kb": 2276, "github_languages": {"Python": 89374, "Shell": 3596}, "github_archived": false, "github_pushed_at": "2025-06-09T04:32:19Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/museg-reinforcing-video-temporal"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "bqUsdBeRjQ", "year": 2025, "status": "rejected", "title": "Personalized Language Modeling from Personalized Human Feedback", "authors": ["Xinyu Li", "Ruiyang Zhou", "Zachary Chase Lipton", "Liu Leqi"], "authorids": ["~Xinyu_Li7", "~Ruiyang_Zhou1", "~Zachary_Chase_Lipton1", "~Liu_Leqi1"], "authors_source": "OpenReview API", "abstract": "Personalized large language models (LLMs)  are designed to tailor responses to individual user preferences. While Reinforcement Learning from Human Feedback (RLHF) is a commonly used framework for aligning LLMs with human preferences, vanilla RLHF assumes that all human preferences share the same distribution, preventing fine-tuned LLMs from generating personalized content when user preferences are diverse. In this work, we propose Personalized-RLHF (P-RLHF), an efficient framework that utilizes a lightweight user model to capture individual user preferences and jointly learns the user model and the personalized LLM from human feedback. P-RLHF exhibits the following three characteristics: (1) It enables an LLM to generate personalized content and scale efficiently with growing number of users. (2) It handles both explicit user preferences described as textual input and implicit user preferences encoded in the feedback data. (3) It eliminates the need for users to fully articulate their preferences, which are normally needed for prompting LLMs to generate personalized content yet are often impractical to obtain in real-world scenarios. Our experimental results show that personalized LLMs trained using P-RLHF generate responses that are more closely aligned with individual user preferences, outperforming vanilla, non-personalized RLHF and prompting-based personalization approaches across different tasks.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=bqUsdBeRjQ", "arxiv_id": "2402.05133", "paper_pdf": "papers/bqUsdBeRjQ.pdf", "paper_pdf_sha256": "b118110a056ec4199f663b2a38668714995b0f8f2c095d91423c13f9104da245", "paper_pdf_bytes": 576712, "paper_pdf_source": "openreview", "code_url": "https://github.com/HumainLab/Personalized_RLHF", "code_repository": "HumainLab/Personalized_RLHF", "code_commit": "ae1f81bf8a344cbdfa6bc4e8dd6ce7823757db9f", "code_archive": "repos/bqUsdBeRjQ.zip", "code_archive_sha256": "12842b18dc934a827ef6d43597fc6ffc46a7009d44ae3dcee2bb69b78d329ce6", "code_archive_bytes": 233944, "code_file_count": 61, "code_extensions": {".py": 43, ".sh": 14, ".ipynb": 4}, "github_disk_usage_kb": 167, "github_languages": {"Python": 326357, "Jupyter Notebook": 56604, "Shell": 14440}, "github_archived": false, "github_pushed_at": "2024-12-09T05:44:30Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/personalized-language-modeling-from"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "gUL6zYN4Uaf", "year": 2023, "status": "rejected", "title": "Cramming: Training a language model on a single GPU in one day", "authors": ["Jonas Geiping", "Tom Goldstein"], "authorids": ["~Jonas_Geiping1", "~Tom_Goldstein1"], "authors_source": "OpenReview API", "abstract": "Recent trends in language modeling have focused on increasing performance through scaling, and have resulted in an environment where training language models is out of reach for most researchers and practitioners.  While most in the community are asking how to push the limits of extreme computation, we ask the opposite question:  \nHow far can we get with a single GPU in just one day? \nWe investigate the downstream performance achievable with a transformer-based language model trained completely from scratch with masked language modeling for a single day on a single consumer GPU.\nAside from re-analyzing nearly all components of the pretraining pipeline for this scenario and providing a modified pipeline with performance close to BERT, we investigate why scaling down is hard, and which modifications actually improve performance in this scenario. We provide evidence that even in this constrained setting, performance closely follows scaling laws observed in large-compute settings. Through the lens of scaling laws, we categorize a range of recent improvements to training and architecture and discuss their merit and practical applicability (or lack thereof) for the limited compute setting.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=gUL6zYN4Uaf", "arxiv_id": "2212.14034", "paper_pdf": "papers/gUL6zYN4Uaf.pdf", "paper_pdf_sha256": "e5b18783401f125b8c5f4cd61469e662064876a5cb0070bfd1fa947bf6be05ff", "paper_pdf_bytes": 706584, "paper_pdf_source": "openreview", "code_url": "https://github.com/JonasGeiping/cramming", "code_repository": "JonasGeiping/cramming", "code_commit": "196c9912d8c5b06a05e9a58edd1521e3d38f7c0c", "code_archive": "repos/gUL6zYN4Uaf.zip", "code_archive_sha256": "71e797eb8a000d855a1cf2924462750fa444c6a298b9417747433783cc4f7b3a", "code_archive_bytes": 228108, "code_file_count": 56, "code_extensions": {".py": 41, ".sh": 15}, "github_disk_usage_kb": 237, "github_languages": {"Python": 375582, "Shell": 135326}, "github_archived": false, "github_pushed_at": "2024-06-13T07:56:48Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/cramming-training-a-language-model-on-a"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "yOBqNg-CqB0", "year": 2022, "status": "rejected", "title": "Re-evaluating Word Mover's Distance", "authors": ["Ryoma Sato", "Makoto Yamada", "Hisashi Kashima"], "authorids": ["~Ryoma_Sato1", "~Makoto_Yamada3", "~Hisashi_Kashima2"], "authors_source": "OpenReview API", "abstract": "The word mover's distance (WMD) is a fundamental technique for measuring the similarity of two documents. As the crux of WMD, it can take advantage of the underlying geometry of the word space by employing an optimal transport formulation. The original study on WMD reported that WMD outperforms classical baselines such as bag-of-words (BOW) and TF-IDF by significant margins in various datasets. In this paper, we point out that the evaluation in the original study could be misleading. We re-evaluate the performances of WMD and the classical baselines and find that the classical baselines are competitive with WMD if we employ an appropriate preprocessing, i.e., L1 normalization. In addition, We introduce an analogy between WMD and L1-normalized BOW and find that not only the performance of WMD but also the distance values resemble those of BOW in high dimensional spaces.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=yOBqNg-CqB0", "arxiv_id": "2105.14403", "paper_pdf": "papers/yOBqNg-CqB0.pdf", "paper_pdf_sha256": "ce33ca71a6c9eda4af42247d277db5846b0b4f3f97a99d32fda30d3e3f574ce1", "paper_pdf_bytes": 459310, "paper_pdf_source": "openreview", "code_url": "https://github.com/joisino/reeval-wmd", "code_repository": "joisino/reeval-wmd", "code_commit": "67d0c8e582c7f4b13befb9a84f128fb818c6b388", "code_archive": "repos/yOBqNg-CqB0.zip", "code_archive_sha256": "077c4acac0a1216c1d20ed81981194fd984c4dbd176cf4aeee071a41d0f7c583", "code_archive_bytes": 563248, "code_file_count": 9, "code_extensions": {".py": 8, ".sh": 1}, "github_disk_usage_kb": 557, "github_languages": {"Python": 36320, "Shell": 256}, "github_archived": false, "github_pushed_at": "2022-06-15T05:00:09Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/re-evaluating-word-mover-s-distance"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "V1N4GEWki_E", "year": 2021, "status": "rejected", "title": "Gradient Flow in Sparse Neural Networks and How Lottery Tickets Win", "authors": ["Utku Evci", "Yani Ioannou", "Cem Keskin", "Yann Dauphin"], "authorids": ["~Utku_Evci1", "~Yani_Ioannou1", "~Cem_Keskin2", "~Yann_Dauphin1"], "authors_source": "OpenReview API", "abstract": "Sparse Neural Networks (NNs) can match the generalization of dense NNs using a fraction of the compute/storage for inference, and also have the potential to enable efficient training. However, naively training unstructured sparse NNs from random initialization results in significantly worse generalization, with the notable exception of Lottery Tickets (LTs) and Dynamic Sparse Training (DST). In this work, we attempt to answer: (1) why training unstructured sparse networks from random initialization performs poorly and; and (2) what makes LTs and DST the exceptions? We show that sparse NNs have poor gradient flow at initialization and propose a modified initialization for unstructured connectivity. Furthermore, we find that DST methods significantly improve gradient flow during training over traditional sparse training methods. Finally, we show that LTs do not improve gradient flow, rather their success lies in re-learning the pruning solution they are derived from — however, this comes at the cost of learning novel solutions.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=V1N4GEWki_E", "arxiv_id": "2010.03533", "paper_pdf": "papers/V1N4GEWki_E.pdf", "paper_pdf_sha256": "7d6e5c1e5e2ea9926f5c7151ed93f3b1d05c06027db2322b623c7885314e6283", "paper_pdf_bytes": 1072138, "paper_pdf_source": "openreview", "code_url": "https://github.com/google-research/rigl", "code_repository": "google-research/rigl", "code_commit": "d39fc7d46505cb3196cb1edeb32ed0b6dd44c0f9", "code_archive": "repos/V1N4GEWki_E.zip", "code_archive_sha256": "0f60b5406e049cb72423bfad48b6549c8bceb61dab16629051be2cef8c6f9d65", "code_archive_bytes": 808678, "code_file_count": 95, "code_extensions": {".py": 88, ".ipynb": 4, ".sh": 3}, "github_disk_usage_kb": 830, "github_languages": {"Python": 791400, "Jupyter Notebook": 71595, "Shell": 2845}, "github_archived": false, "github_pushed_at": "2023-01-26T17:47:14Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/gradient-flow-in-sparse-neural-networks-and-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "SJeS16EKPr", "year": 2020, "status": "rejected", "title": "Learning relevant features for statistical inference", "authors": ["Cédric Bény"], "authorids": ["cedric.beny@gmail.com"], "authors_source": "OpenReview API", "abstract": "We introduce an new technique to learn correlations between two types of data.\nThe learned representation can be used to directly compute the expectations of functions over one type of data conditioned on the other, such as Bayesian estimators and their standard deviations. \nSpecifically, our loss function teaches two neural nets to extract features representing the probability vectors of highest singular value for the stochastic map (set of conditional probabilities) implied by the joint dataset, relative to the inner product defined by the Fisher information metrics evaluated at the marginals.\nWe test the approach using a synthetic dataset, analytical calculations, and inference on occluded MNIST images. \nSurprisingly, when applied to supervised learning (one dataset consists of labels), this approach automatically provides regularization and faster convergence compared to the cross-entropy objective.\nWe also explore using this approach to discover salient independent features of a single dataset. ", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=SJeS16EKPr", "arxiv_id": "1904.10387", "paper_pdf": "papers/SJeS16EKPr.pdf", "paper_pdf_sha256": "a27836b98fbc649ded272a72b1c5c6cab418185f35507c452f7c60bd2d1fbb83", "paper_pdf_bytes": 878876, "paper_pdf_source": "openreview", "code_url": "https://github.com/cbeny/RFA", "code_repository": "cbeny/RFA", "code_commit": "fe4a331f2a7baea1c41892815978d93b6d10b8c4", "code_archive": "repos/SJeS16EKPr.zip", "code_archive_sha256": "c50a2c1ff960ddfd965b025843f70aca34123aed1071d96da3549117a9078fbe", "code_archive_bytes": 527427, "code_file_count": 6, "code_extensions": {".jl": 3, ".py": 2, ".ipynb": 1}, "github_disk_usage_kb": 2070, "github_languages": {"Jupyter Notebook": 749107, "Julia": 18202, "Python": 7560}, "github_archived": false, "github_pushed_at": "2020-05-13T11:37:48Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/relevant-feature-extraction-for-statistical"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "NdbUfhttc1", "year": 2024, "status": "rejected", "title": "Learning to Optimize for Reinforcement Learning", "authors": ["Qingfeng Lan", "A. Rupam Mahmood", "Shuicheng YAN", "Zhongwen Xu"], "authorids": ["~Qingfeng_Lan1", "~A._Rupam_Mahmood1", "~Shuicheng_YAN3", "~Zhongwen_Xu1"], "authors_source": "OpenReview API", "abstract": "In recent years, by leveraging more data, computation, and diverse tasks, learned optimizers have achieved remarkable success in supervised learning, outperforming classical hand-designed optimizers. Reinforcement learning (RL) is essentially different from supervised learning and in practice these learned optimizers do not work well even in simple RL tasks. We investigate this phenomenon and identity three issues. First, the gradients of an RL agent vary across a wide range in logarithms while their absolute values are in a small range, making neural networks hard to obtain accurate parameter updates. Second, the agent-gradient distribution is non-independent and identically distributed, leading to inefficient meta-training. Finally, due to highly stochastic agent-environment interactions, the agent-gradients have high bias and variance, which increase the difficulty of learning an optimizer for RL. We propose gradient processing, pipeline training, and a novel optimizer structure with good inductive bias to address these issues. By applying these techniques, for the first time, we show that learning an optimizer for RL from scratch is possible. Although only trained in toy tasks, our learned optimizer can generalize to unseen complex tasks in Brax.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=NdbUfhttc1", "arxiv_id": "2302.01470", "paper_pdf": "papers/NdbUfhttc1.pdf", "paper_pdf_sha256": "0f9a7e811e1a5386e20500e6a27b8eaa029052afb4b021f98e4ca29b873a77d9", "paper_pdf_bytes": 1647564, "paper_pdf_source": "openreview", "code_url": "https://github.com/sail-sg/optim4rl", "code_repository": "sail-sg/optim4rl", "code_commit": "523d42ecea582d432246e80fce0e88442497a6d0", "code_archive": "repos/NdbUfhttc1.zip", "code_archive_sha256": "8a18f5af557ee4f912410e0687ed05ef2bde44be1df635b28ded45d1741ed836", "code_archive_bytes": 118641, "code_file_count": 35, "code_extensions": {".py": 34, ".sh": 1}, "github_disk_usage_kb": 138, "github_languages": {"Python": 222651, "Shell": 6976}, "github_archived": false, "github_pushed_at": "2024-11-27T01:33:39Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/learning-to-optimize-for-reinforcement"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "bGC7Ai125lR", "year": 2023, "status": "rejected", "title": "Towards Understanding How Machines Can Learn Causal Overhypotheses ", "authors": ["Eliza Kosoy", "David Chan", "Adrian Liu", "Jasmine Collins", "Bryanna Kaufmann", "Sandy Huang", "Jessica B Hamrick", "John Canny", "Nan Rosemary Ke", "Alison Gopnik"], "authorids": ["~Eliza_Kosoy1", "~David_Chan3", "~Adrian_Liu1", "~Jasmine_Collins1", "bryannakaufmann@berkeley.edu", "~Sandy_Huang1", "~Jessica_B_Hamrick1", "~John_Canny1", "~Nan_Rosemary_Ke1", "~Alison_Gopnik1"], "authors_source": "OpenReview API", "abstract": "Recent work in machine learning and cognitive science has suggested that understanding causal information is essential to the development of intelligence. One of the key challenges for current machine learning algorithms is modeling and understanding causal overhypotheses: transferable abstract hypotheses about sets of causal relationships. In contrast, even young children spontaneously learn causal overhypotheses, and use these to guide their exploration or to generalize to new situations. This has been demonstrated in a variety of cognitive science experiments using the “blicket detector” environment. We present a causal learning benchmark adapting the “blicket\" environment for machine learning agents and evaluate a range of state-of-the-art methods in this environment. We find that although most agents have no problem learning causal structures seen during training, they are unable to learn causal overhypotheses from these experiences, and thus cannot generalize to new settings. ", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=bGC7Ai125lR", "arxiv_id": "2206.08353", "paper_pdf": "papers/bGC7Ai125lR.pdf", "paper_pdf_sha256": "e564b76c97a2323eb83270d0c74b3e2cd43424fbd9fc2ba42e97645c5567fa44", "paper_pdf_bytes": 1102829, "paper_pdf_source": "openreview", "code_url": "https://github.com/CannyLab/causal_overhypotheses", "code_repository": "CannyLab/causal_overhypotheses", "code_commit": "d9fe308e33512bd35dd5633b42806c69bd944930", "code_archive": "repos/bGC7Ai125lR.zip", "code_archive_sha256": "803c58868c749d3d483c37bebba43aad17f677535b6cab1e23dc73a6ec16025a", "code_archive_bytes": 244956, "code_file_count": 16, "code_extensions": {".py": 16}, "github_disk_usage_kb": 239, "github_languages": {"Python": 99641}, "github_archived": false, "github_pushed_at": "2022-08-16T19:13:22Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/towards-understanding-how-machines-can-learn"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "RW_GTtTfHJ6", "year": 2022, "status": "rejected", "title": "Causal Reinforcement Learning using Observational and Interventional Data", "authors": ["Maxime Gasse", "Damien GRASSET", "Guillaume Gaudron", "Pierre-Yves Oudeyer"], "authorids": ["~Maxime_Gasse2", "~Damien_GRASSET1", "~Guillaume_Gaudron1", "~Pierre-Yves_Oudeyer1"], "authors_source": "OpenReview API", "abstract": "Learning efficiently a causal model of the environment is a key challenge of model-based RL agents operating in POMDPs. We consider here a scenario where the learning agent has the ability to collect online experiences through direct interactions with the environment (interventional data), but also has access to a large collection of offline experiences, obtained by observing another agent interacting with the environment (observational data). A key ingredient, which makes this situation non-trivial, is that we allow the observed agent to act based on privileged information, hidden from the learning agent. We then ask the following questions: can the online and offline experiences be safely combined for learning a causal transition model ? And can we expect the offline experiences to improve the agent's performances ? To answer these, first we bridge the fields of reinforcement learning and causality, by importing ideas from the well-established causal framework of do-calculus, and expressing model-based reinforcement learning as a causal inference problem. Second, we propose a general yet simple methodology for safely leveraging offline data during learning. In a nutshell, our method relies on learning a latent-based causal transition model that explains both the interventional and observational regimes, and then inferring the standard POMDP transition model via deconfounding using the recovered latent variable. We prove our method is correct and efficient in the sense that it attains better generalization guarantees due to the offline data (in the asymptotic case), and we assess its effectiveness empirically on a series of synthetic toy problems.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=RW_GTtTfHJ6", "arxiv_id": "2106.14421", "paper_pdf": "papers/RW_GTtTfHJ6.pdf", "paper_pdf_sha256": "be8e3b529972c0a9c88bd8302a1127b12ef76c1fac821617b9d2cf5c87d3e52b", "paper_pdf_bytes": 1327916, "paper_pdf_source": "openreview", "code_url": "https://github.com/causal-rl-anonymous/causal-rl", "code_repository": "causal-rl-anonymous/causal-rl", "code_commit": "59fdfee79e579505ed4c43ee627fe7715874508e", "code_archive": "repos/RW_GTtTfHJ6.zip", "code_archive_sha256": "d75e925ff21f923ead7bc3d1d364e716ededbedd89f5dddbfca8f391c35e9a4f", "code_archive_bytes": 53073, "code_file_count": 25, "code_extensions": {".py": 25}, "github_disk_usage_kb": 584, "github_languages": {"Python": 166277}, "github_archived": false, "github_pushed_at": "2022-03-28T21:38:57Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/causal-reinforcement-learning-using"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "jQ0XleVhYuT", "year": 2021, "status": "rejected", "title": "Double Generative Adversarial Networks for Conditional Independence Testing", "authors": ["Chengchun Shi", "Tianlin Xu", "Wicher Pieter Bergsma", "Lexin Li"], "authorids": ["~Chengchun_Shi1", "~Tianlin_Xu1", "~Wicher_Pieter_Bergsma1", "~Lexin_Li1"], "authors_source": "OpenReview API", "abstract": "In this article, we consider the problem of high-dimensional conditional independence testing, which is a key building block in statistics and machine learning. We propose a double generative adversarial networks (GAN)-based inference procedure. We first introduce a double GANs framework to learn two generators, and integrate the two generators to construct a doubly-robust test statistic. We next consider multiple generalized covariance measures, and take their maximum as our test statistic. Finally, we obtain the empirical distribution of our test statistic through multiplier bootstrap. We show that our test controls type-I error, while the\npower approaches one asymptotically. More importantly, these theoretical guarantees are obtained under much weaker and practically more feasible conditions compared to existing tests. We demonstrate the efficacy of our test through both synthetic and real datasets.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=jQ0XleVhYuT", "arxiv_id": "2006.02615", "paper_pdf": "papers/jQ0XleVhYuT.pdf", "paper_pdf_sha256": "0bf1bddb6969e0ef2175aff837aa1142c5ddc33cb81dd960c8b7f135d25e04d7", "paper_pdf_bytes": 3445043, "paper_pdf_source": "openreview", "code_url": "https://github.com/tianlinxu312/dgcit", "code_repository": "tianlinxu312/dgcit", "code_commit": "686ea5ee151d947d42e5836866f8961420090080", "code_archive": "repos/jQ0XleVhYuT.zip", "code_archive_sha256": "8a11fb19b1fc8dfd193baf059a6449569123e94e53887ffa3f5d104cf4abff60", "code_archive_bytes": 863307, "code_file_count": 6, "code_extensions": {".py": 6}, "github_disk_usage_kb": 864, "github_languages": {"Python": 88906}, "github_archived": false, "github_pushed_at": "2023-11-22T18:37:04Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/double-generative-adversarial-networks-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "SJlVn6NKPB", "year": 2020, "status": "rejected", "title": "Representation Learning for Remote Sensing: An Unsupervised Sensor Fusion Approach", "authors": ["Aidan M. Swope", "Xander H. Rudelis", "Kyle T. Story"], "authorids": ["aidanswope@gmail.com", "xander@descarteslabs.com", "kyle@descarteslabs.com"], "authors_source": "OpenReview API", "abstract": "In the application of machine learning to remote sensing, labeled data is often scarce or expensive, which impedes the training of powerful models like deep convolutional neural networks. Although unlabeled data is abundant, recent self-supervised learning approaches are ill-suited to the remote sensing domain. In addition, most remote sensing applications currently use only a small subset of the multi-sensor, multi-channel information available, motivating the need for fused multi-sensor representations. We propose a new self-supervised training objective, Contrastive Sensor Fusion, which exploits coterminous data from multiple sources to learn useful representations of every possible combination of those sources. This method uses information common across multiple sensors and bands by training a single model to produce a representation that remains similar when any subset of its input channels is used. Using a dataset of 47 million unlabeled coterminous image triplets, we train an encoder to produce semantically meaningful representations from any possible combination of channels from the input sensors. These representations outperform fully supervised ImageNet weights on a remote sensing classification task and improve as more sensors are fused.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=SJlVn6NKPB", "arxiv_id": "2108.05094", "paper_pdf": "papers/SJlVn6NKPB.pdf", "paper_pdf_sha256": "62f62632ffd3790ca38c7ac40c7ca1f9bfc379ea5236a28ef80438d6ae9b76ab", "paper_pdf_bytes": 4506565, "paper_pdf_source": "openreview", "code_url": "https://github.com/dlarchives/contrastive_sensor_fusion", "code_repository": "dlarchives/contrastive_sensor_fusion", "code_commit": "f9d02e910e4b6213e0af5ea13b2c3fa99c8c8e1b", "code_archive": "repos/SJlVn6NKPB.zip", "code_archive_sha256": "63b34ac2dde04710f4e181f7b3aaa65dd748bd6c3a5a0a49050281771ff730bf", "code_archive_bytes": 715371, "code_file_count": 17, "code_extensions": {".py": 17}, "github_disk_usage_kb": 2104, "github_languages": {"Python": 63383}, "github_archived": false, "github_pushed_at": "2023-03-24T22:38:24Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/representation-learning-for-remote-sensing-an"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "EAUGN4pszX", "year": 2025, "status": "rejected", "title": "FreRA: A Frequency-Refined Augmentation for Contrastive Learning on Time Series Classification", "authors": ["Tian Tian", "Chunyan Miao", "Hangwei Qian"], "authorids": ["~Tian_Tian7", "~Chunyan_Miao1", "~Hangwei_Qian1"], "authors_source": "OpenReview API", "abstract": "Contrastive learning has emerged as a competent approach for unsupervised representation learning. However, the design of an optimal augmentation strategy, although crucial for contrastive learning, is less explored for time series classification tasks. Existing predefined time-domain augmentation methods are primarily adopted from vision and are not specific to time series data. Consequently, this cross-modality incompatibility may distort the global semantics of time series by introducing mismatched patterns into the data. To address this limitation, we present a novel perspective from the frequency domain and identify three advantages for downstream classification: 1) the frequency component naturally encodes global features, 2) the orthogonal nature of the Fourier basis allows easier isolation and independent modifications of critical and unimportant information, and 3) a compact set of frequency components can preserve semantic integrity. To fully utilize the three properties, we propose the lightweight yet effective Frequency-Refined Augmentation (FreRA) tailored for time series contrastive learning on classification tasks, which can be seamlessly integrated with contrastive learning frameworks in a plug-and-play manner. Specifically, FreRA automatically separates critical and unimportant frequency components. Accordingly, we propose Identity Modification and Self-adaptive Modification to protect global semantics in the critical frequency components and infuse variance to the unimportant ones respectively. \nTheoretically, we prove that FreRA generates semantic-preserving views. Empirically, we conduct extensive experiments on two benchmark datasets including UCR and UEA archives, as well as 5 large-scale datasets on diverse applications. FreRA consistently outperforms 10 leading baselines on time series classification, anomaly detection, and transfer learning tasks, demonstrating superior capabilities in contrastive representation learning and generalization in transfer learning scenarios across diverse datasets.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=EAUGN4pszX", "arxiv_id": "2505.23181", "paper_pdf": "papers/EAUGN4pszX.pdf", "paper_pdf_sha256": "8c8865a5390dc35b02fa4deadf5c3c6415d3bbb33a26954edeb8339e203540c8", "paper_pdf_bytes": 997286, "paper_pdf_source": "openreview", "code_url": "https://github.com/Tian0426/FreRA", "code_repository": "Tian0426/FreRA", "code_commit": "7236fbfc1c665f83ed5f4364cad59093ee283c14", "code_archive": "repos/EAUGN4pszX.zip", "code_archive_sha256": "2f8479ab1e690c17c4acc7aa846eb7dcec31f0739d0d6fd57f31227aa3771fb1", "code_archive_bytes": 158195, "code_file_count": 20, "code_extensions": {".py": 20}, "github_disk_usage_kb": 169, "github_languages": {"Python": 131451}, "github_archived": false, "github_pushed_at": "2025-06-20T05:23:44Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/frera-a-frequency-refined-augmentation-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "xMxHJxp192", "year": 2025, "status": "rejected", "title": "DeltaGNN: Graph Neural Network with Information Flow Control", "authors": ["Kevin Mancini", "Islem Rekik"], "authorids": ["~Kevin_Mancini1", "~Islem_Rekik1"], "authors_source": "OpenReview API", "abstract": "Graph Neural Networks (GNNs) are popular machine learning models designed to process graph-structured data through recursive neighborhood aggregations in the message passing process. When applied to semi-supervised node classification, the message-passing enables GNNs to understand short-range spatial interactions, but also causes them to suffer from over-smoothing and over-squashing. These challenges hinder model expressiveness and prevent the use of deeper models to capture long-range node interactions (LRIs) within the graph. Popular solutions for LRIs detection are either too expensive to process large graphs due to high time complexity or fail to generalize across diverse graph structures. To address these limitations, we propose a mechanism called information flow control, which leverages a novel connectivity measure, called information flow score, to address over-smoothing and over-squashing with linear computational overhead, supported by theoretical evidence. Finally, to prove the efficacy of our methodology we design DeltaGNN, the first scalable and generalizable approach for long-range and short-range interaction detection. \nWe benchmark our model across 10 real-world datasets, including graphs with varying sizes, topologies, densities, and homophilic ratios, showing superior performance with limited computational complexity.", "decision": "Reject", "meta_review": "", "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=xMxHJxp192", "arxiv_id": "2501.06002", "paper_pdf": "papers/xMxHJxp192.pdf", "paper_pdf_sha256": "59ac11372de6ce46ea080ba2fe123929962f371c3502b6e11e310cf44ef7def2", "paper_pdf_bytes": 4961402, "paper_pdf_source": "openreview", "code_url": "https://github.com/basiralab/DeltaGNN", "code_repository": "basiralab/DeltaGNN", "code_commit": "5e0c3bdc74e87e4500c6fdaec575841e7a0ecf00", "code_archive": "repos/xMxHJxp192.zip", "code_archive_sha256": "9f6cbd99461f23545bae753dff47ee13a9916f4168cf9e4e7d64088e76436540", "code_archive_bytes": 191110, "code_file_count": 35, "code_extensions": {".py": 27, ".sh": 8}, "github_disk_usage_kb": 174, "github_languages": {"Python": 252592, "Shell": 156030}, "github_archived": false, "github_pushed_at": "2026-08-26T09:29:12Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/deltagnn-graph-neural-network-with"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "8SPSIfR2e0", "year": 2024, "status": "rejected", "title": "Dissecting Language Models: Machine Unlearning via Selective Pruning", "authors": ["Nicky Pochinkov", "Nandi Schoots"], "authorids": ["~Nicky_Pochinkov1", "~Nandi_Schoots1"], "authors_source": "OpenReview API", "abstract": "Understanding and shaping the behaviour of Large Language Models (LLMs) is increasingly important as applications become more powerful and more frequently adopted.\nThis paper introduces a machine unlearning method specifically designed for LLMs. \nWe introduce a selective pruning method for LLMs that removes neurons based on their relative importance on a targeted capability compared to overall network performance. \nThis approach is a compute- and data-efficient method for identifying and removing neurons that enable specific behaviours.\nOur findings reveal that both feed-forward and attention neurons in LLMs are specialized; \nthat is, for specific tasks, certain neurons are more crucial than others.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=8SPSIfR2e0", "arxiv_id": "2403.01267", "paper_pdf": "papers/8SPSIfR2e0.pdf", "paper_pdf_sha256": "add0e21e1e68fa1e95655332f0e8fbee761c7e882a013eba296af36471c74255", "paper_pdf_bytes": 4096323, "paper_pdf_source": "openreview", "code_url": "https://github.com/nickypro/selective-pruning", "code_repository": "nickypro/selective-pruning", "code_commit": "96109170d686fad48efa5dc1319b2b5c3868976e", "code_archive": "repos/8SPSIfR2e0.zip", "code_archive_sha256": "bce0042b8b3cdc8b55fb4df7bf4d1c66418516e6b52b7be1b04e2df33f7bdf71", "code_archive_bytes": 150062, "code_file_count": 21, "code_extensions": {".py": 16, ".sh": 5}, "github_disk_usage_kb": 143, "github_languages": {"Python": 91819, "Shell": 7426}, "github_archived": false, "github_pushed_at": "2024-07-26T17:23:12Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/dissecting-language-models-machine-unlearning"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "fHT8kZcyyT", "year": 2023, "status": "rejected", "title": "CausalBench: A Large-scale Benchmark for Network Inference from Single-cell Perturbation Data", "authors": ["Mathieu Chevalley", "Yusuf H Roohani", "Arash Mehrjou", "Jure Leskovec", "Patrick Schwab"], "authorids": ["~Mathieu_Chevalley1", "~Yusuf_H_Roohani1", "~Arash_Mehrjou1", "~Jure_Leskovec1", "~Patrick_Schwab1"], "authors_source": "OpenReview API", "abstract": "Mapping biological mechanisms in cellular systems is a fundamental step in early-stage drug discovery that serves to generate hypotheses on what disease-relevant molecular targets may effectively be modulated by pharmacological interventions. With the advent of high-throughput methods for measuring single-cell gene expression under genetic perturbations, we now have effective means for generating evidence for causal gene-gene interactions at scale. However, inferring graphical networks of the size typically encountered in real-world gene-gene interaction networks is difficult in terms of both achieving and evaluating faithfulness to the true underlying causal graph. Moreover, standardised benchmarks for comparing methods for causal discovery in perturbational single-cell data do not yet exist. Here, we introduce CausalBench - a comprehensive benchmark suite for evaluating network inference methods on large-scale perturbational single-cell gene expression data. CausalBench introduces several biologically meaningful performance metrics and operates on two large, curated and openly available benchmark data sets for evaluating methods on the inference of gene regulatory networks from single-cell data generated under perturbations. With real-world datasets consisting of over 200000 training samples under interventions, CausalBench could potentially help facilitate advances in causal network inference by providing what is - to the best of our knowledge - the largest openly available test bed for causal discovery from real-world perturbation data to date.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=fHT8kZcyyT", "arxiv_id": "2210.17283", "paper_pdf": "papers/fHT8kZcyyT.pdf", "paper_pdf_sha256": "3d4254927a0e2599da21a008904d89d4bc01e5149ce7eb90930aeb516d46e57f", "paper_pdf_bytes": 442282, "paper_pdf_source": "openreview", "code_url": "https://github.com/causalbench/causalbench-starter", "code_repository": "causalbench/causalbench-starter", "code_commit": "74a26e31df85ac844960d2b1a2f883eea0b155e9", "code_archive": "repos/fHT8kZcyyT.zip", "code_archive_sha256": "0ab2703f074b319cb1a26cfe53d19a265c3024b95415a3feee179db52cc755cc", "code_archive_bytes": 239268, "code_file_count": 4, "code_extensions": {".py": 3, ".sh": 1}, "github_disk_usage_kb": 243, "github_languages": {"Python": 19878, "Shell": 1266, "Dockerfile": 71}, "github_archived": false, "github_pushed_at": "2023-05-17T12:14:48Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/causalbench-a-large-scale-benchmark-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "SbV8J9JHb6", "year": 2022, "status": "rejected", "title": "Soteria: In search of efficient neural networks for private inference", "authors": ["Anshul Aggarwal", "Trevor E Carlson", "Reza Shokri", "Shruti Tople"], "authorids": ["anshul@comp.nus.edu.sg", "~Trevor_E_Carlson1", "~Reza_Shokri1", "~Shruti_Tople2"], "authors_source": "OpenReview API", "abstract": "In the context of ML as a service, our objective is to protect the confidentiality of the users’ queries and the server's model parameters, with modest computation and communication overhead. Prior solutions primarily propose fine-tuning cryptographic methods to make them efficient for known fixed model architectures. The drawback with this line of approach is that the model itself is never designed to efficiently operate with existing cryptographic computations. We observe that the network architecture, internal functions, and parameters of a model, which are all chosen during training, significantly influence the computation and communication overhead of a cryptographic method, during inference.Thus, we propose SOTERIA — a training method to construct model architectures that are by-design efficient for private inference. We use neural architecture search algorithms with the dual objective of optimizing the accuracy of the model and the overhead of using cryptographic primitives for secure inference. Given the flexibility of modifying a model during training, we find accurate models that are also efficient for private computation. We select garbled circuits as our underlying cryptographic primitive, due to their expressiveness and efficiency. We empirically evaluate SOTERIA on MNIST and CIFAR10 datasets, to compare with the prior work on secure inference. Our results confirm that SOTERIA is indeed effective in balancing performance and accuracy.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=SbV8J9JHb6", "arxiv_id": "2007.12934", "paper_pdf": "papers/SbV8J9JHb6.pdf", "paper_pdf_sha256": "79b376a3900ad37b049427431cca14b7c9a74eadcf5a533325a83a81bb92a394", "paper_pdf_bytes": 219608, "paper_pdf_source": "openreview", "code_url": "https://github.com/privacytrustlab/soteria_private_nn_inference", "code_repository": "privacytrustlab/soteria_private_nn_inference", "code_commit": "4cf28c41f0c73c23bb026cc80a70ed41ee20da97", "code_archive": "repos/SbV8J9JHb6.zip", "code_archive_sha256": "aea2a1735f07a9941937d754617066c6f80711b9abbb26094cbef9acfdbffc8b", "code_archive_bytes": 611948, "code_file_count": 32, "code_extensions": {".py": 31, ".sh": 1}, "github_disk_usage_kb": 611, "github_languages": {"Verilog": 3296657, "SuperCollider": 319125, "SystemVerilog": 210024, "Python": 117109, "Dockerfile": 1417, "Shell": 88}, "github_archived": false, "github_pushed_at": "2020-09-29T15:15:06Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/soteria-in-search-of-efficient-neural"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "WdOCkf4aCM", "year": 2021, "status": "rejected", "title": "CDT: Cascading Decision Trees for Explainable Reinforcement Learning", "authors": ["Zihan Ding", "Pablo Hernandez-Leal", "Gavin Weiguang Ding", "Changjian Li", "Ruitong Huang"], "authorids": ["~Zihan_Ding1", "~Pablo_Hernandez-Leal2", "~Gavin_Weiguang_Ding1", "~Changjian_Li1", "~Ruitong_Huang1"], "authors_source": "OpenReview API", "abstract": "Deep Reinforcement Learning (DRL) has recently achieved significant advances in various domains. However, explaining the policy of RL agents still remains an open problem due to several factors, one being the complexity of explaining neural networks decisions. Recently, a group of works have used decision-tree-based models to learn explainable policies. Soft decision trees (SDTs) and discretized differentiable decision trees (DDTs) have been demonstrated to achieve both good performance and share the benefit of having explainable policies. In this work, we further improve the results for tree-based explainable RL in both performance and explainability. Our proposal, Cascading Decision Trees (CDTs) apply representation learning on the decision path to allow richer expressivity. Empirical results show that in both situations, where CDTs are used as policy function approximators or as imitation learners to explain black-box policies, CDTs can achieve better performances with more succinct and explainable models than SDTs. As a second contribution our study reveals limitations of explaining black-box policies via imitation learning with tree-based explainable models, due to its inherent instability.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=WdOCkf4aCM", "arxiv_id": "2011.07553", "paper_pdf": "papers/WdOCkf4aCM.pdf", "paper_pdf_sha256": "38786a72bcb69f815ba586eef61dcd6fed48071f3d0eb098adaf0b40b876d6dc", "paper_pdf_bytes": 6924375, "paper_pdf_source": "openreview", "code_url": "https://github.com/quantumiracle/Cascading-Decision-Tree", "code_repository": "quantumiracle/Cascading-Decision-Tree", "code_commit": "23e123787f502a1613733922cbf6711db0096f1a", "code_archive": "repos/WdOCkf4aCM.zip", "code_archive_sha256": "1423604ccc0efa8e8269c680c36d5c66baf638c499580622c75475066bf30eb7", "code_archive_bytes": 1789082, "code_file_count": 63, "code_extensions": {".py": 46, ".ipynb": 13, ".sh": 4}, "github_disk_usage_kb": 929, "github_languages": {"Jupyter Notebook": 2405577, "Python": 218614, "Shell": 2155}, "github_archived": false, "github_pushed_at": "2025-10-31T18:47:28Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/cdt-cascading-decision-trees-for-explainable-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "HJe9cR4KvB", "year": 2020, "status": "rejected", "title": "Learning to Contextually Aggregate Multi-Source Supervision for Sequence Labeling", "authors": ["Ouyu Lan*", "Xiao Huang*", "Bill Yuchen Lin", "He Jiang", "Xiang Ren"], "authorids": ["olan@usc.edu", "huan183@usc.edu", "yuchen.lin@usc.edu", "jian567@usc.edu", "xiangren@usc.edu"], "authors_source": "OpenReview API", "abstract": "Sequence labeling is a fundamental framework for various natural language processing problems including part-of-speech tagging and named entity recognition. Its performance is largely influenced by the annotation quality and quantity in supervised learning scenarios.  In many cases, ground truth labels are costly and time-consuming to collect or even non-existent,  while imperfect ones could be easily accessed or transferred from different domains. A typical example is crowd-sourced datasets which have multiple annotations for each sentence which may be noisy or incomplete.   Additionally,  predictions from multiple source models in transfer learning can be seen as a case of multi-source supervision.  In this paper, we propose a novel framework named Consensus Network (CONNET) to conduct training with imperfect annotations from multiple sources.   It learns the representation for every weak supervision source and dynamically aggregates them by a context-aware attention mechanism.  Finally, it leads to a model reflecting the consensus among multiple sources.  We evaluate the proposed framework in two practical settings of multi-source learning:  learning with crowd annotations and unsupervised cross-domain model adaptation. Extensive experimental results show that our model achieves significant improvements over existing methods in both settings.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=HJe9cR4KvB", "arxiv_id": "1910.04289", "paper_pdf": "papers/HJe9cR4KvB.pdf", "paper_pdf_sha256": "280f0278ecab706cc0382edbd49ff7ebd080192ba4601488de1561772595ecb8", "paper_pdf_bytes": 2088972, "paper_pdf_source": "openreview", "code_url": "https://github.com/INK-USC/ConNet", "code_repository": "INK-USC/ConNet", "code_commit": "adb299f160556004561df302c19578200bd3835b", "code_archive": "repos/HJe9cR4KvB.zip", "code_archive_sha256": "7d6cda2a1fd28206c3c5bbff552ac4a9748f2c4b383831c1e662cb98425df7f7", "code_archive_bytes": 2878408, "code_file_count": 67, "code_extensions": {".py": 56, ".sh": 11}, "github_disk_usage_kb": 2242, "github_languages": {"Python": 502939, "Shell": 28229}, "github_archived": false, "github_pushed_at": "2020-04-14T05:00:34Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/learning-to-contextually-aggregate-multi"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "0OUkySEAf0", "year": 2026, "status": "rejected", "title": "DeepTheorem: Advancing LLM Reasoning for Theorem Proving Through Natural Language and Reinforcement Learning", "authors": ["Ziyin Zhang", "Jiahao Xu", "Zhiwei He", "Tian Liang", "Qiuzhi Liu", "Yansi Li", "Linfeng Song", "Zhenwen Liang", "Zhuosheng Zhang", "Rui Wang", "Zhaopeng Tu", "Haitao Mi", "Dong Yu"], "authorids": ["~Ziyin_Zhang2", "~Jiahao_Xu1", "~Zhiwei_He1", "~Tian_Liang4", "~Qiuzhi_Liu1", "~Yansi_Li1", "~Linfeng_Song1", "~Zhenwen_Liang1", "~Zhuosheng_Zhang1", "~Rui_Wang10", "~Zhaopeng_Tu1", "~Haitao_Mi1", "~Dong_Yu2"], "authors_source": "OpenReview API", "abstract": "Theorem proving serves as a major testbed for evaluating complex reasoning abilities in large language models (LLMs). However, traditional automated theorem proving (ATP) approaches rely heavily on formal proof systems that poorly align with LLMs’ strength derived from informal, natural language knowledge acquired during pre-training. In this work, we introduce DeepTheorem, a comprehensive informal theorem-proving suite exploiting natural language to enhance LLM mathematical reasoning. DeepTheorem includes 1) a large-scale dataset of 121K\nhigh-quality IMO-level informal theorems and proofs spanning diverse mathematical domains, rigorously annotated for correctness, difficulty, and topic categories, accompanied by systematically constructed verifiable theorem variants; 2) adaptation of RL-Zero explicitly to informal theorem proving, leveraging the verified theorem variants to incentivize robust mathematical inference; 3) comprehensive outcome and process evaluation metrics examining proof correctness and the quality of reasoning steps; and 4) a new informal theorem proving benchmark consoliadted from three established math competitions, formatted for automatic evaluation. Extensive experimental analyses demonstrate DeepTheorem significantly improves LLM theorem-proving performance compared to existing datasets and supervised fine-tuning protocols, achieving state-of-the-art accuracy and reasoning quality. Our findings highlight DeepTheorem’s potential to fundamentally advance automated informal theorem proving and mathematical exploration.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=0OUkySEAf0", "arxiv_id": "2505.23754", "paper_pdf": "papers/0OUkySEAf0.pdf", "paper_pdf_sha256": "5b3e29585076ac1b4343473081c296b578747e805a141abab2b5949ef9de9c5a", "paper_pdf_bytes": 1647137, "paper_pdf_source": "openreview", "code_url": "https://github.com/Jiahao004/DeepTheorem", "code_repository": "Jiahao004/DeepTheorem", "code_commit": "7a841fd421b34b05f387766893b70773c2c65d4a", "code_archive": "repos/0OUkySEAf0.zip", "code_archive_sha256": "d0f638b6b67667f346ca226f2def28256e7d51e919c71051226fc522fbb8d70e", "code_archive_bytes": 2600525, "code_file_count": 6, "code_extensions": {".py": 5, ".sh": 1}, "github_disk_usage_kb": 2572, "github_languages": {"Python": 12120, "Shell": 3163}, "github_archived": false, "github_pushed_at": "2025-06-10T15:29:59Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/deeptheorem-advancing-llm-reasoning-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ns0KIpfQVy", "year": 2025, "status": "rejected", "title": "Multimodal Banking Dataset: Understanding Client Needs through Event Sequences", "authors": ["Dzhambulat Mollaev", "Alexander Kostin", "Postnova Maria", "Ivan Karpukhin", "Ivan A Kireev", "Gleb Gennadjevich Gusev", "Andrey Savchenko"], "authorids": ["~Dzhambulat_Mollaev1", "~Alexander_Kostin1", "~Postnova_Maria1", "~Ivan_Karpukhin1", "~Ivan_A_Kireev1", "~Gleb_Gennadjevich_Gusev1", "~Andrey_Savchenko1"], "authors_source": "OpenReview API", "abstract": "Financial organizations collect a huge amount of data about clients that typi-\ncally has a temporal (sequential) structure and is collected from multiple sources\n(modalities). However, despite the urgent practical need, developing deep learn-\ning techniques suitable to handle such data is limited by the absence of large open-\nsource multi-source real-world datasets of event sequences. To fill this gap mainly\ncaused by security reasons, we present the industrial-scale publicly available mul-\ntimodal banking dataset, MBD, that contains more than 2M corporate clients with\nseveral data sources: 950M bank transactions, 1B geo position events, 5M em-\nbeddings of dialogues with technical support and monthly aggregated purchases\nof four bank’s products. All entries are properly anonymized from real proprietary\nbank data. Moreover, we introduce a novel multimodal benchmark incorporating\nour MBD and two open-source financial datasets. We provide numerical results\ndemonstrating the superiority of fusion baselines over single-modal techniques\nfor each task. Moreover, our anonymization techniques still save all significant\ninformation for introduced downstream tasks.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=ns0KIpfQVy", "arxiv_id": "2409.17587", "paper_pdf": "papers/ns0KIpfQVy.pdf", "paper_pdf_sha256": "6e35b32e79ea35967450001bde8422d29397d27c233ce0274560840a0ec11e23", "paper_pdf_bytes": 464801, "paper_pdf_source": "openreview", "code_url": "https://github.com/Dzhambo/MBD", "code_repository": "Dzhambo/MBD", "code_commit": "0f8a41b2eca7945656b51dfb447b0d08b624ac77", "code_archive": "repos/ns0KIpfQVy.zip", "code_archive_sha256": "1f4c2672cdef3dc904541aabd6c5201a037e3b4267a09ba8ff7fb56ca221ba92", "code_archive_bytes": 121140, "code_file_count": 39, "code_extensions": {".sh": 23, ".py": 9, ".ipynb": 7}, "github_disk_usage_kb": 177, "github_languages": {"Jupyter Notebook": 138212, "Python": 38762, "Shell": 19999}, "github_archived": false, "github_pushed_at": "2025-03-03T14:38:44Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/multimodal-banking-dataset-understanding"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "r0BcyqWAcj", "year": 2024, "status": "rejected", "title": "Loci-Segmented: Improving Scene Segmentation Learning", "authors": ["Manuel Traub", "Frederic Becker", "Adrian Sauter", "Sebastian Otte", "Martin V. Butz"], "authorids": ["~Manuel_Traub1", "~Frederic_Becker1", "~Adrian_Sauter1", "~Sebastian_Otte1", "~Martin_V._Butz2"], "authors_source": "OpenReview API", "abstract": "Slot-oriented processing approaches for compositional scene representation have recently undergone a tremendous development. We present Loci-Segmented (Loci-s), an advanced scene segmentation neural network that extends the slot-based location and identity tracking architecture Loci (Traub et al., ICLR 2023). The main advancements are (i) the addition of a pre-trained dynamic background module; (ii) a hyper-convolution encoder module, which enables object-focused bottom-up processing; and (iii) a cascaded decoder module, which successively generates object masks, masked depth maps, and masked, depth-map-informed RGB reconstructions. The background module features the learning of both a foreground identifying module and a background re-generator. We further improve performance via (a) the integration of depth information as well as improved slot assignments via (b) slot-location-entity regularization and (b) a prior segmentation network. Even without these latter improvements, the results reveal superior segmentation performance in the MOVi datasets and in another established dataset collection. With all improvements, Loci-s achieves a 32% better intersection over union (IoU) score in MOVi-E than the previous best. We furthermore show that Loci-s generates well-interpretable latent representations. We believe that these representations may serve as a foundation-model-like interpretable basis for solving downstream tasks, such as grounding language and context- and\ngoal-conditioned event processing.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=r0BcyqWAcj", "arxiv_id": "2310.10410", "paper_pdf": "papers/r0BcyqWAcj.pdf", "paper_pdf_sha256": "2011ce3fb1fd19cf1e4aebdd4d1f773e9199de4d7ccc7d6352c59b9a5ecdd3ea", "paper_pdf_bytes": 2157267, "paper_pdf_source": "openreview", "code_url": "https://github.com/CognitiveModeling/Loci-Segmented", "code_repository": "CognitiveModeling/Loci-Segmented", "code_commit": "b4abc3a83a2de825da6bde33cb684b7d9daceaa2", "code_archive": "repos/r0BcyqWAcj.zip", "code_archive_sha256": "677753b5522830bcb4ae079f0c01af69c860076743add987a16ee434da3be6a2", "code_archive_bytes": 161551, "code_file_count": 52, "code_extensions": {".py": 49, ".sh": 3}, "github_disk_usage_kb": 144, "github_languages": {"Python": 452780, "Shell": 34659}, "github_archived": false, "github_pushed_at": "2025-05-23T07:38:45Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/loci-segmented-improving-scene-segmentation"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ToYi8C6fetv", "year": 2023, "status": "rejected", "title": "Personalized Subgraph Federated Learning", "authors": ["Jinheon Baek", "Wonyong Jeong", "Jiongdao Jin", "Jaehong Yoon", "Sung Ju Hwang"], "authorids": ["~Jinheon_Baek1", "~Wonyong_Jeong1", "~Jiongdao_Jin1", "~Jaehong_Yoon1", "~Sung_Ju_Hwang1"], "authors_source": "OpenReview API", "abstract": "In real-world scenarios, subgraphs of a larger global graph may be distributed across multiple devices or institutions, and only locally accessible due to privacy restrictions, although there may be links between them. Recently proposed subgraph Federated Learning (FL) methods deal with those missing links across private local subgraphs while distributively training Graph Neural Networks (GNNs) on them. However, they have overlooked the inevitable heterogeneity among subgraphs, caused by subgraphs comprising different communities of a global graph, therefore, consequently collapsing the incompatible knowledge from local GNN models trained on heterogeneous graph distributions. To overcome such a limitation, we introduce a new subgraph FL problem, personalized subgraph FL, which focuses on the joint improvement of the interrelated local GNN models rather than learning a single global GNN model, and propose a novel framework, FEDerated Personalized sUBgraph learning (FED-PUB), to tackle it. A crucial challenge in personalized subgraph FL is that the server does not know which subgraph each client has. FED-PUB thus utilizes functional embeddings of the local GNNs using random graphs as inputs to compute similarities between them, and use them to perform weighted averaging for server-side aggregation. Further, it learns a personalized sparse mask at each client to select and update only the subgraph-relevant subset of the aggregated parameters. We validate FED-PUB for its subgraph FL performance on six datasets, considering both non-overlapping and overlapping subgraphs, on which ours largely outperforms relevant baselines.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=ToYi8C6fetv", "arxiv_id": "2206.10206", "paper_pdf": "papers/ToYi8C6fetv.pdf", "paper_pdf_sha256": "62be180d31ff55ad879003a3ad720e6c6649f5993e63b93423a1d0e433af7c6f", "paper_pdf_bytes": 1086170, "paper_pdf_source": "openreview", "code_url": "https://github.com/JinheonBaek/FED-PUB", "code_repository": "JinheonBaek/FED-PUB", "code_commit": "302980ded3bb2b719fd0864e29607031372a3e6d", "code_archive": "repos/ToYi8C6fetv.zip", "code_archive_sha256": "857ae2a25785a72970fe9b20424ea53217ec3d2249d808deb3108d55f11fd6ae", "code_archive_bytes": 191607, "code_file_count": 21, "code_extensions": {".py": 18, ".sh": 2, ".ipynb": 1}, "github_disk_usage_kb": 265, "github_languages": {"Jupyter Notebook": 234354, "Python": 67609, "Shell": 1432}, "github_archived": false, "github_pushed_at": "2023-07-02T07:20:18Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/personalized-subgraph-federated-learning"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "RVdN1-eDZ1b", "year": 2022, "status": "rejected", "title": "Plug-In Inversion: Model-Agnostic Inversion for Vision with Data Augmentations", "authors": ["Amin Ghiasi", "Hamid Kazemi", "Steven Reich", "Chen Zhu", "Micah Goldblum", "Tom Goldstein"], "authorids": ["~Amin_Ghiasi1", "~Hamid_Kazemi1", "~Steven_Reich1", "~Chen_Zhu2", "~Micah_Goldblum1", "~Tom_Goldstein1"], "authors_source": "OpenReview API", "abstract": "Existing techniques for model inversion typically rely on hard-to-tune regularizers, such as total variation or feature regularization, which must be individually calibrated for each network in order to produce adequate images. In this work, we introduce Plug-In Inversion, which relies on a simple set of augmentations and does not require excessive hyper-parameter tuning.  Under our proposed augmentation-based scheme, the same set of augmentation hyper-parameters can be used for inverting a wide range of image classification models, regardless of input dimensions or the architecture. We illustrate the practicality of our approach by inverting Vision Transformers (ViTs) and Multi-Layer Perceptrons (MLPs) trained on the ImageNet dataset, tasks which to the best of our knowledge have not been successfully accomplished by any previous works. ", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=RVdN1-eDZ1b", "arxiv_id": "2201.12961", "paper_pdf": "papers/RVdN1-eDZ1b.pdf", "paper_pdf_sha256": "759cb6ece49c24d5b93639674fb8d7ff304f925ba008e9ec70f9470df059c14e", "paper_pdf_bytes": 23229715, "paper_pdf_source": "openreview", "code_url": "https://github.com/youranonymousefriend/plugininversion", "code_repository": "youranonymousefriend/plugininversion", "code_commit": "1141f08f79f713b5ecc2b77fb9814b08fe762c34", "code_archive": "repos/RVdN1-eDZ1b.zip", "code_archive_sha256": "7f1e1118891fb8f7aafa66adb713a5f2d38c057c67edae7b126cc8b90dc14892", "code_archive_bytes": 742652, "code_file_count": 179, "code_extensions": {".py": 113, ".sh": 54, ".js": 12}, "github_disk_usage_kb": 640, "github_languages": {"Python": 139276, "Shell": 136312}, "github_archived": false, "github_pushed_at": "2021-10-23T21:16:44Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/plug-in-inversion-model-agnostic-inversion-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "DE0MSwKv32y", "year": 2021, "status": "rejected", "title": "Trust, but verify: model-based exploration in sparse reward environments", "authors": ["Konrad Czechowski", "Tomasz Odrzygóźdź", "Michał Izworski", "Marek Zbysiński", "Łukasz Kuciński", "Piotr Miłoś"], "authorids": ["~Konrad_Czechowski1", "tomaszo@impan.pl", "m.izworski@student.uw.edu.pl", "marek.zbysinski@gmail.com", "~Łukasz_Kuciński1", "~Piotr_Miłoś1"], "authors_source": "OpenReview API", "abstract": "We propose $\\textit{trust-but-verify}$ (TBV) mechanism, a new method which uses model uncertainty estimates to guide  exploration. The mechanism augments graph search planning algorithms by the capacity to deal with learned model's imperfections. We identify certain type of frequent model errors, which we dub $\\textit{false loops}$, and which are particularly dangerous for graph search algorithms in discrete environments. These errors impose falsely pessimistic expectations and thus hinder exploration. We confirm this experimentally and show that TBV can effectively alleviate them. TBV combined with MCTS or Best First Search forms an effective model-based reinforcement learning solution, which is able to robustly solve sparse reward problems. ", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=DE0MSwKv32y", "arxiv_id": null, "paper_pdf": "papers/DE0MSwKv32y.pdf", "paper_pdf_sha256": "57ea382757f5b3a937230f60da27babb629196363ad8d2b7a1fe1442e8767562", "paper_pdf_bytes": 656461, "paper_pdf_source": "openreview", "code_url": "https://github.com/ComradeMisha/TrustButVerify", "code_repository": "ComradeMisha/TrustButVerify", "code_commit": "c410b44e18157fcf1f7ed82745cd4719e623416a", "code_archive": "repos/DE0MSwKv32y.zip", "code_archive_sha256": "d3e57575504c6e39bc9898937977dc2b74b56305df8eef3f529e1c17384b43f1", "code_archive_bytes": 981650, "code_file_count": 75, "code_extensions": {".py": 74, ".js": 1}, "github_disk_usage_kb": 942, "github_languages": {"Python": 487273, "JavaScript": 13049, "CSS": 1198, "HTML": 778}, "github_archived": false, "github_pushed_at": "2021-02-10T20:35:18Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/trust-but-verify-model-based-exploration-in"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "B1lqDertwr", "year": 2020, "status": "rejected", "title": "Regularization Matters in Policy Optimization", "authors": ["Zhuang Liu", "Xuanlin Li", "Bingyi Kang", "Trevor Darrell"], "authorids": ["zhuangl@berkeley.edu", "xuanlinli17@berkeley.edu", "kang@u.nus.edu", "trevor@eecs.berkeley.edu"], "authors_source": "OpenReview API", "abstract": "Deep Reinforcement Learning (Deep RL) has been receiving increasingly more attention  thanks to its encouraging performance on a variety of control tasks. Yet, conventional regularization techniques in training neural networks (e.g., $L_2$ regularization, dropout) have been largely ignored in RL methods, possibly because agents are typically trained and evaluated in the same environment. In this work, we present the first comprehensive study of regularization techniques with multiple policy optimization algorithms on continuous control tasks. Interestingly, we find conventional regularization techniques on the policy networks can often bring large improvement on the task performance, and the improvement is typically more significant when the task is more difficult. We also compare with the widely used entropy regularization and find $L_2$ regularization is generally better. Our findings are further confirmed to be robust against the choice of training hyperparameters. We also study the effects of regularizing different components and find that only regularizing the policy network is typically enough. We hope our study provides guidance for future practices in regularizing policy optimization algorithms.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=B1lqDertwr", "arxiv_id": "1910.09191", "paper_pdf": "papers/B1lqDertwr.pdf", "paper_pdf_sha256": "90f7bef954753b20d6341470db00a6a47c64e6325bba9c1c2203a7ac53cbb74f", "paper_pdf_bytes": 16829849, "paper_pdf_source": "openreview", "code_url": "https://github.com/xuanlinli17/iclr2021_rlreg", "code_repository": "xuanlinli17/iclr2021_rlreg", "code_commit": "1fbecb9d75b35b89f239e4bf5739c8aba8ce51b2", "code_archive": "repos/B1lqDertwr.zip", "code_archive_sha256": "94acdc54dac9ef2400a095ec2ebf2672cec794f5ac4aac73e0e82c8086176ae7", "code_archive_bytes": 1432869, "code_file_count": 124, "code_extensions": {".py": 124}, "github_disk_usage_kb": 2242, "github_languages": {"Python": 577494}, "github_archived": false, "github_pushed_at": "2021-11-01T06:02:18Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/regularization-matters-in-policy-optimization-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "yJiMANMsz2", "year": 2026, "status": "rejected", "title": "DRESS: Disentangled Representation-based Self-Supervised Meta-Learning for Diverse Tasks", "authors": ["Wei Cui", "Tongzi Wu", "Jesse C. Cresswell", "Yi Sui", "Keyvan Golestan"], "authorids": ["~Wei_Cui4", "~Tongzi_Wu1", "~Jesse_C._Cresswell1", "~Yi_Sui1", "~Keyvan_Golestan1"], "authors_source": "OpenReview API", "abstract": "Meta-learning represents a strong class of approaches for solving few-shot learning tasks. Nonetheless, recent research suggests that simply pre-training a generic encoder can potentially surpass meta-learning algorithms. In this paper, we hypothesize that the reason meta-learning fails to stand out in popular few-shot learning benchmarks is the lack of diversity among the few-shot learning tasks. We propose DRESS, a task-agnostic Disentangled REpresentation-based Self-Supervised meta-learning approach that enables fast model adaptation on highly diversified few-shot learning tasks. Specifically, DRESS utilizes disentangled representation learning to create self-supervised tasks that can fuel the meta-training process. We validate the effectiveness of DRESS through experiments on datasets with multiple factors of variation and varying complexity. The results suggest that DRESS is able to outperform competing methods on the majority of the datasets and task setups. Through this paper, we advocate for a re-examination of how task adaptation studies are conducted, and aim to reignite interest in the potential of meta-learning for solving few-shot learning tasks via disentangled representations.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=yJiMANMsz2", "arxiv_id": "2503.09679", "paper_pdf": "papers/yJiMANMsz2.pdf", "paper_pdf_sha256": "5e74067d12025fcb090e97c40e8bc03d4b4dbe8d19bc5777891b4ed77d01779a", "paper_pdf_bytes": 5112783, "paper_pdf_source": "openreview", "code_url": "https://github.com/layer6ai-labs/DRESS", "code_repository": "layer6ai-labs/DRESS", "code_commit": "692d3d6231ecc38e02e4751c55cf6909c67aa119", "code_archive": "repos/yJiMANMsz2.zip", "code_archive_sha256": "06521407470ff41a2736a08ba80a1a0fae5d929dcb77f6e23bb9108698b96799", "code_archive_bytes": 2646029, "code_file_count": 47, "code_extensions": {".py": 41, ".sh": 6}, "github_disk_usage_kb": 2913, "github_languages": {"Python": 258853, "Shell": 76821}, "github_archived": false, "github_pushed_at": "2026-06-17T01:59:36Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/dress-disentangled-representation-based-self"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "kMT8ujhYbA", "year": 2025, "status": "rejected", "title": "Catastrophic Cyber Capabilities Benchmark (3CB): Robustly Evaluating LLM Agent Cyber Offense Capabilities", "authors": ["Andrey Anurin", "Jonathan Ng", "Kibo Schaffer", "Ziyue Wang", "Jason Hoelscher-Obermaier", "Esben Kran"], "authorids": ["~Andrey_Anurin1", "~Jonathan_Ng1", "~Kibo_Schaffer1", "~Ziyue_Wang6", "~Jason_Hoelscher-Obermaier1", "~Esben_Kran1"], "authors_source": "OpenReview API", "abstract": "LLM agents have the potential to revolutionize defensive cyber operations, but their offensive capabilities are not yet fully understood. To prepare for emerging threats, model developers and governments are evaluating the cyber capabilities of foundation models. However, these assessments often lack transparency and a comprehensive focus on offensive capabilities. In response, we introduce the Catastrophic Cyber Capabilities Benchmark (3CB), a novel framework designed to rigorously assess the real-world offensive capabilities of LLM agents. Our evaluation of modern LLMs on 3CB reveals that frontier models, such as GPT-4o and Claude 3.5 Sonnet, can perform offensive tasks such as reconnaissance and exploitation across domains ranging from binary analysis to web technologies. Conversely, smaller open-source models exhibit limited offensive capabilities. Our software solution and the corresponding benchmark provides a critical tool to reduce the gap between rapidly improving capabilities and robustness of cyber offense evaluations, aiding in the safer deployment and regulation of these powerful technologies.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=kMT8ujhYbA", "arxiv_id": "2410.09114", "paper_pdf": "papers/kMT8ujhYbA.pdf", "paper_pdf_sha256": "8a7756a2dcef05742d474779cfd0d052845eaf2907b5888c406391a9ef42b4ac", "paper_pdf_bytes": 499013, "paper_pdf_source": "openreview", "code_url": "https://github.com/apartresearch/3cb", "code_repository": "apartresearch/3cb", "code_commit": "d87effdeaceeac153517ede88a5b57b885c7f5cc", "code_archive": "repos/kMT8ujhYbA.zip", "code_archive_sha256": "c5161fc62328fa4062841ff26e10aa81dd1725065ec7f10f89622e4dfdd951de", "code_archive_bytes": 191715, "code_file_count": 12, "code_extensions": {".py": 11, ".ipynb": 1}, "github_disk_usage_kb": 194, "github_languages": {"Python": 50123, "Jupyter Notebook": 17921, "Mako": 510}, "github_archived": false, "github_pushed_at": "2024-10-30T12:42:57Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/catastrophic-cyber-capabilities-benchmark-3cb"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "piWvNRR0Ym", "year": 2024, "status": "rejected", "title": "Towards Minimal Targeted Updates of Language Models with Targeted Negative Training", "authors": ["Lily H Zhang", "Rajesh Ranganath", "Arya Tafvizi"], "authorids": ["~Lily_H_Zhang1", "~Rajesh_Ranganath2", "~Arya_Tafvizi1"], "authors_source": "OpenReview API", "abstract": "Generative models of language exhibit impressive capabilities but still place non-negligible probability mass over undesirable outputs. In this work, we address the task of updating a model to avoid unwanted outputs while minimally changing model behavior otherwise, a challenge we refer to as a minimal targeted update. We first formalize the notion of a minimal targeted update and propose a method to achieve such updates using negative examples from a model's generations. Our proposed Targeted Negative Training (TNT) results in updates that keep the new distribution close to the original, unlike existing losses for negative signal which push down probability but do not control what the updated distribution will be. In experiments, we demonstrate that TNT yields a better trade-off between reducing unwanted behavior and preserving model generation behavior than baselines, paving the way towards a modeling paradigm based on iterative training updates that constrain models from generating undesirable outputs while preserving their impressive capabilities.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=piWvNRR0Ym", "arxiv_id": "2406.13660", "paper_pdf": "papers/piWvNRR0Ym.pdf", "paper_pdf_sha256": "2e7fad090acf793c5aa2befa33e5206140e352f4a9892e572350e491f1580b2b", "paper_pdf_bytes": 779365, "paper_pdf_source": "openreview", "code_url": "https://github.com/google/t5patches", "code_repository": "google/t5patches", "code_commit": "09285b378b7e36cd428738b3e7c5614cd3dde330", "code_archive": "repos/piWvNRR0Ym.zip", "code_archive_sha256": "3789371e1e12e5285ff1a8a2854fc26153d6d05c8a80fffab6e4a7fcfa4a093e", "code_archive_bytes": 57367, "code_file_count": 11, "code_extensions": {".py": 11}, "github_disk_usage_kb": 144, "github_languages": {"Python": 231087}, "github_archived": true, "github_pushed_at": "2024-05-31T20:55:18Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/towards-minimal-targeted-updates-of-language"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "cddqs4kvC20", "year": 2023, "status": "rejected", "title": "Deep Transformer Q-Networks for Partially Observable Reinforcement Learning", "authors": ["Kevin Esslinger", "Robert Platt", "Christopher Amato"], "authorids": ["~Kevin_Esslinger1", "~Robert_Platt2", "~Christopher_Amato1"], "authors_source": "OpenReview API", "abstract": "Real-world reinforcement learning tasks often involve some form of partial observability where the observations only give a partial or noisy view of the true state of the world. Such tasks typically require some form of memory, where the agent has access to multiple past observations, in order to perform well. One popular way to incorporate memory is by using a recurrent neural network to access the agent's history. However, recurrent neural networks in reinforcement learning are often fragile and difficult to train, susceptible to catastrophic forgetting and sometimes fail completely as a result. In this work, we propose Deep Transformer Q-Networks (DTQN), a novel architecture utilizing transformers and self-attention to encode an agent's history. DTQN is designed modularly, and we compare results against several modifications to our base model. Our experiments demonstrate the transformer can solve partially observable tasks faster and more stably than previous recurrent approaches.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=cddqs4kvC20", "arxiv_id": "2206.01078", "paper_pdf": "papers/cddqs4kvC20.pdf", "paper_pdf_sha256": "58fb3af6988cce055d6661a74a557f537555bdcfb1a8d9274ff314fc9a930df2", "paper_pdf_bytes": 1434690, "paper_pdf_source": "openreview", "code_url": "https://github.com/kevslinger/DTQN", "code_repository": "kevslinger/DTQN", "code_commit": "9f8d72bff5f0d607691989439ae93b9393b68a63", "code_archive": "repos/cddqs4kvC20.zip", "code_archive_sha256": "13776ed625ae110dcf4b68f175d2acf01b699977dbb022e37da4ba0533369dce", "code_archive_bytes": 52707, "code_file_count": 27, "code_extensions": {".py": 27}, "github_disk_usage_kb": 276, "github_languages": {"Python": 131947}, "github_archived": false, "github_pushed_at": "2024-07-07T08:44:37Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/deep-transformer-q-networks-for-partially"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "7pZiaojaVGU", "year": 2022, "status": "rejected", "title": "An Equivalence Between Data Poisoning and Byzantine Gradient Attacks", "authors": ["Sadegh Farhadkhani", "Rachid Guerraoui", "Lê-Nguyên Hoang", "Oscar Villemaud"], "authorids": ["~Sadegh_Farhadkhani1", "~Rachid_Guerraoui1", "~Lê-Nguyên_Hoang1", "~Oscar_Villemaud1"], "authors_source": "OpenReview API", "abstract": "To address the resilience of distributed learning, the ``Byzantine\" literature considers a strong threat model where workers can report arbitrary gradients to the parameter server. While this model helped generate several fundamental results, it has however sometimes been considered unrealistic, when the workers are mostly trustworthy machines. In this paper, we show a surprising equivalence between this model and data poisoning, a threat considered much more realistic. More specifically, we prove that any gradient attack can be reduced to data poisoning in a personalized federated learning system that provides PAC guarantees (which we show are both desirable and realistic in various personalized federated learning contexts such as linear regression and classification). Maybe most importantly, we derive a simple and practical attack that may be constructed against classical personalized federated learning models, and we show both theoretically and empirically the effectiveness of this attack.", "decision": "Reject", "meta_review": null, "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=7pZiaojaVGU", "arxiv_id": "2202.08578", "paper_pdf": "papers/7pZiaojaVGU.pdf", "paper_pdf_sha256": "056fd0b7945253687a4b3140cd1abcaafc9aaed9d7ab872080cebe9637c8d4b8", "paper_pdf_bytes": 1479640, "paper_pdf_source": "openreview", "code_url": "https://github.com/LPD-EPFL/Attack_Equivalence", "code_repository": "LPD-EPFL/Attack_Equivalence", "code_commit": "953769950fc9a25d212ecae3df93b735d61658af", "code_archive": "repos/7pZiaojaVGU.zip", "code_archive_sha256": "e965d8e050fd63762a33d2f397ab7ccaa8770c083d1a91afdad7947f3384f035", "code_archive_bytes": 389595, "code_file_count": 6, "code_extensions": {".ipynb": 6}, "github_disk_usage_kb": 712, "github_languages": {"Jupyter Notebook": 969609}, "github_archived": false, "github_pushed_at": "2022-06-17T11:03:34Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/an-equivalence-between-data-poisoning-and-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "JeweO9-QqV-", "year": 2021, "status": "rejected", "title": "SoGCN: Second-Order Graph Convolutional Networks", "authors": ["Peihao Wang", "Yuehao Wang", "Hua Lin", "Jianbo Shi"], "authorids": ["~Peihao_Wang1", "~Yuehao_Wang1", "~Hua_Lin1", "~Jianbo_Shi1"], "authors_source": "OpenReview API", "abstract": "We introduce a second-order graph convolution (SoGC), a maximally localized kernel, that can express a polynomial spectral filter with arbitrary coefficients.  We contrast our SoGC with vanilla GCN, first-order (one-hop) aggregation, and higher-order (multi-hop) aggregation by analyzing graph convolutional layers via generalized filter space.  We argue that SoGC is a simple design capable of forming the basic building block of graph convolution, playing the same role as $3 \\times 3$ kernels in CNNs.   We build purely topological Second-Order Graph Convolutional Networks (SoGCN) and demonstrate that SoGCN consistently achieves state-of-the-art performance on the latest benchmark. Moreover, we introduce the Gated Recurrent Unit (GRU) to spectral GCNs. This explorative attempt further improves our experimental results.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=JeweO9-QqV-", "arxiv_id": "2110.07141", "paper_pdf": "papers/JeweO9-QqV-.pdf", "paper_pdf_sha256": "0c885175c58932e549cf8507cf33d498ef9d5fc110a328c69d561a4f0e54c5ca", "paper_pdf_bytes": 836166, "paper_pdf_source": "openreview", "code_url": "https://github.com/yuehaowang/SoGCN", "code_repository": "yuehaowang/SoGCN", "code_commit": "bd65b2d8667791b79d6174a1dd2ac13b7bd50db5", "code_archive": "repos/JeweO9-QqV-.zip", "code_archive_sha256": "a2d39793aa704a24926dfa0d53afe29ea46958d99f611f78d32c705025e12b77", "code_archive_bytes": 892049, "code_file_count": 69, "code_extensions": {".py": 50, ".sh": 11, ".ipynb": 8}, "github_disk_usage_kb": 945, "github_languages": {"Jupyter Notebook": 1529440, "Python": 253660, "Shell": 10230}, "github_archived": false, "github_pushed_at": "2021-12-18T06:36:29Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/sogcn-second-order-graph-convolutional-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "BJe-unNYPr", "year": 2020, "status": "rejected", "title": "Accelerated Information Gradient flow", "authors": ["Yifei Wang", "Wuchen Li"], "authorids": ["zackwang24@pku.edu.cn", "wcli@math.ucla.edu"], "authors_source": "OpenReview API", "abstract": "We present a systematic framework for the Nesterov's accelerated gradient flows in the spaces of probabilities embedded with information metrics. Here two metrics are considered, including both the Fisher-Rao metric and the Wasserstein-$2$ metric. For the Wasserstein-$2$ metric case, we prove the convergence properties of the accelerated gradient flows, and introduce their formulations in Gaussian families. Furthermore, we propose a practical discrete-time algorithm in particle implementations with an adaptive restart technique.  We formulate a novel bandwidth selection method, which learns the Wasserstein-$2$ gradient direction from Brownian-motion samples. Experimental results including Bayesian inference show the strength of the current method compared with the state-of-the-art.", "decision": "Reject", "meta_review": null, "num_reviews": 2, "openreview_url": "https://openreview.net/forum?id=BJe-unNYPr", "arxiv_id": "1909.02102", "paper_pdf": "papers/BJe-unNYPr.pdf", "paper_pdf_sha256": "26deafeaf2524b327f94c0fe9c2cb46752075d0bafe28e1a63c9aeaf9efd9a6c", "paper_pdf_bytes": 1838086, "paper_pdf_source": "openreview", "code_url": "https://github.com/YiifeiWang/Accelerated-Information-Gradient-flow", "code_repository": "YiifeiWang/Accelerated-Information-Gradient-flow", "code_commit": "fb5e485066fd2658cb93787da44203b18df2f193", "code_archive": "repos/BJe-unNYPr.zip", "code_archive_sha256": "20a32c2d56865f4bc1c2977e026e81eaca3def10de62926f34e6f3d827026ea6", "code_archive_bytes": 330285, "code_file_count": 42, "code_extensions": {".m": 23, ".py": 9, ".sh": 7, ".ipynb": 3}, "github_disk_usage_kb": 2333, "github_languages": {"MATLAB": 67109, "Python": 60402, "Jupyter Notebook": 6000, "Shell": 3447, "M": 2051}, "github_archived": false, "github_pushed_at": "2020-06-02T02:51:19Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/accelerated-information-gradient-flow"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "GsLvRPAfoU", "year": 2026, "status": "rejected", "title": "Improving Fine-Grained Control via Aggregation of Multiple Diffusion Models", "authors": ["Conghan Yue", "Zhengwei Peng", "Shiyan Du", "Zhi Ji", "Chuangjian Cai", "Le Wan", "Dongyu Zhang"], "authorids": ["~Conghan_Yue2", "~Zhengwei_Peng1", "~Shiyan_Du1", "~Zhi_Ji1", "~Chuangjian_Cai2", "~Le_Wan1", "~Dongyu_Zhang1"], "authors_source": "OpenReview API", "abstract": "While many diffusion models perform well when controlling particular aspects such as style, character, and interaction, they struggle with fine-grained control due to dataset limitations and intricate model architecture design. This paper introduces a novel training-free algorithm for fine-grained generation, called Aggregation of Multiple Diffusion Models (AMDM). The algorithm integrates features in the latent data space from multiple diffusion models within the same ecosystem into a specified model, thereby activating particular features and enabling fine-grained control. Experimental results demonstrate that AMDM significantly improves fine-grained control without training, validating its effectiveness. Additionally, it reveals that diffusion models initially focus on features such as position, attributes, and style, with later stages improving generation quality and consistency. AMDM offers a new perspective for tackling the challenges of fine-grained conditional generation in diffusion models. Specifically, it allows us to fully utilize existing or develop new conditional diffusion models that control specific aspects, and then aggregate them using the AMDM algorithm. This eliminates the need for constructing complex datasets, designing intricate model architectures, and incurring high training costs.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=GsLvRPAfoU", "arxiv_id": "2410.01262", "paper_pdf": "papers/GsLvRPAfoU.pdf", "paper_pdf_sha256": "7c4723c82fd9ddd32cda5b405e4f05691cf5ffb1de87b51b9a113720d5dfdf4a", "paper_pdf_bytes": 11034682, "paper_pdf_source": "openreview", "code_url": "https://github.com/Hammour-steak/AMDM", "code_repository": "Hammour-steak/AMDM", "code_commit": "076ec6a6df136f6917acb3e1000c1a23f95b7581", "code_archive": "repos/GsLvRPAfoU.zip", "code_archive_sha256": "db16ff3ef751c050d5f964b5971619e9a999a3d34556b8f79b3aa4b30a3462c5", "code_archive_bytes": 3036572, "code_file_count": 68, "code_extensions": {".py": 58, ".js": 9, ".ipynb": 1}, "github_disk_usage_kb": 2924, "github_languages": {"Python": 594009, "JavaScript": 47637, "HTML": 29571, "CSS": 11526, "Jupyter Notebook": 8885}, "github_archived": false, "github_pushed_at": "2024-12-17T06:37:21Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/aggregation-of-multi-diffusion-models-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "kXYuzGcfTT", "year": 2026, "status": "rejected", "title": "Overclocking LLM Reasoning: Monitoring and Controlling Thinking Path Lengths in LLMs", "authors": ["Roy Eisenstadt", "Itamar Zimerman", "Lior Wolf"], "authorids": ["~Roy_Eisenstadt1", "~Itamar_Zimerman1", "~Lior_Wolf1"], "authors_source": "OpenReview API", "abstract": "Recently, techniques such as explicit structured reasoning have demonstrated strong test-time scaling behavior by enforcing a separation between the model’s internal \"thinking\" process and the final response. A key factor influencing answer quality in this setting is the length of the thinking stage. When the reasoning is too short, the model may fail to capture the complexity of the task. Conversely, when it is too long, the model may overthink, leading to unnecessary computation and degraded performance. This paper explores and exploits the underlying mechanisms by which LLMs understand and regulate the length of their reasoning during explicit thought processes. First, we show that LLMs encode their progress through the reasoning process and introduce an interactive progress bar visualization, which is then used to reveal insights on the model's planning dynamics. Second, we manipulate the internal progress encoding during inference to reduce unnecessary steps and generate a more concise and decisive chain of thoughts. Our empirical results demonstrate that this ``overclocking'' method mitigates overthinking, improves answer accuracy, and reduces inference latency. Our code is attached as supplementary material.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=kXYuzGcfTT", "arxiv_id": "2506.07240", "paper_pdf": "papers/kXYuzGcfTT.pdf", "paper_pdf_sha256": "6da118bc0f3d1041a132070b12dff57f2fc6e740dec277d4a87a543f0596a3c2", "paper_pdf_bytes": 5687995, "paper_pdf_source": "openreview", "code_url": "https://github.com/royeisen/reasoning_loading_bar", "code_repository": "royeisen/reasoning_loading_bar", "code_commit": "98922eeb34eadcfdbfdeb6727c940e4b87b55c3c", "code_archive": "repos/kXYuzGcfTT.zip", "code_archive_sha256": "4fa51e51dc76815ad39cfab26aa655bb2e34061199d73bb371bcab5d95fc5190", "code_archive_bytes": 2375970, "code_file_count": 5, "code_extensions": {".py": 5}, "github_disk_usage_kb": 2318, "github_languages": {"Python": 50795}, "github_archived": false, "github_pushed_at": "2025-07-07T12:47:27Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/overclocking-llm-reasoning-monitoring-and"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Aemqy6Hjdj", "year": 2024, "status": "rejected", "title": "Enhancing Compositional Generalization via Compositional Feature Alignment", "authors": ["Haoxiang Wang", "Haozhe Si", "Huajie Shao", "Han Zhao"], "authorids": ["~Haoxiang_Wang1", "~Haozhe_Si1", "~Huajie_Shao1", "~Han_Zhao1"], "authors_source": "OpenReview API", "abstract": "Real-world applications of machine learning (ML) models often confront data distribution shifts, wherein discrepancies exist between the training and test data distributions. In the common multi-domain multi-class setup, as the number of classes and domains scales up, it becomes infeasible to gather training data for every domain-class combination. This challenge naturally leads the quest for models with Compositional Generalization (CG) ability, where models can generalize to unseen domain-class combinations. To delve into the CG challenge, we develop CG-Bench, a suite of CG benchmarks derived from existing real-world image datasets, and observe that the prevalent pretraining-finetuning paradigm on foundational models, such as CLIP and DINOv2, struggles with the challenge. To address this challenge, we propose Compositional Feature Alignment (CFA), a simple two-stage finetuning technique that i) learns two orthogonal linear heads on a pretrained encoder with respect to class and domain labels, and ii) fine-tunes the encoder with the newly learned head frozen. We theoretically and empirically justify that CFA encourages compositional feature learning of pretrained models. We further conduct extensive experiments on CG-Bench for CLIP and DINOv2, two powerful pretrained vision foundation models. Experiment results show that CFA outperforms common finetuning techniques in compositional generalization, corroborating CFA's efficacy in compositional feature learning.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=Aemqy6Hjdj", "arxiv_id": "2402.02851", "paper_pdf": "papers/Aemqy6Hjdj.pdf", "paper_pdf_sha256": "5f59cf258c7160728f91144659ed1274cd047e8a61a75d8a88fd1dde86f70603", "paper_pdf_bytes": 5692695, "paper_pdf_source": "openreview", "code_url": "https://github.com/Haoxiang-Wang/Compositional-Feature-Alignment", "code_repository": "Haoxiang-Wang/Compositional-Feature-Alignment", "code_commit": "70a12c1d5014fffd8b277db3cfaa023c75af28dc", "code_archive": "repos/Aemqy6Hjdj.zip", "code_archive_sha256": "18e106c3d0d5108d5abae0bd24eba827a9d0da8df3ff45f789ea6ed5a38c1d6d", "code_archive_bytes": 175358, "code_file_count": 78, "code_extensions": {".py": 78}, "github_disk_usage_kb": 147, "github_languages": {"Python": 413530}, "github_archived": false, "github_pushed_at": "2024-02-04T05:30:20Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/enhancing-compositional-generalization-via"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "rjYUBo_uWEs", "year": 2023, "status": "rejected", "title": "Manifold Characteristics That Predict Downstream Task Performance", "authors": ["Ruan Henry Van der Merwe", "Gregory Newman", "Etienne Barnard"], "authorids": ["~Ruan_Henry_Van_der_Merwe1", "greg.newman@bytefuse.ai", "~Etienne_Barnard1"], "authors_source": "OpenReview API", "abstract": "Pretraining methods are typically compared by evaluating the accuracy of linear classifiers, transfer learning performance, or visually inspecting the representation manifold's (RM) lower-dimensional projections. We show that the differences between methods can be understood more clearly by investigating the RM directly, which allows for a more detailed comparison. To this end, we propose a framework and new metric to measure and compare different RMs. We also investigate and report on the RM characteristics for various pretraining methods. These characteristics are measured by applying sequentially larger local alterations to the input data, using white noise injections and Projected Gradient Descent (PGD) adversarial attacks, and then tracking each datapoint. We calculate the total distance moved for each datapoint and the relative change in distance between successive alterations. We show that self-supervised methods learn an RM where alterations lead to large but constant size changes, indicating a smoother RM than fully supervised methods. We then combine these measurements into one metric, the Representation Manifold Quality Metric (RMQM), where larger values indicate larger and less variable step sizes, and show that RMQM correlates positively with performance on downstream tasks.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=rjYUBo_uWEs", "arxiv_id": "2205.07477", "paper_pdf": "papers/rjYUBo_uWEs.pdf", "paper_pdf_sha256": "bdea4d6ee46e3d256c7607daf565db0545bf95b017325621c053c3c3eb47a1fa", "paper_pdf_bytes": 1030493, "paper_pdf_source": "openreview", "code_url": "https://github.com/ByteFuse/representation-manifold-quality-metric", "code_repository": "ByteFuse/representation-manifold-quality-metric", "code_commit": "92a3e925b1456848659505728b704844ba065801", "code_archive": "repos/rjYUBo_uWEs.zip", "code_archive_sha256": "651828f4ac09a876ae0cfa9b976ce6c8d6d4cb94d7f53b6be128b5081f774486", "code_archive_bytes": 245046, "code_file_count": 21, "code_extensions": {".py": 19, ".ipynb": 2}, "github_disk_usage_kb": 290, "github_languages": {"Jupyter Notebook": 385298, "Python": 138912}, "github_archived": false, "github_pushed_at": "2022-07-28T11:03:23Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/manifold-characteristics-that-predict"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Rivn22SJjg9", "year": 2022, "status": "rejected", "title": "Contrastive Embeddings for Neural Architectures", "authors": ["Daniel Hesslow", "Iacopo Poli"], "authorids": ["~Daniel_Hesslow1", "~Iacopo_Poli1"], "authors_source": "OpenReview API", "abstract": "The performance of algorithms for neural architecture search strongly depends on the parametrization of the search space. We use contrastive learning to identify networks across different initializations based on their data Jacobians and their number of parameters, and automatically produce the first architecture embeddings independent from the parametrization of the search space. Using our contrastive embeddings, we show that traditional black-box optimization algorithms, without modification, can reach state-of-the-art performance in Neural Architecture Search. As our method provides a unified embedding space, we successfully perform transfer learning between search spaces. Finally, we show the evolution of embeddings during training, motivating future studies into using embeddings at different training stages to gain a deeper understanding of the networks in a search space.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=Rivn22SJjg9", "arxiv_id": "2102.04208", "paper_pdf": "papers/Rivn22SJjg9.pdf", "paper_pdf_sha256": "36c094f7d14283a8afa57a775ab4bf0079814bdd91a67031172354e60a3ab5ee", "paper_pdf_bytes": 5772111, "paper_pdf_source": "openreview", "code_url": "https://github.com/lightonai/contrastive-embeddings-for-neural-architectures", "code_repository": "lightonai/contrastive-embeddings-for-neural-architectures", "code_commit": "8606a4a4d7ea9ed9c01964660e8b2afb1e003480", "code_archive": "repos/Rivn22SJjg9.zip", "code_archive_sha256": "18da13d12b2033eb40a0d09c44d849df76e823b6c9007c65ecd184c93ad1f52a", "code_archive_bytes": 807640, "code_file_count": 9, "code_extensions": {".py": 9}, "github_disk_usage_kb": 784, "github_languages": {"Python": 59168}, "github_archived": true, "github_pushed_at": "2021-03-17T10:46:23Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/contrastive-embeddings-for-neural"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "K5a_QFEUzA1", "year": 2021, "status": "rejected", "title": "Cross-model Back-translated Distillation for Unsupervised Machine Translation", "authors": ["Phi Xuan Nguyen", "Shafiq Joty", "Kui Wu", "AiTi Aw"], "authorids": ["~Phi_Xuan_Nguyen1", "~Shafiq_Joty1", "wuk@i2r.a-star.edu.sg", "~AiTi_Aw1"], "authors_source": "OpenReview API", "abstract": "Recent unsupervised machine translation (UMT) systems usually employ three main principles: initialization, language modeling and iterative back-translation, though they may apply them differently. Crucially, iterative back-translation and denoising auto-encoding for language modeling provide data diversity to train the UMT systems. However, these diversification processes may have reached their limit. We introduce a novel component to the standard UMT framework called Cross-model Back-translated Distillation (CBD), that is aimed to induce another level of data diversification that existing principles lack. CBD is applicable to all previous UMT approaches. In our experiments, it boosts the performance of the standard UMT methods by 1.5-2.0 BLEU. In particular, in WMT'14 English-French, WMT'16 German-English and English-Romanian, CBD outperforms cross-lingual masked language model (XLM) by 2.3, 2.2 and 1.6 BLEU, respectively. It also yields 1.5-3.3 BLEU improvements in IWSLT English-French and English-German tasks. Through extensive experimental analyses, we show that CBD is effective because it embraces data diversity while other similar variants do not.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=K5a_QFEUzA1", "arxiv_id": "2006.02163", "paper_pdf": "papers/K5a_QFEUzA1.pdf", "paper_pdf_sha256": "e2b1a8ab2032f2460291f063d3fb79b212ba084fe8709afa34aea5f295a664de", "paper_pdf_bytes": 275372, "paper_pdf_source": "openreview", "code_url": "https://github.com/nxphi47/multiagent_crosstranslate", "code_repository": "nxphi47/multiagent_crosstranslate", "code_commit": "041c30d4a67fb1dba309a8cd555bc6d3b40464eb", "code_archive": "repos/K5a_QFEUzA1.zip", "code_archive_sha256": "1c45bfa85ceaf031a9cabacb402d40fa6b7a630318a0b989c8e4aa71fc2054f0", "code_archive_bytes": 968363, "code_file_count": 52, "code_extensions": {".py": 38, ".sh": 14}, "github_disk_usage_kb": 952, "github_languages": {"Python": 694863, "Shell": 57470, "Perl": 5233}, "github_archived": false, "github_pushed_at": "2023-06-16T12:32:41Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/multi-agent-cross-translated-diversification"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "S1l66nNFvB", "year": 2020, "status": "rejected", "title": "Graph Warp Module: an Auxiliary Module for Boosting the Power of Graph Neural Networks in Molecular Graph Analysis", "authors": ["Katsuhiko Ishiguro", "Shin-ichi Maeda", "Masanori Koyama"], "authorids": ["k.ishiguro.jp@ieee.org", "ichi@preferred.jp", "masomatics@preferred.jp"], "authors_source": "OpenReview API", "abstract": "Graph Neural Network (GNN) is a popular architecture for the analysis of chemical molecules, and it has numerous applications in material and medicinal science.\nCurrent lines of GNNs developed for molecular analysis, however, do not fit well on the training set, and their performance does not scale well with the complexity of the network. \nIn this paper, we propose an auxiliary module to be attached to a GNN that can boost the representation power of the model without hindering the original GNN architecture. \nOur auxiliary module can improve the representation power and the generalization ability of a wide variety of GNNs, including those that are used commonly in biochemical applications. ", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=S1l66nNFvB", "arxiv_id": "1902.01020", "paper_pdf": "papers/S1l66nNFvB.pdf", "paper_pdf_sha256": "f54de0d1cac26a7fb23921694a47d9230dde32d25190d40429e5b4314e5a5513", "paper_pdf_bytes": 2138141, "paper_pdf_source": "openreview", "code_url": "https://github.com/chainer/chainer-chemistry", "code_repository": "chainer/chainer-chemistry", "code_commit": "efe323aa21f63a815130d673781e7cca1ccb72d2", "code_archive": "repos/S1l66nNFvB.zip", "code_archive_sha256": "b352b724079084a6a79d90372346354c4fedfc6e29dd02812a65b8993fa4c7a0", "code_archive_bytes": 734901, "code_file_count": 316, "code_extensions": {".py": 296, ".sh": 18, ".ipynb": 2}, "github_disk_usage_kb": 2388, "github_languages": {"Python": 1022006, "Dockerfile": 11627, "Shell": 3086}, "github_archived": false, "github_pushed_at": "2023-04-20T05:13:47Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/graph-warp-module-an-auxiliary-module-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "h1ZEMXxSz1", "year": 2024, "status": "rejected", "title": "Segment Anything Model is a Good Teacher for Local Feature Learning", "authors": ["Jingqian Wu", "Rongtao Xu", "Zach Wood-Doughty", "Changwei Wang"], "authorids": ["~Jingqian_Wu1", "~Rongtao_Xu1", "~Zach_Wood-Doughty1", "~Changwei_Wang2"], "authors_source": "OpenReview API", "abstract": "Local feature detection and description play an important role in many computer vision tasks, which are designed to detect and describe keypoints in \"any scene'' and \"any downstream task''. Data-driven local feature learning methods need to rely on pixel-level correspondence for training, which is challenging to acquire at scale, thus hindering further performance improvement. In this paper, we propose SAMFeat to introduce SAM (segment anything model), a fundamental model trained on 10 million images, as a teacher to guide local feature learning and thus inspire higher performance on limited datasets. First, we construct an auxiliary task of Pixel Semantic Relational Distillation (PSRD) for distilling feature relations with category-agnostic semantic information learned by the SAM encoder into a local feature learning network, hence improving local feature description using semantic discrimination. Second, we develop a technique called Weakly Supervised Contrastive Learning Based on Semantic Grouping (WSC), which utilizes semantic groupings derived from SAM as weakly supervised signals to optimize the metric space of local descriptors. Third, we design an Edge Attention Guidance Module (EAGM) to further improve the accuracy of local feature detection and description by prompting the network to pay more attention to the edge region guided by SAM. SAMFeat's performance in experiments conducted on various tasks such as image matching on HPatches, and long-term, extensive visual localization datasets like Aachen Day-Night showcases its superiority over previous local features. The release code is available at supplementary material.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=h1ZEMXxSz1", "arxiv_id": "2309.16992", "paper_pdf": "papers/h1ZEMXxSz1.pdf", "paper_pdf_sha256": "464da62bd874d158b39e919df74dd9f7a68a1407b9f7fdcf7ac0f997b66e7138", "paper_pdf_bytes": 2676790, "paper_pdf_source": "openreview", "code_url": "https://github.com/vignywang/SAMFeat", "code_repository": "vignywang/SAMFeat", "code_commit": "38be900236c1506db6c0c82018ba637b6b6eaea8", "code_archive": "repos/h1ZEMXxSz1.zip", "code_archive_sha256": "bc355f6f9f6c180945598cd55a7d25db87a222490387e2f9fa552f97760ecb36", "code_archive_bytes": 85886, "code_file_count": 22, "code_extensions": {".py": 21, ".sh": 1}, "github_disk_usage_kb": 141, "github_languages": {"Python": 281271, "Shell": 364}, "github_archived": false, "github_pushed_at": "2025-06-08T15:17:52Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/segment-anything-model-is-a-good-teacher-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "7dmsy2Vd5h", "year": 2025, "status": "rejected", "title": "Comparing and Contrasting Deep Learning Weather Prediction Backbones on Navier-Stokes and Atmospheric Dynamics", "authors": ["Matthias Karlbauer", "Danielle C. Maddix", "Abdul Fatir Ansari", "Boran Han", "Gaurav Gupta", "Bernie Wang", "Andrew Stuart", "Michael W. Mahoney"], "authorids": ["~Matthias_Karlbauer1", "~Danielle_C._Maddix1", "~Abdul_Fatir_Ansari2", "~Boran_Han1", "~Gaurav_Gupta2", "~Bernie_Wang1", "~Andrew_Stuart2", "~Michael_W._Mahoney1"], "authors_source": "OpenReview API", "abstract": "Remarkable progress in the development of Deep Learning Weather Prediction (DLWP) models positions  them  to become competitive with traditional numerical weather prediction (NWP) models. Indeed, a wide number of DLWP architectures---based on various backbones, including U-Net, Transformer, Graph Neural Network (GNN), and Fourier Neural Operator (FNO)---have demonstrated their potential at forecasting atmospheric states. However, due to differences in training protocols, forecast horizons, and data choices, it remains unclear which (if any) of these methods and architectures are most suitable for weather forecasting and for future model development. Here, we step back and provide a detailed empirical analysis, under controlled conditions, comparing and contrasting the most prominent DLWP models, along with their backbones. We accomplish this by predicting  synthetic two-dimensional incompressible Navier-Stokes and real-world global weather dynamics. In terms of accuracy, memory consumption, and runtime, our results illustrate various tradeoffs. For example, on synthetic data, we observe favorable performance of FNO; and on the real-world WeatherBench dataset, our results demonstrate the suitability of ConvLSTM and SwinTransformer for short-to-mid-ranged forecasts. For long-ranged weather rollouts of up to 365 days, we observe superior stability and physical soundness in architectures that formulate a spherical data representation, i.e., GraphCast and Spherical FNO. In addition, we observe that all of these model backbones ``saturate,'' i.e., none of them exhibit so-called neural scaling, which highlights an important direction for future work on these and related models. The code is available at \\url{https://anonymous.4open.science/r/dlwp-benchmark-F88C}.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=7dmsy2Vd5h", "arxiv_id": "2407.14129", "paper_pdf": "papers/7dmsy2Vd5h.pdf", "paper_pdf_sha256": "a6665ffde5a288a16346d80b4488daa7088ed1bbd077abfaeb94ce839f0823f9", "paper_pdf_bytes": 12397055, "paper_pdf_source": "openreview", "code_url": "https://github.com/amazon-science/dlwp-benchmark", "code_repository": "amazon-science/dlwp-benchmark", "code_commit": "b935e10b46a84fa7d012a0afef4ae2b4e32110c8", "code_archive": "repos/7dmsy2Vd5h.zip", "code_archive_sha256": "476ff0c28d25c2840c4efef1d725dcc3ebf4ab4e2025564c7d94437d57cbd9ff", "code_archive_bytes": 336166, "code_file_count": 105, "code_extensions": {".py": 105}, "github_disk_usage_kb": 204, "github_languages": {"Python": 995937}, "github_archived": false, "github_pushed_at": "2024-08-09T09:00:38Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/comparing-and-contrasting-deep-learning"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "kkQSwtx0p3", "year": 2024, "status": "rejected", "title": "Leveraging Task Structures for Improved Identifiability in Neural Network Representations", "authors": ["Wenlin Chen", "Julien Horwood", "Juyeon Heo", "José Miguel Hernández-Lobato"], "authorids": ["~Wenlin_Chen2", "~Julien_Horwood1", "~Juyeon_Heo1", "~José_Miguel_Hernández-Lobato1"], "authors_source": "OpenReview API", "abstract": "This work extends the theory of identifiability in supervised learning by considering the consequences of having access to a distribution of tasks. In such cases, we show that identifiability is achievable even in the case of regression, extending prior work restricted to linear identifiability in the single-task classification case. Furthermore, we show that the existence of a task distribution which defines a conditional prior over latent factors reduces the equivalence class for identifiability to permutations and scaling, a much stronger and more useful result than linear identifiability. When we further assume a causal structure over these tasks, our approach enables simple maximum marginal likelihood optimization together with downstream applicability to causal representation learning. Empirically, we validate that our model outperforms more general unsupervised models in recovering canonical representations for both synthetic and real-world molecular data.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=kkQSwtx0p3", "arxiv_id": "2306.14861", "paper_pdf": "papers/kkQSwtx0p3.pdf", "paper_pdf_sha256": "444affb7351004ef2364172b7983043b3edf9845b8b4f603bb1f1bc84864d128", "paper_pdf_bytes": 528109, "paper_pdf_source": "openreview", "code_url": "https://github.com/jdhorwood/mtlcm", "code_repository": "jdhorwood/mtlcm", "code_commit": "ea559ae662d08605ca65ff3c02b1f52fd1267513", "code_archive": "repos/kkQSwtx0p3.zip", "code_archive_sha256": "996bfec1d15394f6b48e975a55d20de1019f0066f84246977d171cdce8b17035", "code_archive_bytes": 91116, "code_file_count": 16, "code_extensions": {".py": 16}, "github_disk_usage_kb": 147, "github_languages": {"Python": 77092}, "github_archived": false, "github_pushed_at": "2024-08-28T01:30:18Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/leveraging-task-structures-for-improved"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "3TfSOxiRiFH", "year": 2023, "status": "rejected", "title": "On a Built-in Conflict between Deep Learning and Systematic Generalization", "authors": ["Yuanpeng Li"], "authorids": ["~Yuanpeng_Li2"], "authors_source": "OpenReview API", "abstract": "Out-of-distribution or systematic generalization is a desirable property that most deep learning algorithms lack. In this paper, we hypothesize that internal function sharing is one of the reasons to weaken systematic generalization in deep learning for classification tasks. Under equivalent prediction, a model partitions an input space into multiple parts separated by boundaries. The function sharing prefers to reuse boundaries, leading to fewer parts for new outputs, which conflicts with systematic generalization. We show such phenomena in standard deep learning models, such as fully connected, convolutional, residual networks, LSTMs, and (Vision) Transformers. We hope this study provides novel insights and forms a basis for new research directions to improve systematic generalization.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=3TfSOxiRiFH", "arxiv_id": "2208.11633", "paper_pdf": "papers/3TfSOxiRiFH.pdf", "paper_pdf_sha256": "d8d0b4487ad6cd050cf254d96ab38be8e732df42ee8452975efee5ad65d87a38", "paper_pdf_bytes": 1142934, "paper_pdf_source": "openreview", "code_url": "https://github.com/yuanpeng16/BCDLSG", "code_repository": "yuanpeng16/BCDLSG", "code_commit": "08763a62d60e97cf2a765c7b8069f62e785ddda2", "code_archive": "repos/3TfSOxiRiFH.zip", "code_archive_sha256": "06d26134860a069f0fa20e2c4971fae2d6f85805012ad03217fe3e94d34dacfd", "code_archive_bytes": 55675, "code_file_count": 46, "code_extensions": {".sh": 32, ".py": 14}, "github_disk_usage_kb": 305, "github_languages": {"Python": 110383, "Shell": 23618}, "github_archived": false, "github_pushed_at": "2024-11-25T05:19:26Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/on-a-built-in-conflict-between-deep-learning"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "xf0B7-7MRo6", "year": 2022, "status": "rejected", "title": "AIR-Net: Adaptive and Implicit Regularization Neural Network for matrix completion", "authors": ["Zhemin Li", "Hongxia Wang"], "authorids": ["~Zhemin_Li1", "~Hongxia_Wang1"], "authors_source": "OpenReview API", "abstract": "Conventionally, the matrix completion (MC) model aims to recover a matrix from partially observed elements. Accurate recovery necessarily requires a regularization encoding priors of the unknown matrix/signal properly. However, encoding the priors accurately for the complex natural signal is difficult, and even then, the model might not generalize well outside the particular matrix type. This work combines adaptive and implicit low-rank regularization that captures the prior dynamically according to the current recovered matrix. Furthermore, we aim to answer the question: how does adaptive regularization affect implicit regularization? We utilize neural networks to represent Adaptive and Implicit Regularization and named the proposed model \\textit{AIR-Net}. Theoretical analyses show that the adaptive part of the AIR-Net enhances implicit regularization. In addition, the adaptive regularizer vanishes at the end, thus can avoid saturation issues. Numerical experiments for various data demonstrate the effectiveness of AIR-Net, especially when the locations of missing elements are not randomly chosen. With complete flexibility to select neural networks for matrix representation, AIR-Net can be extended to solve more general inverse problems.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=xf0B7-7MRo6", "arxiv_id": "2110.07557", "paper_pdf": "papers/xf0B7-7MRo6.pdf", "paper_pdf_sha256": "e64eb006c1d752e830a60f503fde87f744093117f0e2cddfcd8a4b4984b0f9ab", "paper_pdf_bytes": 3322931, "paper_pdf_source": "openreview", "code_url": "https://github.com/lizhemin15/AIR-Net", "code_repository": "lizhemin15/AIR-Net", "code_commit": "2771c101586e39b0b20748ca0eb0aca35a4e5f47", "code_archive": "repos/xf0B7-7MRo6.zip", "code_archive_sha256": "900f9a4a9cc86230931b1be98cb6746cac3630a8afba0de5d77b5545bf542770", "code_archive_bytes": 1260764, "code_file_count": 18, "code_extensions": {".py": 16, ".ipynb": 2}, "github_disk_usage_kb": 785, "github_languages": {"Jupyter Notebook": 1157062, "Python": 72059}, "github_archived": false, "github_pushed_at": "2023-02-03T03:01:02Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/air-net-adaptive-and-implicit-regularization-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "qoTcTS9-IZ-", "year": 2021, "status": "rejected", "title": "Dynamically Stable Infinite-Width Limits of Neural Classifiers", "authors": ["Eugene Golikov"], "authorids": ["~Eugene_Golikov1"], "authors_source": "OpenReview API", "abstract": "Recent research has been focused on two different approaches to studying neural networks training in the limit of infinite width (1) a mean-field (MF) and (2) a constant neural tangent kernel (NTK) approximations. These two approaches have different scaling of hyperparameters with the width of a network layer and as a result, different infinite-width limit models. Restricting ourselves to single hidden layer nets with zero-mean initialization trained for binary classification with SGD, we propose a general framework to study how the limit behavior of neural models depends on the scaling of hyperparameters with network width. Our framework allows us to derive scaling for existing MF and NTK limits, as well as an uncountable number of other scalings that lead to a dynamically stable limit behavior of corresponding models. However, only a finite number of distinct limit models are induced by these scalings. Each distinct limit model corresponds to a unique combination of such properties as boundedness of logits and tangent kernels at initialization or stationarity of tangent kernels. Existing MF and NTK limit models, as well as one novel limit model, satisfy most of the properties demonstrated by finite-width models. We also propose a novel initialization-corrected mean-field limit that satisfies all properties noted above, and its corresponding model is a simple modification for a finite-width model.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=qoTcTS9-IZ-", "arxiv_id": "2006.06574", "paper_pdf": "papers/qoTcTS9-IZ-.pdf", "paper_pdf_sha256": "11de9c6887984a4b35709cf0d2974cef56e196eb6ddb5c9802e60f44bc247c0f", "paper_pdf_bytes": 1271126, "paper_pdf_source": "openreview", "code_url": "https://github.com/deeppavlov/research", "code_repository": "deeppavlov/research", "code_commit": "31a6c98099ef974d1ec7cb09186ca80dc3a95b34", "code_archive": "repos/qoTcTS9-IZ-.zip", "code_archive_sha256": "4b1952854a8bc02d92c5a0fbeddb97d94b2b259b3a0f701bd67a740a99d32716", "code_archive_bytes": 974047, "code_file_count": 19, "code_extensions": {".py": 16, ".ipynb": 3}, "github_disk_usage_kb": 955, "github_languages": {"Jupyter Notebook": 1309113, "Python": 138067}, "github_archived": false, "github_pushed_at": "2020-10-22T12:12:35Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/dynamically-stable-infinite-width-limits-of"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ByeSYa4KPS", "year": 2020, "status": "rejected", "title": "Sparse Networks from Scratch: Faster Training without Losing Performance", "authors": ["Tim Dettmers", "Luke Zettlemoyer"], "authorids": ["dettmers@cs.washington.edu", "lsz@cs.washington.edu"], "authors_source": "OpenReview API", "abstract": "We demonstrate the possibility of what we call sparse learning: accelerated training of deep neural networks that maintain sparse weights throughout training while achieving dense performance levels. We accomplish this by developing sparse momentum, an algorithm which uses exponentially smoothed gradients (momentum) to identify layers and weights which reduce the error efficiently. Sparse momentum redistributes pruned weights across layers according to the mean momentum magnitude of each layer. Within a layer, sparse momentum grows weights according to the momentum magnitude of zero-valued weights. We demonstrate state-of-the-art sparse performance on MNIST, CIFAR-10, and ImageNet, decreasing the mean error by a relative 8%, 15%, and 6% compared to other sparse algorithms. Furthermore, we show that sparse momentum reliably reproduces dense performance levels while providing up to 5.61x faster training. In our analysis, ablations show that the benefits of momentum redistribution and growth increase with the depth and size of the network. ", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=ByeSYa4KPS", "arxiv_id": "1907.04840", "paper_pdf": "papers/ByeSYa4KPS.pdf", "paper_pdf_sha256": "3f495c6b76cf154ce24b0b04d2bdfb2e78ddea539b5890e5291a4d433c752e4a", "paper_pdf_bytes": 437250, "paper_pdf_source": "openreview", "code_url": "https://github.com/TimDettmers/sparse_learning", "code_repository": "TimDettmers/sparse_learning", "code_commit": "e19ff16c475462d50ad247ec2927113617662a4e", "code_archive": "repos/ByeSYa4KPS.zip", "code_archive_sha256": "a66ab02038a6b4c6c228e245fbc1391b8c2236287f5e41ec9e68862afaf46d8c", "code_archive_bytes": 2272589, "code_file_count": 33, "code_extensions": {".py": 22, ".sh": 11}, "github_disk_usage_kb": 2438, "github_languages": {"Python": 233428, "Shell": 3473}, "github_archived": false, "github_pushed_at": "2022-06-06T18:16:54Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/sparse-networks-from-scratch-faster-training"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "s6q6zX45F8", "year": 2025, "status": "rejected", "title": "Uncertainty-Based Extensible Codebook for Discrete Federated Learning in Heterogeneous Data Silos", "authors": ["Tianyi Zhang", "Yu Cao", "Dianbo Liu"], "authorids": ["~Tianyi_Zhang18", "~Yu_Cao10", "~Dianbo_Liu2"], "authors_source": "OpenReview API", "abstract": "Federated learning (FL), aimed at leveraging vast distributed datasets, confronts a crucial challenge: the heterogeneity of data across different silos. While previous studies have explored discrete representations to enhance model generalization across minor distributional shifts, these approaches often struggle to adapt to new data silos with significantly divergent distributions. In response, we have identified that models derived from FL exhibit markedly increased uncertainty when applied to data silos with unfamiliar distributions. Consequently, we propose an innovative yet straightforward iterative framework, termed Uncertainty-Based Extensible-Codebook Federated Learning (UEFL). This framework dynamically maps latent features to trainable discrete vectors, assesses the uncertainty, and specifically extends the discretization dictionary or codebook for silos exhibiting high uncertainty. Our approach aims to simultaneously enhance accuracy and reduce uncertainty by explicitly addressing the diversity of data distributions, all while maintaining minimal computational overhead in environments characterized by heterogeneous data silos. Through experiments conducted on six datasets, our method has demonstrated its superiority, achieving significant improvements in accuracy (by 3%--22.1%) and uncertainty reduction (by 38.83%--96.24%), thereby outperforming contemporary state-of-the-art methods.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=s6q6zX45F8", "arxiv_id": "2402.18888", "paper_pdf": "papers/s6q6zX45F8.pdf", "paper_pdf_sha256": "1ee10d8baabed5432f656c6f784893207945598475d35d82d2d1adb90023f38d", "paper_pdf_bytes": 1849144, "paper_pdf_source": "openreview", "code_url": "https://github.com/destiny301/uefl", "code_repository": "destiny301/uefl", "code_commit": "833571a8d0739120229aacf49d7303f557176aea", "code_archive": "repos/s6q6zX45F8.zip", "code_archive_sha256": "5281df8a2ad092d8d3efc957a86471e7c73f0fb3ff17f9947dc2b431d460850b", "code_archive_bytes": 170884, "code_file_count": 11, "code_extensions": {".py": 11}, "github_disk_usage_kb": 222, "github_languages": {"Python": 56378}, "github_archived": false, "github_pushed_at": "2024-03-28T04:55:49Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/uncertainty-based-extensible-codebook-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ZkEsEFFUyo", "year": 2024, "status": "rejected", "title": "Pushing the Limits of Pre-training for Time Series Forecasting in the CloudOps Domain", "authors": ["Gerald Woo", "Chenghao Liu", "Akshat Kumar", "Doyen Sahoo"], "authorids": ["~Gerald_Woo1", "~Chenghao_Liu1", "~Akshat_Kumar2", "~Doyen_Sahoo1"], "authors_source": "OpenReview API", "abstract": "Time series has been left behind in the era of pre-training and transfer learning. While research in the fields of natural language processing and computer vision are enjoying progressively larger datasets to train massive models, the most popular time series datasets consist of only tens of thousands of time steps, limiting our ability to study the effectiveness of pre-training and scaling. Recent studies have also cast doubt on the need for expressive models and scale. To alleviate these issues, we  introduce three large-scale time series forecasting datasets from the cloud operations (CloudOps) domain, the largest having billions of observations, enabling further study into pre-training and scaling of time series models. We build the empirical groundwork for studying pre-training and scaling of time series models and pave the way for future research by identifying a promising candidate architecture. We show that it is a strong zero-shot baseline and benefits from further scaling, both in model and dataset size. Accompanying these datasets and results is a suite of comprehensive benchmark results comparing classical and deep learning baselines to our pre-trained method -- achieving a 27% reduction in error on the largest dataset. Code and datasets will be made publicly available.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=ZkEsEFFUyo", "arxiv_id": "2310.05063", "paper_pdf": "papers/ZkEsEFFUyo.pdf", "paper_pdf_sha256": "2f52a6ecbba40b79974bf316614458e104aa9116c537e11de2aa351f34f127c8", "paper_pdf_bytes": 1466322, "paper_pdf_source": "openreview", "code_url": "https://github.com/SalesforceAIResearch/pretrain-time-series-cloudops", "code_repository": "SalesforceAIResearch/pretrain-time-series-cloudops", "code_commit": "983c1b9515b72721856a2f4b3de9107080c1dc51", "code_archive": "repos/ZkEsEFFUyo.zip", "code_archive_sha256": "fa44def132bcbf4bbfaf9d3afc14ffeac2bae7cb05a0824ecbac7a27b598ce1e", "code_archive_bytes": 227523, "code_file_count": 94, "code_extensions": {".py": 94}, "github_disk_usage_kb": 148, "github_languages": {"Python": 761347}, "github_archived": true, "github_pushed_at": "2025-09-09T17:45:25Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/pushing-the-limits-of-pre-training-for-time"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "bSuY3hSRJPP", "year": 2023, "status": "rejected", "title": "SpectraNet: multivariate forecasting and imputation under distribution shifts and missing data", "authors": ["Cristian Ignacio Challu", "Peihong Jiang", "Ying Nian Wu", "Laurent Callot"], "authorids": ["~Cristian_Ignacio_Challu1", "jpeihong@amazon.com", "~Ying_Nian_Wu1", "~Laurent_Callot1"], "authors_source": "OpenReview API", "abstract": "In this work, we tackle two widespread challenges in real applications for time-series forecasting that have been largely understudied: distribution shifts and missing data. We propose SpectraNet, a novel multivariate time-series forecasting model that dynamically infers a latent space spectral decomposition to capture current temporal dynamics and correlations on the recent observed history. A Convolution Neural Network maps the learned representation by sequentially mixing its components and refining the output. Our proposed approach can simultaneously produce forecasts and interpolate past observations and can, therefore, greatly simplify production systems by unifying imputation and forecasting tasks into a single model. SpectraNet achieves SoTA performance simultaneously on both tasks on five benchmark datasets, compared to forecasting and imputation models, with up to 92% fewer parameters and comparable training times. On settings with up to 80% missing data, SpectraNet has average performance improvements of almost 50% over the second-best alternative.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=bSuY3hSRJPP", "arxiv_id": "2210.12515", "paper_pdf": "papers/bSuY3hSRJPP.pdf", "paper_pdf_sha256": "d2d8025216197478f16ec840ab8c8ca3d45851d7e85c9d8dcfd12233e301b047", "paper_pdf_bytes": 3414267, "paper_pdf_source": "openreview", "code_url": "https://github.com/cchallu/spectranet", "code_repository": "cchallu/spectranet", "code_commit": "f476444d048793b259f0eb65995f66f292784e2b", "code_archive": "repos/bSuY3hSRJPP.zip", "code_archive_sha256": "99c91d9aa03ce0065a93eee6b92777d23cff43fcdf8c1ba2ca43b0faef374831", "code_archive_bytes": 310192, "code_file_count": 4, "code_extensions": {".py": 4}, "github_disk_usage_kb": 312, "github_languages": {"Python": 34864}, "github_archived": false, "github_pushed_at": "2022-10-10T15:17:38Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/spectranet-multivariate-forecasting-and"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Vq_QHT5kcAK", "year": 2022, "status": "rejected", "title": "Greedy Bayesian Posterior Approximation with Deep Ensembles", "authors": ["Aleksei Tiulpin", "Matthew B. Blaschko"], "authorids": ["~Aleksei_Tiulpin1", "~Matthew_B._Blaschko1"], "authors_source": "OpenReview API", "abstract": "Ensembles of independently trained neural networks are a state-of-the-art approach to estimate predictive uncertainty in Deep Learning, and can be interpreted as an approximation of the posterior distribution via a mixture of delta functions. The training of ensembles relies on non-convexity of the loss landscape and random initialization of their individual members, making the resulting posterior approximation uncontrolled. This paper proposes a novel and principled method to tackle this limitation, minimizing an $f$-divergence between the true posterior and a kernel density estimator in a function space. We analyze this objective from a combinatorial point of view, and show that it is submodular with respect to mixture components for any $f$. Subsequently, we consider the problem of greedy ensemble construction, and from the marginal gain of the total objective, we derive a novel diversity term for ensemble methods. The performance of our approach is demonstrated on computer vision out-of-distribution detection benchmarks in a range of architectures trained on multiple datasets. The source code of our method is made publicly available.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=Vq_QHT5kcAK", "arxiv_id": "2105.14275", "paper_pdf": "papers/Vq_QHT5kcAK.pdf", "paper_pdf_sha256": "481ae9e437852cba1ba4c1385072a2ae9ca8be37418be32c13fd34f9e1a8264f", "paper_pdf_bytes": 1008569, "paper_pdf_source": "openreview", "code_url": "https://github.com/imedslab/greedy_ensembles_training", "code_repository": "imedslab/greedy_ensembles_training", "code_commit": "387dd8678c6cf3d4b12463eb9b4096007c80b316", "code_archive": "repos/Vq_QHT5kcAK.zip", "code_archive_sha256": "b2c7b10ec69ea04203d802bbf01a97b1d42b37df23a98b6232d9b7ad7ac2ec25", "code_archive_bytes": 328488, "code_file_count": 17, "code_extensions": {".py": 16, ".sh": 1}, "github_disk_usage_kb": 805, "github_languages": {"Python": 61865, "Shell": 1478}, "github_archived": false, "github_pushed_at": "2022-07-18T12:52:19Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/greedy-bayesian-posterior-approximation-with"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Y-Wl1l0Va-", "year": 2021, "status": "rejected", "title": "Shortest-Path Constrained Reinforcement Learning for Sparse Reward Tasks", "authors": ["Sungryull Sohn", "Sungtae Lee", "Jongwook Choi", "Harm van Seijen", "Honglak Lee", "Mehdi Fatemi"], "authorids": ["~Sungryull_Sohn1", "~Sungtae_Lee1", "~Jongwook_Choi1", "~Harm_van_Seijen1", "~Honglak_Lee2", "~Mehdi_Fatemi1"], "authors_source": "OpenReview API", "abstract": "We propose the k-Shortest-Path (k-SP) constraint: a novel constraint on the agent’s trajectory that improves the sample-efficiency in sparse-reward MDPs. We show that any optimal policy necessarily satisfies the k-SP constraint. Notably, the k-SP constraint prevents the policy from exploring state-action pairs along the non-k-SP trajectories (e.g., going back and forth).  However, in practice, excluding state-action pairs may hinder convergence of many RL algorithms. To overcome this, we propose a novel cost function that penalizes the policy violating SP constraint, instead of completely excluding it.  Our numerical experiment in a tabular RL setting demonstrate that the SP constraint can significantly reduce the trajectory space of policy. As a result, our constraint enables more sample efficient learning by suppressing redundant exploration and exploitation. Our empirical experiment results on MiniGrid and DeepMind Lab show that the proposed method significantly improves proximal policy optimization (PPO) and outperforms existing novelty-seeking exploration methods including count-based exploration, indicating that it improves the sample efficiency by preventing the agent from taking redundant actions.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=Y-Wl1l0Va-", "arxiv_id": "2107.06405", "paper_pdf": "papers/Y-Wl1l0Va-.pdf", "paper_pdf_sha256": "b85c620475245eabf55d17161de0183016a3850a3d5ddd2627d6f96f6ff6b26f", "paper_pdf_bytes": 45923891, "paper_pdf_source": "openreview", "code_url": "https://github.com/srsohn/shortest-path-rl", "code_repository": "srsohn/shortest-path-rl", "code_commit": "60094c4ce116bfd63506e48ab40ea686314bb913", "code_archive": "repos/Y-Wl1l0Va-.zip", "code_archive_sha256": "b36ac25302339b55c2f39663dcf9260ce4946db2a9122e2d5141029123326a39", "code_archive_bytes": 1000583, "code_file_count": 22, "code_extensions": {".py": 17, ".sh": 5}, "github_disk_usage_kb": 957, "github_languages": {"Python": 116594, "Starlark": 31093, "Shell": 3933}, "github_archived": false, "github_pushed_at": "2021-07-19T03:17:17Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/shortest-path-constrained-reinforcement-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "r1xwA34KDB", "year": 2020, "status": "rejected", "title": "Learning Invariants through Soft Unification", "authors": ["Nuri Cingillioglu", "Alessandra Russo"], "authorids": ["nuri.cingillioglu13@imperial.ac.uk", "a.russo@imperial.ac.uk"], "authors_source": "OpenReview API", "abstract": "Human reasoning involves recognising common underlying principles across many examples by utilising variables. The by-products of such reasoning are invariants that capture patterns across examples such as \"if someone went somewhere then they are there\" without mentioning specific people or places. Humans learn what variables are and how to use them at a young age, and the question this paper addresses is whether machines can also learn and use variables solely from examples without requiring human pre-engineering. We propose Unification Networks that incorporate soft unification into neural networks to learn variables and by doing so lift examples into invariants that can then be used to solve a given task. We evaluate our approach on four datasets to demonstrate that learning invariants captures patterns in the data and can improve performance over baselines.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=r1xwA34KDB", "arxiv_id": "1909.07328", "paper_pdf": "papers/r1xwA34KDB.pdf", "paper_pdf_sha256": "44c2f65c7c10ef5b30a6fe26efbcc40143464f94fd43e3e6c156aab03cb82cf6", "paper_pdf_bytes": 434864, "paper_pdf_source": "openreview", "code_url": "https://github.com/nuric/softuni", "code_repository": "nuric/softuni", "code_commit": "86dc865dddddb1445caea61f666d0e1abc669067", "code_archive": "repos/r1xwA34KDB.zip", "code_archive_sha256": "6b4ccb9fbedb24c76e4c87d3d590ccd52cb42baeb02e455cd460c7289c1e9689", "code_archive_bytes": 1357338, "code_file_count": 11, "code_extensions": {".py": 7, ".ipynb": 3, ".sh": 1}, "github_disk_usage_kb": 2559, "github_languages": {"Jupyter Notebook": 1855523, "Python": 114820, "Shell": 548}, "github_archived": false, "github_pushed_at": "2020-10-17T11:35:39Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/learning-invariants-through-soft-unification"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "TJbfURjxyL", "year": 2026, "status": "rejected", "title": "SAISA: Towards Multimodal Large Language Models with Both Training and Inference Efficiency", "authors": ["Qianhao Yuan", "yanjiang liu", "Mo guozhao", "Yaojie Lu", "Hongyu Lin", "Jia Zheng", "Ben He", "Xianpei Han", "Le Sun"], "authorids": ["~Qianhao_Yuan1", "~yanjiang_liu1", "~Mo_guozhao1", "~Yaojie_Lu1", "~Hongyu_Lin1", "~Jia_Zheng3", "~Ben_He1", "~Xianpei_Han1", "~Le_Sun1"], "authors_source": "OpenReview API", "abstract": "Multimodal Large Language Models (MLLMs) mainly fall into two architectures, each involving a trade-off between training and inference efficiency: embedding space alignment (e.g. LLaVA) is inefficient during inference, while cross-attention space alignment (e.g. Flamingo) is inefficient in training. In this paper, we compare these two architectures and identify key factors for building efficient MLLMs. A primary difference between them lies in how attention is applied to visual tokens, particularly in their interactions with each other. To investigate whether attention among visual tokens is necessary, we propose a new self-attention mechanism, NAAViT (No Attention Among Visual Tokens), which eliminates this type of attention. Our pilot experiment on LLaVA-1.5 shows that attention among visual tokens is highly redundant. Based on these insights, we introduce SAISA (Self-Attention Input Space Alignment), a novel architecture that enhances both training and inference efficiency. SAISA directly aligns visual features with the input spaces of NAAViT self-attention blocks, reducing computational overhead in both self-attention blocks and feed-forward networks (FFNs). Compared with the LLaVA-1.5 architecture, SAISA reduces the inference FLOPs by 66% and the training budget by 26%, while achieving superior performance in terms of accuracy. Comprehensive ablation studies further validate the effectiveness of SAISA across various LLMs and visual encoders. The code and models will be publicly available.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=TJbfURjxyL", "arxiv_id": "2502.02458", "paper_pdf": "papers/TJbfURjxyL.pdf", "paper_pdf_sha256": "c0f37edf930ae9b6e7e0fc5012251be69fc4e8ebbab95b9ae6fc49f7d4c45ada", "paper_pdf_bytes": 5135429, "paper_pdf_source": "openreview", "code_url": "https://github.com/icip-cas/SAISA", "code_repository": "icip-cas/SAISA", "code_commit": "53a3670ab4c0ca92f57cf9c3bfce66a67b1089ef", "code_archive": "repos/TJbfURjxyL.zip", "code_archive_sha256": "b8e2042eff582bbc9596e4ec5ae522908e83f8219b5c1b2f27eaef3e478208ec", "code_archive_bytes": 3574035, "code_file_count": 467, "code_extensions": {".py": 437, ".sh": 29, ".js": 1}, "github_disk_usage_kb": 3082, "github_languages": {"Python": 3445654, "Shell": 29328, "JavaScript": 9991, "HTML": 7669, "CSS": 1822}, "github_archived": false, "github_pushed_at": "2025-04-08T07:45:12Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/saisa-towards-multimodal-large-language"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "06B23UkNid", "year": 2025, "status": "rejected", "title": "MV-CLAM: Multi-View Molecular Interpretation with Cross-Modal Projection via Language Model", "authors": ["Sumin Ha", "Jun Hyeong Kim", "Yinhua Piao", "Sun Kim"], "authorids": ["~Sumin_Ha1", "~Jun_Hyeong_Kim3", "~Yinhua_Piao1", "~Sun_Kim2"], "authors_source": "OpenReview API", "abstract": "Large language models (LLMs) have shown significant potential in the biomolecular domain, particularly by demonstrating that effective adaptation of molecular representations for LLMs can greatly improve the quality of molecular captions. Most previous works have focused on aligning unimodal molecular structures with text, overlooking the diversity of modalities. Naive approaches to aligning multi-modal molecular structures with text often lead to (1) separately aligned embeddings, (2) inconsistent textual representations, and (3) increased computational overhead. To address these challenges, we propose LLM framework MV-CLAM equipped with MQ-Former, a novel multi-querying transformer. This architecture introduces a cross-model projector facilitating the simultaneous alignment of 2D and 3D molecular representations to a unified text token. By employing a shared self-attention layer, MQ-Former preserves rich molecular embeddings across different dimensions while consolidating them into a universal molecular token. Our approach outperforms baseline models in both molecule-text retrieval and molecule captioning tasks. Additionally, our framework shows promising results for zero-shot molecule editing and molecule-related question answering. By effectively integrating multi-view molecular data into a format conducive to LLMs, our method serves as a valuable tool for enhancing the characterization and understanding of chemical structures, facilitating a more seamless transition from molecular data to textual descriptions. The source code of MV-CLAM is available in https://anonymous.4open.science/r/mv-clam-4827.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=06B23UkNid", "arxiv_id": "2503.04780", "paper_pdf": "papers/06B23UkNid.pdf", "paper_pdf_sha256": "8229880c9dff7240555a729dce46d075d1c776b342664db8941544a7b6396529", "paper_pdf_bytes": 1545670, "paper_pdf_source": "openreview", "code_url": "https://github.com/sumin124/mv-clam", "code_repository": "sumin124/mv-clam", "code_commit": "d537bf334cdcf438bbd0d343ca0eea6bfdb96837", "code_archive": "repos/06B23UkNid.zip", "code_archive_sha256": "a75bc12fc904c48c80a5a9d73434b290fdfa9019180ac2a5f0c84f4e4a986524", "code_archive_bytes": 103490, "code_file_count": 20, "code_extensions": {".py": 20}, "github_disk_usage_kb": 224, "github_languages": {"Python": 333095}, "github_archived": false, "github_pushed_at": "2025-08-21T05:55:31Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/mv-clam-multi-view-molecular-interpretation"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "468KWV14ll", "year": 2024, "status": "rejected", "title": "Exploration and Anti-Exploration with Distributional Random Network Distillation", "authors": ["Kai Yang", "Jian Tao", "Jiafei Lyu", "Xiu Li"], "authorids": ["~Kai_Yang6", "~Jian_Tao5", "~Jiafei_Lyu1", "~Xiu_Li1"], "authors_source": "OpenReview API", "abstract": "Exploration remains a critical issue in deep reinforcement learning for an agent to attain high returns in unknown environments. Although the prevailing exploration Random Network Distillation (RND) algorithm has been demonstrated to be effective in numerous environments, it often needs more discriminative power in bonus allocation. This paper highlights the ''bonus inconsistency'' issue within RND, pinpointing its primary limitation. To address this issue, we introduce the Distributional RND (DRND), a derivative of the RND. DRND enhances the exploration process by distilling a distribution of random networks and implicitly incorporates pseudo counts to improve the precision of bonus allocation. This refinement encourages agents to engage in more extensive exploration. Our method effectively mitigates the inconsistency issue without introducing significant computational overhead. Both theoretical analysis and experimental results demonstrate the superiority of our approach over the original RND algorithm. Our method excels in challenging online exploration scenarios and effectively serves as an anti-exploration mechanism in D4RL offline tasks.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=468KWV14ll", "arxiv_id": "2401.09750", "paper_pdf": "papers/468KWV14ll.pdf", "paper_pdf_sha256": "4b3a6122f775cd7284d4aee2097dcf1ac6abac83995f6d011f4e684b29d3582b", "paper_pdf_bytes": 13044600, "paper_pdf_source": "openreview", "code_url": "https://github.com/yk7333/DRND", "code_repository": "yk7333/DRND", "code_commit": "e1ea983c9c12603804e6d268efb212c208663aac", "code_archive": "repos/468KWV14ll.zip", "code_archive_sha256": "c8de4118f7eb4b3aaccf02f71c77da4c94ce3f10ef3e4dee119a7ab5d83276ca", "code_archive_bytes": 135866, "code_file_count": 13, "code_extensions": {".py": 12, ".sh": 1}, "github_disk_usage_kb": 150, "github_languages": {"Python": 87011, "Shell": 923}, "github_archived": false, "github_pushed_at": "2024-10-12T06:27:07Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/exploration-and-anti-exploration-with"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "C9uEwyfklBE", "year": 2023, "status": "rejected", "title": "Pareto Manifold Learning: Tackling multiple tasks via ensembles of single-task models", "authors": ["Nikolaos Dimitriadis", "Pascal Frossard", "François Fleuret"], "authorids": ["~Nikolaos_Dimitriadis1", "~Pascal_Frossard1", "~François_Fleuret2"], "authors_source": "OpenReview API", "abstract": "In Multi-Task Learning, tasks may compete and limit the performance achieved on each other rather than guiding the optimization trajectory to a common solution, superior to its single-task counterparts. There is often not a single solution that is optimal for all tasks, leading practitioners to balance tradeoffs between tasks' performance, and to resort to optimality in the Pareto sense. Current Multi-Task Learning methodologies either completely neglect this aspect of functional diversity, and produce one solution in the Pareto Front predefined by their optimization schemes, or produce diverse but discrete solutions, each requiring a separate training run. In this paper, we conjecture that there exist Pareto Subspaces, i.e., weight subspaces where multiple optimal functional solutions lie. We propose Pareto Manifold Learning, an ensembling method in weight space that is able to discover such a parameterization and produces a continuous Pareto Front in a single training run, allowing practitioners to modulate the performance on each task during inference on the fly. We validate the proposed method on a diverse set of multi-task learning benchmarks, ranging from image classification to tabular datasets and scene understanding, and show that Pareto Manifold Learning outperforms state-of-the-art algorithms.\n", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=C9uEwyfklBE", "arxiv_id": "2210.09759", "paper_pdf": "papers/C9uEwyfklBE.pdf", "paper_pdf_sha256": "30ceebbfacd9d11d5c62de0d6cc2b6d547598d842226e3f2d177bc7e57db9cc4", "paper_pdf_bytes": 23785593, "paper_pdf_source": "openreview", "code_url": "https://github.com/nik-dim/pamal", "code_repository": "nik-dim/pamal", "code_commit": "456a338bac579977ddb7100c2bae36f2768df0c6", "code_archive": "repos/C9uEwyfklBE.zip", "code_archive_sha256": "f37acbc245fbec724186f83cfdd57c0b2b78b8a34582cb532c7a2f793815cc10", "code_archive_bytes": 347349, "code_file_count": 64, "code_extensions": {".py": 63, ".sh": 1}, "github_disk_usage_kb": 315, "github_languages": {"Python": 325702, "Shell": 1980, "Dockerfile": 484}, "github_archived": false, "github_pushed_at": "2026-06-30T16:37:52Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/pareto-manifold-learning-tackling-multiple"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "WZR7ckBkzPY", "year": 2022, "status": "rejected", "title": "Variational Wasserstein gradient flow", "authors": ["Jiaojiao Fan", "Amirhossein Taghvaei", "Yongxin Chen"], "authorids": ["~Jiaojiao_Fan1", "~Amirhossein_Taghvaei1", "~Yongxin_Chen1"], "authors_source": "OpenReview API", "abstract": "The gradient flow of a function over the space of probability densities with respect to the Wasserstein metric often exhibits nice properties and has been utilized in several machine learning applications. The standard approach to compute the Wasserstein gradient flow is the finite difference which discretizes the underlying space over a grid, and is not scalable. In this work, we propose a scalable proximal gradient type algorithm for Wasserstein gradient flow. The key of our method is a variational formulation of the objective function, which makes it possible to realize the JKO proximal map through a primal-dual optimization. This primal-dual problem can be efficiently solved by alternatively updating the parameters in the inner and outer loops. Our framework covers all the classical Wasserstein gradient flows including the heat equation and the porous medium equation. We demonstrate the performance and scalability of our algorithm with several numerical examples.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=WZR7ckBkzPY", "arxiv_id": "2112.02424", "paper_pdf": "papers/WZR7ckBkzPY.pdf", "paper_pdf_sha256": "3656e16fe03ca9803436ebbea2c5a062cb6dd5aadb6afc2f5ac5d86b756bae23", "paper_pdf_bytes": 4584986, "paper_pdf_source": "openreview", "code_url": "https://github.com/sbyebss/variational_wgf", "code_repository": "sbyebss/variational_wgf", "code_commit": "b0e22909b0b145989088af8df174eccba189a23e", "code_archive": "repos/WZR7ckBkzPY.zip", "code_archive_sha256": "96640698eb10a64618d2b79266f1d282e5c63e167b2063586bfc4068c4d226b1", "code_archive_bytes": 899253, "code_file_count": 90, "code_extensions": {".py": 85, ".sh": 4, ".ipynb": 1}, "github_disk_usage_kb": 823, "github_languages": {"Jupyter Notebook": 409440, "Python": 295837, "Shell": 1775}, "github_archived": false, "github_pushed_at": "2022-10-17T05:45:38Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/variational-wasserstein-gradient-flow-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "lXW6Sk1075v", "year": 2021, "status": "rejected", "title": "FORK: A FORward-looKing Actor for Model-Free Reinforcement Learning", "authors": ["Honghao Wei", "Lei Ying"], "authorids": ["~Honghao_Wei2", "~Lei_Ying1"], "authors_source": "OpenReview API", "abstract": "In this paper, we propose a new type of Actor, named forward-looking Actor or FORK for short, for Actor-Critic algorithms. FORK can be easily integrated into a model-free Actor-Critic algorithm. Our experiments on six Box2D and MuJoCo environments with continuous state and action spaces demonstrate significant performance improvement FORK can bring to the state-of-the-art algorithms. A variation of FORK can further solve BipedalWalkerHardcore in as few as four hours using a single GPU.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=lXW6Sk1075v", "arxiv_id": "2010.01652", "paper_pdf": "papers/lXW6Sk1075v.pdf", "paper_pdf_sha256": "6fc657cc12ab0ee13bf84ea5b096427c3afd5fb9b81b8ee61e136248c9601e5b", "paper_pdf_bytes": 1937556, "paper_pdf_source": "openreview", "code_url": "https://github.com/honghaow/FORK", "code_repository": "honghaow/FORK", "code_commit": "12e002d24ba7e5718e34dd20d4888a1dae1b4774", "code_archive": "repos/lXW6Sk1075v.zip", "code_archive_sha256": "06e0ff9896c612157241912b6b5f45084ce2449b3a014cc4e8bfc4dcc6a6ab4c", "code_archive_bytes": 750821, "code_file_count": 15, "code_extensions": {".py": 10, ".sh": 4, ".ipynb": 1}, "github_disk_usage_kb": 991, "github_languages": {"Jupyter Notebook": 705939, "Python": 61936, "Shell": 4444}, "github_archived": false, "github_pushed_at": "2022-03-31T23:32:42Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/fork-a-forward-looking-actor-for-model-free-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "HJlk-eHFwH", "year": 2020, "status": "rejected", "title": "AdaGAN: Adaptive GAN for Many-to-Many Non-Parallel Voice Conversion", "authors": ["Maitreya Patel", "Mirali Purohit", "Mihir Parmar", "Nirmesh J. Shah", "Hemant A. Patil"], "authorids": ["maitreya_patel@daiict.ac.in", "purohit_mirali@daiict.ac.in", "mihirparmar@asu.edu", "nirmesh88_shah@daiict.ac.in", "hemant_patil@daiict.ac.in"], "authors_source": "OpenReview API", "abstract": "Voice Conversion (VC) is a task of converting perceived speaker identity from a source speaker to a particular target speaker. Earlier approaches in the literature primarily find a mapping between the given source-target speaker-pairs. Developing mapping techniques for many-to-many VC using non-parallel data, including zero-shot learning remains less explored areas in VC. Most of the many-to-many VC architectures require training data from all the target speakers for whom we want to convert the voices. In this paper, we propose a novel style transfer architecture, which can also be extended to generate voices even for target speakers whose data were not used in the training (i.e., case of zero-shot learning). In particular, propose Adaptive Generative Adversarial Network (AdaGAN), new architectural training procedure help in learning normalized speaker-independent latent representation, which will be used to generate speech with different speaking styles in the context of VC. We compare our results with the state-of-the-art StarGAN-VC architecture. In particular, the AdaGAN achieves 31.73%, and 10.37% relative improvement compared to the StarGAN in MOS tests for speech quality and speaker similarity, respectively. The key strength of the proposed architectures is that it yields these results with less computational complexity. AdaGAN is 88.6% less complex than StarGAN-VC in terms of FLoating Operation Per Second (FLOPS), and 85.46% less complex in terms of trainable parameters.  ", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=HJlk-eHFwH", "arxiv_id": null, "paper_pdf": "papers/HJlk-eHFwH.pdf", "paper_pdf_sha256": "7bd4784b8a86d45ef4760cc198c7212a0cf5b9eba7065139ae7175b833b19ff1", "paper_pdf_bytes": 629151, "paper_pdf_source": "openreview", "code_url": "https://github.com/liusongxiang/StarGAN-Voice-Conversion", "code_repository": "liusongxiang/StarGAN-Voice-Conversion", "code_commit": "a4633d8b4888e0fa77d6b67f9d94da603da2af31", "code_archive": "repos/HJlk-eHFwH.zip", "code_archive_sha256": "7c0feefee240c81972d7964c1ac5fa5293cc571f5c3d8cc37b8b9b8608f7966a", "code_archive_bytes": 2760494, "code_file_count": 8, "code_extensions": {".py": 8}, "github_disk_usage_kb": 2702, "github_languages": {"Python": 50459}, "github_archived": false, "github_pushed_at": "2019-10-11T02:49:58Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/adagan-adaptive-gan-for-many-to-many-non"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "2XP4HD0NBy", "year": 2026, "status": "rejected", "title": "SegRet: An Efficient Design for Semantic Segmentation with Retentive Network", "authors": ["Zhiyuan Li", "Yi Chang", "Yuan Wu"], "authorids": ["~Zhiyuan_Li19", "~Yi_Chang4", "~Yuan_Wu2"], "authors_source": "OpenReview API", "abstract": "With the rapid evolution of autonomous driving technology and intelligent transportation systems, semantic segmentation has become increasingly critical. Precise interpretation and analysis of real-world environments are indispensable for these advanced applications. However, traditional semantic segmentation approaches frequently face challenges in balancing model performance with computational efficiency, especially regarding the volume of model parameters. To address these constraints, we propose SegRet, a novel model employing the Retentive Network (RetNet) architecture coupled with a lightweight residual decoder that integrates zero-initialization. SegRet offers three distinctive advantages: (1) Lightweight Residual Decoder: by embedding a zero-initialization layer within the residual network structure, the decoder remains computationally streamlined without sacrificing essential information propagation; (2) Robust Feature Extraction: adopting RetNet as its backbone enables SegRet to effectively capture hierarchical image features, thereby enriching the representation quality of extracted features; (3) Parameter Efficiency: SegRet attains state-of-the-art (SOTA) segmentation performance while markedly decreasing the number of parameters, ensuring high accuracy without imposing additional computational burdens. Comprehensive empirical evaluations on prominent benchmarks, such as ADE20K, Citycapes, and COCO-Stuff, highlight the effectiveness and superiority of our method.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=2XP4HD0NBy", "arxiv_id": "2502.14014", "paper_pdf": "papers/2XP4HD0NBy.pdf", "paper_pdf_sha256": "1f84b41df82cfac1258bce7480e673aabce9825d25b8720479cf82f7b57f28a9", "paper_pdf_bytes": 1921993, "paper_pdf_source": "openreview", "code_url": "https://github.com/ZhiyuanLi218/segret", "code_repository": "ZhiyuanLi218/segret", "code_commit": "99c409054cf74eac7a08d610525df89090af35b7", "code_archive": "repos/2XP4HD0NBy.zip", "code_archive_sha256": "261afc109c8aa4ef6b068d68cfea8ff2066c8881bb8b15a40d83dc111f15aa04", "code_archive_bytes": 3424375, "code_file_count": 683, "code_extensions": {".py": 677, ".sh": 4, ".ipynb": 2}, "github_disk_usage_kb": 3085, "github_languages": {"Python": 1203518, "Shell": 2351, "Dockerfile": 750}, "github_archived": false, "github_pushed_at": "2025-09-22T02:07:24Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/segret-an-efficient-design-for-semantic"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "r1cbFEH0Df", "year": 2025, "status": "rejected", "title": "How to Correctly Do Semantic Backpropagation on Language-based Agentic Systems", "authors": ["Wenyi Wang", "Hisham Abdullah Alyahya", "Dylan R. Ashley", "Oleg Serikov", "Dmitrii Khizbullin", "Francesco Faccio", "Jürgen Schmidhuber"], "authorids": ["~Wenyi_Wang1", "~Hisham_Abdullah_Alyahya1", "~Dylan_R._Ashley1", "~Oleg_Serikov1", "~Dmitrii_Khizbullin2", "~Francesco_Faccio1", "~Jürgen_Schmidhuber1"], "authors_source": "OpenReview API", "abstract": "Language-based agentic systems have shown great promise in recent years, transitioning from solving small-scale research problems to being deployed in challenging real-world tasks. However, optimizing these systems often requires substantial manual labor. Recent studies have demonstrated that these systems can be represented as computational graphs, enabling automatic optimization. Despite these advancements, most current efforts in Graph-based Agentic System Optimization (GASO) fail to properly assign feedback to the system’s components given feedback on the system’s output. To address this challenge, we formalize the concept of semantic backpropagation with semantic gradients—a generalization that aligns several key optimization techniques, including reverse-mode automatic differentiation and the more recent TextGrad by exploiting the relationship among nodes with a common successor. This serves as a method for computing directional information about how changes to each component of an agentic system might improve the system’s output. To use these gradients, we propose a method called semantic gradient descent which enables us to solve GASO effectively. Our results on both BIG-Bench Hard and GSM8K show that our approach outperforms existing state-of-the-art methods for solving GASO problems. A detailed ablation study on the LIAR dataset demonstrates the parsimonious nature of our method.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=r1cbFEH0Df", "arxiv_id": "2412.03624", "paper_pdf": "papers/r1cbFEH0Df.pdf", "paper_pdf_sha256": "2e4f7e2de54d6ded9e3029eaf0dfc60abe00390310a4d7e967d3c50d3ea417ec", "paper_pdf_bytes": 746052, "paper_pdf_source": "openreview", "code_url": "https://github.com/HishamAlyahya/semantic_backprop", "code_repository": "HishamAlyahya/semantic_backprop", "code_commit": "aad0c724625f5f1e304b9b8a0c67bb1e8f929d81", "code_archive": "repos/r1cbFEH0Df.zip", "code_archive_sha256": "bb034dab83dcdfaf43e02944985c48305a1f0059719eeb7deaf47c406dbc36a3", "code_archive_bytes": 241655, "code_file_count": 22, "code_extensions": {".py": 18, ".sh": 4}, "github_disk_usage_kb": 229, "github_languages": {"Python": 49654, "Shell": 1812}, "github_archived": false, "github_pushed_at": "2024-12-06T18:18:09Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/how-to-correctly-do-semantic-backpropagation"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "eddd0YTCiq", "year": 2024, "status": "rejected", "title": "Graph-level Representation Learning with Joint-Embedding Predictive Architectures", "authors": ["Geri Skenderi", "Hang Li", "Jiliang Tang", "Marco Cristani"], "authorids": ["~Geri_Skenderi1", "~Hang_Li10", "~Jiliang_Tang1", "~Marco_Cristani1"], "authors_source": "OpenReview API", "abstract": "Joint-Embedding Predictive Architectures (JEPAs) have recently emerged as a novel and powerful technique for self-supervised representation learning. They aim to learn an energy-based model by predicting the latent representation of a target signal $y$ from a context signal $x$. JEPAs bypass the need for data augmentation and negative samples, which are typically required by contrastive learning, while avoiding the overfitting issues associated with generative-based pretraining. In this paper, we show that graph-level representations can be effectively modeled using this paradigm and propose Graph-JEPA, the first JEPA for the graph domain. In particular, we employ masked modeling to learn embeddings for different subgraphs of the input graph. To endow the representations with the implicit hierarchy that is often present in graph-level concepts, we devise an alternative training objective that consists of predicting the coordinates of the encoded subgraphs on the unit hyperbola in the 2D plane. Extensive validation shows that Graph-JEPA can learn representations that are expressive and competitive in both graph classification and regression problems. The implementation will be available upon acceptance.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=eddd0YTCiq", "arxiv_id": "2309.16014", "paper_pdf": "papers/eddd0YTCiq.pdf", "paper_pdf_sha256": "6c31c175acee0b77e8d1e788df2f69e1cf1e8646c24329c1fb33ea5d0ef7158d", "paper_pdf_bytes": 595303, "paper_pdf_source": "openreview", "code_url": "https://github.com/geriskenderi/graph-jepa", "code_repository": "geriskenderi/graph-jepa", "code_commit": "72df1b7704921001ea012a21f840300fbc792cdd", "code_archive": "repos/eddd0YTCiq.zip", "code_archive_sha256": "05daaf1f38416f82a4cd2ae89c0c8cec6824602f58186b5495988d1cbf42ea88", "code_archive_bytes": 162648, "code_file_count": 32, "code_extensions": {".py": 31, ".sh": 1}, "github_disk_usage_kb": 153, "github_languages": {"Python": 130078, "Shell": 531}, "github_archived": false, "github_pushed_at": "2025-01-17T11:19:37Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/graph-level-representation-learning-with"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "RNnKhz25N1O", "year": 2022, "status": "rejected", "title": "Low-Cost Algorithmic Recourse for Users With Uncertain Cost Functions", "authors": ["Prateek Yadav", "Peter Hase", "Mohit Bansal"], "authorids": ["~Prateek_Yadav1", "~Peter_Hase1", "~Mohit_Bansal2"], "authors_source": "OpenReview API", "abstract": "The problem of identifying algorithmic recourse for people affected by machine learning model decisions has received much attention recently. Existing approaches for recourse generation obtain solutions using properties like diversity, proximity, sparsity, and validity. Yet, these objectives are only heuristics for what we truly care about, which is whether a user is satisfied with the recourses offered to them. Some recent works try to model user-incurred cost, which is more directly linked to user satisfaction. But they assume a single global cost function that is shared across all users. This is an unrealistic assumption when users have dissimilar preferences about their willingness to act upon a feature and different costs associated with changing that feature. In this work, we formalize the notion of user-specific cost functions and introduce a new method for identifying actionable recourses for users. By default, we assume that users' cost functions are hidden from the recourse method, though our framework allows users to partially or completely specify their preferences or cost function. We propose an objective function, Expected Minimum Cost (EMC), based on two key ideas: (1) when presenting a set of options to a user, it is vital that there is at least one low-cost solution the user could adopt; (2) when we do not know the user's true cost function, we can approximately optimize for user satisfaction by first sampling plausible cost functions, then finding a set that achieves a good cost for the user in expectation. We optimize EMC with a novel discrete optimization algorithm, Cost-Optimized Local Search (COLS), which is guaranteed to improve the recourse set quality over iterations. Experimental evaluation on popular real-world datasets with simulated user costs demonstrates that our method satisfies up to 25.89 percentage points more users compared to strong baseline methods. Using standard fairness metrics, we also show that our method can provide more fair solutions across demographic groups than comparable methods, and we verify that our method is robust to misspecification of the cost function distribution. ", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=RNnKhz25N1O", "arxiv_id": "2111.01235", "paper_pdf": "papers/RNnKhz25N1O.pdf", "paper_pdf_sha256": "4de0195e99ecd2486440e6240ac732a910e5540ff9fd106ed97a6ec268aac7d5", "paper_pdf_bytes": 574461, "paper_pdf_source": "openreview", "code_url": "https://github.com/prateeky2806/EMC-COLS-recourse", "code_repository": "prateeky2806/EMC-COLS-recourse", "code_commit": "dd2c04fccd3c85ddaed0b4705aa98133d8f1b48a", "code_archive": "repos/RNnKhz25N1O.zip", "code_archive_sha256": "30560e796f6ff481493a3a1b702f1040a2532c158aed84131426088a1182c995", "code_archive_bytes": 916938, "code_file_count": 18, "code_extensions": {".py": 16, ".sh": 2}, "github_disk_usage_kb": 886, "github_languages": {"Python": 231463, "Shell": 33197}, "github_archived": false, "github_pushed_at": "2021-11-03T21:09:47Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/low-cost-algorithmic-recourse-for-users-with-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "LcPefbNSwx_", "year": 2021, "status": "rejected", "title": "Factor Normalization for Deep Neural Network Models", "authors": ["Haobo Qi", "Jing Zhou", "Hansheng Wang"], "authorids": ["qihaobo_gsm@pku.edu.cn", "~Jing_Zhou3", "hansheng@pku.edu.cn"], "authors_source": "OpenReview API", "abstract": "Deep neural network (DNN) models often involve features of high dimensions. In most cases, the high-dimensional features can be decomposed into two parts. The first part is a low-dimensional factor. The second part is the residual feature, with much-reduced variability and inter-feature correlation. This leads to a number of interesting theoretical findings for deep neural network training. Accordingly, we are inspired to develop a new factor normalization method for better performance. The proposed method leads to a new deep learning model with two important features. First, it allows factor related feature extraction. Second, it allows adaptive learning rates for factors and residuals, respectively. This leads to fast convergence speed on both training and validation datsets. A number of empirical experiments are presented to demonstrate its superior performance. The code is available at https://github.com/HazardNeo4869/FactorNormalization", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=LcPefbNSwx_", "arxiv_id": null, "paper_pdf": "papers/LcPefbNSwx_.pdf", "paper_pdf_sha256": "43d7270f45c137f1949c4a7ba41494ebe3b749c0d699008bea735a30351f41b2", "paper_pdf_bytes": 298805, "paper_pdf_source": "openreview", "code_url": "https://github.com/HazardNeo4869/FactorNormalization", "code_repository": "HazardNeo4869/FactorNormalization", "code_commit": "e762cdd6ed19d8d9252c08210573285c5f17e918", "code_archive": "repos/LcPefbNSwx_.zip", "code_archive_sha256": "37bfb8a6437d72d4f37828673fa2b143155caebf3a536e7672d613735ab8dd1e", "code_archive_bytes": 1027837, "code_file_count": 5, "code_extensions": {".py": 5}, "github_disk_usage_kb": 1018, "github_languages": {"Python": 73337}, "github_archived": false, "github_pushed_at": "2020-11-23T15:11:06Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/factor-normalization-for-deep-neural-network"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "SJgw51HFDr", "year": 2020, "status": "rejected", "title": "Sparse Weight Activation Training", "authors": ["Md Aamir Raihan", "Tor M. Aamodt"], "authorids": ["araihan@ece.ubc.ca", "aamodt@ece.ubc.ca"], "authors_source": "OpenReview API", "abstract": "Training convolutional neural networks (CNNs) is time consuming. Prior work has explored how to reduce the computational demands of training by eliminating gradients with relatively small magnitude. We show that eliminating small magnitude components has limited impact on the direction of high-dimensional vectors. However, in the context of training a CNN, we find that eliminating small magnitude components of weight and activation vectors allows us to train deeper networks on more complex datasets versus eliminating small magnitude components of gradients. We propose Sparse Weight Activation Training (SWAT), an algorithm that embodies these observations. SWAT reduces computations by 50% to 80% with better accuracy at a given level of sparsity versus the Dynamic Sparse Graph algorithm. SWAT also reduces memory footprint by 23% to 37% for activations and 50% to 80% for weights.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=SJgw51HFDr", "arxiv_id": "2001.01969", "paper_pdf": "papers/SJgw51HFDr.pdf", "paper_pdf_sha256": "be25473e6a5ffc0f0535bfd3b211245da24f2396203f45c2642242dca82d8b32", "paper_pdf_bytes": 1285260, "paper_pdf_source": "openreview", "code_url": "https://github.com/AamirRaihan/SWAT", "code_repository": "AamirRaihan/SWAT", "code_commit": "bbe271afa6e5ead3c02b20409594e17b6265b958", "code_archive": "repos/SJgw51HFDr.zip", "code_archive_sha256": "dd860cae2d548df0e132ab355590f717f3caec3bc87f4416f0059197db5fc488", "code_archive_bytes": 2911205, "code_file_count": 52, "code_extensions": {".py": 40, ".sh": 11, ".cpp": 1}, "github_disk_usage_kb": 2813, "github_languages": {"Python": 231631, "C++": 7587, "Shell": 1707}, "github_archived": false, "github_pushed_at": "2021-07-23T10:25:37Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/sparse-weight-activation-training-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "KJNgtPNxtv", "year": 2026, "status": "rejected", "title": "PFMBench: Protein Foundation Model Benchmark", "authors": ["Zhangyang Gao", "Hao Wang", "Cheng Tan", "Chenrui XU", "Mengdi Liu", "Bozhen Hu", "Linlin Chao", "Xiaoming Zhang", "Stan Z. Li"], "authorids": ["~Zhangyang_Gao1", "~Hao_Wang136", "~Cheng_Tan1", "~Chenrui_XU3", "~Mengdi_Liu1", "~Bozhen_Hu1", "~Linlin_Chao1", "~Xiaoming_Zhang2", "~Stan_Z._Li2"], "authors_source": "OpenReview API", "abstract": "This study investigates the current landscape and future directions of protein foundation model research.  While recent advancements have transformed protein science and engineering, the field lacks a comprehensive benchmark for fair evaluation and in-depth understanding. Since ESM-1B, numerous protein foundation models have emerged, each with unique datasets and methodologies. However, evaluations often focus on limited tasks tailored to specific models, hindering insights into broader generalization and limitations. Specifically, researchers struggle to understand the relationships between tasks, assess how well current models perform across them, and determine the criteria in developing new foundation models.  To fill this gap, we present PFMBench, a comprehensive benchmark evaluating protein foundation models across 38 tasks spanning 8 key areas of protein science. Through hundreds of experiments on 17 state-of-the-art models across 38 tasks, PFMBench reveals the inherent correlations between tasks, identifies top-performing models, and provides a streamlined evaluation protocol. Code will be released upon acceptance.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=KJNgtPNxtv", "arxiv_id": "2506.14796", "paper_pdf": "papers/KJNgtPNxtv.pdf", "paper_pdf_sha256": "a1f3e770e11eb3898ddf5a8d939558beddc2c256f4cfe4e08e437382c3748bc9", "paper_pdf_bytes": 1891782, "paper_pdf_source": "openreview", "code_url": "https://github.com/biomap-research/PFMBench", "code_repository": "biomap-research/PFMBench", "code_commit": "53758ffcbdf1d79b5d125383e4dd52d6fd59d2a1", "code_archive": "repos/KJNgtPNxtv.zip", "code_archive_sha256": "16607a6cff798f162961d6a3f0f8254cfbf6e8bc28b4ff45b5a97f846e00dfac", "code_archive_bytes": 3172484, "code_file_count": 181, "code_extensions": {".py": 180, ".sh": 1}, "github_disk_usage_kb": 3107, "github_languages": {"Python": 1495687, "Shell": 9029}, "github_archived": false, "github_pushed_at": "2025-08-01T05:25:58Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/pfmbench-protein-foundation-model-benchmark"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "NEEtm5laNK1", "year": 2023, "status": "rejected", "title": "CHiLS: Zero-Shot Image Classification with Hierarchical Label Sets", "authors": ["Zachary Novack", "Saurabh Garg", "Zachary Chase Lipton"], "authorids": ["~Zachary_Novack1", "~Saurabh_Garg3", "~Zachary_Chase_Lipton1"], "authors_source": "OpenReview API", "abstract": "Open vocabulary models (e.g. CLIP) have shown strong performance on zeroshot classification through their ability generate embeddings for each class based on their (natural language) names. Prior work has focused on improving the accuracy of these models through prompt engineering or by incorporating a small amount of labeled downstream data (via finetuning). In this paper, we propose Classification with Hierarchical Label Sets (or CHiLS), an alternative strategy that proceeds in three steps: (i) for each class, produce a set of subclasses, using either existing label hierarchies or by querying GPT-3; (ii) perform the standard zero-shot CLIP procedure as though these subclasses were the labels of interest; (iii) map the predicted subclass back to its parent to produce the final prediction. Across numerous datasets, CHiLS leads to improved accuracy yielding gains of over 30% in situations where known hierarchies are available and more modest gains when they are not. CHiLS is simple to implement within existing CLIP pipelines and requires no additional training cost.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=NEEtm5laNK1", "arxiv_id": "2302.02551", "paper_pdf": "papers/NEEtm5laNK1.pdf", "paper_pdf_sha256": "ecce8401113a2293776e89aa54bf0c86402b873aa6c75620417a27b53ed98f23", "paper_pdf_bytes": 2756789, "paper_pdf_source": "openreview", "code_url": "https://github.com/acmi-lab/CHILS", "code_repository": "acmi-lab/CHILS", "code_commit": "c90ce90db72e1c0f01994c00f40b0f4c3a6983c2", "code_archive": "repos/NEEtm5laNK1.zip", "code_archive_sha256": "657600cd66c08ecba190063398f16f5cdcca35c3d941fe8b2c22ab263fc2e7bb", "code_archive_bytes": 358625, "code_file_count": 13, "code_extensions": {".py": 13}, "github_disk_usage_kb": 336, "github_languages": {"Python": 109978}, "github_archived": false, "github_pushed_at": "2023-06-04T19:32:03Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/chils-zero-shot-image-classification-with"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "gzeruP-0J29", "year": 2022, "status": "rejected", "title": "Revisiting and Advancing Fast Adversarial Training Through the lens of Bi-Level Optimization", "authors": ["Yihua Zhang", "Guanhua Zhang", "Prashant Khanduri", "Mingyi Hong", "Shiyu Chang", "Sijia Liu"], "authorids": ["~Yihua_Zhang1", "~Guanhua_Zhang1", "~Prashant_Khanduri1", "~Mingyi_Hong1", "~Shiyu_Chang2", "~Sijia_Liu1"], "authors_source": "OpenReview API", "abstract": "Adversarial training (AT) has become a widely recognized defense mechanism to improve the robustness of deep neural networks against adversarial attacks. It is originated from solving a min-max optimization problem, where the minimizer (i.e., defender) seeks a robust model to minimize the worst-case training loss at the presence of adversarial examples crafted by the maximizer (i.e., attacker). However,the min-max nature makes AT computationally intensive and thus difficult to scale. Thus, the problem of FAST-AT arises. Nearly all the recent progress is achieved based on the following simplification: The iterative attack generation method used in the maximization step of AT is replaced by the simplest one-shot gradient sign-based PGD method. Nevertheless, FAST-AT is far from satisfactory, and it lacks theoretically-grounded design. For example, a FAST-AT method may suffer from robustness catastrophic overfitting when training with strong adversaries.\n\nIn this paper, we foster a technological breakthrough for designing FAST-AT through the lens of bi-level optimization (BLO) instead of min-max optimization. First, we theoretically show that the most commonly-used algorithmic specification of FAST-AT is equivalent to the linearized BLO along the direction given by the sign of input gradient. Second, with the aid of BLO, we develop a new systematic and effective fast bi-level AT framework, termed FAST-BAT, whose algorithm is rigorously derived by leveraging the theory of implicit gradient. In contrast to FAST-AT, FAST-BAT has the least restriction to placing the tradeoff between computation efficiency and adversarial robustness. For example, it is capable of defending sign-based projected gradient descent (PGD) attacks without calling any gradient sign method and explicit robust regularization during training. Furthermore, we empirically show that our method outperforms state-of-the-art FAST-AT baselines. In particular, FAST-BAT can achieve superior model robustness without inducing robustness catastrophic overfitting and losing standard accuracy.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=gzeruP-0J29", "arxiv_id": "2112.12376", "paper_pdf": "papers/gzeruP-0J29.pdf", "paper_pdf_sha256": "aff32c2a8be4979176a71a4b9a26aa17a1897699e8ecea6d146a56b29051a094", "paper_pdf_bytes": 969338, "paper_pdf_source": "openreview", "code_url": "https://github.com/OPTML-Group/Fast-BAT", "code_repository": "OPTML-Group/Fast-BAT", "code_commit": "926599de5b7a99166f725519db6c5042405f9251", "code_archive": "repos/gzeruP-0J29.zip", "code_archive_sha256": "81466f24243f5f3933feb4c691bb23125f10277efb9bf26dad9bd81061462d36", "code_archive_bytes": 744999, "code_file_count": 43, "code_extensions": {".py": 40, ".sh": 3}, "github_disk_usage_kb": 889, "github_languages": {"Shell": 4338110, "Python": 300694}, "github_archived": false, "github_pushed_at": "2022-10-25T21:01:44Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/revisiting-and-advancing-fast-adversarial-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "OJiM1R3jAtZ", "year": 2021, "status": "rejected", "title": "AWAC: Accelerating Online Reinforcement Learning with Offline Datasets", "authors": ["Ashvin Nair", "Murtaza Dalal", "Abhishek Gupta", "Sergey Levine"], "authorids": ["~Ashvin_Nair1", "~Murtaza_Dalal1", "~Abhishek_Gupta1", "~Sergey_Levine1"], "authors_source": "OpenReview API", "abstract": "Reinforcement learning provides an appealing formalism for learning control policies from experience. However, the classic active formulation of reinforcement learning necessitates a lengthy active exploration process for each behavior, making it difficult to apply in real-world settings. If we can instead allow reinforcement learning to effectively use previously collected data to aid the online learning process, where the data could be expert demonstrations or more generally any prior experience, we could make reinforcement learning a substantially more practical tool. While a number of recent methods have sought to learn offline from previously collected data, it remains exceptionally difficult to train a policy with offline data and improve it further with online reinforcement learning. In this paper we systematically analyze why this problem is so challenging, and propose an algorithm that combines sample-efficient dynamic programming with maximum likelihood policy updates, providing a simple and effective framework that is able to leverage large amounts of offline data and then quickly perform online fine-tuning of reinforcement learning policies. We show that our method enables rapid learning of skills with a combination of prior demonstration data and online experience across a suite of difficult dexterous manipulation and benchmark tasks.", "decision": "Reject", "meta_review": null, "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=OJiM1R3jAtZ", "arxiv_id": "2006.09359", "paper_pdf": "papers/OJiM1R3jAtZ.pdf", "paper_pdf_sha256": "668b265c09b89d596ad7a980bc55a88bd41c7ccd897500364752ccd1ac2a195e", "paper_pdf_bytes": 3787956, "paper_pdf_source": "openreview", "code_url": "https://github.com/rail-berkeley/rlkit", "code_repository": "rail-berkeley/rlkit", "code_commit": "ac45a9db24b89d97369bef302487273bcc3e3d84", "code_archive": "repos/OJiM1R3jAtZ.zip", "code_archive_sha256": "7d0c4965c00baaeb3ed746a9e351a1471517b0eef699230055d20c5cf7711ecb", "code_archive_bytes": 681913, "code_file_count": 190, "code_extensions": {".py": 190}, "github_disk_usage_kb": 1052, "github_languages": {"Python": 861329, "Dockerfile": 3338}, "github_archived": false, "github_pushed_at": "2024-06-17T17:33:51Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/accelerating-online-reinforcement-learning"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "S1xLuRVFvr", "year": 2020, "status": "rejected", "title": "Visual Explanation for Deep Metric Learning", "authors": ["Sijie Zhu", "Taojiannan Yang", "Chen Chen"], "authorids": ["szhu3@uncc.edu", "tyang30@uncc.edu", "chen.chen@uncc.edu"], "authors_source": "OpenReview API", "abstract": "This work explores the visual explanation for deep metric learning and its applications. As an important problem for learning representation, metric learning has attracted much attention recently, while the interpretation of such model is not as well studied as classification. To this end, we propose an intuitive idea to show where contributes the most to the overall similarity of two input images by decomposing the final activation. Instead of only providing the overall activation map of each image, we propose to generate point-to-point activation intensity between two images so that the relationship between different regions is uncovered. We show that the proposed framework can be directly deployed to a large range of metric learning applications and provides valuable information for understanding the model. Furthermore, our experiments show its effectiveness on two potential applications, i.e. cross-view pattern discovery and interactive retrieval. ", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=S1xLuRVFvr", "arxiv_id": "1909.12977", "paper_pdf": "papers/S1xLuRVFvr.pdf", "paper_pdf_sha256": "84808360cc28fe67885d247b9bf46e6133c3293f951efd695db6baf46bacc77e", "paper_pdf_bytes": 3031392, "paper_pdf_source": "openreview", "code_url": "https://github.com/Jeff-Zilence/Explain_Metric_Learning", "code_repository": "Jeff-Zilence/Explain_Metric_Learning", "code_commit": "c289f8d1d690dd8c14c2aca3542ca7d714c74d6b", "code_archive": "repos/S1xLuRVFvr.zip", "code_archive_sha256": "7cd7c94672149c8d14ba71482a22e8dcf8cfb1ae6082516779d22fd97e17fa37", "code_archive_bytes": 3115704, "code_file_count": 12, "code_extensions": {".py": 12}, "github_disk_usage_kb": 3030, "github_languages": {"Python": 113280}, "github_archived": false, "github_pushed_at": "2022-10-30T18:01:22Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/visual-explanation-for-deep-metric-learning"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "3SAXoxEdCK", "year": 2026, "status": "rejected", "title": "On Path to Multimodal Historical Reasoning: HistBench and HistAgent", "authors": ["Jiahao Qiu", "Fulian Xiao", "Yimin Wang", "Yuchen Mao", "Yijia Chen", "Xinzhe Juan", "Siran Wang", "Xuan Qi", "Tongcheng Zhang", "Zixin Yao", "Jiacheng Guo", "Yifu Lu", "Charles Argon", "Jundi Cui", "Daixin Chen", "Junran Zhou", "Shuyao Zhou", "Zhanpeng Zhou", "Ling Yang", "Shilong Liu", "Hongru WANG", "Kaixuan Huang", "xun jiang", "Xi Gao", "Mengdi Wang"], "authorids": ["~Jiahao_Qiu1", "~Fulian_Xiao1", "~Yimin_Wang5", "~Yuchen_Mao3", "~Yijia_Chen5", "~Xinzhe_Juan1", "~Siran_Wang1", "~Xuan_Qi3", "~Tongcheng_Zhang1", "~Zixin_Yao2", "~Jiacheng_Guo1", "~Yifu_Lu1", "~Charles_Argon1", "~Jundi_Cui1", "~Daixin_Chen1", "~Junran_Zhou1", "~Shuyao_Zhou1", "~Zhanpeng_Zhou1", "~Ling_Yang1", "~Shilong_Liu1", "~Hongru_WANG1", "~Kaixuan_Huang1", "~xun_jiang3", "~Xi_Gao2", "~Mengdi_Wang1"], "authors_source": "OpenReview API", "abstract": "Recent advances in large language models (LLMs) have led to remarkable progress across various domains, yet their capabilities in the humanities, particularly history, remain underexplored. Historical reasoning poses unique challenges for LLMs, involving multimodal source interpretation, temporal inference, and cross-linguistic analysis. Existing general-purpose agents perform well on many current benchmarks but lack the domain expertise needed to address complex historical questions.\nTo address this gap, we introduce HistBench, a new benchmark of 414 high-quality and carefully-reviewed questions stratified by difficulty and designed to evaluate LLM's capacity for historical reasoning. The tasks span a wide range of historical problems—from factual retrieval based on primary sources to interpretive analysis of manuscripts and images, to interdisciplinary challenges involving archaeology, linguistics, or cultural history. Furthermore, the benchmark dataset spans 29 ancient and modern languages and covers a wide range of historical periods and world regions. Finding the poor performance of LLMs and other agents on HistBench, we further present HistAgent, a history-specific agent equipped with carefully designed tools for OCR, translation, archival search, and image understanding in History. On HistBench, HistAgent based on GPT-4o achieves an accuracy of 27.54% pass@1 and 36.47% pass@2, significantly outperforming LLMs with online search and generalist agents, including GPT-4o (18.60%), DeepSeek-R1(14.49%), Grok 3(17.63%) and Open Deep Research by smolagents(20.29% pass@1 and 25.12% pass@2). These results highlight the limitations of existing LLMs and generalist agents and demonstrate the advantages of HistAgent for historical reasoning. Notably, HistAgent also achieves 60.00% pass@1 accuracy on the GAIA benchmark, showing that domain-specific customization doesn't hinder HistAgent's competitive performance on real-world general tasks.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=3SAXoxEdCK", "arxiv_id": "2505.20246", "paper_pdf": "papers/3SAXoxEdCK.pdf", "paper_pdf_sha256": "389f8ec78baac39220a6e3a94ab599aabfe60b4aa9f154c558db5b827cbae847", "paper_pdf_bytes": 1836555, "paper_pdf_source": "openreview", "code_url": "https://github.com/CharlesQ9/HistAgent", "code_repository": "CharlesQ9/HistAgent", "code_commit": "47bbe21dc81618489f5d5929358032883a3fe448", "code_archive": "repos/3SAXoxEdCK.zip", "code_archive_sha256": "c483fe2c8466b8f7f51ea33f20d9950bbe58834aa502e00d7d7d31df373519dd", "code_archive_bytes": 2545677, "code_file_count": 56, "code_extensions": {".py": 55, ".js": 1}, "github_disk_usage_kb": 3152, "github_languages": {"Python": 906511, "JavaScript": 31059}, "github_archived": false, "github_pushed_at": "2025-10-18T01:15:07Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/on-path-to-multimodal-historical-reasoning"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "34xYxTTiM0", "year": 2025, "status": "rejected", "title": "Optimizing Calibration by Gaining Aware of Prediction Correctness", "authors": ["Yuchi Liu", "Lei Wang", "Yuli Zou", "James Zou", "Liang Zheng"], "authorids": ["~Yuchi_Liu1", "~Lei_Wang20", "~Yuli_Zou1", "~James_Zou1", "~Liang_Zheng4"], "authors_source": "OpenReview API", "abstract": "Model calibration aims to align confidence with prediction correctness. The Cross-Entropy (CE) loss is widely used for calibrator training, which enforces the model to increase confidence on the ground truth class. However, we find the CE loss has intrinsic limitations. For example, for a narrow misclassification, a calibrator trained by the CE loss often produces high confidence on the wrongly predicted class (e.g., a test sample is wrongly classified and its softmax score on the ground truth class is around 0.4), which is undesirable. In this paper, we propose a new post-hoc calibration objective derived from the aim of calibration. Intuitively, the proposed objective function asks that the calibrator decrease model confidence on wrongly predicted samples and increase confidence on correctly predicted samples. \nBecause a sample itself has insufficient ability to indicate correctness, we use its transformed versions (e.g., rotated, greyscaled, and color-jittered) during calibrator training. Trained on an in-distribution validation set and tested with isolated, individual test samples, \nour method achieves competitive calibration performance on both in-distribution and out-of-distribution test sets compared with the state of the art. Further, our analysis points out the difference between our method and commonly used objectives such as CE loss and Mean Square Error (MSE) loss, where the latters sometimes deviates from the calibration aim.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=34xYxTTiM0", "arxiv_id": "2404.13016", "paper_pdf": "papers/34xYxTTiM0.pdf", "paper_pdf_sha256": "9489674b98bee80691e307b0d2c9341498857dfbba96d439d93edb72850682c0", "paper_pdf_bytes": 3912860, "paper_pdf_source": "openreview", "code_url": "https://github.com/liuyvchi/Correctness-aware-calibration", "code_repository": "liuyvchi/Correctness-aware-calibration", "code_commit": "6159521ed5ba759b6669372da6990cd4e83abf34", "code_archive": "repos/34xYxTTiM0.zip", "code_archive_sha256": "a27f9edc46742905cb499f07e602e5a55b567f1774fb390f63879b80814e6b6d", "code_archive_bytes": 210809, "code_file_count": 10, "code_extensions": {".py": 10}, "github_disk_usage_kb": 233, "github_languages": {"Python": 96437}, "github_archived": false, "github_pushed_at": "2025-02-12T05:15:03Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/optimizing-calibration-by-gaining-aware-of"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "RtOTTdWbZd", "year": 2024, "status": "rejected", "title": "Fine-Tuning Language Models with Advantage-Induced Policy Alignment", "authors": ["Banghua Zhu", "Hiteshi Sharma", "Felipe Vieira Frujeri", "Shi Dong", "Chenguang Zhu", "Michael Jordan", "Jiantao Jiao"], "authorids": ["~Banghua_Zhu1", "~Hiteshi_Sharma1", "~Felipe_Vieira_Frujeri1", "~Shi_Dong1", "~Chenguang_Zhu1", "~Michael_Jordan1", "~Jiantao_Jiao1"], "authors_source": "OpenReview API", "abstract": "Reinforcement learning from human feedback (RLHF) has emerged as a reliable approach to aligning large language models (LLMs) to human preferences.\nAmong the plethora of RLHF techniques, proximal policy optimization (PPO) is of the most widely used methods. Despite its popularity, however, PPO may suffer from mode collapse, instability, and poor sample efficiency.\nWe show that these issues can be alleviated by a novel algorithm that we refer to as Advantage-Induced Policy Alignment (APA), which leverages a squared error loss function based on the estimated advantages.\nWe demonstrate empirically that APA consistently outperforms PPO in language tasks by a large margin, when a separate reward model is employed as the evaluator.\nIn addition, compared with PPO, APA offers a more stable form of control over the deviation from the model's initial policy, ensuring that the model improves its performance without collapsing to deterministic output.\nIn addition to empirical results, we also provide a theoretical justification supporting the design of our loss function.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=RtOTTdWbZd", "arxiv_id": "2306.02231", "paper_pdf": "papers/RtOTTdWbZd.pdf", "paper_pdf_sha256": "83514553cd0be92ef16ed199f6e16be07bbc108eb298e68cc027da105dc7ecf7", "paper_pdf_bytes": 1540874, "paper_pdf_source": "openreview", "code_url": "https://github.com/microsoft/RLHF-APA", "code_repository": "microsoft/RLHF-APA", "code_commit": "cfe7da20dbc6cf9753825a6acd9c3367fd22f7fa", "code_archive": "repos/RtOTTdWbZd.zip", "code_archive_sha256": "848edbbee8d0170d54844d66a473176887102a8c8e36480a779f763610e7d7d5", "code_archive_bytes": 146722, "code_file_count": 50, "code_extensions": {".py": 50}, "github_disk_usage_kb": 157, "github_languages": {"Python": 374040}, "github_archived": true, "github_pushed_at": "2023-08-11T20:43:30Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/fine-tuning-language-models-with-advantage"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "3tYvDb4dwab", "year": 2023, "status": "rejected", "title": "Understanding Self-Supervised Pretraining with Part-Aware Representation Learning", "authors": ["Jiyang Qi", "Jie Zhu", "Mingyu Ding", "Xiaokang Chen", "Ping Luo", "Leye Wang", "Xinggang Wang", "Wenyu Liu", "Jingdong Wang"], "authorids": ["~Jiyang_Qi1", "~Jie_Zhu3", "~Mingyu_Ding1", "~Xiaokang_Chen1", "~Ping_Luo2", "~Leye_Wang1", "~Xinggang_Wang1", "~Wenyu_Liu3", "~Jingdong_Wang1"], "authors_source": "OpenReview API", "abstract": "In this paper, we are interested in understanding self-supervised pretraining through studying the capability that self-supervised representation pretraining methods learn part-aware representations. The study is mainly motivated by that random views, used in contrastive learning, and random masked (visible) patches, used in masked image modeling, are often about object parts.\n\nWe explain that masked image modeling is a part-to-part task: the masked patches of the object are hallucinated from the visible patches, and that contrastive learning is a part-to-whole task: the projection layer hallucinates the whole object representation from the object part representation learned from the encoder. The explanation suggests that the self-supervised pretrained encoder is required to understand the object part. We empirically compare the off-the-shelf encoders pretrained with several representative methods on object-level recognition and part-level recognition. The results show that the fully-supervised model outperforms self-supervised models for object-level recognition, and most self-supervised contrastive learning and masked image modeling methods outperform the fully-supervised method for part-level recognition. It is observed that the combination of contrastive learning and masked image modeling further improves the performance.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=3tYvDb4dwab", "arxiv_id": "2301.11915", "paper_pdf": "papers/3tYvDb4dwab.pdf", "paper_pdf_sha256": "dcb1fcbf7550f9f1bd5121ba5f2a965d435cd4029e65151650b3a87b8e073c70", "paper_pdf_bytes": 7287188, "paper_pdf_source": "openreview", "code_url": "https://github.com/JiePKU/understand-ssl-part-aware", "code_repository": "JiePKU/understand-ssl-part-aware", "code_commit": "ffea818117952525aa7cb05d793c07e94b2f34fb", "code_archive": "repos/3tYvDb4dwab.zip", "code_archive_sha256": "c966c3b925f0ce8ae36631b7defdfeb6f824a4afda2399fbbe054beb9ef39dca", "code_archive_bytes": 448303, "code_file_count": 226, "code_extensions": {".py": 221, ".sh": 5}, "github_disk_usage_kb": 336, "github_languages": {"Python": 1089473, "Shell": 3884, "HTML": 3482}, "github_archived": false, "github_pushed_at": "2024-01-03T03:55:05Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/understanding-self-supervised-pretraining"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "WnOLO1f50MH", "year": 2022, "status": "rejected", "title": "Exploiting Redundancy: Separable Group Convolutional Networks on Lie Groups", "authors": ["David M Knigge", "David W. Romero", "Erik J Bekkers"], "authorids": ["~David_M_Knigge1", "~David_W._Romero1", "~Erik_J_Bekkers1"], "authors_source": "OpenReview API", "abstract": "Group convolutional neural networks (G-CNNs) have been shown to increase parameter efficiency and model accuracy by incorporating geometric inductive biases. In this work, we investigate the properties of representations learned by regular G-CNNs, and show considerable parameter redundancy in group convolution kernels. This finding motivates further weight-tying by sharing convolution kernels over subgroups. To this end, we introduce convolution kernels that are separable over the subgroup and channel dimensions. In order to obtain equivariance to arbitrary affine Lie groups we provide a continuous parameterisation of separable convolution kernels. We evaluate our approach across several vision datasets, and show that our weight sharing leads to improved performance and computational efficiency. In many settings, separable G-CNNs outperform their non-separable counterpart, while only using a fraction of their training time. In addition, thanks to the increase in computational efficiency, we are able to implement G-CNNs equivariant to the $\\mathrm{Sim(2)}$ group; the group of dilations, rotations and translations.  $\\mathrm{Sim(2)}$-equivariance further improves performance on all tasks considered.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=WnOLO1f50MH", "arxiv_id": "2110.13059", "paper_pdf": "papers/WnOLO1f50MH.pdf", "paper_pdf_sha256": "2d18a95128579de97dd96b3db9b84d9940890c57d68c73acd37cc4546124a4a5", "paper_pdf_bytes": 4150341, "paper_pdf_source": "openreview", "code_url": "https://github.com/david-knigge/separable-group-convolutional-networks", "code_repository": "david-knigge/separable-group-convolutional-networks", "code_commit": "e2de2bdb1c382b76bcaf3f24ad21fe35ac2674e8", "code_archive": "repos/WnOLO1f50MH.zip", "code_archive_sha256": "869d013c682133f25f787da1432854efafd17b7ad3af5332c84ce0f13402c048", "code_archive_bytes": 915204, "code_file_count": 71, "code_extensions": {".py": 69, ".ipynb": 2}, "github_disk_usage_kb": 890, "github_languages": {"Python": 252936}, "github_archived": false, "github_pushed_at": "2022-07-20T18:21:58Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/exploiting-redundancy-separable-group-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "m0ECRXO6QlP", "year": 2021, "status": "rejected", "title": "Supervision Accelerates Pre-training in Contrastive Semi-Supervised Learning of Visual Representations", "authors": ["Mido Assran", "Nicolas Ballas", "Lluis Castrejon", "Michael Rabbat"], "authorids": ["~Mido_Assran1", "~Nicolas_Ballas1", "~Lluis_Castrejon1", "~Michael_Rabbat1"], "authors_source": "OpenReview API", "abstract": "We investigate a strategy for improving the efficiency of contrastive learning of visual representations by leveraging a small amount of supervised information during pre-training. We propose a semi-supervised loss, SuNCEt, based on noise-contrastive estimation and neighbourhood component analysis, that aims to distinguish examples of different classes in addition to the self-supervised instance-wise pretext tasks. On ImageNet, we find that SuNCEt can be used to match the semi-supervised learning accuracy of previous contrastive approaches while using less than half the amount of pre-training and compute. Our main insight is that leveraging even a small amount of labeled data during pre-training, and not only during fine-tuning, provides an important signal that can significantly accelerate contrastive learning of visual representations.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=m0ECRXO6QlP", "arxiv_id": "2006.10803", "paper_pdf": "papers/m0ECRXO6QlP.pdf", "paper_pdf_sha256": "cd1343e72e6e4395d0ee7e3105d10365d94918251410fd9996da846a23532e41", "paper_pdf_bytes": 5101968, "paper_pdf_source": "openreview", "code_url": "https://github.com/facebookresearch/suncet", "code_repository": "facebookresearch/suncet", "code_commit": "731547d727b8c94d06c08a7848b4955de3a70cea", "code_archive": "repos/m0ECRXO6QlP.zip", "code_archive_sha256": "8baf35507e1864a2854a87ff854df59b0e351706535e6498e80b4894b5fab416", "code_archive_bytes": 993386, "code_file_count": 14, "code_extensions": {".py": 14}, "github_disk_usage_kb": 1130, "github_languages": {"Python": 148301}, "github_archived": true, "github_pushed_at": "2023-04-28T12:54:13Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/recovering-petaflops-in-contrastive-semi"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "B1gXR3NtwS", "year": 2020, "status": "rejected", "title": "Deep Bayesian Structure Networks", "authors": ["Zhijie Deng", "Yucen Luo", "Jun Zhu", "Bo Zhang"], "authorids": ["dzj17@mails.tsinghua.edu.cn", "luoyc15@mails.tsinghua.edu.cn", "dcszj@tsinghua.edu.cn", "dcszb@tsinghua.edu.cn"], "authors_source": "OpenReview API", "abstract": "Bayesian neural networks (BNNs) introduce uncertainty estimation to deep networks by performing Bayesian inference on network weights. However, such models bring the challenges of inference, and further BNNs with weight uncertainty rarely achieve superior performance to standard models. In this paper, we investigate a new line of Bayesian deep learning by performing Bayesian reasoning on the structure of deep neural networks. Drawing inspiration from the neural architecture search, we define the network structure as random weights on the redundant operations between computational nodes, and apply stochastic variational inference techniques to learn the structure distributions of networks. Empirically, the proposed method substantially surpasses the advanced deep neural networks across a range of classification and segmentation tasks. More importantly, our approach also preserves benefits of Bayesian principles, producing improved uncertainty estimation than the strong baselines including MC dropout and variational BNNs algorithms (e.g. noisy EK-FAC). ", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=B1gXR3NtwS", "arxiv_id": null, "paper_pdf": "papers/B1gXR3NtwS.pdf", "paper_pdf_sha256": "ab15cbdaf85181ef806aaebc9f7d61cc4d3bbbcd55349164de9114499965db72", "paper_pdf_bytes": 1276567, "paper_pdf_source": "openreview", "code_url": "https://github.com/anonymousest/DBSN", "code_repository": "anonymousest/DBSN", "code_commit": "79f8cac7f3794e7bddda1902e14d365d58f50f11", "code_archive": "repos/B1gXR3NtwS.zip", "code_archive_sha256": "a8fa61537adf75c786973a3906d89bdeac63b6eb346d877009e03b80945f7a38", "code_archive_bytes": 3240763, "code_file_count": 59, "code_extensions": {".py": 59}, "github_disk_usage_kb": 3107, "github_languages": {"Python": 455485}, "github_archived": false, "github_pushed_at": "2019-11-13T02:54:07Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/deep-bayesian-structure-networks"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "4TFfiG17ec", "year": 2026, "status": "rejected", "title": "Thanos: A Block-wise Pruning Algorithm for Efficient Large Language Model Compression", "authors": ["Ivan Ilin", "Peter Richtárik"], "authorids": ["~Ivan_Ilin1", "~Peter_Richtárik1"], "authors_source": "OpenReview API", "abstract": "This paper presents Thanos, a novel weight-pruning algorithm designed to reduce the memory footprint and enhance the computational efficiency of large language models (LLMs) by removing redundant weights while maintaining accuracy. Thanos introduces a block-wise pruning strategy with adaptive masks that dynamically adjust to weight importance, enabling flexible sparsity patterns and structured formats, such as n:m sparsity, optimized for hardware acceleration. Experimental evaluations demonstrate that Thanos achieves state-of-the-art performance in structured pruning and outperforms existing methods in unstructured pruning. By providing an efficient and adaptable approach to model compression, Thanos offers a practical solution for deploying large models in resource-constrained environments. The algorithm is publicly available for further research and application.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=4TFfiG17ec", "arxiv_id": "2504.05346", "paper_pdf": "papers/4TFfiG17ec.pdf", "paper_pdf_sha256": "b4f2dc64593f347c399837d1ea68d226bd715ac7b80c14429e311c0a5f44b770", "paper_pdf_bytes": 4680307, "paper_pdf_source": "openreview", "code_url": "https://github.com/vectozavr/thanos", "code_repository": "vectozavr/thanos", "code_commit": "ee4910891ecf071719e2a4406fbdfe53488a5e61", "code_archive": "repos/4TFfiG17ec.zip", "code_archive_sha256": "5461d179ca168ecdf29c27316877d17274b13df39bbe0f5c02082241559e56f9", "code_archive_bytes": 45851, "code_file_count": 17, "code_extensions": {".py": 16, ".sh": 1}, "github_disk_usage_kb": 3195, "github_languages": {"Python": 132502, "Shell": 664}, "github_archived": false, "github_pushed_at": "2026-08-06T17:21:30Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/thanos-a-block-wise-pruning-algorithm-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "VHGZjZmzsO", "year": 2025, "status": "rejected", "title": "Memory-Enhanced Neural Solvers for Efficient Adaptation in Combinatorial Optimization", "authors": ["Felix Chalumeau", "Refiloe Shabe", "Noah De Nicola", "Arnu Pretorius", "Thomas D Barrett", "Nathan Grinsztajn"], "authorids": ["~Felix_Chalumeau1", "~Refiloe_Shabe1", "~Noah_De_Nicola1", "~Arnu_Pretorius1", "~Thomas_D_Barrett1", "~Nathan_Grinsztajn1"], "authors_source": "OpenReview API", "abstract": "Combinatorial Optimization is crucial to numerous real-world applications, yet still presents challenges due to its (NP-)hard nature. Amongst existing approaches, heuristics often offer the best trade-off between quality and scalability, making them suitable for industrial use. While Reinforcement Learning (RL) offers a flexible framework for designing heuristics, its adoption over handcrafted heuristics remains incomplete within industrial solvers. Existing learned methods still lack the ability to adapt to specific instances and fully leverage the available computational budget. The current best methods either rely on a collection of pre-trained policies, or on data-inefficient fine-tuning; hence failing to fully utilize newly available information within the constraints of the budget. In response, we present MEMENTO, an approach that leverages memory to improve the adaptation of neural solvers at inference time. MEMENTO enables updating the action distribution dynamically based on the outcome of previous decisions. We validate its effectiveness on benchmark problems, in particular Traveling Salesman and Capacitated Vehicle Routing, demonstrating its superiority over tree-search and policy-gradient fine-tuning; and showing it can be zero-shot combined with diversity-based solvers. We successfully train all RL auto-regressive solvers on large instances, and show that MEMENTO can scale and is data-efficient. Overall, MEMENTO enables to push the state-of-the-art on 11 out of 12 evaluated tasks.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=VHGZjZmzsO", "arxiv_id": "2406.16424", "paper_pdf": "papers/VHGZjZmzsO.pdf", "paper_pdf_sha256": "695c555e19ee2061f26614e1315cb93ebeae051d3251f729c0e8f23a8151ee06", "paper_pdf_bytes": 1336591, "paper_pdf_source": "openreview", "code_url": "https://github.com/instadeepai/memento", "code_repository": "instadeepai/memento", "code_commit": "10a65e2f09be81155bfa24ba78ebbbb78863fc28", "code_archive": "repos/VHGZjZmzsO.zip", "code_archive_sha256": "df47efb7df63d48027874b5d51ff0d567a38f8bc41c7ff3870148c483a9f4995", "code_archive_bytes": 190368, "code_file_count": 35, "code_extensions": {".py": 31, ".sh": 4}, "github_disk_usage_kb": 236, "github_languages": {"Python": 144938, "Dockerfile": 3904, "Shell": 3654, "Makefile": 1642}, "github_archived": false, "github_pushed_at": "2026-08-21T13:55:35Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/memory-enhanced-neural-solvers-for-efficient"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Dq7iJqKKM7", "year": 2024, "status": "rejected", "title": "Rethinking Independent Cross-Entropy Loss For Graph-Structured Data", "authors": ["Rui Miao", "Kaixiong Zhou", "Yili Wang", "Ninghao Liu", "Ying Wang", "Xin Wang"], "authorids": ["~Rui_Miao2", "~Kaixiong_Zhou1", "~Yili_Wang2", "~Ninghao_Liu2", "~Ying_Wang13", "~Xin_Wang54"], "authors_source": "OpenReview API", "abstract": "Graph neural networks (GNNs) have exhibited prominent performance in learning graph-structured data. Considering node classification task, the individual label distribution conditioned on node representation is used to predict its classes. Based on the i.i.d assumption among node labels, the traditional supervised learning simply sums up cross-entropy losses of the independent training nodes and applies the average loss to optimize GNNs' weights. But different from other data formats, the nodes are naturally connected and their classes are correlated to neighbors at the same cluster. It is found that the independent distribution modeling of node labels restricts GNNs' capability to generalize over the entire graph and defend adversarial attacks. In this work, we propose a new framework, termed joint-cluster supervised learning, to model the joint distribution of each node with its corresponding cluster. Rather than assuming the node labels are independent, we learn the joint distribution of node and cluster labels conditioned on their representations, and train GNNs with the obtained joint loss. In this way, the data-label reference signals extracted from the local cluster explicitly strengthen the discrimination ability on the target node. The extensive experiments on 12 benchmark datasets and 7 backbone models demonstrate that our joint-cluster supervised learning can effectively bolster GNNs' node classification accuracy. Furthermore, being benefited from the reference signals which may be free from spiteful interference, our learning paradigm significantly protects the node classification from being affected by the adversarial attack.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=Dq7iJqKKM7", "arxiv_id": "2405.15564", "paper_pdf": "papers/Dq7iJqKKM7.pdf", "paper_pdf_sha256": "1aad97e8dd8901ad0dade9ee74cde8dce8fececfe00bfd774e9c85ad5e8671fd", "paper_pdf_bytes": 823465, "paper_pdf_source": "openreview", "code_url": "https://github.com/MR9812/Joint-Cluster-Supervised-Learning", "code_repository": "MR9812/Joint-Cluster-Supervised-Learning", "code_commit": "bfa9dccd2f24be018a23c94152804831e2a35cf7", "code_archive": "repos/Dq7iJqKKM7.zip", "code_archive_sha256": "5776367ef7995273e21272934d74da7347b2dcea0f795c4745e7308c539d38c1", "code_archive_bytes": 99382, "code_file_count": 37, "code_extensions": {".py": 34, ".sh": 3}, "github_disk_usage_kb": 162, "github_languages": {"Python": 238618, "Shell": 5059}, "github_archived": false, "github_pushed_at": "2025-10-28T09:42:12Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/rethinking-independent-cross-entropy-loss-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "YdFkY-QHkPl", "year": 2023, "status": "rejected", "title": "Provably Learning Diverse Features in Multi-View Data with Midpoint Mixup", "authors": ["Muthu Chidambaram", "Xiang Wang", "Chenwei Wu", "Rong Ge"], "authorids": ["~Muthu_Chidambaram1", "~Xiang_Wang1", "~Chenwei_Wu1", "~Rong_Ge1"], "authors_source": "OpenReview API", "abstract": "Mixup is a data augmentation technique that relies on training using random convex combinations of data points and their labels. In recent years, Mixup has become a standard primitive used in the training of state-of-the-art image classification models due to its demonstrated benefits over empirical risk minimization with regards to generalization and robustness. In this work, we try to explain some of this success from a feature learning perspective. We focus our attention on classification problems in which each class may have multiple associated features (or views) that can be used to predict the class correctly. Our main theoretical results demonstrate that, for a non-trivial class of data distributions with two features per class, training a 2-layer convolutional network using empirical risk minimization can lead to learning only one feature for almost all classes while training with a specific instantiation of Mixup succeeds in learning both features for every class. We also show empirically that these theoretical insights extend to the practical settings of image benchmarks modified to have additional synthetic features.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=YdFkY-QHkPl", "arxiv_id": "2210.13512", "paper_pdf": "papers/YdFkY-QHkPl.pdf", "paper_pdf_sha256": "d4c1fee17ce9cb685a56995df52747a7a5184055fd3c05ef7f1a9af1ad29873b", "paper_pdf_bytes": 807669, "paper_pdf_source": "openreview", "code_url": "https://github.com/2014mchidamb/midpoint-mixup-multi-view-icml", "code_repository": "2014mchidamb/midpoint-mixup-multi-view-icml", "code_commit": "e81305d0d48a61d8690530fff785dc09daa6d6e2", "code_archive": "repos/YdFkY-QHkPl.zip", "code_archive_sha256": "40ae5582d81ae55bd881d0fba028fd7bbe597204cc9a0436dd6f0d1c2306221a", "code_archive_bytes": 356559, "code_file_count": 9, "code_extensions": {".py": 8, ".sh": 1}, "github_disk_usage_kb": 343, "github_languages": {"Python": 36732, "Shell": 799}, "github_archived": false, "github_pushed_at": "2023-05-27T16:58:50Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/provably-learning-diverse-features-in-multi"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "xs-tJn58XKv", "year": 2022, "status": "rejected", "title": "Learning Stable Classifiers by Transferring Unstable Features", "authors": ["Yujia Bao", "Shiyu Chang", "Regina Barzilay"], "authorids": ["~Yujia_Bao1", "~Shiyu_Chang2", "~Regina_Barzilay1"], "authors_source": "OpenReview API", "abstract": "While unbiased machine learning models are essential for many applications, bias is a human-defined concept that can vary across tasks. Given only input-label pairs, algorithms may lack sufficient information to distinguish stable (causal) features from unstable (spurious) features. However, related tasks often share similar biases -- an observation we may leverage to develop stable classifiers in the transfer setting. In this work, we explicitly inform the target classifier about unstable features in the source tasks. Specifically, we derive a representation that encodes the unstable features by contrasting different data environments in the source task. We achieve robustness by clustering data of the target task according to this representation and minimizing the worst-case risk across these clusters. We evaluate our method on both text and image classifications. Empirical results demonstrate that our algorithm is able to maintain robustness on the target task, outperforming the best baseline by 22.9% in absolute accuracy across 12 transfer settings. Our code and data will be publicly available.", "decision": "Reject", "meta_review": null, "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=xs-tJn58XKv", "arxiv_id": "2106.07847", "paper_pdf": "papers/xs-tJn58XKv.pdf", "paper_pdf_sha256": "83267c3304738c0a79a8b40b8535124cc8a971a4ce0deeb935c898fb72f4f20f", "paper_pdf_bytes": 4384556, "paper_pdf_source": "openreview", "code_url": "https://github.com/YujiaBao/tofu", "code_repository": "YujiaBao/tofu", "code_commit": "94d38031300c80764ee921fc2bfb02040f5459dd", "code_archive": "repos/xs-tJn58XKv.zip", "code_archive_sha256": "b53f65d51dfa703e1738fde1bafd1ba66f099c274788c4a8b2ef30d8d4b3fb7d", "code_archive_bytes": 1186397, "code_file_count": 25, "code_extensions": {".py": 19, ".sh": 6}, "github_disk_usage_kb": 1154, "github_languages": {"Python": 73879, "Shell": 1901}, "github_archived": false, "github_pushed_at": "2022-07-24T22:13:12Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/learning-stable-classifiers-by-transferring"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "7K0UUL9y9lE", "year": 2021, "status": "rejected", "title": "You Only Sample (Almost) Once: Linear Cost Self-Attention Via Bernoulli Sampling", "authors": ["Zhanpeng Zeng", "Yunyang Xiong", "Sathya N. Ravi", "Shailesh Acharya", "Glenn Fung", "Vikas Singh"], "authorids": ["~Zhanpeng_Zeng1", "~Yunyang_Xiong2", "~Sathya_N._Ravi1", "sachary1@amfam.com", "~Glenn_Fung2", "~Vikas_Singh1"], "authors_source": "OpenReview API", "abstract": "Transformer-based models have come to dominate the landscape in a wide range of natural language processing (NLP) applications. The heart of the transformer model is the self-attention mechanism, which captures the interactions of token pairs in the input sequences and consequently, depends quadratically on the input sequence length. It is known that training such models on longer sequences is quite expensive, and often, prohibitively so. We show that a Bernoulli sampling attention mechanism based on Locality Sensitive Hashing (LSH), decreases the quadratic complexity to linear. We bypass the quadratic cost by considering self-attention as a sum of individual tokens associated with Bernoulli random variables that can, in principle, be sampled at once by a single hash (although in practice, this number may be a small constant). This leads to an efficient sampling scheme to estimate self-attention which relies on specific modifications of LSH (based on feasibility of deployment on GPU architectures). We evaluate our proposed algorithm on the GLUE benchmark with standard 512 sequence length and our method achieves comparable or even slightly better performance than a standard pretrained Transformer. To evaluate whether our method can indeed handle longer sequences, we conduct experiments on long sequence (4096) language model pretraining and achieve consistent results as standard self-attention, while observing sizable inference speed-ups and memory savings.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=7K0UUL9y9lE", "arxiv_id": "2111.09714", "paper_pdf": "papers/7K0UUL9y9lE.pdf", "paper_pdf_sha256": "46c61e2b234128f89e3617dce9eb90b4a7d81e64f3090f4cbb2725407b886bed", "paper_pdf_bytes": 2251447, "paper_pdf_source": "openreview", "code_url": "https://github.com/mlpen/YOSO", "code_repository": "mlpen/YOSO", "code_commit": "6000487d9cd8e34519aa47650011f458ca0db64c", "code_archive": "repos/7K0UUL9y9lE.zip", "code_archive_sha256": "09db5c010e79e83a2185b31817876897f144428f5efeb3ce48edb3ebb04c0883", "code_archive_bytes": 1102033, "code_file_count": 69, "code_extensions": {".py": 38, ".h": 13, ".cu": 10, ".cpp": 5, ".sh": 3}, "github_disk_usage_kb": 1153, "github_languages": {"Python": 190445, "Cuda": 101422, "C": 13489, "C++": 11218, "Shell": 2922}, "github_archived": false, "github_pushed_at": "2021-07-01T17:17:33Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/you-only-sample-almost-once-linear-cost-self-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "S1xRnxSYwS", "year": 2020, "status": "rejected", "title": "Goten: GPU-Outsourcing Trusted Execution of Neural Network Training and Prediction", "authors": ["Lucien K.L. Ng", "Sherman S.M. Chow", "Anna P.Y. Woo", "Donald P. H. Wong", "Yongjun Zhao"], "authorids": ["nkl018@ie.cuhk.edu.hk", "smchow@ie.cuhk.edu.hk", "woopuiyung@gmail.com", "foreverjun.zhao@gmail.com"], "authors_source": "OpenReview API", "abstract": "Before we saw worldwide collaborative efforts in training machine-learning models or widespread deployments of prediction-as-a-service, we need to devise an efficient privacy-preserving mechanism which guarantees the privacy of all stakeholders (data contributors, model owner, and queriers). Slaom (ICLR ’19) preserves privacy only for prediction by leveraging both trusted environment (e.g., Intel SGX) and untrusted GPU. The challenges for enabling private training are explicitly left open – its pre-computation technique does not hide the model weights and fails to support dynamic quantization corresponding to the large changes in weight magnitudes during training. Moreover, it is not a truly outsourcing solution since (offline) pre-computation for a job takes as much time as computing the job locally by SGX, i.e., it only works before all pre-computations are exhausted.\n\nWe propose Goten, a privacy-preserving framework supporting both training and prediction. We tackle all the above challenges by proposing a secure outsourcing protocol which 1) supports dynamic quantization, 2) hides the model weight from GPU, and 3) performs better than a pure-SGX solution even if we perform the precomputation online. Our solution leverages a non-colluding assumption which is often employed by cryptographic solutions aiming for practical efficiency (IEEE SP ’13, Usenix Security ’17, PoPETs ’19). We use three servers, which can be reduced to two if the pre-computation is done offline. Furthermore, we implement our tailor-made memory-aware measures for minimizing the overhead when the SGX memory limit is exceeded (cf., EuroSys ’17, Usenix ATC ’19). Compared to a pure-SGX solution, our experiments show that Goten can speed up linear-layer computations in VGG up to 40×, and overall speed up by 8.64× on VGG11.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=S1xRnxSYwS", "arxiv_id": null, "paper_pdf": "papers/S1xRnxSYwS.pdf", "paper_pdf_sha256": "b65c5957a7d73dce54f572e8b1ba09afd52ad92b14b4dd0746c842085cdd0d0e", "paper_pdf_bytes": 658765, "paper_pdf_source": "openreview", "code_url": "https://github.com/goten-team/Goten", "code_repository": "goten-team/Goten", "code_commit": "690f1429b62c70caec72f4010ee5b7a9786f0d25", "code_archive": "repos/S1xRnxSYwS.zip", "code_archive_sha256": "e4409347b81b3ae3eabffd6cfc9e76b2cfc5a51eda7857af46eb96d65d02a305", "code_archive_bytes": 5487816, "code_file_count": 2036, "code_extensions": {".h": 942, ".cpp": 891, ".f": 51, ".cmake": 48, ".py": 37, ".cu": 25, ".c": 21, ".hpp": 7, ".sh": 6, ".cc": 5, ".js": 3}, "github_disk_usage_kb": 3217, "github_languages": {"C++": 18147133, "Fortran": 1326307, "C": 453077, "CMake": 352722, "Cuda": 263716, "Python": 259349, "Assembly": 39018, "HTML": 19500, "Shell": 16801, "Makefile": 9381}, "github_archived": false, "github_pushed_at": "2021-01-01T14:42:15Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/goten-gpu-outsourcing-trusted-execution-of"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "sp1Wr40egd", "year": 2026, "status": "rejected", "title": "To Trust Or Not To Trust Your Vision-Language Model's Prediction", "authors": ["Hao Dong", "Moru Liu", "Jian Liang", "Eleni Chatzi", "Olga Fink"], "authorids": ["~Hao_Dong4", "~Moru_Liu1", "~Jian_Liang1", "~Eleni_Chatzi1", "~Olga_Fink1"], "authors_source": "OpenReview API", "abstract": "Vision-Language Models (VLMs) have demonstrated strong capabilities in aligning visual and textual modalities, enabling a wide range of applications in multimodal understanding and generation. While they excel in zero-shot and transfer learning scenarios, VLMs remain susceptible to misclassification, often yielding confident yet incorrect predictions. This limitation poses a significant risk in safety-critical domains, where erroneous predictions can lead to severe consequences. In this work, we introduce TrustVLM, a training-free framework designed to address the critical challenge of estimating when VLM’s predictions can be trusted. Motivated by the observed modality gap in VLMs and the insight that certain concepts are more distinctly represented in the image embedding space, we propose a novel confidence-scoring function that leverages this space to improve misclassification detection. We rigorously evaluate our approach across 17 diverse datasets, employing 4 architectures and 2 VLMs, and demonstrate state-of-the-art performance, with improvements of up to 51.87% in AURC, 9.14% in AUROC, and 32.42% in FPR95 compared to existing baselines. By improving the reliability of the model without requiring retraining, TrustVLM paves the way for safer deployment of VLMs in real-world applications. The code is available in Supplementary Material.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=sp1Wr40egd", "arxiv_id": "2505.23745", "paper_pdf": "papers/sp1Wr40egd.pdf", "paper_pdf_sha256": "8378b06fee2da516ef0d045432daa4c13042bb62926a44ffdc6d336f9ea36061", "paper_pdf_bytes": 5224967, "paper_pdf_source": "openreview", "code_url": "https://github.com/EPFL-IMOS/TrustVLM", "code_repository": "EPFL-IMOS/TrustVLM", "code_commit": "0a0caa60fba35624cde966c2a4b8c71c55ba5038", "code_archive": "repos/sp1Wr40egd.zip", "code_archive_sha256": "e44415f703baa2100a4f146c3fdd550d1f5e6ffa6f8b4d2e0eff6493b07c74c3", "code_archive_bytes": 3275633, "code_file_count": 19, "code_extensions": {".py": 17, ".sh": 2}, "github_disk_usage_kb": 3267, "github_languages": {"Python": 166413, "Shell": 2757}, "github_archived": false, "github_pushed_at": "2025-05-30T01:56:37Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/to-trust-or-not-to-trust-your-vision-language"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "QByW8EYEtt", "year": 2025, "status": "rejected", "title": "Worse than Random? An Embarrassingly Simple Probing Evaluation of Large Multimodal Models in Medical VQA", "authors": ["Qianqi Yan", "Xuehai He", "Xiang Yue", "Xin Eric Wang"], "authorids": ["~Qianqi_Yan1", "~Xuehai_He1", "~Xiang_Yue1", "~Xin_Eric_Wang2"], "authors_source": "OpenReview API", "abstract": "Large Multimodal Models (LMMs) have shown remarkable progress in medical Visual Question Answering (Med-VQA), achieving high accuracy on existing benchmarks. However, their reliability under robust evaluation is questionable. This study reveals that state-of-the-art models perform worse than random guessing on medical diagnosis questions when subjected to simple probing evaluation. To address this critical evaluation problem, we introduce the Probing Evaluation for Medical Diagnosis (ProbMed) dataset to rigorously assess LMM performance in medical imaging through probing evaluation and procedural diagnosis. Particularly, probing evaluation features pairing original questions with negation questions with hallucinated attributes, while procedural diagnosis requires reasoning across various diagnostic dimensions for each image, including modality recognition, organ identification, clinical findings, abnormalities, and positional grounding. Our evaluation reveals that top-performing models like GPT-4o, GPT-4V, and Gemini Pro perform worse than random guessing on specialized diagnostic questions, indicating significant limitations in handling fine-grained medical inquiries. We further investigate the underperformance of open-source models (e.g., LLaVA, LLaVA-Med, and Med-Flamingo) through an ablation study. This study reveals that poor visual understanding is a primary bottleneck, which can be mitigated by adding visual descriptions generated by GPT-4o, leading to an average performance improvement of 9.44%. These findings underscore the urgent need for more robust evaluation methods and domain-specific expertise to ensure LMM reliability in critical medical fields.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=QByW8EYEtt", "arxiv_id": "2405.20421", "paper_pdf": "papers/QByW8EYEtt.pdf", "paper_pdf_sha256": "4f73495ccc14076f934defa2a09a0db80d1d6d7cc7fc35356e3187e7f6182fff", "paper_pdf_bytes": 2811135, "paper_pdf_source": "openreview", "code_url": "https://github.com/UCSB-AI/ProbMed", "code_repository": "UCSB-AI/ProbMed", "code_commit": "36f91fbc865fd3cd1f1522c4472d56f5cb4df642", "code_archive": "repos/QByW8EYEtt.zip", "code_archive_sha256": "6a053df3b420be79a4daa017a88e0225693b8fcf26693d25798fc27c0abfb468", "code_archive_bytes": 206000, "code_file_count": 13, "code_extensions": {".py": 11, ".sh": 2}, "github_disk_usage_kb": 237, "github_languages": {"Python": 61789, "Shell": 5075}, "github_archived": false, "github_pushed_at": "2026-05-12T08:00:40Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/worse-than-random-an-embarrassingly-simple"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "zIJFG7wW2d", "year": 2024, "status": "rejected", "title": "Agent Instructs Large Language Models to be General Zero-Shot Reasoners", "authors": ["Nicholas R Crispino", "Kyle Montgomery", "Fankun Zeng", "Dawn Song", "Chenguang Wang"], "authorids": ["~Nicholas_R_Crispino1", "~Kyle_Montgomery1", "~Fankun_Zeng1", "~Dawn_Song1", "~Chenguang_Wang1"], "authors_source": "OpenReview API", "abstract": "We introduce a method to improve the zero-shot reasoning abilities of large language models on general language understanding tasks. Unlike existing zero-shot reasoning approaches that are often suboptimal for general tasks, we build an autonomous agent to generate task-specific instructions to optimize the reasoning performance of large language models. We show our agent instructions further unleash the zero-shot reasoning abilities of large language models to more tasks. We study the performance of our method on a wide set of datasets spanning generation, classification, and reasoning. We show that our method generalizes to most tasks and obtains state-of-the-art zero-shot performance on 20 of the 29 datasets that we evaluate. For instance, our method boosts the performance of state-of-the-art large language models by a large margin, including Vicuna-13b (13.3%), Llama-2-70b-chat (23.2%), and GPT-3.5 Turbo (17.0%). Compared to zero-shot chain of thought, our improvement in reasoning is striking, with an average increase of 10.4%. With our method, Llama-2-70b-chat outperforms zero-shot GPT-3.5 Turbo by 10.2%. The code is available at https://anonymous.4open.science/r/AgentInstruct_ICLR2024.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=zIJFG7wW2d", "arxiv_id": "2310.03710", "paper_pdf": "papers/zIJFG7wW2d.pdf", "paper_pdf_sha256": "5ab79234bea60aed36995a89da6b2a2d8eb8e3223d472f35f13a5b90255fb1ca", "paper_pdf_bytes": 1892943, "paper_pdf_source": "openreview", "code_url": "https://github.com/wang-research-lab/agentinstruct", "code_repository": "wang-research-lab/agentinstruct", "code_commit": "0006a4571280be73c84b05b51a94aedc8ea4768f", "code_archive": "repos/zIJFG7wW2d.zip", "code_archive_sha256": "65ad91611d47f5ef4bd36f0b23ec92939fe9c9d2dda4a6baf850cfdaea728907", "code_archive_bytes": 220542, "code_file_count": 54, "code_extensions": {".py": 45, ".sh": 9}, "github_disk_usage_kb": 162, "github_languages": {"Python": 483273, "Shell": 9321}, "github_archived": false, "github_pushed_at": "2025-10-20T22:45:07Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/agent-instructs-large-language-models-to-be"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "P63GxgD7LIl", "year": 2023, "status": "rejected", "title": "TransFool: An Adversarial Attack against Neural Machine Translation Models", "authors": ["Sahar Sadrizadeh", "Pascal Frossard", "Ljiljana Dolamic"], "authorids": ["~Sahar_Sadrizadeh1", "~Pascal_Frossard1", "~Ljiljana_Dolamic1"], "authors_source": "OpenReview API", "abstract": "Deep neural networks have been shown to be vulnerable to small perturbations of their inputs known as adversarial attacks. In this paper, we consider the particular task of Neural Machine Translation (NMT), where security is often critical. We investigate the vulnerability of NMT models to adversarial attacks and propose a new attack algorithm called TransFool. It builds on a multi-term optimization problem and a gradient projection step to compute adversarial examples that fool NMT models. By integrating the embedding representation of a language model in the proposed attack, we generate fluent adversarial examples in the source language that maintain a high level of semantic similarity with the clean samples and render the attack largely undetectable. Experimental results demonstrate that, for multiple translation tasks and different NMT architectures, our white-box attack can severely degrade the translation quality for more than 60% of the sentences while the semantic similarity between the original sentence and the adversarial example stays very high. Moreover, we show that the proposed attack is transferable to unknown target models and can fool those quite easily. Finally, our method leads to improvement in terms of success rate, semantic similarity, and fluency compared to the existing attack strategies both in white-box and black-box settings. Hence, TransFool permits to better characterize the vulnerability of NMT systems and outlines the necessity to design strong defense mechanisms and more robust NMT systems for real-life applications.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=P63GxgD7LIl", "arxiv_id": "2302.00944", "paper_pdf": "papers/P63GxgD7LIl.pdf", "paper_pdf_sha256": "e4a0eee24850a30ad1abcb089275af40c10362e7acfb35a67396ba9f305697d1", "paper_pdf_bytes": 2069662, "paper_pdf_source": "openreview", "code_url": "https://github.com/sssadrizadeh/TransFool", "code_repository": "sssadrizadeh/TransFool", "code_commit": "8b0d9de5c6f60c4017e5ddcc81c8f14dd5c9a85b", "code_archive": "repos/P63GxgD7LIl.zip", "code_archive_sha256": "f9ea93a3097aab4c296a260a78e39ab8d168afa04ae76eea1a07fc1b47819e5b", "code_archive_bytes": 373518, "code_file_count": 19, "code_extensions": {".py": 19}, "github_disk_usage_kb": 350, "github_languages": {"Python": 153487}, "github_archived": false, "github_pushed_at": "2024-06-26T12:59:17Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/transfool-an-adversarial-attack-against"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "qSTEPv2uLR8", "year": 2022, "status": "rejected", "title": "Physics Informed Convex Artificial Neural Networks (PICANNs) for Optimal Transport based Density Estimation", "authors": ["Amanpreet Singh", "Martin Bauer", "Sarang Joshi"], "authorids": ["~Amanpreet_Singh3", "~Martin_Bauer2", "~Sarang_Joshi1"], "authors_source": "OpenReview API", "abstract": "Optimal Mass Transport (OMT) is a well studied problem with a variety of applications in a  diverse set of fields ranging from Physics to Computer Vision and in particular Statistics and Data Science. Since the original formulation of Monge in 1781 significant theoretical progress been made on the existence, uniqueness and properties of the optimal transport maps. The actual numerical computation of the transport maps, particularly in high dimensions, remains a challenging problem. In the past decade several neural network based algorithms have been proposed to tackle this task. In this paper, building on recent developments of input convex neural networks and physics informed neural networks for solving PDE's, we propose a new Deep Learning approach to solve the continuous OMT problem. Our framework is based on Brenier's theorem, which reduces the continuous OMT problem to that of solving a non-linear PDE of Monge-Ampere type whose solution is a convex function. To demonstrate the accuracy of our framework we compare our method to several other deep learning based algorithms. We then focus on applications to the ubiquitous density estimation and generative modeling tasks in statistics and machine learning. Finally as an example we present how our framework can be incorporated with an autoencoder to estimate an effective probabilistic generative model.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=qSTEPv2uLR8", "arxiv_id": "2104.01194", "paper_pdf": "papers/qSTEPv2uLR8.pdf", "paper_pdf_sha256": "bb006b47bb4a52f4e3ea441a4edf779a4cff57f9bd7aa4da993f61ca1a286a33", "paper_pdf_bytes": 1660285, "paper_pdf_source": "openreview", "code_url": "https://github.com/4m4npr33t/PICANNs", "code_repository": "4m4npr33t/PICANNs", "code_commit": "a1d52dab402427414f6f38ce0b5d9e47fafb58fa", "code_archive": "repos/qSTEPv2uLR8.zip", "code_archive_sha256": "291338b716a7d0084f09c094d2ce76b7f64fc6aadc994738b83263c560eacecd", "code_archive_bytes": 1176871, "code_file_count": 18, "code_extensions": {".py": 17, ".sh": 1}, "github_disk_usage_kb": 1199, "github_languages": {"Python": 86788, "Shell": 494}, "github_archived": false, "github_pushed_at": "2021-04-02T18:50:14Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/physics-informed-convex-artificial-neural"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "sAX7Z7uIJ_Y", "year": 2021, "status": "rejected", "title": "Calibrated Adversarial Refinement for Stochastic Semantic Segmentation", "authors": ["Elias Kassapis", "Georgi Dikov", "Deepak Gupta", "Cedric Nugteren"], "authorids": ["~Elias_Kassapis1", "~Georgi_Dikov1", "~Deepak_Gupta2", "~Cedric_Nugteren1"], "authors_source": "OpenReview API", "abstract": "Ambiguities in images or unsystematic annotation can lead to multiple valid solutions in semantic segmentation. To learn a distribution over predictions, recent work has explored the use of probabilistic networks. However, these do not necessarily capture the empirical distribution accurately. In this work, we aim to learn a calibrated multimodal predictive distribution, where the empirical frequency of the sampled predictions closely reflects that of the corresponding labels in the training set. To this end, we propose a novel two-stage, cascaded strategy for calibrated adversarial refinement. In the first stage, we explicitly model the data with a categorical likelihood. In the second, we train an adversarial network to sample from it an arbitrary number of coherent predictions. The model can be used independently or integrated into any black-box segmentation framework to facilitate learning of calibrated stochastic mappings. We demonstrate the utility and versatility of the approach by attaining state-of-the-art results on the multigrader LIDC dataset and a modified Cityscapes dataset. In addition, we use a toy regression dataset to show that our framework is not confined to semantic segmentation, and the core design can be adapted to other tasks requiring learning a calibrated predictive distribution.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=sAX7Z7uIJ_Y", "arxiv_id": "2006.13144", "paper_pdf": "papers/sAX7Z7uIJ_Y.pdf", "paper_pdf_sha256": "c01566ad73252c3b012ea4082f09a0f1728ce82e0232db33f06fc1728ad75329", "paper_pdf_bytes": 7897455, "paper_pdf_source": "openreview", "code_url": "https://github.com/EliasKassapis/CARSSS", "code_repository": "EliasKassapis/CARSSS", "code_commit": "ff7ec86aab68c4b9ff8aea171244991bd132d487", "code_archive": "repos/sAX7Z7uIJ_Y.zip", "code_archive_sha256": "f9ae38b3ae77d2843bd67d3019f7238f9b4c4e0b5075304e42c35571a09528f1", "code_archive_bytes": 1204432, "code_file_count": 46, "code_extensions": {".py": 46}, "github_disk_usage_kb": 1158, "github_languages": {"Python": 188372}, "github_archived": false, "github_pushed_at": "2022-12-08T10:10:23Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/calibrated-adversarial-refinement-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "B1gi0TEFDB", "year": 2020, "status": "rejected", "title": "Understanding Top-k Sparsification in Distributed Deep Learning", "authors": ["Shaohuai Shi", "Xiaowen Chu", "Ka Chun Cheung", "Simon See"], "authorids": ["csshshi@comp.hkbu.edu.hk", "chxw@comp.hkbu.edu.hk", "chcheung@nvidia.com", "ssee@nvidia.com"], "authors_source": "OpenReview API", "abstract": "Distributed stochastic gradient descent (SGD) algorithms are widely deployed in training large-scale deep learning models, while the communication overhead among workers becomes the new system bottleneck. Recently proposed gradient sparsification techniques, especially Top-$k$ sparsification with error compensation (TopK-SGD), can significantly reduce the communication traffic without obvious impact on the model accuracy. Some theoretical studies have been carried out to analyze the convergence property of TopK-SGD. However, existing studies do not dive into the details of Top-$k$ operator in gradient sparsification and use relaxed bounds (e.g., exact bound of Random-$k$) for analysis; hence the derived results cannot well describe the real convergence performance of TopK-SGD. To this end, we first study the gradient distributions of TopK-SGD during training process through extensive experiments. We then theoretically derive a tighter bound for the Top-$k$ operator. Finally, we exploit the property of gradient distribution to propose an approximate top-$k$ selection algorithm, which is computing-efficient for GPUs, to improve the scaling efficiency of TopK-SGD by significantly reducing the computing overhead.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=B1gi0TEFDB", "arxiv_id": "1911.08772", "paper_pdf": "papers/B1gi0TEFDB.pdf", "paper_pdf_sha256": "5238e3088a30bc860167a8a7cf2f13d99905c30d37e5e764f574a4d15363319f", "paper_pdf_bytes": 3598631, "paper_pdf_source": "openreview", "code_url": "https://github.com/hclhkbu/GaussianK-SGD", "code_repository": "hclhkbu/GaussianK-SGD", "code_commit": "3da1fb3b62e4067828a965687d5cc7ae7892e742", "code_archive": "repos/B1gi0TEFDB.zip", "code_archive_sha256": "ea37273ea50b3f82c205a0ccc15ae66d0afd9a66cff5dd14eb0545e779c676ae", "code_archive_bytes": 85610, "code_file_count": 36, "code_extensions": {".py": 35, ".sh": 1}, "github_disk_usage_kb": 3729, "github_languages": {"Python": 205644, "Shell": 746}, "github_archived": false, "github_pushed_at": "2019-11-15T18:42:40Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/understanding-top-k-sparsification-in-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "RBktryANQ9", "year": 2026, "status": "rejected", "title": "Dataset Distillers Are Good Label Denoisers In the Wild", "authors": ["Lechao Cheng", "KaifengChen", "Jiyang Li", "Zerun Liu", "Shengeng Tang", "Shufei Zhang", "Zhihui Yang"], "authorids": ["~Lechao_Cheng2", "~KaifengChen2", "~Jiyang_Li1", "~Zerun_Liu1", "~Shengeng_Tang1", "~Shufei_Zhang1", "~Zhihui_Yang3"], "authors_source": "OpenReview API", "abstract": "Dataset distillation aims to synthesize a small set of informative samples that preserve the generalization ability of large datasets. However, its behavior under noisy conditions remains underexplored. In this paper, we systematically study dataset distillation under three representative noise types: symmetric, asymmetric, and natural noise. We first discover that, when the noise ratio exceeds a critical threshold, mainstream distillation methods consistently outperform training on the full noisy dataset using significantly fewer distilled samples. In contrast, under asymmetric noise, the structured label corruption often entangles with semantic features, making it difficult for distilled samples to recover the clean data distribution. We further validate the effectiveness of dataset distillation on real-world noisy datasets, highlighting its robustness under high noise but degraded performance in low-noise settings due to over-compression. To provide theoretical insights, we derive upper and lower bounds on the required images per class (IPC) under each noise type, grounded in information theory and PAC-Bayes analysis. Our findings offer both empirical and theoretical guidelines for effective distillation in noisy learning scenarios.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=RBktryANQ9", "arxiv_id": "2411.11924", "paper_pdf": "papers/RBktryANQ9.pdf", "paper_pdf_sha256": "8aaa694f3d4ec23318fb3542734657d5c28741454b6718999006a563b2385e82", "paper_pdf_bytes": 1163235, "paper_pdf_source": "openreview", "code_url": "https://github.com/Kciiiman/DD_LNL", "code_repository": "Kciiiman/DD_LNL", "code_commit": "d59a5a59637188c7a16121cc9a9be39544556cd1", "code_archive": "repos/RBktryANQ9.zip", "code_archive_sha256": "383d95f05b1c79f879b1a581cbb6f59225ce0386aa4ac515aac8ab4e199f5043", "code_archive_bytes": 4374038, "code_file_count": 70, "code_extensions": {".py": 70}, "github_disk_usage_kb": 3418, "github_languages": {"Python": 794394}, "github_archived": false, "github_pushed_at": "2025-01-12T05:45:22Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/dataset-distillers-are-good-label-denoisers"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "M8xtZuxqC5", "year": 2025, "status": "rejected", "title": "It Takes Two: On the Seamlessness between Reward and Policy Model in RLHF", "authors": ["TaiMing Lu", "Lingfeng Shen", "Xinyu Yang", "Weiting Tan", "Ryumei Nakada", "Linjun Zhang", "Beidi Chen", "Huaxiu Yao"], "authorids": ["~TaiMing_Lu1", "~Lingfeng_Shen1", "~Xinyu_Yang4", "~Weiting_Tan1", "~Ryumei_Nakada1", "~Linjun_Zhang1", "~Beidi_Chen1", "~Huaxiu_Yao1"], "authors_source": "OpenReview API", "abstract": "Reinforcement Learning from Human Feedback (RLHF) involves training policy models (PMs) and reward models (RMs) to align language models with human preferences. Instead of focusing solely on PMs and RMs independently, we propose to examine their interactions during fine-tuning, introducing the concept of \\textbf{seamlessness}. Our study starts with observing the saturation phenomenon, where continual improvements in RM and PM do not translate into RLHF progress. Our analysis shows that RMs fail to assign proper scores to PM responses, resulting in a 35% mismatch rate with human preferences, highlighting a significant discrepancy between PM and RM. To measure seamlessness between PM and RM without human effort, we propose an automatic metric, SEAM. SEAM quantifies the discrepancies between PM and RM judgments induced by data samples. We validate the effectiveness of SEAM in data selection and model augmentation. Our experiments demonstrate that (1) using SEAM-filtered data for RL training improves RLHF performance by 4.5%, and (2) SEAM-guided model augmentation results in a 4% performance improvement over standard augmentation methods.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=M8xtZuxqC5", "arxiv_id": "2406.07971", "paper_pdf": "papers/M8xtZuxqC5.pdf", "paper_pdf_sha256": "a6181ad1689846cf792d235ea2e27e3367176bf71cd3bba1477d9b5f3a42373f", "paper_pdf_bytes": 1017798, "paper_pdf_source": "openreview", "code_url": "https://github.com/TaiMingLu/seamless", "code_repository": "TaiMingLu/seamless", "code_commit": "9a704f2a28800e162be53e61aee0a11ce4146841", "code_archive": "repos/M8xtZuxqC5.zip", "code_archive_sha256": "f76e0c4acce6fdd5adf45fcfd7ae19bc7a0981b3a84248f701bb8e1dc0eafeb3", "code_archive_bytes": 245646, "code_file_count": 11, "code_extensions": {".py": 11}, "github_disk_usage_kb": 243, "github_languages": {"Python": 54869}, "github_archived": false, "github_pushed_at": "2024-08-26T20:05:53Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/it-takes-two-on-the-seamlessness-between"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "oNkYPgnfHt", "year": 2024, "status": "rejected", "title": "Learning to Intervene on Concept Bottlenecks", "authors": ["David Steinmann", "Wolfgang Stammer", "Felix Friedrich", "Kristian Kersting"], "authorids": ["~David_Steinmann1", "~Wolfgang_Stammer1", "~Felix_Friedrich1", "~Kristian_Kersting1"], "authors_source": "OpenReview API", "abstract": "While deep learning models often lack interpretability, concept bottleneck models (CBMs) provide inherent explanations via their concept representations. Moreover, they allow users to perform interventional interactions on these concepts by updating the concept values and thus correcting the predictive output of the model. Up to this point, these interventions were typically applied to the model just once and then discarded. To rectify this, we present concept bottleneck memory models (CB2Ms), which keep a memory of past interventions. Specifically, CB2Ms leverage a two-fold, differentiable memory to generalize interventions to appropriate novel situations, enabling the model to identify errors and reapply previous interventions. This way, a CB2M learns to automatically improve model performance from a few initially obtained interventions. If no prior human interventions are available, a CB2M can detect potential mistakes of the CBM bottleneck and request targeted interventions. Our experimental evaluations on challenging scenarios like handling distribution shifts and confounded data demonstrate that CB2Ms are able to successfully generalize interventions to unseen data and can indeed identify wrongly inferred concepts. Hence, CB2Ms are a valuable tool for users to provide interactive feedback on CBMs, e.g., by guiding a user’s interaction and requiring fewer interventions.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=oNkYPgnfHt", "arxiv_id": "2308.13453", "paper_pdf": "papers/oNkYPgnfHt.pdf", "paper_pdf_sha256": "efc4342a1a81161c983a12c369902374fb3289a48d06e054e1fa0ec2164f0e62", "paper_pdf_bytes": 469831, "paper_pdf_source": "openreview", "code_url": "https://github.com/ml-research/CB2M", "code_repository": "ml-research/CB2M", "code_commit": "17bb3d5c0682dbcb90d67ae9b3d5ac7f12bb71dd", "code_archive": "repos/oNkYPgnfHt.zip", "code_archive_sha256": "0ba0ee4727973b2556879d127bdd443a5a2e5158589399e0a22545e8e3a12417", "code_archive_bytes": 180534, "code_file_count": 24, "code_extensions": {".py": 23, ".sh": 1}, "github_disk_usage_kb": 170, "github_languages": {"Python": 175767, "Shell": 16336}, "github_archived": false, "github_pushed_at": "2024-05-27T09:55:52Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/learning-to-intervene-on-concept-bottlenecks"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "lrzX-rNuRvw", "year": 2023, "status": "rejected", "title": "Understanding Rare Spurious Correlations in Neural Networks", "authors": ["Yao-Yuan Yang", "Chi-Ning Chou", "Kamalika Chaudhuri"], "authorids": ["~Yao-Yuan_Yang1", "~Chi-Ning_Chou1", "~Kamalika_Chaudhuri1"], "authors_source": "OpenReview API", "abstract": "Neural networks are known to use spurious correlations such as background information for classification. While prior work has looked at spurious correlations that are widespread in the training data, in this work, we investigate how sensitive neural networks are to $rare$ spurious correlations, which may be harder to detect and correct, and may lead to privacy leaks. We introduce spurious patterns correlated with a fixed class to a few training examples and find that it takes only a handful of such examples for the network to learn the correlation. Furthermore, these rare spurious correlations also impact accuracy and privacy. We empirically and theoretically analyze different factors involved in rare spurious correlations and propose mitigation methods accordingly. Specifically, we observe that $\\ell_2$ regularization and adding Gaussian noise to inputs can reduce the undesirable effects.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=lrzX-rNuRvw", "arxiv_id": "2202.05189", "paper_pdf": "papers/lrzX-rNuRvw.pdf", "paper_pdf_sha256": "e2cc2471fa226504b443e31d33a2be91230d32b369561a85f417926cc0bb46cf", "paper_pdf_bytes": 1342183, "paper_pdf_source": "openreview", "code_url": "https://github.com/yangarbiter/rare-spurious-correlation", "code_repository": "yangarbiter/rare-spurious-correlation", "code_commit": "af036940cfcabb14ff70fe1849dcf3f03edb22aa", "code_archive": "repos/lrzX-rNuRvw.zip", "code_archive_sha256": "070323fc40b8ca4e97b1430ea021164c9a196aabbac59bec5579d72a702732c7", "code_archive_bytes": 373759, "code_file_count": 29, "code_extensions": {".py": 24, ".ipynb": 5}, "github_disk_usage_kb": 357, "github_languages": {"Jupyter Notebook": 1180515, "Python": 104529}, "github_archived": false, "github_pushed_at": "2022-06-05T04:37:43Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/understanding-rare-spurious-correlations-in"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "YmONQIWli--", "year": 2022, "status": "rejected", "title": "Gotta Go Fast When Generating Data with Score-Based Models", "authors": ["Alexia Jolicoeur-Martineau", "Ke Li", "Rémi Piché-Taillefer", "Tal Kachman", "Ioannis Mitliagkas"], "authorids": ["~Alexia_Jolicoeur-Martineau1", "~Ke_Li1", "~Rémi_Piché-Taillefer1", "~Tal_Kachman1", "~Ioannis_Mitliagkas1"], "authors_source": "OpenReview API", "abstract": "Score-based (denoising diffusion) generative models have recently gained a lot of success in generating realistic and diverse data. These approaches define a forward diffusion process for transforming data to noise and generate data by reversing it (thereby going from noise to data). Unfortunately, current score-based models generate data very slowly due to the sheer number of score network evaluations required by numerical SDE solvers. \n   \nIn this work, we aim to accelerate this process by devising a more efficient SDE solver. Existing approaches rely on the Euler-Maruyama (EM) solver, which uses a fixed step size. We found that naively replacing it with other SDE solvers fares poorly - they either result in low-quality samples or become slower than EM. To get around this issue, we carefully devise an SDE solver with adaptive step sizes tailored to score-based generative models piece by piece. Our solver requires only two score function evaluations, rarely rejects samples, and leads to high-quality samples. Our approach generates data 2 to 10 times faster than EM while achieving better or equal sample quality. For high-resolution images, our method leads to significantly higher quality samples than all other methods tested. Our SDE solver has the benefit of requiring no step size tuning.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=YmONQIWli--", "arxiv_id": "2105.14080", "paper_pdf": "papers/YmONQIWli--.pdf", "paper_pdf_sha256": "dedf9d52686bacce2e4f6f10ddb437cedf99a4165593118417213a1e02f0e81a", "paper_pdf_bytes": 27898084, "paper_pdf_source": "openreview", "code_url": "https://github.com/AlexiaJM/score_sde_fast_sampling", "code_repository": "AlexiaJM/score_sde_fast_sampling", "code_commit": "5da8f3fe103ee5ac3c3a336f16cc06c9541f0ed9", "code_archive": "repos/YmONQIWli--.zip", "code_archive_sha256": "4bfa837e1daf0c4793bddb3b8fc3c976839c971d45fd7e59060f980218da6d84", "code_archive_bytes": 1164621, "code_file_count": 68, "code_extensions": {".py": 67, ".sh": 1}, "github_disk_usage_kb": 1240, "github_languages": {"Python": 313793, "Shell": 40262}, "github_archived": false, "github_pushed_at": "2021-11-20T21:59:57Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/gotta-go-fast-when-generating-data-with-score"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "avBunqDXFS", "year": 2021, "status": "rejected", "title": "Memory-Efficient Semi-Supervised Continual Learning: The World is its Own Replay Buffer", "authors": ["James Smith", "Jonathan C Balloch", "Yen-Chang Hsu", "Zsolt Kira"], "authorids": ["~James_Smith1", "~Jonathan_C_Balloch1", "~Yen-Chang_Hsu1", "~Zsolt_Kira1"], "authors_source": "OpenReview API", "abstract": "Rehearsal is a critical component for class-incremental continual learning, yet it requires a substantial memory budget. Our work investigates whether we can significantly reduce this memory budget by leveraging unlabeled data from an agent's environment in a realistic and challenging continual learning paradigm. Specifically, we explore and formalize a novel semi-supervised continual learning (SSCL) setting, where labeled data is scarce yet non-i.i.d. unlabeled data from the agent's environment is plentiful. Importantly, data distributions in the SSCL setting are realistic and therefore reflect object class correlations between, and among, the labeled and unlabeled data distributions. We show that a strategy built on pseudo-labeling, consistency regularization, Out-of-Distribution (OoD) detection, and knowledge distillation reduces forgetting in this setting. Our approach, DistillMatch, increases performance over the state-of-the-art by no less than 8.7% average task accuracy and up to a 54.5% increase in average task accuracy in SSCL CIFAR-100 experiments. Moreover, we demonstrate that DistillMatch can save up to 0.23 stored images per processed unlabeled image compared to the next best method which only saves 0.08. Our results suggest that focusing on realistic correlated distributions is a significantly new perspective, which accentuates the importance of leveraging the world's structure as a continual learning strategy.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=avBunqDXFS", "arxiv_id": "2101.09536", "paper_pdf": "papers/avBunqDXFS.pdf", "paper_pdf_sha256": "a243c7f2ef13199a5776f9ad2e7dd477f99429e39b6f68bba1fe84eaedcf2df6", "paper_pdf_bytes": 1635086, "paper_pdf_source": "openreview", "code_url": "https://github.com/GT-RIPL/DistillMatch-SSCL", "code_repository": "GT-RIPL/DistillMatch-SSCL", "code_commit": "e572671fd6994b3c43ad6e46e9efb3588804524c", "code_archive": "repos/avBunqDXFS.zip", "code_archive_sha256": "938dfa7e51feb599dec6be7a12a78c0ef6d0f0f7c9f2ba3f71c63438d9892c8c", "code_archive_bytes": 1195938, "code_file_count": 22, "code_extensions": {".py": 12, ".sh": 10}, "github_disk_usage_kb": 1163, "github_languages": {"Python": 165881, "Shell": 36248}, "github_archived": false, "github_pushed_at": "2022-10-17T11:46:40Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/memory-efficient-semi-supervised-continual-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "B1lda1HtvB", "year": 2020, "status": "rejected", "title": "Feature Selection using Stochastic Gates", "authors": ["Yutaro Yamada", "Ofir Lindenbaum", "Sahand Negahban", "Yuval Kluger"], "authorids": ["yutaro.yamada@yale.edu", "ofirlin@gmail.com", "sahand.negahban@yale.edu", "yuval.kluger@yale.edu"], "authors_source": "OpenReview API", "abstract": "Feature selection problems have been extensively studied in the setting of\nlinear estimation, for instance LASSO, but less emphasis has been placed on\nfeature selection for non-linear functions. In this study, we propose a method\nfor feature selection in high-dimensional non-linear function estimation\nproblems. The new procedure is based on directly penalizing the $\\ell_0$ norm of\nfeatures, or the count of the number of selected features. Our $\\ell_0$ based regularization relies on a continuous relaxation of the Bernoulli distribution, which\nallows our model to learn the parameters of the approximate Bernoulli\ndistributions via gradient descent. The proposed framework simultaneously learns\na non-linear regression or classification function while selecting a small\nsubset of features. We provide an information-theoretic justification for\nincorporating Bernoulli distribution into our approach. Furthermore, we evaluate\nour method using synthetic and real-life data and demonstrate that our approach\noutperforms other embedded methods in terms of predictive performance and feature selection.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=B1lda1HtvB", "arxiv_id": "1810.04247", "paper_pdf": "papers/B1lda1HtvB.pdf", "paper_pdf_sha256": "f241c78e43d9e65868668ebf948b1bd04eedcd13d7f2e1df7ccdc753d3ac9b9c", "paper_pdf_bytes": 3344409, "paper_pdf_source": "openreview", "code_url": "https://github.com/runopti/stg", "code_repository": "runopti/stg", "code_commit": "8b60e349a70d2f392b13d2505c316cdd8868a067", "code_archive": "repos/B1lda1HtvB.zip", "code_archive_sha256": "84f6db32300a48241aa75982bf36ee191e7966b6388698ff70f1edebd874ceb8", "code_archive_bytes": 2535326, "code_file_count": 17, "code_extensions": {".py": 12, ".ipynb": 3, ".r": 2}, "github_disk_usage_kb": 3889, "github_languages": {"Python": 51920, "R": 3571}, "github_archived": false, "github_pushed_at": "2022-02-18T16:44:06Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/feature-selection-using-stochastic-gates"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "jbGGPSI9aO", "year": 2026, "status": "rejected", "title": "AdaSCALE: Adaptive Scaling for OOD detection", "authors": ["Sudarshan Regmi"], "authorids": ["~Sudarshan_Regmi1"], "authors_source": "OpenReview API", "abstract": "The ability of the deep learning model to recognize when a sample falls outside its learned distribution is critical for safe and reliable deployment. Recent state-of-the-art out-of-distribution (OOD) detection methods leverage activation shaping to improve the separation between in-distribution (ID) and OOD inputs. These approaches resort to sample-specific scaling but apply a static percentile threshold across all samples regardless of their nature. In this work, we propose AdaSCALE, an adaptive scaling procedure that dynamically adjusts the percentile threshold based on a sample's estimated OOD likelihood. This estimation leverages our key observation that OOD samples exhibit significantly more pronounced activation shifts at high-magnitude activations under minor perturbation compared to ID samples. AdaSCALE enables stronger scaling for likely ID samples and weaker scaling for likely OOD samples, creating highly separable energy scores. Our approach achieves state-of-the-art OOD detection performance, outperforming the latest rival OptFS by **14.94%** in near-OOD and **21.67%** in far-OOD datasets in average FPR@95 metric in the ImageNet-1k benchmark across eight diverse architectures.", "decision": "Reject", "meta_review": "", "num_reviews": 6, "openreview_url": "https://openreview.net/forum?id=jbGGPSI9aO", "arxiv_id": "2503.08023", "paper_pdf": "papers/jbGGPSI9aO.pdf", "paper_pdf_sha256": "5b2bb73018cfc8df543d8c24727e55fb236ef6264b6134aed40fb7a90eef68f6", "paper_pdf_bytes": 3368404, "paper_pdf_source": "openreview", "code_url": "https://github.com/sudarshanregmi/AdaSCALE", "code_repository": "sudarshanregmi/AdaSCALE", "code_commit": "ed5f639e10520a04f6a83f30a32c060d6f012ea8", "code_archive": "repos/jbGGPSI9aO.zip", "code_archive_sha256": "7ddee61d37e57e3a1bf8d08863d92a3b5538494c1cabcf7ffe649b0528545fa0", "code_archive_bytes": 3913697, "code_file_count": 681, "code_extensions": {".sh": 427, ".py": 250, ".ipynb": 4}, "github_disk_usage_kb": 3484, "github_languages": {"Python": 919586, "Shell": 313563, "Jupyter Notebook": 300629}, "github_archived": false, "github_pushed_at": "2026-05-02T23:12:19Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/adascale-adaptive-scaling-for-ood-detection"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "TjvSFVJdzJ", "year": 2025, "status": "rejected", "title": "Reinforced In-Context Black-Box Optimization", "authors": ["Lei Song", "Chen-Xiao Gao", "Ke Xue", "Chenyang Wu", "Dong Li", "Jianye HAO", "Zongzhang Zhang", "Chao Qian"], "authorids": ["~Lei_Song4", "~Chen-Xiao_Gao1", "~Ke_Xue1", "~Chenyang_Wu1", "~Dong_Li10", "~Jianye_HAO1", "~Zongzhang_Zhang1", "~Chao_Qian1"], "authors_source": "OpenReview API", "abstract": "Black-Box Optimization (BBO) has found successful applications in many fields of science and engineering. Recently, there has been a growing interest in meta-learning particular components of BBO algorithms to speed up optimization and get rid of tedious hand-crafted heuristics. As an extension, learning the entire algorithm from data requires the least labor from experts and can provide the most flexibility. In this paper, we propose RIBBO, a method to reinforce-learn a BBO algorithm from offline data in an end-to-end fashion. RIBBO employs expressive sequence models to learn the optimization histories produced by multiple behavior algorithms and tasks, leveraging the in-context learning ability of large models to extract task information and make decisions accordingly. Central to our method is to augment the optimization histories with *regret-to-go* tokens, which are designed to represent the performance of an algorithm based on cumulative regret over the future part of the histories. The integration of regret-to-go tokens enables RIBBO to automatically generate sequences of query points that satisfy the user-desired regret, which is verified by its universally good empirical performance on diverse problems, including BBO benchmark functions, hyper-parameter optimization and robot control problems.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=TjvSFVJdzJ", "arxiv_id": "2402.17423", "paper_pdf": "papers/TjvSFVJdzJ.pdf", "paper_pdf_sha256": "605cf117362bf50c1b81eef9816c537674d6bc69f5af0c03c825ae0dc3781bdc", "paper_pdf_bytes": 1587027, "paper_pdf_source": "openreview", "code_url": "https://github.com/lamda-bbo/RIBBO", "code_repository": "lamda-bbo/RIBBO", "code_commit": "1b42a7da42f6ceba6e92299d5d8dc6583cdd380e", "code_archive": "repos/TjvSFVJdzJ.zip", "code_archive_sha256": "e90c92edd7b9eb3e646ff09aaa14bb6c3e459c025078377ef947f400b869cd5a", "code_archive_bytes": 136125, "code_file_count": 97, "code_extensions": {".py": 92, ".sh": 5}, "github_disk_usage_kb": 247, "github_languages": {"Python": 353462, "Shell": 4728}, "github_archived": false, "github_pushed_at": "2025-05-19T06:21:44Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/reinforced-in-context-black-box-optimization"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Rt6btdXS2b", "year": 2024, "status": "rejected", "title": "Continuous Indeterminate Probability Neural Network", "authors": ["Tao Yang"], "authorids": ["~Tao_Yang19"], "authors_source": "OpenReview API", "abstract": "Currently, there is no mathematical analytical form for a general posterior, however, Indeterminate Probability Theory has now discovered a way to address this issue. This is a big discovery in the field of probability and it is applicable in various fields.\nThis paper introduces a general model called CIPNN - Continuous Indeterminate Probability Neural Network, which is an analytical probability neural network with continuous latent random variables.  \nOur contributions are Four-fold. First, we apply the analytical form of the posterior for continuous latent random variables and propose a general classification model (CIPNN). Second, we propose a general auto-encoder called CIPAE - Continuous Indeterminate Probability Auto-Encoder, instead of using a neural network as the decoder component, we first employ a probabilistic equation. Third, we propose a new method to visualize the latent random variables, we use one of N dimensional latent variables as a decoder to reconstruct the input image, which can work even for classification tasks, in this way, we can see what each latent variable has learned. Fourth, IPNN has shown great classification capability, CIPNN has pushed this classification capability to infinity.\nTheoretical advantages are reflected in experimental results.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=Rt6btdXS2b", "arxiv_id": "2303.12964", "paper_pdf": "papers/Rt6btdXS2b.pdf", "paper_pdf_sha256": "c6560aef1007ed830b67529c45dafb14f4861b8e46d53571a12f27907c4bd2f8", "paper_pdf_bytes": 3261823, "paper_pdf_source": "openreview", "code_url": "https://github.com/Starfruit007/cipnn", "code_repository": "Starfruit007/cipnn", "code_commit": "e77f2062c3183d2cba0fdf620298f90cfbf49f17", "code_archive": "repos/Rt6btdXS2b.zip", "code_archive_sha256": "009015230b68621a796f9ad1f601a087f7408ba93e53f0221e149b77b8dc1e3a", "code_archive_bytes": 81969, "code_file_count": 10, "code_extensions": {".py": 9, ".sh": 1}, "github_disk_usage_kb": 180, "github_languages": {"Python": 52551, "Shell": 893}, "github_archived": false, "github_pushed_at": "2023-10-15T01:39:20Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/continuous-indeterminate-probability-neural"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "KVYq2Ea90PC", "year": 2022, "status": "rejected", "title": "A Study of Face Obfuscation in ImageNet", "authors": ["Kaiyu Yang", "Jacqueline Yau", "Li Fei-Fei", "Jia Deng", "Olga Russakovsky"], "authorids": ["~Kaiyu_Yang1", "~Jacqueline_Yau1", "~Li_Fei-Fei1", "~Jia_Deng1", "~Olga_Russakovsky1"], "authors_source": "OpenReview API", "abstract": "Face obfuscation (blurring, mosaicing, etc.) has been shown to be effective for privacy protection; nevertheless, object recognition research typically assumes access to complete, unobfuscated images. In this paper, we explore the effects of face obfuscation on the popular ImageNet challenge visual recognition benchmark. Most categories in the ImageNet challenge are not people categories; however, many incidental people appear in the images, and their privacy is a concern. We first annotate faces in the dataset. Then we demonstrate that face blurring and overlaying---two typical obfuscation techniques---have minimal impact on the accuracy of recognition models. Concretely, we benchmark multiple deep neural networks on face-obfuscated images and observe that the overall recognition accuracy drops only slightly (<= 1.0%). Further, we experiment with transfer learning to 4 downstream tasks (object recognition, scene recognition, face attribute classification, and object detection) and show that features learned on face-obfuscated images are equally transferable. Our work demonstrates the feasibility of privacy-aware visual recognition, improves the highly-used ImageNet challenge benchmark, and suggests an important path for future visual datasets. ", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=KVYq2Ea90PC", "arxiv_id": "2103.06191", "paper_pdf": "papers/KVYq2Ea90PC.pdf", "paper_pdf_sha256": "609078997326405e7d1f2738eb0a37506ab450d4b037387a2500e41cb2b2e92a", "paper_pdf_bytes": 4251919, "paper_pdf_source": "openreview", "code_url": "https://github.com/princetonvisualai/imagenet-face-obfuscation", "code_repository": "princetonvisualai/imagenet-face-obfuscation", "code_commit": "83f027962c41e11517d40f9c9f183621bedc4ed4", "code_archive": "repos/KVYq2Ea90PC.zip", "code_archive_sha256": "388e42614bcb64c2c1568442fbc3f20f325be9ef1b4e0a19943833c65dfff3e2", "code_archive_bytes": 1408607, "code_file_count": 9, "code_extensions": {".py": 9}, "github_disk_usage_kb": 1420, "github_languages": {"Python": 42871, "HTML": 32150}, "github_archived": false, "github_pushed_at": "2023-07-08T16:18:42Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/a-study-of-face-obfuscation-in-imagenet"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "NfZ6g2OmXEk", "year": 2021, "status": "rejected", "title": "Prioritized Level Replay", "authors": ["Minqi Jiang", "Edward Grefenstette", "Tim Rocktäschel"], "authorids": ["~Minqi_Jiang1", "~Edward_Grefenstette1", "~Tim_Rocktäschel1"], "authors_source": "OpenReview API", "abstract": "Simulated environments with procedurally generated content have become popular benchmarks for testing systematic generalization of reinforcement learning agents. Every level in such an environment is algorithmically created, thereby exhibiting a unique configuration of underlying factors of variation, such as layout, positions of entities, asset appearances, or even the rules governing environment transitions. Fixed sets of training levels can be determined to aid comparison and reproducibility, and test levels can be held out to evaluate the generalization and robustness of agents. While prior work samples training levels in a direct way (e.g.~uniformly) for the agent to learn from, we investigate the hypothesis that different levels provide different learning progress for an agent at specific times during training. We introduce Prioritized Level Replay, a general framework for estimating the future learning potential of a level given the current state of the agent's policy. We find that temporal-difference (TD) errors, while previously used to selectively sample past transitions, also prove effective for scoring a level's future learning potential when the agent replays (that is, revisits) that level to generate entirely new episodes of experiences from it. We report significantly improved sample-efficiency and generalization on the majority of Procgen Benchmark environments as well as two challenging MiniGrid environments. Lastly, we present a qualitative analysis showing that Prioritized Level Replay induces an implicit curriculum, taking the agent gradually from easier to harder levels.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=NfZ6g2OmXEk", "arxiv_id": "2010.03934", "paper_pdf": "papers/NfZ6g2OmXEk.pdf", "paper_pdf_sha256": "076d72332256d46f38873fa3e1f3b03f172dbd302a887a3306246a863aa83277", "paper_pdf_bytes": 5922362, "paper_pdf_source": "openreview", "code_url": "https://github.com/facebookresearch/level-replay", "code_repository": "facebookresearch/level-replay", "code_commit": "ccecf452ee3342217ece964aaf10c2831625f9b3", "code_archive": "repos/NfZ6g2OmXEk.zip", "code_archive_sha256": "fd9f0ed1624113f9dbce62d1255eddab9b2dec52025cb70d4d96847927d9d255", "code_archive_bytes": 1201087, "code_file_count": 15, "code_extensions": {".py": 15}, "github_disk_usage_kb": 1168, "github_languages": {"Python": 98573}, "github_archived": true, "github_pushed_at": "2021-06-11T01:47:58Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/prioritized-level-replay-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "HJlU-AVtvS", "year": 2020, "status": "rejected", "title": "A Fine-Grained Spectral Perspective on Neural Networks", "authors": ["Greg Yang", "Hadi Salman"], "authorids": ["gregyang@microsoft.com", "hadicsalman@gmail.com"], "authors_source": "OpenReview API", "abstract": "Are neural networks biased toward simple functions?\nDoes depth always help learn more complex features?\nIs training the last layer of a network as good as training all layers?\nThese questions seem unrelated at face value, but in this work we give all of them a common treatment from the spectral perspective.\nWe will study the spectra of the *Conjugate Kernel, CK,* (also called the *Neural Network-Gaussian Process Kernel*), and the *Neural Tangent Kernel, NTK*.\nRoughly, the CK and the NTK tell us respectively ``\"what a network looks like at initialization\" and \"``what a network looks like during and after training.\"\nTheir spectra then encode valuable information about the initial distribution and the training and generalization properties of neural networks.\nBy analyzing the eigenvalues, we lend novel insights into the questions put forth at the beginning, and we verify these insights by extensive experiments of neural networks.\nWe believe the computational tools we develop here for analyzing the spectra of CK and NTK serve as a solid foundation for future studies of deep neural networks.\nWe have open-sourced the code for it and for generating the plots in this paper at github.com/jxVmnLgedVwv6mNcGCBy/NNspectra.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=HJlU-AVtvS", "arxiv_id": "1907.10599", "paper_pdf": "papers/HJlU-AVtvS.pdf", "paper_pdf_sha256": "cd7662738181efa64a240c89e100c9b716d04cc10c1c97e0f44bb4a01fda8e17", "paper_pdf_bytes": 3439067, "paper_pdf_source": "openreview", "code_url": "https://github.com/thegregyang/NNspectra", "code_repository": "thegregyang/NNspectra", "code_commit": "8c71181e93a46cdcabfafdf71ae5a58830cbb27d", "code_archive": "repos/HJlU-AVtvS.zip", "code_archive_sha256": "015fc770003bae3db824c98cd88c3a901f9c561bbee89d549512c73d13b58f6b", "code_archive_bytes": 4214439, "code_file_count": 5, "code_extensions": {".ipynb": 3, ".py": 2}, "github_disk_usage_kb": 4160, "github_languages": {"Jupyter Notebook": 3800945, "Python": 15442}, "github_archived": false, "github_pushed_at": "2019-09-19T23:57:11Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/a-fine-grained-spectral-perspective-on-neural"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ZSDmfoVvCs", "year": 2026, "status": "rejected", "title": "ROTATE: Regret-driven Open-ended Training for Ad Hoc Teamwork", "authors": ["Caroline Wang", "Arrasy Rahman", "Jiaxun Cui", "Yoonchang Sung", "Peter Stone"], "authorids": ["~Caroline_Wang1", "~Arrasy_Rahman1", "~Jiaxun_Cui1", "~Yoonchang_Sung1", "~Peter_Stone1"], "authors_source": "OpenReview API", "abstract": "Learning to collaborate with previously unseen partners is a fundamental generalization challenge in multi-agent learning, known as Ad Hoc Teamwork (AHT). \nExisting AHT approaches often adopt a two-stage pipeline, where first, a fixed population of teammates is generated with the idea that they should be representative of the teammates that will be seen at deployment time, and second, an AHT agent is trained to collaborate well with agents in the population. \nTo date, the research community has focused on designing separate algorithms for each stage. This separation has led to algorithms that generate teammates with limited coverage of possible behaviors, and that ignore whether the generated teammates are easy to learn from for the AHT agent. \nFurthermore, algorithms for training AHT agents typically treat the set of training teammates as static, thus attempting to generalize to previously unseen partner agents without assuming any control over the set of training teammates.\nThis paper presents a unified framework for AHT by reformulating the problem as an open-ended learning process between an AHT agent and an adversarial teammate generator. \nWe introduce ROTATE, a regret-driven, open-ended training algorithm that alternates between improving the AHT agent and generating teammates that probe its deficiencies. \nExperiments across diverse two-player environments demonstrate that ROTATE significantly outperforms baselines at generalizing to an unseen set of evaluation teammates, thus establishing a new standard for robust and generalizable teamwork.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=ZSDmfoVvCs", "arxiv_id": "2505.23686", "paper_pdf": "papers/ZSDmfoVvCs.pdf", "paper_pdf_sha256": "f0bc17e1c3a661066f0f7bb5e6e4362e2caa9d8ab806ba646a916aac4f745429", "paper_pdf_bytes": 1577273, "paper_pdf_source": "openreview", "code_url": "https://github.com/carolinewang01/rotate", "code_repository": "carolinewang01/rotate", "code_commit": "68dbf05f49b77096b2b662853b7ac7dfef97aaae", "code_archive": "repos/ZSDmfoVvCs.zip", "code_archive_sha256": "23df00704ec505fb5a58e555e329d2ce098d9bdef43bc2572386c6fd8992af8a", "code_archive_bytes": 669506, "code_file_count": 84, "code_extensions": {".py": 79, ".sh": 4, ".ipynb": 1}, "github_disk_usage_kb": 3523, "github_languages": {"Python": 888697, "Jupyter Notebook": 86576, "Shell": 7124}, "github_archived": false, "github_pushed_at": "2025-10-22T02:47:21Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/rotate-regret-driven-open-ended-training-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "MHP4jGMN2E", "year": 2025, "status": "rejected", "title": "The Impact of Element Ordering on LM Agent Performance", "authors": ["Wayne Chi", "Ameet Talwalkar", "Chris Donahue"], "authorids": ["~Wayne_Chi2", "~Ameet_Talwalkar1", "~Chris_Donahue1"], "authors_source": "OpenReview API", "abstract": "There has been a surge of interest in language model agents that can navigate virtual environments such as the web or desktop. To navigate such environments, agents benefit from information on the various elements (e.g., buttons, text, or images) present. However, it remains unclear which element attributes have the greatest impact on agent performance, especially in environments that only provide a graphical representation (i.e., pixels). Here we find that the ordering in which elements are presented to the language model is surprisingly impactful—randomizing element ordering in webpages compromises average agent performance to a degree comparable to removing all visible text from webpages. While web agents benefit from the semantic hierarchical ordering of elements available via the browser, agents that parse elements directly from pixels do not have access to any such ordering. Here we endeavor to derive effective orderings and investigate the impact of various element ordering methods in web and desktop environments. We find that dimensionality reduction provides a viable ordering for pixel-only environments. We train a UI element detection model to derive elements from pixels and apply our findings to an agent benchmark—OmniACT—where we only have access to pixels. Our method completes more than two times as many tasks on average relative to the previous state-of-the-art.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=MHP4jGMN2E", "arxiv_id": "2409.12089", "paper_pdf": "papers/MHP4jGMN2E.pdf", "paper_pdf_sha256": "3a2774b7b1f9a431b383f08350bd80475c3120a4df251d0af71931491329af4f", "paper_pdf_bytes": 3166130, "paper_pdf_source": "openreview", "code_url": "https://github.com/waynchi/gui-agent", "code_repository": "waynchi/gui-agent", "code_commit": "31f972daa507b3225a20dcc68e36e92483f47557", "code_archive": "repos/MHP4jGMN2E.zip", "code_archive_sha256": "12922954691006e90aee8fa9ca7b4f309daa3a71241c5241a6d0763350d12733", "code_archive_bytes": 364840, "code_file_count": 88, "code_extensions": {".py": 81, ".sh": 7}, "github_disk_usage_kb": 250, "github_languages": {"Python": 417822, "Shell": 9397}, "github_archived": false, "github_pushed_at": "2024-09-30T16:38:09Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/the-impact-of-element-ordering-on-lm-agent"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Qvoe4wXWFi", "year": 2024, "status": "rejected", "title": "NeuralFuse: Learning to Recover the Accuracy of Access-Limited Neural Network Inference in Low-Voltage Regimes", "authors": ["Hao-Lun Sun", "Lei Hsiung", "Nandhini Chandramoorthy", "Pin-Yu Chen", "Tsung-Yi Ho"], "authorids": ["~Hao-Lun_Sun1", "~Lei_Hsiung1", "~Nandhini_Chandramoorthy1", "~Pin-Yu_Chen1", "~Tsung-Yi_Ho2"], "authors_source": "OpenReview API", "abstract": "Deep neural networks (DNNs) have become ubiquitous in machine learning, but their energy consumption remains a notable issue. Lowering the supply voltage is an effective strategy for reducing energy consumption. However, aggressively scaling down the supply voltage can lead to accuracy degradation due to random bit flips in static random access memory (SRAM) where model parameters are stored. To address this challenge, we introduce NeuralFuse, a novel add-on module that addresses the accuracy-energy tradeoff in low-voltage regimes by learning input transformations to generate error-resistant data representations. NeuralFuse protects DNN accuracy in both nominal and low-voltage scenarios. Moreover, NeuralFuse is easy to implement and can be readily applied to DNNs with limited access, such as non-configurable hardware or remote access to cloud-based APIs. Experimental results demonstrate that, at a 1% bit error rate, NeuralFuse can reduce SRAM memory access energy by up to 24% while recovering accuracy by up to 57%. To the best of our knowledge, this is the first model-agnostic approach (i.e., no model retraining) to address low-voltage-induced bit errors.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=Qvoe4wXWFi", "arxiv_id": "2306.16869", "paper_pdf": "papers/Qvoe4wXWFi.pdf", "paper_pdf_sha256": "85c4fcdd41bc0c4bb4d87bc9265415dadddb1030ec27d14f83c78e28ef166bff", "paper_pdf_bytes": 5099181, "paper_pdf_source": "openreview", "code_url": "https://github.com/IBM/NeuralFuse", "code_repository": "IBM/NeuralFuse", "code_commit": "8bfdbaf5c969fc78b866f48d5ab9c724258a8c33", "code_archive": "repos/Qvoe4wXWFi.zip", "code_archive_sha256": "f199813f55785ae5d4940cb9148179c84b8e0b67aa355a9d2de860004dfe8493", "code_archive_bytes": 189571, "code_file_count": 19, "code_extensions": {".py": 19}, "github_disk_usage_kb": 183, "github_languages": {"Python": 188756}, "github_archived": false, "github_pushed_at": "2025-09-18T00:19:41Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/neuralfuse-learning-to-improve-the-accuracy"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Krk0Gnft2Zc", "year": 2023, "status": "rejected", "title": "Discrete State-Action Abstraction via the Successor Representation", "authors": ["Amnon Attali", "Pedro Cisneros-Velarde", "Marco Morales", "Nancy Amato"], "authorids": ["~Amnon_Attali1", "~Pedro_Cisneros-Velarde1", "~Marco_Morales1", "~Nancy_Amato1"], "authors_source": "OpenReview API", "abstract": "While the difficulty of reinforcement learning problems is typically related to the complexity of their state spaces, Abstraction proposes that solutions often lie in simpler underlying latent spaces. Prior works have focused on learning either a continuous or dense abstraction, or require a human to provide one. Information-dense representations capture features irrelevant for solving tasks, and continuous spaces can struggle to represent discrete objects. In this work we automatically learn a sparse discrete abstraction of the underlying environment. We do so using a simple end-to-end trainable model based on the successor representation and max-entropy regularization. We describe an algorithm to apply our model, named Discrete State-Action Abstraction (DSAA), which computes an action abstraction in the form of temporally extended actions, i.e., Options, to transition between discrete abstract states. Empirically, we demonstrate the effects of different exploration schemes on our resulting abstraction, and show that it is efficient for solving downstream tasks.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=Krk0Gnft2Zc", "arxiv_id": "2206.03467", "paper_pdf": "papers/Krk0Gnft2Zc.pdf", "paper_pdf_sha256": "384519ddb592fc816f23367572cbe0ec49988cc01190be8cac574b9edae717ee", "paper_pdf_bytes": 1267863, "paper_pdf_source": "openreview", "code_url": "https://github.com/amnonattali/dsaa", "code_repository": "amnonattali/dsaa", "code_commit": "a82a3afb45eb6ba33bdc7f5230d0f275daf205fb", "code_archive": "repos/Krk0Gnft2Zc.zip", "code_archive_sha256": "b6d90f86b31915300fcf546b6034f9edbd64ced93332c6d9d03f35f456caeaf3", "code_archive_bytes": 341352, "code_file_count": 20, "code_extensions": {".py": 20}, "github_disk_usage_kb": 407, "github_languages": {"Python": 199174}, "github_archived": false, "github_pushed_at": "2023-01-11T18:59:16Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/discrete-state-action-abstraction-via-the"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "OgCcfc1m0TO", "year": 2022, "status": "rejected", "title": "Learning to Prompt for Vision-Language Models", "authors": ["Kaiyang Zhou", "Jingkang Yang", "Chen Change Loy", "Ziwei Liu"], "authorids": ["~Kaiyang_Zhou1", "~Jingkang_Yang1", "~Chen_Change_Loy2", "~Ziwei_Liu1"], "authors_source": "OpenReview API", "abstract": "Vision-language pre-training has recently emerged as a promising alternative for representation learning. It shifts from the tradition of using images and discrete labels for learning a fixed set of weights, seen as visual concepts, to aligning images and raw text for two separate encoders. Such a paradigm benefits from a broader source of supervision and allows zero-shot transfer to downstream tasks since visual concepts can be diametrically generated from natural language, known as prompt. In this paper, we identify that a major challenge of deploying such models in practice is prompt engineering. This is because designing a proper prompt, especially for context words surrounding a class name, requires domain expertise and typically takes a significant amount of time for words tuning since a slight change in wording could have a huge impact on performance. Moreover, different downstream tasks require specific designs, further hampering the efficiency of deployment. To overcome this challenge, we propose a novel approach named \\emph{context optimization (CoOp)}. The main idea is to model context in prompts using continuous representations and perform end-to-end learning from data while keeping the pre-trained parameters fixed. In this way, the design of task-relevant prompts can be fully automated. Experiments on 11 datasets show that CoOp effectively turns pre-trained vision-language models into data-efficient visual learners, requiring as few as one or two shots to beat hand-crafted prompts with a decent margin and able to gain significant improvements when using more shots (e.g., at 16 shots the average gain is around 17\\% with the highest reaching over 50\\%). CoOp also exhibits strong robustness to distribution shift.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=OgCcfc1m0TO", "arxiv_id": "2109.01134", "paper_pdf": "papers/OgCcfc1m0TO.pdf", "paper_pdf_sha256": "9fa2eb3a60029d39448d790e4cd660d8c69bae7ce18b3a270fd2c5705be08ab2", "paper_pdf_bytes": 2147082, "paper_pdf_source": "openreview", "code_url": "https://github.com/KaiyangZhou/CoOp", "code_repository": "KaiyangZhou/CoOp", "code_commit": "ff61507c790454bce7c5052c3ac39e60772f1f89", "code_archive": "repos/OgCcfc1m0TO.zip", "code_archive_sha256": "c3c3d7a6b0d2d5e63ec0f561ed06554c77c5f56f9c1715c9753aea2861825fec", "code_archive_bytes": 1438659, "code_file_count": 38, "code_extensions": {".py": 29, ".sh": 9}, "github_disk_usage_kb": 1442, "github_languages": {"Python": 134456, "Shell": 5875}, "github_archived": false, "github_pushed_at": "2024-05-20T16:58:40Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/learning-to-prompt-for-vision-language-models"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "oev4KdikGjy", "year": 2021, "status": "rejected", "title": "FMix: Enhancing Mixed Sample Data Augmentation", "authors": ["Ethan Harris", "Antonia Marcu", "Matthew Painter", "Mahesan Niranjan", "Adam Prugel-Bennett", "Jonathon Hare"], "authorids": ["~Ethan_Harris1", "am1g15@ecs.soton.ac.uk", "mp2u16@ecs.soton.ac.uk", "~Mahesan_Niranjan1", "~Adam_Prugel-Bennett1", "~Jonathon_Hare1"], "authors_source": "OpenReview API", "abstract": "Mixed Sample Data Augmentation (MSDA) has received increasing attention in recent years, with many successful variants such as MixUp and CutMix. We analyse MSDA from an information theoretic perspective, characterising learned models in terms of how they impact the models’ perception of the data.  Ultimately, our analyses allow us to decouple two complementary properties of augmentations that are useful for reasoning about MSDA. From insight on the efficacy of CutMix in particular, we subsequently propose FMix, an MSDA that uses binary masks obtained by applying a threshold to low frequency images sampled from Fourier space.  FMix improves performance over MixUp and CutMix for a number of models across a range of data sets and problem settings,  obtaining new state-of-the-art results on CIFAR-10 and Fashion-MNIST.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=oev4KdikGjy", "arxiv_id": "2002.12047", "paper_pdf": "papers/oev4KdikGjy.pdf", "paper_pdf_sha256": "cdfbd4c6cfe9b56067cdd8b6ec6f591ea6bc33ef3fb497ed60a8aeecf83d46db", "paper_pdf_bytes": 602932, "paper_pdf_source": "openreview", "code_url": "https://github.com/ecs-vlc/FMix", "code_repository": "ecs-vlc/FMix", "code_commit": "e5991dca018882734c8ea63599f10dfbe67fa0ae", "code_archive": "repos/oev4KdikGjy.zip", "code_archive_sha256": "162c921728edae993a503bdf6715d944b6521484cdc7a97cba7c2cc9874056d2", "code_archive_bytes": 1008157, "code_file_count": 59, "code_extensions": {".py": 49, ".sh": 8, ".ipynb": 2}, "github_disk_usage_kb": 1170, "github_languages": {"Jupyter Notebook": 449087, "Python": 245318, "Shell": 3473}, "github_archived": false, "github_pushed_at": "2021-03-26T19:00:33Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/understanding-and-enhancing-mixed-sample-data"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ryg7vA4tPB", "year": 2020, "status": "rejected", "title": "Rigging the Lottery: Making All Tickets Winners", "authors": ["Utku Evci", "Erich Elsen", "Pablo Castro", "Trevor Gale"], "authorids": ["ue225@nyu.edu", "eriche@google.com", "tgale@google.com", "psc@google.com"], "authors_source": "OpenReview API", "abstract": "Sparse neural networks have been shown to yield computationally efficient networks with improved inference times.  There is a large body of work on training dense networks to yield sparse networks for inference (Molchanov et al., 2017;Zhu & Gupta, 2018; Louizos et al., 2017; Li et al., 2016; Guo et al., 2016).  This limits the size of the largest trainable sparse model to that of the largest trainable dense model. In this paper we introduce a method to train sparse neural networks with a fixed parameter count and a fixed computational cost throughout training, without sacrificing accuracy relative to existing dense-to-sparse training methods. Our method updates the topology of the network during training by using parameter magnitudes and infrequent gradient calculations. We show that this approach requires less floating-point operations (FLOPs) to achieve a given level of accuracy compared to prior techniques. We demonstrate state-of-the-art sparse training results with ResNet-50, MobileNet v1 and MobileNet v2 on the ImageNet-2012 dataset. Finally,  we  provide  some  insights  into  why  allowing  the  topology  to change during the optimization can overcome local minima encountered when the topology remains static.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=ryg7vA4tPB", "arxiv_id": "1911.11134", "paper_pdf": "papers/ryg7vA4tPB.pdf", "paper_pdf_sha256": "31921ef6c611ae38f3e02b701fee3b472d0e97e364a84dd857c8c7a43b1907c2", "paper_pdf_bytes": 671480, "paper_pdf_source": "openreview", "code_url": "https://github.com/varun19299/rigl-reproducibility", "code_repository": "varun19299/rigl-reproducibility", "code_commit": "97443beac90e03f899652943594695e5152c2b09", "code_archive": "repos/ryg7vA4tPB.zip", "code_archive_sha256": "50a3c4c4086b27e76dded723123beb58060a856c0fbefca2350cc7452116b1cd", "code_archive_bytes": 3443931, "code_file_count": 53, "code_extensions": {".py": 39, ".js": 14}, "github_disk_usage_kb": 4955, "github_languages": {"Python": 201476, "Makefile": 12540, "Batchfile": 799}, "github_archived": false, "github_pushed_at": "2022-01-06T15:28:04Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/rigging-the-lottery-making-all-tickets-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "sV2qR9QIEo", "year": 2026, "status": "rejected", "title": "SkipVAR: Accelerating Visual Autoregressive Modeling via Adaptive Frequency-Aware Skipping", "authors": ["Jiajun Li", "Yue Ma", "Xinyu Zhang", "Qingyan Wei", "Songhua Liu", "Linfeng Zhang"], "authorids": ["~Jiajun_Li16", "~Yue_Ma2", "~Xinyu_Zhang53", "~Qingyan_Wei1", "~Songhua_Liu2", "~Linfeng_Zhang2"], "authors_source": "OpenReview API", "abstract": "Recent studies on Visual Autoregressive (VAR) models have highlighted that high-frequency components, or later steps, in the generation process contribute disproportionately to inference latency. However, the underlying computational redundancy involved in these steps has yet to be thoroughly investigated. In this paper, we conduct an in-depth analysis of the VAR inference process and identify two primary sources of inefficiency: ***step redundancy*** and ***unconditional branch redundancy***. To address step redundancy, we propose an automatic step-skipping strategy that selectively omits unnecessary generation steps to improve efficiency. For unconditional branch redundancy, we observe that the information gap between the conditional and unconditional branches is minimal. Leveraging this insight, we introduce unconditional branch replacement, a technique that bypasses the unconditional branch to reduce computational cost. Notably, we observe that the effectiveness of acceleration strategies varies significantly across different samples. Motivated by this, we propose **SkipVAR**, a sample-adaptive framework that leverages frequency information to dynamically select the most suitable acceleration strategy for each instance. To evaluate the role of high-frequency information, we further introduce multiple high-variation benchmark datasets that evaluate the performance in terms of fine details. Extensive experiments show that SkipVAR achieves over 0.88 average SSIM with up to 1.81$\\times$ overall acceleration and 2.62$\\times$ lossless speedup on the GenEval benchmark.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=sV2qR9QIEo", "arxiv_id": "2506.08908", "paper_pdf": "papers/sV2qR9QIEo.pdf", "paper_pdf_sha256": "34bd9cd94f8898e3e75bc7eec10b2335ead80c908b6401c8dd9553ec30a33341", "paper_pdf_bytes": 50916108, "paper_pdf_source": "openreview", "code_url": "https://github.com/fakerone-li/SkipVAR", "code_repository": "fakerone-li/SkipVAR", "code_commit": "607be0efea96a91458ce1e6f6623a8b2ad0b01e6", "code_archive": "repos/sV2qR9QIEo.zip", "code_archive_sha256": "e468556324b0c1175a6f1011b3913a68bbdf206fbe0cf0fbdacb384877df8363", "code_archive_bytes": 4009344, "code_file_count": 53, "code_extensions": {".py": 51, ".sh": 2}, "github_disk_usage_kb": 3587, "github_languages": {"Python": 488231, "Shell": 9109}, "github_archived": false, "github_pushed_at": "2025-06-15T09:06:50Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/skipvar-accelerating-visual-autoregressive"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "LUcdXA8hAa", "year": 2024, "status": "rejected", "title": "Identifiability Matters: Revealing the Hidden Recoverable Condition in Unbiased Learning to Rank", "authors": ["Mouxiang Chen", "Chenghao Liu", "Zemin Liu", "Zhuo Li", "Jianling Sun"], "authorids": ["~Mouxiang_Chen1", "~Chenghao_Liu1", "~Zemin_Liu1", "lizhuo@zju.edu.cn", "~Jianling_Sun2"], "authors_source": "OpenReview API", "abstract": "The application of Unbiased Learning to Rank (ULTR) is widespread in modern systems for training unbiased ranking models from biased click logs. The key is to explicitly model a generation process for user behavior and fit click data based on examination hypothesis. Previous research found empirically that the true latent relevance can be recovered in most cases as long as the clicks are perfectly fitted. However, we demonstrate that this is not always achievable, resulting in a significant reduction in ranking performance. In this work, we aim to answer if or when the true relevance can be recovered from click data, which is a foundation issue for ULTR field. We first define a ranking model as identifiable if it can recover the true relevance up to a scaling transformation, which is enough for pairwise ranking objective. Then we explore an equivalent condition for identifiability that can be novely expressed as a graph connectivity test problem: if and only if a graph (namely identifiability graph, or IG) constructed on the underlying structure of the dataset is connected, we can guarantee that the relevance can be correctly recovered. When the IG is not connected, there may be bad cases leading to poor ranking performance. To address this issue, we propose two methods, namely node intervention and node merging, to modify the dataset and restore connectivity of the IG. Empirical results obtained on a simulation dataset and two LTR benchmark datasets confirm the validity of our proposed theorems and show the effectiveness of our methods in mitigating data bias when the relevance model is unidentifiable.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=LUcdXA8hAa", "arxiv_id": "2309.15560", "paper_pdf": "papers/LUcdXA8hAa.pdf", "paper_pdf_sha256": "664040d4c83342ecb7a25338f5439d32457061dee97440eaccd1b3eb0d33283d", "paper_pdf_bytes": 714069, "paper_pdf_source": "openreview", "code_url": "https://github.com/Keytoyze/ULTR-identifiability", "code_repository": "Keytoyze/ULTR-identifiability", "code_commit": "08e623547d4ac09698a0c0d7c4ed166b39d72ddb", "code_archive": "repos/LUcdXA8hAa.zip", "code_archive_sha256": "e3f22b88cb5ef3b965ee8846bbce6ae5d38e6dfc3e0ce6cfd8a5a5bdb1d2232d", "code_archive_bytes": 440385, "code_file_count": 8, "code_extensions": {".py": 8}, "github_disk_usage_kb": 204, "github_languages": {"Gherkin": 588900, "Python": 50958}, "github_archived": false, "github_pushed_at": "2024-05-18T09:52:43Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/identifiability-matters-revealing-the-hidden"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "4NLyCJQR3ZR", "year": 2023, "status": "rejected", "title": "Optimal Neural Network Approximation of Wasserstein Gradient Direction via Convex Optimization", "authors": ["Yifei Wang", "Peng Chen", "Mert Pilanci", "Wuchen Li"], "authorids": ["~Yifei_Wang2", "~Peng_Chen1", "~Mert_Pilanci3", "~Wuchen_Li1"], "authors_source": "OpenReview API", "abstract": "The computation of Wasserstein gradient direction is essential for posterior sampling problems and scientific computing. The approximation of the Wasserstein gradient with finite samples requires solving a variational problem. We study the variational problem in the family of two-layer networks with squared-ReLU activations, towards which we derive a semi-definite programming (SDP) relaxation. This SDP can be viewed as an approximation of the Wasserstein gradient in a broader function family including two-layer networks. By solving the convex SDP, we obtain the optimal approximation of the Wasserstein gradient direction in this class of functions. Numerical experiments including PDE-constrained Bayesian inference and parameter estimation in COVID-19 modeling demonstrate the effectiveness of the proposed method.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=4NLyCJQR3ZR", "arxiv_id": "2205.13098", "paper_pdf": "papers/4NLyCJQR3ZR.pdf", "paper_pdf_sha256": "adf05e3e2554827a520fe53080e15dc3b34065c9dcaa7421726f75e414182fe6", "paper_pdf_bytes": 593672, "paper_pdf_source": "openreview", "code_url": "https://github.com/ai-submit/OptimalWasserstein", "code_repository": "ai-submit/OptimalWasserstein", "code_commit": "d78e87664d2b6e90b7e0925883957809f5417f1c", "code_archive": "repos/4NLyCJQR3ZR.zip", "code_archive_sha256": "622af8bd567fd3e45a6b5a469b355ae24a159fc7076125e1ecd5dfcd3a254b44", "code_archive_bytes": 499459, "code_file_count": 146, "code_extensions": {".py": 132, ".cpp": 7, ".h": 4, ".ipynb": 2, ".hpp": 1}, "github_disk_usage_kb": 410, "github_languages": {"Python": 1529282, "C++": 39813, "Jupyter Notebook": 2886}, "github_archived": false, "github_pushed_at": "2022-05-25T21:14:13Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/optimal-neural-network-approximation-of"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "SuKTLF9stD", "year": 2022, "status": "rejected", "title": "Data-Efficient Augmentation for Training Neural Networks", "authors": ["Tian Yu Liu", "Baharan Mirzasoleiman"], "authorids": ["~Tian_Yu_Liu2", "~Baharan_Mirzasoleiman4"], "authors_source": "OpenReview API", "abstract": "Data augmentation is essential to achieve state-of-the-art performance in many deep learning applications. However, modern data augmentation techniques become computationally prohibitive for large datasets. To address this, we propose a rigorous technique to select subsets of data points that when augmented, closely capture the training dynamics of full data augmentation. We first show that data augmentation, modeled as additive perturbations, speeds up learning by enlarging the smaller singular values of the network Jacobian. Then, we propose a framework to iteratively extract small subsets of training data that when augmented, closely capture the alignment of the fully augmented Jacobian with label/residual vector. We prove that stochastic gradient descent applied to augmented subsets found by our approach have similar training dynamics to that of fully augmented data. Our experiments demonstrate that our method outperforms state-of-the-art max-loss strategy by 7.7% on CIFAR10 while achieving 6.3x speedup, and by 4.7% on SVHN while achieving 2.2x speedup, using 10% and 30% subsets, respectively.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=SuKTLF9stD", "arxiv_id": "2210.08363", "paper_pdf": "papers/SuKTLF9stD.pdf", "paper_pdf_sha256": "d0ef9e2cc5bbed07e1ba995525730f3a97352b34e593cf65f0d05bc1a466572b", "paper_pdf_bytes": 1240777, "paper_pdf_source": "openreview", "code_url": "https://github.com/tianyu139/data-efficient-augmentation", "code_repository": "tianyu139/data-efficient-augmentation", "code_commit": "3abb2f7d3fa610fe2394063872c55927bbccde4a", "code_archive": "repos/SuKTLF9stD.zip", "code_archive_sha256": "9b3892e237a60be61bab31e12bc1840e7e4bbb48b554d261dbb4da4451f4e734", "code_archive_bytes": 1565834, "code_file_count": 21, "code_extensions": {".py": 19, ".sh": 2}, "github_disk_usage_kb": 1517, "github_languages": {"Python": 110230, "Shell": 482}, "github_archived": false, "github_pushed_at": "2023-07-20T04:42:40Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/data-efficient-augmentation-for-training-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "wl0Kr_jqM2a", "year": 2021, "status": "rejected", "title": "Testing Robustness Against Unforeseen Adversaries", "authors": ["Daniel Kang", "Yi Sun", "Dan Hendrycks", "Tom B Brown", "Jacob Steinhardt"], "authorids": ["~Daniel_Kang1", "~Yi_Sun3", "~Dan_Hendrycks1", "~Tom_B_Brown1", "~Jacob_Steinhardt1"], "authors_source": "OpenReview API", "abstract": "Most existing adversarial defenses only measure robustness to $L_p$ adversarial attacks. Not only are adversaries unlikely to exclusively create small $L_p$ perturbations, adversaries are unlikely to remain fixed. Adversaries adapt and evolve their attacks; hence adversarial defenses must be robust to a broad range of unforeseen attacks. We address this discrepancy between research and reality by proposing a new evaluation framework called ImageNet-UA. Our framework enables the research community to test ImageNet model robustness against attacks not encountered during training. To create ImageNet-UA's diverse attack suite, we introduce a total of four novel adversarial attacks. We also demonstrate that, in comparison to ImageNet-UA, prevailing $L_\\infty$ robustness assessments give a narrow account of adversarial robustness. By evaluating current defenses with ImageNet-UA, we find they provide little robustness to unforeseen attacks. We hope the greater variety and realism of ImageNet-UA enables development of more robust defenses which can generalize beyond attacks seen during training.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=wl0Kr_jqM2a", "arxiv_id": "1908.08016", "paper_pdf": "papers/wl0Kr_jqM2a.pdf", "paper_pdf_sha256": "893b51cc924334850f18fef87f9947ec32fd44865d634df179c3089c73e6ea10", "paper_pdf_bytes": 2441782, "paper_pdf_source": "openreview", "code_url": "https://github.com/ddkang/advex-uar", "code_repository": "ddkang/advex-uar", "code_commit": "fd9e0efa1155cedb6d15942ed6e22e1bf1252043", "code_archive": "repos/wl0Kr_jqM2a.zip", "code_archive_sha256": "ab713cf26ce1e92ac0eb2b776b55f5d589a72047cf8242639eff317eb983f3cb", "code_archive_bytes": 1215649, "code_file_count": 34, "code_extensions": {".py": 34}, "github_disk_usage_kb": 1176, "github_languages": {"Python": 132078}, "github_archived": false, "github_pushed_at": "2024-07-25T10:16:03Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/190808016"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "HkxcUxrFPS", "year": 2020, "status": "rejected", "title": "Improving Visual Relation Detection using Depth Maps", "authors": ["Sahand Sharifzadeh", "Sina Moayed Baharlou", "Max Berrendorf", "Rajat Koner", "Volker Tresp"], "authorids": ["sharifzadeh@dbs.ifi.lmu.de", "sina.baharlou@gmail.com", "berrendorf@dbs.ifi.lmu.de", "koner@dbs.ifi.lmu.de", "volker.tresp@siemens.com"], "authors_source": "OpenReview API", "abstract": "State of the art visual relation detection methods mostly rely on object information extracted from RGB images such as predicted class probabilities, 2D bounding boxes and feature maps. In this paper, we argue that the 3D positions of objects in space can provide additional valuable information about object relations. This information helps not only to detect spatial relations, such as \\textit{standing behind}, but also non-spatial relations, such as \\textit{holding}. Since 3D information of a scene is not easily accessible, we propose incorporating a pre-trained RGB-to-Depth model within visual relation detection frameworks. We discuss different feature extraction strategies from depth maps and show their critical role in relation detection.\nOur experiments confirm that the performance of state-of-the-art visual relation detection approaches can significantly be improved by utilizing depth map information.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=HkxcUxrFPS", "arxiv_id": "1905.00966", "paper_pdf": "papers/HkxcUxrFPS.pdf", "paper_pdf_sha256": "ec4722f6b8da47217fbd3165ee451d63a949df7d0d8b57e85faace8070832e7d", "paper_pdf_bytes": 3907913, "paper_pdf_source": "openreview", "code_url": "https://github.com/Sina-Baharlou/Depth-VRD", "code_repository": "Sina-Baharlou/Depth-VRD", "code_commit": "3b416110d2da6534124526c899f21e4242a0ad24", "code_archive": "repos/HkxcUxrFPS.zip", "code_archive_sha256": "cc1259a924dfac11c424893290f935b1ba62ffedbf6f6c4f0690165219f7c423", "code_archive_bytes": 2337576, "code_file_count": 87, "code_extensions": {".py": 56, ".sh": 17, ".h": 6, ".c": 5, ".cu": 3}, "github_disk_usage_kb": 5131, "github_languages": {"Python": 403672, "Shell": 45767, "Cuda": 30757, "C": 10295, "Cython": 6273, "Makefile": 1058}, "github_archived": false, "github_pushed_at": "2022-07-24T13:13:03Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/improving-visual-relation-detection-using"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ZaB2fV3TYo", "year": 2026, "status": "rejected", "title": "Grounding Bodily Awareness in Visual Representations for Efficient Policy Learning", "authors": ["Junlin Wang", "Zhiyun Lin"], "authorids": ["~Junlin_Wang3", "~Zhiyun_Lin1"], "authors_source": "OpenReview API", "abstract": "Learning effective visual representations for robotic manipulation remains a fundamental challenge due to the complex body dynamics involved in action execution. In this paper, we study how visual representations that carry body-relevant cues can enable efficient policy learning for downstream robotic manipulation tasks. We present $\\textbf{I}$nter-token $\\textbf{Con}$trast ($\\textbf{ICon}$), a contrastive learning method applied to the token-level representations of Vision Transformers (ViTs). ICon enforces a separation in the feature space between agent-specific and environment-specific tokens, resulting in agent-centric visual representations that embed body-specific inductive biases. This framework can be seamlessly integrated into end-to-end policy learning by incorporating the contrastive loss as an auxiliary objective. Our experiments show that ICon not only improves policy performance across various manipulation tasks but also facilitates policy transfer across different robots. The project website: https://anonymous.4open.science/w/ICon/", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=ZaB2fV3TYo", "arxiv_id": "2505.18487", "paper_pdf": "papers/ZaB2fV3TYo.pdf", "paper_pdf_sha256": "4f5c0efafd59e9f85ab1fcde44f5501fc3efbabd73b85899991a205cfb674c43", "paper_pdf_bytes": 3336127, "paper_pdf_source": "openreview", "code_url": "https://github.com/HenryWJL/icon", "code_repository": "HenryWJL/icon", "code_commit": "b1957a4da27f2cd21467dc6516a8da3d6aa411cc", "code_archive": "repos/ZaB2fV3TYo.zip", "code_archive_sha256": "efef1b1baabecfda738b0630f32f30fd03def04337819dbd0ce29f4fe0d51c37", "code_archive_bytes": 249326, "code_file_count": 27, "code_extensions": {".py": 27}, "github_disk_usage_kb": 3618, "github_languages": {"Python": 139209}, "github_archived": false, "github_pushed_at": "2026-02-14T16:31:10Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/grounding-bodily-awareness-in-visual"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "JaRihIHbZm", "year": 2025, "status": "rejected", "title": "VideoAgent: Self-Improving Video Generation", "authors": ["Achint Soni", "Sreyas Venkataraman", "Abhranil Chandra", "Sebastian Fischmeister", "Percy Liang", "Bo Dai", "Sherry Yang"], "authorids": ["~Achint_Soni1", "~Sreyas_Venkataraman1", "~Abhranil_Chandra1", "~Sebastian_Fischmeister1", "~Percy_Liang1", "~Bo_Dai1", "~Sherry_Yang1"], "authors_source": "OpenReview API", "abstract": "Video generation has been used to generate visual plans for controlling robotic systems. Given an image observation and a language instruction, previous work has generated video plans which are then converted to robot controls to be executed. However, a major bottleneck in leveraging video generation for control lies in the quality of the generated videos, which often suffer from hallucinatory content and unrealistic physics, resulting in low task success when control actions are extracted from the generated videos. While scaling up dataset and model size provides a partial solution, integrating external feedback is both natural and essential for grounding video generation in the real world. With this observation, we propose VideoAgent for self-improving generated video plans based on external feedback. Instead of directly executing the generated video plan, VideoAgent first refines the generated video plans using a novel procedure which we call self-conditioning consistency, utilizing feedback from a pretrained vision-language model (VLM). As the refined video plan is being executed, VideoAgent collects additional data from the environment to further improve video plan generation. Experiments in simulated robotic manipulation from MetaWorld and iTHOR show that VideoAgent drastically reduces hallucination, thereby boosting success rate of downstream manipulation tasks. We further illustrate that VideoAgent can effectively refine real-robot videos, providing an early indicator that robotics can be an effective tool in grounding video generation in the physical world.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=JaRihIHbZm", "arxiv_id": "2410.10076", "paper_pdf": "papers/JaRihIHbZm.pdf", "paper_pdf_sha256": "535cd2df0e8c785274cafb2060b2aebd51a8fdaf2eecac2b9571960d73dace63", "paper_pdf_bytes": 32572005, "paper_pdf_source": "openreview", "code_url": "https://github.com/Video-as-Agent/VideoAgent", "code_repository": "Video-as-Agent/VideoAgent", "code_commit": "108548efee6a598da6f99a81244ea97fe8d7dd9e", "code_archive": "repos/JaRihIHbZm.zip", "code_archive_sha256": "4fdef1796cd77262f974555104eca5cdac9564b5a5460b302af1f226bc6f09ea", "code_archive_bytes": 263563, "code_file_count": 32, "code_extensions": {".py": 31, ".sh": 1}, "github_disk_usage_kb": 254, "github_languages": {"Python": 490130, "Shell": 782}, "github_archived": false, "github_pushed_at": "2025-04-25T19:28:31Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/videoagent-self-improving-video-generation"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "LhNZqkuVte", "year": 2024, "status": "rejected", "title": "HyperMask: Adaptive Hypernetwork-based Masks for Continual Learning", "authors": ["Kamil Książek", "Przemysław Spurek"], "authorids": ["~Kamil_Książek1", "~Przemysław_Spurek1"], "authors_source": "OpenReview API", "abstract": "Artificial neural networks suffer catastrophic forgetting when they are sequentially trained on multiple tasks. To overcome this problem, there exist many continual learning strategies. One of the most effective is the hypernetwork-based approach.  The hypernetwork generates the weights of a target model based on the task's identity. The model's main limitation is that hypernetwork can produce completely different nests for each task. Consequently, each task is solved separately. \nThe model does not use information from the network dedicated to previous tasks. \nWe practically produce a new architecture when we update the prior task.\nTo solve such a problem, we use the lottery ticket hypothesis, which postulates the existence of sparse subnetworks, named winning tickets, that preserve the performance of a full network.\n\nIn the paper, we propose a method called HyperMask, which trains a single network for all tasks. Hypernetwork produces semi-binary masks to obtain target subnetworks dedicated to new tasks. This solution inherits the ability of the hypernetwork to adapt to new tasks with minimal forgetting. Moreover, due to the lottery ticket hypothesis, we can use a single network with weighted subnets dedicated to each task.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=LhNZqkuVte", "arxiv_id": "2310.00113", "paper_pdf": "papers/LhNZqkuVte.pdf", "paper_pdf_sha256": "16243ebef640821e79a26a623537654d7c0d091289098b893c28abbd256e5eab", "paper_pdf_bytes": 5385063, "paper_pdf_source": "openreview", "code_url": "https://github.com/gmum/HyperMask", "code_repository": "gmum/HyperMask", "code_commit": "8cceace567e27cd1348e42df0a07d0887a62bcdd", "code_archive": "repos/LhNZqkuVte.zip", "code_archive_sha256": "5864ac6bf7bfa3e7023af3588963ce6ab626ddd3d58399f30c6983eff600a3ea", "code_archive_bytes": 199176, "code_file_count": 13, "code_extensions": {".py": 13}, "github_disk_usage_kb": 207, "github_languages": {"Python": 337353}, "github_archived": false, "github_pushed_at": "2026-08-14T15:54:16Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/hypermask-adaptive-hypernetwork-based-masks"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "erHaiO9gz3m", "year": 2023, "status": "rejected", "title": "A Kernel-Based View of Language Model Fine-Tuning", "authors": ["Sadhika Malladi", "Alexander Wettig", "Dingli Yu", "Danqi Chen", "Sanjeev Arora"], "authorids": ["~Sadhika_Malladi2", "~Alexander_Wettig1", "~Dingli_Yu1", "~Danqi_Chen1", "~Sanjeev_Arora1"], "authors_source": "OpenReview API", "abstract": "It has become standard to solve NLP tasks by  fine-tuning pre-trained language models (LMs), especially in low-data settings. There is minimal theoretical understanding of empirical success, e.g., why fine-tuning a model with $10^8$ or more parameters on a couple dozen training points does not result in overfitting. We investigate whether the Neural Tangent Kernel (NTK)--which originated as a model to study the gradient descent dynamics of infinitely wide networks with suitable random initialization--describes fine-tuning of pre-trained LMs.  This study was inspired by the decent performance of NTK  for computer vision tasks (Wei et al., 2022). We also extend the NTK formalism to  fine-tuning with Adam.  We present extensive experiments  that suggest that once the task is formulated as a masked language modeling problem through prompting, the NTK lens can often reasonably describe the model updates during fine-tuning with both SGD and Adam.\nThis kernel view also suggests an explanation for success of parameter-efficient subspace-based fine-tuning methods. Finally, we suggest a path toward a formal explanation for our findings via Tensor Programs (Yang, 2020).", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=erHaiO9gz3m", "arxiv_id": "2210.05643", "paper_pdf": "papers/erHaiO9gz3m.pdf", "paper_pdf_sha256": "56f80e4a4bcd8c4ef503a8746e2bd9a69d9b67a666094a7215842fa269850b88", "paper_pdf_bytes": 480136, "paper_pdf_source": "openreview", "code_url": "https://github.com/princeton-nlp/LM-Kernel-FT", "code_repository": "princeton-nlp/LM-Kernel-FT", "code_commit": "81450f619004ac1f981b31c3a12b4baf78c673c1", "code_archive": "repos/erHaiO9gz3m.zip", "code_archive_sha256": "5d7dc10e149f2261cd76f3a4025559bcb5e1f675ef9b4cdd198c78bcd2ab5f67", "code_archive_bytes": 47298, "code_file_count": 12, "code_extensions": {".py": 10, ".sh": 2}, "github_disk_usage_kb": 411, "github_languages": {"Python": 166036, "Shell": 4988}, "github_archived": false, "github_pushed_at": "2023-09-04T22:20:31Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/a-kernel-based-view-of-language-model-fine"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "g4E6SAAvACo", "year": 2021, "status": "rejected", "title": "Neural Architecture Search without Training", "authors": ["Joseph Mellor", "Jack Turner", "Amos Storkey", "Elliot J. Crowley"], "authorids": ["joe.mellor@ed.ac.uk", "~Jack_Turner1", "~Amos_Storkey1", "~Elliot_J._Crowley1"], "authors_source": "OpenReview API", "abstract": "The time and effort involved in hand-designing deep neural networks is immense. This has prompted the development of Neural Architecture Search (NAS) techniques to automate this design. However, NAS algorithms tend to be slow and expensive; they need to train vast numbers of candidate networks to inform the search process. This could be remedied if we could infer a network's trained accuracy from its initial state. In this work, we examine the correlation of linear maps induced by augmented versions of a single image in untrained networks and motivate how this can be used to give a measure which is highly indicative of a network’s trained performance. We incorporate this measure into a simple algorithm that allows us to search for powerful networks without any training in a matter of seconds on a single GPU, and verify its effectiveness on NAS-Bench-101 and NAS-Bench-201. Finally, we show that our approach can be readily combined with more expensive search methods for added value: we modify regularised evolutionary search to produce a novel algorithm that outperforms its predecessor.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=g4E6SAAvACo", "arxiv_id": "2006.04647", "paper_pdf": "papers/g4E6SAAvACo.pdf", "paper_pdf_sha256": "80a642fdbdb3d79517358dd3e16eb656b5c07c89c4520cfa8c90e12856072661", "paper_pdf_bytes": 8921132, "paper_pdf_source": "openreview", "code_url": "https://github.com/BayesWatch/nas-without-training", "code_repository": "BayesWatch/nas-without-training", "code_commit": "b3a82a6642564df115f989ff940ec6b8ef9ca9d3", "code_archive": "repos/g4E6SAAvACo.zip", "code_archive_sha256": "0c1fabbe716170a08a10c9cc9480e0349062c5155c222f92a404b0a72db0e2ae", "code_archive_bytes": 358568, "code_file_count": 120, "code_extensions": {".py": 119, ".sh": 1}, "github_disk_usage_kb": 1188, "github_languages": {"Python": 767072, "Shell": 3774}, "github_archived": false, "github_pushed_at": "2021-08-06T15:44:44Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/neural-architecture-search-without-training"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "HkeQ6ANYDB", "year": 2020, "status": "rejected", "title": "Blending Diverse Physical Priors with Neural Networks", "authors": ["Yunhao Ba", "Guangyuan Zhao", "Achuta Kadambi"], "authorids": ["yhba@ucla.edu", "zhaoguangyuan@ucla.edu", "achuta@ee.ucla.edu"], "authors_source": "OpenReview API", "abstract": "Rethinking physics in the era of deep learning is an increasingly important topic. This topic is special because, in addition to data, one can leverage a vast library of physical prior models (e.g. kinematics, fluid flow, etc) to perform more robust inference. The nascent sub-field of physics-based learning (PBL) studies this problem of blending neural networks with physical priors. While previous PBL algorithms have been applied successfully to specific tasks, it is hard to generalize existing PBL methods to a wide range of physics-based problems. Such generalization would require an architecture that can adapt to variations in the correctness of the physics, or in the quality of training data. No such architecture exists. In this paper, we aim to generalize PBL, by making a first attempt to bring neural architecture search (NAS) to the realm of PBL. We introduce a new method known as physics-based neural architecture search (PhysicsNAS) that is a top-performer across a diverse range of quality in the physical model and the dataset. ", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=HkeQ6ANYDB", "arxiv_id": "1910.00201", "paper_pdf": "papers/HkeQ6ANYDB.pdf", "paper_pdf_sha256": "b4c5b4d33d20f6db3af248d3570108e5f4760551775eb042dc7dc5af51c45453", "paper_pdf_bytes": 1239374, "paper_pdf_source": "openreview", "code_url": "https://github.com/PhysicsNAS/PhysicsNAS", "code_repository": "PhysicsNAS/PhysicsNAS", "code_commit": "9f9690ce064d1eca4a1f3f89cc761c484e7e594a", "code_archive": "repos/HkeQ6ANYDB.zip", "code_archive_sha256": "de4043c7a31fba559c4d156f770980bc632a7884d2acd033079756998dd71447", "code_archive_bytes": 4778446, "code_file_count": 24, "code_extensions": {".ipynb": 14, ".py": 10}, "github_disk_usage_kb": 5427, "github_languages": {"Jupyter Notebook": 24849050, "Python": 20662}, "github_archived": false, "github_pushed_at": "2019-10-01T04:15:16Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/blending-diverse-physical-priors-with-neural"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "5K9XW9gQ5r", "year": 2026, "status": "rejected", "title": "UniRL: Self-Improving Unified Multimodal Models via Supervised and Reinforcement Learning", "authors": ["Weijia Mao", "Zhenheng Yang", "Mike Zheng Shou"], "authorids": ["~Weijia_Mao1", "~Zhenheng_Yang3", "~Mike_Zheng_Shou1"], "authors_source": "OpenReview API", "abstract": "Unified multimodal large language models such as Show-o and Janus have achieved strong performance across both generation and understanding tasks. However, these models typically rely on large-scale datasets and require substantial computation during the pretraining stage. In addition, several post-training methods have been proposed, but they often depend on external data or are limited to task-specific customization. In this work, we introduce UniRL, a self-improving post-training approach. Our approach enables the model to generate images from prompts and use them as training data in each iteration, without relying on any external image data. Moreover, it enables the two tasks to enhance each other: the generated images are used for understanding, and the understanding results are used to supervise generation. We explore supervised fine-tuning (SFT) and Group Relative Policy Optimization (GRPO) to optimize the models. UniRL offers three key advantages: (1) it requires no external image data, as all training samples are generated by the model itself during training; (2) it not only improves individual task performance, but also reduces the imbalance between generation and understanding; and (3) it requires only several additional training steps during the post-training stage. We evaluate UniRL on top of Show-o and Janus, achieving a GenEval score of 0.77 for Show-o and 0.65 for Janus. Code and models will be released.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=5K9XW9gQ5r", "arxiv_id": "2505.23380", "paper_pdf": "papers/5K9XW9gQ5r.pdf", "paper_pdf_sha256": "fd5aa5544b49e60416dd95ddc5431a569cfab78fab503917764fe49ed7a4b16f", "paper_pdf_bytes": 1377316, "paper_pdf_source": "openreview", "code_url": "https://github.com/showlab/UniRL", "code_repository": "showlab/UniRL", "code_commit": "66ee4843cabb311832e5b36d530b6d70273359f6", "code_archive": "repos/5K9XW9gQ5r.zip", "code_archive_sha256": "f44aecda1a5c007b64c555de186e768695927df85a7728afd9767aded6a4ab50", "code_archive_bytes": 3764255, "code_file_count": 40, "code_extensions": {".py": 34, ".sh": 6}, "github_disk_usage_kb": 3631, "github_languages": {"Python": 479251, "Shell": 22939}, "github_archived": false, "github_pushed_at": "2025-05-30T01:47:25Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/unirl-self-improving-unified-multimodal"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "XW4Xnx0xlH", "year": 2025, "status": "rejected", "title": "Second-Order Forward-Mode Automatic Differentiation for Optimization", "authors": ["Adam D. Cobb", "Atilim Gunes Baydin", "Barak A. Pearlmutter", "Susmit Jha"], "authorids": ["~Adam_D._Cobb1", "~Atilim_Gunes_Baydin1", "~Barak_A._Pearlmutter1", "~Susmit_Jha1"], "authors_source": "OpenReview API", "abstract": "Forward gradient methods offer a promising alternative to backpropagation. Optimization that only requires forward passes could simplify hardware implementation, improve parallelism, lower memory cost, and allow for more biologically plausible learning models. This has motivated recent forward-mode automated differentiation (AD) methods. This paper presents a novel second-order forward-mode AD method for optimization that generalizes a second-order line search to a $K$-dimensional hyperplane. Unlike recent work that relies on directional derivatives (or Jacobian–Vector Products, JVPs), we use hyper-dual numbers to jointly evaluate both directional derivatives and their second-order quadratic terms. As a result, we introduce forward-mode weight perturbation with Hessian information for K-dimensional hyper-plane search (FoMoH-$K$D). We derive the convergence properties of FoMoH-$K$D and show how it generalizes to Newton’s method for $K = D$. We demonstrate this generalization empirically, and compare the performance of FoMoH-$K$D to forward gradient descent (FGD) on three case studies: Rosenbrock function used widely for evaluating optimization methods, logistic regression with 7,850 parameters, and learning a CNN classifier with 431,080 parameters. Our experiments show that FoMoH-$K$D not only achieves better performance and accuracy, but also converges faster, thus, empirically verifying our theoretical results.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=XW4Xnx0xlH", "arxiv_id": "2408.10419", "paper_pdf": "papers/XW4Xnx0xlH.pdf", "paper_pdf_sha256": "096b1621c54f164d1ee7bf7c2ec4878d4d158a477bdbb9d88f6357cec15d87ec", "paper_pdf_bytes": 9699584, "paper_pdf_source": "openreview", "code_url": "https://github.com/SRI-CSL/fomoh", "code_repository": "SRI-CSL/fomoh", "code_commit": "f0da7ae90ac9d93ea4c704462b6c75d74ef31434", "code_archive": "repos/XW4Xnx0xlH.zip", "code_archive_sha256": "780f5c473b03dbea6af439861056995d33b0935dcaa0a978352e647fd1036acd", "code_archive_bytes": 236853, "code_file_count": 17, "code_extensions": {".py": 11, ".sh": 4, ".ipynb": 2}, "github_disk_usage_kb": 232, "github_languages": {"Jupyter Notebook": 221228, "Python": 125546, "Shell": 2928}, "github_archived": false, "github_pushed_at": "2024-08-26T14:34:02Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/second-order-forward-mode-automatic"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Gq1Zjhovjr", "year": 2024, "status": "rejected", "title": "Consistency Regularization for Domain Generalization with Logit Attribution Matching", "authors": ["Han Gao", "Kaican Li", "Weiyan Xie", "Zhi LIN", "Yongxiang Huang", "Luning Wang", "Caleb Chen Cao", "Nevin L. Zhang"], "authorids": ["~Han_Gao5", "~Kaican_Li1", "~Weiyan_Xie1", "~Zhi_LIN1", "~Yongxiang_Huang1", "~Luning_Wang1", "~Caleb_Chen_Cao1", "~Nevin_L._Zhang1"], "authors_source": "OpenReview API", "abstract": "Domain generalization (DG) is about training models that generalize well to unseen domains that follow different distributions than the training domains. It has recently been shown that an effective way to achieve good DG performance is targeted data augmentation, which randomizes spurious factors while preserving robustly predictive factors in training examples. Data augmentation (DA) naturally leads to paired training examples that share the same semantic contents, which can be utilized via consistency regularization (CR). In this paper, we show that CR can further boost DG performance on top of targeted DA. We also propose a novel CR-based DG method called Logit Attribution Matching (LAM). In comparison with previous CR-based DG methods, a key advantage of LAM is that it leverages class labels  often associated with semantic sharing (SS) pairs.  Empirically we find that LAM consistently outperforms previous CR-based DG methods on benchmarks with multiple classes.  In fact, it is the only one that can consistently improve the model DG performance over the targeted DA on all evaluated datasets. To justify the CR-based approach to DG theoretically, we establish conditions for optimal DG in a causal framework and explain how CR is related to those conditions.", "decision": "Reject", "meta_review": "", "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=Gq1Zjhovjr", "arxiv_id": "2305.07888", "paper_pdf": "papers/Gq1Zjhovjr.pdf", "paper_pdf_sha256": "cc4346a681d2d3b2072589ebb8d214dcaa78ea9ba646e19db4ed5c28337572c3", "paper_pdf_bytes": 32265961, "paper_pdf_source": "openreview", "code_url": "https://github.com/Gaohan123/LAM", "code_repository": "Gaohan123/LAM", "code_commit": "e13b95c28976a1d5179167db2510fd6a398f0cd2", "code_archive": "repos/Gq1Zjhovjr.zip", "code_archive_sha256": "7a55d40331601e195aa1f42e296627358770eb000bf12062994a35fe1c75f6d6", "code_archive_bytes": 161446, "code_file_count": 57, "code_extensions": {".py": 57}, "github_disk_usage_kb": 215, "github_languages": {"Python": 458987}, "github_archived": false, "github_pushed_at": "2024-06-13T14:00:20Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/contrastive-domain-generalization-via-logit"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "eiuj6cNv4iI", "year": 2023, "status": "rejected", "title": "WebBrain: Learning to Generate Factually Correct Articles for Queries by Grounding on Large Web Corpus", "authors": ["Hongjin Qian", "Yutao Zhu", "Zhicheng Dou", "Haoqi Gu", "Xinyu Zhang", "Zheng Liu", "Ruofei Lai", "Zhao Cao", "Jian-Yun Nie", "Ji-Rong Wen"], "authorids": ["~Hongjin_Qian1", "~Yutao_Zhu1", "~Zhicheng_Dou1", "guhaoqi1@huawei.com", "~Xinyu_Zhang6", "liuzheng107@huawei.com", "lairuofei@huawei.com", "~Zhao_Cao1", "~Jian-Yun_Nie1", "~Ji-Rong_Wen1"], "authors_source": "OpenReview API", "abstract": "In this paper, we introduce a new NLP task – generating short factual articles for queries by mining supporting evidence from the Web. In this task, called WebBrain, the ultimate goal is to generate a fluent, informative, and factually-correct short article (e.g., Wiki article) for a factual query unseen in Wikipedia. To enable experiments on WebBrain, we construct a large-scale dataset WebBrain-Raw by extracting English Wikipedia articles and their crawlable Wiki references. WebBrain-Raw is ten times larger than the previous biggest peer dataset, which can greatly benefit the research community. Besides, we empirically analyze the performances of the current state-of-the-art NLP techniques on WebBrain and introduce a new framework ReGen, which enhances the generation factualness by improved evidence retrieval and task-specific pre-training for generation. Experiment results show that ReGen outperforms all baselines in both automatic and human evaluations.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=eiuj6cNv4iI", "arxiv_id": "2304.04358", "paper_pdf": "papers/eiuj6cNv4iI.pdf", "paper_pdf_sha256": "dbe08c549a562fb0fc4f10b109d72229cc374c4f43bdd0e4c7f56fa554826253", "paper_pdf_bytes": 793685, "paper_pdf_source": "openreview", "code_url": "https://github.com/qhjqhj00/WebBrain", "code_repository": "qhjqhj00/WebBrain", "code_commit": "d67b53e00bd9438f187080758dffc565d29bb649", "code_archive": "repos/eiuj6cNv4iI.zip", "code_archive_sha256": "becca44f5f1c9df861cb3ac84dc46e66a9c9c36f4011ebdeeea80eb87b4e9e5f", "code_archive_bytes": 402392, "code_file_count": 18, "code_extensions": {".py": 16, ".sh": 2}, "github_disk_usage_kb": 419, "github_languages": {"Python": 52602, "Shell": 529}, "github_archived": false, "github_pushed_at": "2023-06-07T09:43:17Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/webbrain-learning-to-generate-factually"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "fyLvrx9M9YP", "year": 2022, "status": "rejected", "title": "Towards Unsupervised Content Disentanglement in Sentence Representations via Syntactic Roles", "authors": ["Ghazi Felhi", "Joseph Le Roux", "Djamé Seddah"], "authorids": ["~Ghazi_Felhi1", "~Joseph_Le_Roux1", "~Djamé_Seddah1"], "authors_source": "OpenReview API", "abstract": "Linking neural representations to linguistic factors is crucial in order to build and analyze NLP models interpretable by humans. Among these factors, syntactic roles (e.g. subjects, direct objects,$\\dots$)  and their realizations are essential markers since they can be understood as a decomposition of predicative structures and thus the meaning of sentences. Starting from a deep probabilistic generative model with attention, we measure the interaction between latent variables and realizations of syntactic roles, and show that it is possible to obtain, without supervision, representations of sentences where different syntactic roles correspond to clearly identified different latent variables. The probabilistic model we propose is an Attention-Driven Variational Autoencoder (ADVAE). Drawing inspiration from Transformer-based machine translation models, ADVAEs enable the analysis of the interactions between latent variables and input tokens through attention. We also develop an evaluation protocol to measure disentanglement with regard to the realizations of syntactic roles. This protocol is based on attention maxima for the encoder and on disturbing individual latent variables for the decoder. Our experiments on raw English text from the SNLI dataset show that $\\textit{i)}$ disentanglement of syntactic roles can be induced without supervision, $\\textit{ii)}$  ADVAE separates more syntactic roles than classical sequence VAEs, $\\textit{iii)}$ realizations of syntactic roles can be separately modified in sentences by mere intervention on the associated latent variables. Our work constitutes a first step towards unsupervised controllable content generation.  The code for our work is publicly available.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=fyLvrx9M9YP", "arxiv_id": "2206.11184", "paper_pdf": "papers/fyLvrx9M9YP.pdf", "paper_pdf_sha256": "49c26a7bd524b87c146e5fa39e5a4bd1bc1e95f12835d2f8c3dc13fa9b4ef078", "paper_pdf_bytes": 731872, "paper_pdf_source": "openreview", "code_url": "https://github.com/ghazi-f/ADVAE", "code_repository": "ghazi-f/ADVAE", "code_commit": "cbf1c9d542eb826c7599425b32bd701e1a105bd3", "code_archive": "repos/fyLvrx9M9YP.zip", "code_archive_sha256": "f3cccd3558957a005e74745293c243fc84f7612968967682fe9fbe30363802e4", "code_archive_bytes": 59477, "code_file_count": 12, "code_extensions": {".py": 11, ".sh": 1}, "github_disk_usage_kb": 1719, "github_languages": {"Python": 365572, "Shell": 3848}, "github_archived": false, "github_pushed_at": "2022-07-05T09:22:25Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/towards-unsupervised-content-disentanglement-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "rJlUhhVYvS", "year": 2020, "status": "rejected", "title": "Understanding Isomorphism Bias in Graph Data Sets ", "authors": ["Ivanov Sergey", "Sviridov Sergey", "Evgeny Burnaev"], "authorids": ["ivanovserg990@gmail.com", "sergei.sviridov@gmail.com", "e.burnaev@skoltech.ru"], "authors_source": "OpenReview API", "abstract": "In recent years there has been a rapid increase in classification methods on graph structured data. Both in graph kernels and graph neural networks, one of the implicit assumptions of successful state-of-the-art models was that incorporating graph isomorphism features into the architecture leads to better empirical performance. However, as we discover in this work, commonly used data sets for graph classification have repeating instances which cause the problem of isomorphism bias, i.e. artificially increasing the accuracy of the models by memorizing target information from the training set. This prevents fair competition of the algorithms and raises a question of the validity of the obtained results. We analyze 54 data sets, previously extensively used for graph-related tasks, on the existence of isomorphism bias, give a set of recommendations to machine learning practitioners to properly set up their models, and open source new data sets for the future experiments. ", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=rJlUhhVYvS", "arxiv_id": "1910.12091", "paper_pdf": "papers/rJlUhhVYvS.pdf", "paper_pdf_sha256": "515f5c67fbbcb82182b619125237be087af6016c3b1ad49fd3171968a87cdd25", "paper_pdf_bytes": 429170, "paper_pdf_source": "openreview", "code_url": "https://github.com/nd7141/iso_bias", "code_repository": "nd7141/iso_bias", "code_commit": "a235a71814c4b0e644a7c0d66a16980332005327", "code_archive": "repos/rJlUhhVYvS.zip", "code_archive_sha256": "84c13ad3773c57d7b15e31c808f62c92b4e0e937b72d02fe531efb5a9af31580", "code_archive_bytes": 4457038, "code_file_count": 19, "code_extensions": {".py": 13, ".c": 3, ".h": 1, ".cc": 1, ".ipynb": 1}, "github_disk_usage_kb": 5877, "github_languages": {"Jupyter Notebook": 2735667, "Python": 91041, "C++": 31197, "C": 15150, "Makefile": 267}, "github_archived": false, "github_pushed_at": "2019-12-12T16:43:05Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/understanding-isomorphism-bias-in-graph-data"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "bhPaXhWVKG", "year": 2026, "status": "rejected", "title": "MermaidFlow: Redefining Agentic Workflow Generation via Safety-Constrained Evolutionary Programming", "authors": ["Chengqi Zheng", "Yewen Pu", "Jianda Chen", "Yueming Lyu", "Wen zheng terence Ng", "Haopeng Zhang", "Yew-Soon Ong", "Ivor Tsang", "Haiyan Yin"], "authorids": ["~Chengqi_Zheng1", "~Yewen_Pu1", "~Jianda_Chen1", "~Yueming_Lyu1", "~Wen_zheng_terence_Ng1", "~Haopeng_Zhang3", "~Yew-Soon_Ong1", "~Ivor_Tsang1", "~Haiyan_Yin1"], "authors_source": "OpenReview API", "abstract": "Despite the promise of autonomous agentic reasoning, existing workflow generation methods frequently produce fragile, unexecutable plans due to unconstrained LLM-driven construction. We propose MermaidFlow, a framework that redefines the agentic search space through safety-constrained graph evolution. At its core, MermaidFlow represent workflows as a verifiable intermediate representation using Mermaid, a structured and human-interpretable graph language. We formulate domain-aware evolutionary operators, i.e., crossover,  mutation, insertion, and deletion, to preserve semantic correctness while promoting structural diversity, enabling efficient exploration of a high-quality, statically verifiable workflow space. Without modifying task settings or evaluation protocols, MermaidFlow achieves consistent improvements in success rates and faster convergence to executable plans on the agent reasoning benchmark. The experimental results demonstrate that safety-constrained graph evolution offers a scalable, modular foundation for robust and interpretable agentic reasoning systems.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=bhPaXhWVKG", "arxiv_id": "2505.22967", "paper_pdf": "papers/bhPaXhWVKG.pdf", "paper_pdf_sha256": "4bca0c0f42e49ba134dad241f8c50fa0fb8fbd1acac577bd3c30ee849421850f", "paper_pdf_bytes": 3199007, "paper_pdf_source": "openreview", "code_url": "https://github.com/chengqiArchy/MermaidFlow", "code_repository": "chengqiArchy/MermaidFlow", "code_commit": "cd2478dff0b0f1e01137be1bef8b04ea2aa94289", "code_archive": "repos/bhPaXhWVKG.zip", "code_archive_sha256": "c7bfca16649a87041552940958efa6759adc6054bb7de236a80a6df186989385", "code_archive_bytes": 4045407, "code_file_count": 60, "code_extensions": {".py": 60}, "github_disk_usage_kb": 3766, "github_languages": {"Python": 298080, "Mermaid": 3280}, "github_archived": false, "github_pushed_at": "2025-12-12T06:43:03Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/mermaidflow-redefining-agentic-workflow"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "RLzeoy4FzP", "year": 2025, "status": "rejected", "title": "Hybrid Preferences: Learning to Route Instances for Human vs. AI Feedback", "authors": ["Lester James Validad Miranda", "Yizhong Wang", "Yanai Elazar", "Sachin Kumar", "Valentina Pyatkin", "Faeze Brahman", "Noah A. Smith", "Hannaneh Hajishirzi", "Pradeep Dasigi"], "authorids": ["~Lester_James_Validad_Miranda1", "~Yizhong_Wang2", "~Yanai_Elazar1", "~Sachin_Kumar1", "~Valentina_Pyatkin1", "~Faeze_Brahman1", "~Noah_A._Smith2", "~Hannaneh_Hajishirzi1", "~Pradeep_Dasigi1"], "authors_source": "OpenReview API", "abstract": "Learning from human feedback has enabled the alignment of language models (LMs) with human preferences. However, directly collecting human preferences can be expensive, time-consuming, and can have high variance. An appealing alternative is to distill preferences from LMs as a source of synthetic annotations as they are more consistent, cheaper, and scale better than human annotation; however, they are also prone to biases and errors. In this work, we introduce a routing framework that combines inputs from humans and LMs to achieve better annotation quality, while reducing the total cost of human annotation. The crux of our approach is to identify preference instances that will benefit from human annotations. We formulate this as an optimization problem: given a preference dataset and an evaluation metric, we train a performance prediction model to predict a reward model's performance on an arbitrary combination of human and LM annotations and employ a routing strategy that selects a combination that maximizes predicted performance. We train the performance prediction model on MultiPref, a new preference dataset with 10K instances paired with human and LM labels. We show that the selected hybrid mixture of LM and direct human preferences using our routing framework achieves better reward model performance compared to using either one exclusively. We simulate selective human preference collection on three other datasets and show that our method generalizes well to all three. We analyze features from the routing model to identify characteristics of instances that can benefit from human feedback, e.g., prompts with a moderate safety concern or moderate intent complexity. We release the dataset, annotation platform, and source code used in this study to foster more efficient and accurate preference collection in the future.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=RLzeoy4FzP", "arxiv_id": "2410.19133", "paper_pdf": "papers/RLzeoy4FzP.pdf", "paper_pdf_sha256": "8bcb1b99f131b0b493983f8831f6e30ab8c5ea39a04012459adb87e13ed5b1b3", "paper_pdf_bytes": 2960681, "paper_pdf_source": "openreview", "code_url": "https://github.com/allenai/hybrid-preferences", "code_repository": "allenai/hybrid-preferences", "code_commit": "4d96b27b989828b0b69a430f68d25b6b36e9015a", "code_archive": "repos/RLzeoy4FzP.zip", "code_archive_sha256": "fb6a53ea5b4b56f2714b96d0c610ebaa56144131639f30ea3f25f6113eb9b60e", "code_archive_bytes": 88133, "code_file_count": 26, "code_extensions": {".py": 15, ".sh": 11}, "github_disk_usage_kb": 271, "github_languages": {"Python": 171361, "Shell": 13112, "Dockerfile": 2390, "Makefile": 525}, "github_archived": false, "github_pushed_at": "2025-07-23T03:36:45Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/hybrid-preferences-learning-to-route"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "JAKcnjzQI3", "year": 2024, "status": "rejected", "title": "MaSS: Multi-attribute Selective Suppression for Utility-preserving Data Transformation from an Information-theoretic Perspective", "authors": ["Yizhuo Chen", "Chun-Fu Chen", "Hsiang Hsu", "Shaohan Hu", "Marco Pistoia", "Tarek F. Abdelzaher"], "authorids": ["~Yizhuo_Chen2", "~Chun-Fu_Chen1", "~Hsiang_Hsu1", "~Shaohan_Hu2", "~Marco_Pistoia2", "~Tarek_F._Abdelzaher1"], "authors_source": "OpenReview API", "abstract": "The growing richness of large-scale datasets has been a crucial driving force behind the rapid advancement and wide adoption of machine learning technologies. The massive collection and usage of data, however, pose an increasing risk for people’s private and sensitive information due to either inadvertent mishandling or malicious exploitation. Besides legislative solutions, many technical approaches have been proposed towards data privacy protection. However, they bear various limitations such as leading to degraded data availability and utility, or relying on heuristics and lacking solid theoretical bases. To overcome these limitations, we propose a formal information-theoretic definition for this utility-preserving privacy protection problem, and design a data-driven learnable data transformation framework that is capable of selectively suppressing sensitive attributes from target datasets while preserving the other useful attributes, regardless of whether or not they are known in advance or explicitly annotated for preservation. We provide rigorous theoretical analyses on the operational bounds for our framework, and carry out comprehensive experimental evaluations using datasets of a variety of modalities, including facial images, voice audio clips, and human activity motion sensor signals. Results demonstrate the effectiveness and generalizability of our method on different tasks and configurations.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=JAKcnjzQI3", "arxiv_id": "2405.14981", "paper_pdf": "papers/JAKcnjzQI3.pdf", "paper_pdf_sha256": "bc025ab9a5d9beab17e33812b06072106fd2e9a3aca80076251b3b69e077fc89", "paper_pdf_bytes": 683556, "paper_pdf_source": "openreview", "code_url": "https://github.com/jpmorganchase/MaSS", "code_repository": "jpmorganchase/MaSS", "code_commit": "6fbe9be1ff155b3fdd2677de7977d801d10000b5", "code_archive": "repos/JAKcnjzQI3.zip", "code_archive_sha256": "9ed8b9f55c5a55c545af2227c9fd5d9b81e99bd585b6c93e2d7525f5d72dfbdb", "code_archive_bytes": 228313, "code_file_count": 25, "code_extensions": {".py": 23, ".sh": 2}, "github_disk_usage_kb": 217, "github_languages": {"Python": 104454, "Shell": 6389}, "github_archived": true, "github_pushed_at": "2024-07-25T10:20:11Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/mass-multi-attribute-selective-suppression-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "cVEOiBz2Em", "year": 2023, "status": "rejected", "title": "Deep Bayesian Active Learning for Accelerating Stochastic Simulation", "authors": ["Dongxia Wu", "Ruijia Niu", "Matteo Chinazzi", "Alessandro Vespignani", "Yian Ma", "Rose Yu"], "authorids": ["~Dongxia_Wu1", "~Ruijia_Niu1", "~Matteo_Chinazzi1", "~Alessandro_Vespignani1", "~Yian_Ma1", "~Rose_Yu1"], "authors_source": "OpenReview API", "abstract": "Stochastic simulations such as large-scale, spatiotemporal, age-structured epidemic models are computationally expensive at fine-grained resolution. While deep surrogate models can speed up the simulations, doing so for stochastic simulations and with active learning approaches is an underexplored area. We propose Interactive Neural Process (INP), a deep Bayesian active learning framework for learning deep surrogate models to accelerate stochastic simulations. INP consists of two components, a spatiotemporal surrogate model built upon Neural Process (NP) family and an acquisition function for active learning. For surrogate modeling, we develop Spatiotemporal Neural Process (STNP) to mimic the simulator dynamics. For active learning, we propose a novel acquisition function, Latent Information Gain (LIG), calculated in the latent space of NP based models. We perform a theoretical analysis and demonstrate that LIG reduces sample complexity compared with random sampling in high dimensions. We also conduct empirical studies on two complex spatiotemporal simulators for reaction diffusion and infectious disease. The results demonstrate that STNP outperforms the baselines in the offline learning setting and LIG achieves the state-of-the-art for Bayesian active learning.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=cVEOiBz2Em", "arxiv_id": "2106.02770", "paper_pdf": "papers/cVEOiBz2Em.pdf", "paper_pdf_sha256": "5c30d4664d0bb945718e3cfd0116d3e80cc1cea0c28146e009d57c0a33144037", "paper_pdf_bytes": 3432688, "paper_pdf_source": "openreview", "code_url": "https://github.com/Rose-STL-Lab/Interactive-Neural-Process", "code_repository": "Rose-STL-Lab/Interactive-Neural-Process", "code_commit": "962e4431364e8175253801dfdc28755131e5f14e", "code_archive": "repos/cVEOiBz2Em.zip", "code_archive_sha256": "dcf31bb6be9c0f8bdd46ba00ad4fbb95b794d24b9466c46cb619a357672ce426", "code_archive_bytes": 491101, "code_file_count": 29, "code_extensions": {".py": 29}, "github_disk_usage_kb": 433, "github_languages": {"Python": 303961}, "github_archived": false, "github_pushed_at": "2024-10-09T23:11:25Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/accelerating-stochastic-simulation-with"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "bUAdXW8wN6", "year": 2022, "status": "rejected", "title": "Domain Invariant Adversarial Learning", "authors": ["Matan Levi", "Idan Attias", "Aryeh Kontorovich"], "authorids": ["~Matan_Levi1", "~Idan_Attias1", "~Aryeh_Kontorovich1"], "authors_source": "OpenReview API", "abstract": "The phenomenon of adversarial examples illustrates one of the most basic vulnerabilities of deep neural networks. Among the variety of techniques introduced to surmount this inherent weakness, adversarial training has emerged as the most effective strategy to achieve robustness. Typically, this is achieved by balancing robust and natural objectives. In this work, we aim to further reduce the trade-off between robust and standard accuracy by enforcing a domain-invariant feature representation. We present a new adversarial training method, Domain Invariant Adversarial Learning (DIAL), which learns a feature representation which is both robust and domain invariant. DIAL uses a variant of Domain Adversarial Neural Network (DANN) on the natural domain and its corresponding adversarial domain. In a case where the source domain consists of natural examples and the target domain is the adversarially perturbed examples, our method learns a feature representation constrained not to discriminate between the natural and adversarial examples, and can therefore achieve a more robust representation. Our experiments indicate that our method improves both robustness and standard accuracy, when compared to other state-of-the-art adversarial training methods.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=bUAdXW8wN6", "arxiv_id": "2104.00322", "paper_pdf": "papers/bUAdXW8wN6.pdf", "paper_pdf_sha256": "b146067d64014f407811f8ee0b8f00b940dc6fa6785c829c3bbb360342cd4eab", "paper_pdf_bytes": 4161643, "paper_pdf_source": "openreview", "code_url": "https://github.com/matanle51/DIAL", "code_repository": "matanle51/DIAL", "code_commit": "a94066d530cbd39c02957af0b6924d5e01ac66e6", "code_archive": "repos/bUAdXW8wN6.zip", "code_archive_sha256": "b6f34adab7914bb26df2058558a9bc93be285452478d9f1817b7c7ef531d84d9", "code_archive_bytes": 2111918, "code_file_count": 23, "code_extensions": {".py": 23}, "github_disk_usage_kb": 1764, "github_languages": {"Python": 83326}, "github_archived": false, "github_pushed_at": "2022-09-11T10:33:37Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/domain-invariant-adversarial-learning"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "1toB0Fo9CZy", "year": 2021, "status": "rejected", "title": "Neural Architecture Search of SPD Manifold Networks", "authors": ["Rhea Sanjay Sukthanker", "Zhiwu Huang", "Suryansh Kumar", "Erik Goron", "Yan Wu", "Luc Van Gool"], "authorids": ["~Rhea_Sanjay_Sukthanker1", "~Zhiwu_Huang1", "~Suryansh_Kumar1", "~Erik_Goron1", "~Yan_Wu4", "~Luc_Van_Gool1"], "authors_source": "OpenReview API", "abstract": "In this paper, we propose a new neural architecture search (NAS) problem of Symmetric Positive Definite (SPD) manifold networks. Unlike the conventional NAS problem, our problem requires to search for a unique computational cell called the SPD cell. This SPD cell serves as a basic building block of SPD neural architectures. An efficient solution to our problem is important to minimize the extraneous manual effort in the SPD neural architecture design. To accomplish this goal, we first introduce a geometrically rich and diverse SPD neural architecture search space for an efficient SPD cell design. Further, we model our new NAS problem using the supernet strategy, which models the architecture search problem as a one-shot training process of a single supernet. Based on the supernet modeling, we exploit a differentiable NAS algorithm on our relaxed continuous search space for SPD neural architecture search. Statistical evaluation of our method on drone, action, and emotion recognition tasks mostly provides better results than the state-of-the-art SPD networks and NAS algorithms. Empirical results show that our algorithm excels in discovering better SPD network design and providing models that are more than 3 times lighter than searched by state-of-the-art NAS algorithms.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=1toB0Fo9CZy", "arxiv_id": "2010.14535", "paper_pdf": "papers/1toB0Fo9CZy.pdf", "paper_pdf_sha256": "cc57aebd2f196ffa35c84fd367b468b1a0335bf6f8b83689463eccf5164e7917", "paper_pdf_bytes": 1645915, "paper_pdf_source": "openreview", "code_url": "https://github.com/rheasukthanker/SPDNetNAS", "code_repository": "rheasukthanker/SPDNetNAS", "code_commit": "0f82e735f03603c0d355e2d8bf8fc570ac225ddb", "code_archive": "repos/1toB0Fo9CZy.zip", "code_archive_sha256": "90ce892caee48bd59c5f53f622d255b2f35eacd133e5a5dd9e92d17877288e82", "code_archive_bytes": 891545, "code_file_count": 39, "code_extensions": {".py": 39}, "github_disk_usage_kb": 1206, "github_languages": {"Python": 337356}, "github_archived": false, "github_pushed_at": "2022-02-18T01:32:19Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/neural-architecture-search-of-spd-manifold-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "r1gzoaNtvr", "year": 2020, "status": "rejected", "title": "Emergence of Compositional Language with Deep Generational Transmission", "authors": ["Michael Cogswell", "Jiasen Lu", "Stefan Lee", "Devi Parikh", "Dhruv Batra"], "authorids": ["cogswell@gatech.edu", "jiasenlu@gatech.edu", "steflee@gatech.edu", "parikh@gatech.edu", "dbatra@gatech.edu"], "authors_source": "OpenReview API", "abstract": "Recent work has studied the emergence of language among deep reinforcement learning agents that must collaborate to solve a task. Of particular interest are the factors that cause language to be compositional---i.e., express meaning by combining words which themselves have meaning. Evolutionary linguists have found that in addition to structural priors like those already studied in deep learning, the dynamics of transmitting language from generation to generation contribute significantly to the emergence of  compositionality. In this paper, we introduce these cultural evolutionary dynamics into language emergence by periodically replacing agents in a population to create a knowledge gap, implicitly inducing cultural transmission of language. We show that this implicit cultural transmission encourages the resulting languages to exhibit better compositional generalization.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=r1gzoaNtvr", "arxiv_id": "1904.09067", "paper_pdf": "papers/r1gzoaNtvr.pdf", "paper_pdf_sha256": "3d50add8df81ba45f44bea6f2ea167f3079074e42746373f8a72aa7163fb0059", "paper_pdf_bytes": 5518103, "paper_pdf_source": "openreview", "code_url": "https://github.com/mcogswell/evolang", "code_repository": "mcogswell/evolang", "code_commit": "12034f8e8395c95b012af74ce7d121775c7cf4c3", "code_archive": "repos/r1gzoaNtvr.zip", "code_archive_sha256": "1b1fb13df28ad394fb3a40c83ba85eb551ca6c4bc85353efb66d1c06b83c9a78", "code_archive_bytes": 6195247, "code_file_count": 13, "code_extensions": {".py": 12, ".ipynb": 1}, "github_disk_usage_kb": 6042, "github_languages": {"Jupyter Notebook": 9872698, "Python": 161765}, "github_archived": false, "github_pushed_at": "2020-05-20T04:03:23Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/emergence-of-compositional-language-with-deep"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "GM9H3iM7VJ", "year": 2026, "status": "rejected", "title": "LLMail-Inject: A Dataset from a Realistic Adaptive Prompt Injection Challenge", "authors": ["Sahar Abdelnabi", "Aideen Fay", "Ahmed Salem", "Egor Zverev", "Kai-Chieh Liao", "Chi-Huang Liu", "Chun Chih Kuo", "Jannis Weigend", "Danyael Manlangit", "Alex Apostolov", "Haris Umair", "João Donato", "Masayuki Kawakita", "Athar Mahboob", "Tran Huu Bach", "Tsun-Han Chiang", "Myeongjin Cho", "Hajin Choi", "Byeonghyeon Kim", "HyeonjinLee", "Benjamin Pannell", "Conor McCauley", "Mark Russinovich", "Andrew Paverd", "Giovanni Cherubin"], "authorids": ["~Sahar_Abdelnabi1", "~Aideen_Fay1", "~Ahmed_Salem2", "~Egor_Zverev1", "~Kai-Chieh_Liao1", "~Chi-Huang_Liu1", "~Chun_Chih_Kuo1", "~Jannis_Weigend1", "~Danyael_Manlangit1", "~Alex_Apostolov1", "~Haris_Umair1", "~João_Donato1", "~Masayuki_Kawakita1", "~Athar_Mahboob1", "~Tran_Huu_Bach1", "~Tsun-Han_Chiang2", "~Myeongjin_Cho1", "~Hajin_Choi1", "~Byeonghyeon_Kim1", "~HyeonjinLee1", "~Benjamin_Pannell1", "~Conor_McCauley1", "~Mark_Russinovich1", "~Andrew_Paverd1", "~Giovanni_Cherubin1"], "authors_source": "OpenReview API", "abstract": "Indirect Prompt Injection attacks exploit a fundamental weakness of large language models (LLMs): the inability to reliably separate instructions from data. This vulnerability poses critical real-world security risks, yet systematic evaluation against adaptive adversaries remains largely unexplored. We introduce LLMail-Inject, the first large-scale public challenge simulating a realistic email-assistant environment—a high-value attack surface in practice. Involving 839 participants, the challenge produced 208,095 unique attack prompts across multiple LLM architectures and retrieval configurations. Unlike prior benchmarks, LLMail-Inject requires end-to-end compromise: attacks must be retrieved, adaptively evade defenses, trigger unauthorized tool calls with correct formatting, and exfiltrate contextual data.\nOur findings reveal a stark gap between perceived and actual robustness: while state-of-the-art models achieve <5% success on existing benchmarks, LLMail-Inject drives success rates to 32%, exposing the fragility of current defenses under realistic conditions. We release the dataset, code, and analysis to catalyze research toward structural, practical defenses against prompt injection.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=GM9H3iM7VJ", "arxiv_id": "2506.09956", "paper_pdf": "papers/GM9H3iM7VJ.pdf", "paper_pdf_sha256": "53f7a2ce688f04179bd963f9d30964a9beb8cf94d1cb5f5f0e03c1f05c0a13e0", "paper_pdf_bytes": 4227128, "paper_pdf_source": "openreview", "code_url": "https://github.com/microsoft/llmail-inject-challenge", "code_repository": "microsoft/llmail-inject-challenge", "code_commit": "ad115315c1cb34381d20875d6675a6cfe6ca80fa", "code_archive": "repos/GM9H3iM7VJ.zip", "code_archive_sha256": "217d591012cd46d1481fda280e10d3eb14f4d2ab123e101ba6298aa1362c6a71", "code_archive_bytes": 4676926, "code_file_count": 83, "code_extensions": {".py": 61, ".vue": 15, ".ts": 7}, "github_disk_usage_kb": 4113, "github_languages": {"Python": 202588, "Vue": 29962, "TypeScript": 9695, "TypeSpec": 7873, "CSS": 4921, "HTML": 2645}, "github_archived": false, "github_pushed_at": "2026-04-09T07:37:33Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/llmail-inject-a-dataset-from-a-realistic"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "XLt0eudh8t", "year": 2025, "status": "rejected", "title": "Efficient Neural Common Neighbor for Temporal Graph Link Prediction", "authors": ["Xiaohui Zhang", "Yanbo Wang", "Xiyuan Wang", "Muhan Zhang"], "authorids": ["~Xiaohui_Zhang9", "~Yanbo_Wang2", "~Xiyuan_Wang1", "~Muhan_Zhang1"], "authors_source": "OpenReview API", "abstract": "Temporal graphs are ubiquitous in real-world scenarios, such as social network, trade and transportation. Predicting dynamic links between nodes in a temporal graph is of vital importance. \\textcolor{blue}{\nTraditional memory-based methods typically leverage the temporal neighborhood of interaction histories to generate node embeddings, which are then aggregated to predict links between source and target nodes. However, these methods primarily focus on learning individual node representations and often neglect the nature of pairwise representation learning aspect. While some recent methods attempt to capture pairwise features, they are less emphasized in large-scale datasets like TGB. Meanwhile, most of these \nexisting methods tend to suffer from high computational complexity due to the repeated calculation of node embeddings.\n}\nMotivated by the success of Neural Common Neighbor (NCN) for static graph link prediction, we propose \\textbf{TNCN}, a temporal version of NCN for link prediction in temporal graphs. Based on a memory-based backbone instead of traditional static graph neural network, TNCN dynamically updates a temporal neighbor dictionary for each node, and utilizes multi-hop common neighbors between the source and target node to learn a more effective pairwise representation. We validate our model on five large-scale real-world datasets from the Temporal Graph Benchmark (TGB), and find that it achieves new state-of-the-art performance on three of them. Additionally, TNCN demonstrates excellent scalability on large datasets, outperforming popular GNN baselines by up to 6.4 times in speed.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=XLt0eudh8t", "arxiv_id": "2406.07926", "paper_pdf": "papers/XLt0eudh8t.pdf", "paper_pdf_sha256": "86d0e9925d3f58f0f3e7aa3b9aab31ab58621bfd3871f1f92db5fc8a80bc8d55", "paper_pdf_bytes": 625900, "paper_pdf_source": "openreview", "code_url": "https://github.com/GraphPKU/TNCN", "code_repository": "GraphPKU/TNCN", "code_commit": "4c93a2a6c67c4033bab50a2c0a6833cbf8f24b5b", "code_archive": "repos/XLt0eudh8t.zip", "code_archive_sha256": "e9548cd9debcdf4ab87091e993456d48f3341ccda68da61a695fc46d80dc048b", "code_archive_bytes": 167128, "code_file_count": 60, "code_extensions": {".py": 60}, "github_disk_usage_kb": 272, "github_languages": {"Python": 250960}, "github_archived": false, "github_pushed_at": "2025-05-04T17:21:55Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/efficient-neural-common-neighbor-for-temporal"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "iT1ttQXwOg", "year": 2024, "status": "rejected", "title": "Equivariant Deep Weight Space Alignment", "authors": ["Aviv Navon", "Aviv Shamsian", "Ethan Fetaya", "Gal Chechik", "Nadav Dym", "Haggai Maron"], "authorids": ["~Aviv_Navon1", "~Aviv_Shamsian1", "~Ethan_Fetaya1", "~Gal_Chechik1", "~Nadav_Dym1", "~Haggai_Maron1"], "authors_source": "OpenReview API", "abstract": "Permutation symmetries of deep networks make simple operations like model averaging and similarity estimation challenging. In many cases, aligning the weights of the networks, i.e., finding optimal permutations between their weights, is necessary. More generally, weight alignment is essential for a wide range of applications, from model merging, through exploring the optimization landscape of deep neural networks, to defining meaningful distance functions between neural networks. Unfortunately, weight alignment is an NP-hard problem. Prior research has mainly focused on solving relaxed versions of the alignment problem, leading to either time-consuming methods or sub-optimal solutions. To accelerate the alignment process and improve its quality, we propose a novel framework aimed at learning to solve the weight alignment problem, which we name DEEP-ALIGN. To that end,  we first demonstrate that weight alignment adheres to two fundamental symmetries and then, propose a deep architecture that respects these symmetries. Notably, our framework does not require any labeled data. We provide a theoretical analysis of our approach and evaluate DEEP-ALIGN on several types of network architectures and learning setups. Our experimental results indicate that a feed-forward pass with DEEP-ALIGN produces better or equivalent alignments compared to those produced by current optimization algorithms. Additionally, our alignments can be used as an initialization for other methods to gain even better solutions with a significant speedup in convergence.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=iT1ttQXwOg", "arxiv_id": "2310.13397", "paper_pdf": "papers/iT1ttQXwOg.pdf", "paper_pdf_sha256": "30d39a410e5c50490f90a6635e41c8cb40031b6c887e1713dd15d9668116efdf", "paper_pdf_bytes": 3896857, "paper_pdf_source": "openreview", "code_url": "https://github.com/AvivNavon/deep-align", "code_repository": "AvivNavon/deep-align", "code_commit": "7712044ddbf3d5648b8f7bdf31221f775895ed78", "code_archive": "repos/iT1ttQXwOg.zip", "code_archive_sha256": "64b9bb906a72857ab254b8eb681211eb0b78788da53eba5292d6dd8c4e055c69", "code_archive_bytes": 263315, "code_file_count": 29, "code_extensions": {".py": 28, ".ipynb": 1}, "github_disk_usage_kb": 225, "github_languages": {"Python": 292292, "Jupyter Notebook": 20866}, "github_archived": false, "github_pushed_at": "2024-08-18T18:25:32Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/equivariant-deep-weight-space-alignment"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Siln8xpTMrZ", "year": 2023, "status": "rejected", "title": "DADAO: Decoupled Accelerated Decentralized Asynchronous Optimization", "authors": ["Adel Nabli", "Edouard Oyallon"], "authorids": ["~Adel_Nabli1", "~Edouard_Oyallon1"], "authors_source": "OpenReview API", "abstract": "DADAO is a novel decentralized asynchronous stochastic first order algorithm to minimize a sum of $L$-smooth and $\\mu$-strongly convex functions distributed over a time-varying connectivity network of size $n$.  We model the local gradient updates and gossip communication procedures with separate independent Poisson Point Processes, decoupling the computation and communication steps in addition to making the whole approach completely asynchronous. Our method employs primal gradients and does not use a multi-consensus inner loop nor other ad-hoc mechanisms such as Error Feedback, Gradient Tracking, or a Proximal operator. By relating the inverse of the smallest positive eigenvalue $\\chi^*_1$ and the effective resistance $\\chi_2^*$ of our graph to a necessary minimal communication rate between nodes of the network, we show that our algorithm requires $\\mathcal{O}(n\\sqrt{\\frac{L}{\\mu}}\\log \\epsilon)$ local gradients and only  $\\mathcal{O}(n\\sqrt{\\chi_1^*\\chi_2^*}\\sqrt{\\frac{L}{\\mu}}\\log \\epsilon)$ communications to reach a precision $\\epsilon$. If SGD with uniform noise $\\sigma^2$ is used, we reach a precision $\\epsilon$ with same speed, up to a bias term in $\\mathcal{O}(\\frac{\\sigma^2}{\\sqrt{\\mu L}})$. This improves upon the bounds obtained with current state-of-the-art approaches, our simulations validating the strength of our relatively unconstrained method.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=Siln8xpTMrZ", "arxiv_id": "2208.00779", "paper_pdf": "papers/Siln8xpTMrZ.pdf", "paper_pdf_sha256": "feda7224d95cb90a607cdf032996ec5681d8ac8977415b1947af41f155dada1c", "paper_pdf_bytes": 596633, "paper_pdf_source": "openreview", "code_url": "https://github.com/AdelNabli/DADAO", "code_repository": "AdelNabli/DADAO", "code_commit": "eed1ed1b21a66da393f55433c26f7920b465eabe", "code_archive": "repos/Siln8xpTMrZ.zip", "code_archive_sha256": "0a549ae8cafbdceebcee58dc6c2f6dcc9d08119b7ab40bfda0634e445003f1e1", "code_archive_bytes": 457184, "code_file_count": 8, "code_extensions": {".py": 7, ".ipynb": 1}, "github_disk_usage_kb": 464, "github_languages": {"Jupyter Notebook": 592835, "Python": 91733}, "github_archived": false, "github_pushed_at": "2023-10-07T00:04:22Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/dadao-decoupled-accelerated-decentralized"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "LT0KSFnQDWF", "year": 2021, "status": "rejected", "title": "Improving Graph Neural Network Expressivity via Subgraph Isomorphism Counting", "authors": ["Giorgos Bouritsas", "Fabrizio Frasca", "Stefanos Zafeiriou", "Michael M. Bronstein"], "authorids": ["~Giorgos_Bouritsas1", "~Fabrizio_Frasca1", "~Stefanos_Zafeiriou1", "~Michael_M._Bronstein1"], "authors_source": "OpenReview API", "abstract": "While Graph Neural Networks (GNNs) have achieved remarkable results in a variety of applications, recent studies exposed important shortcomings in their ability to capture the structure of the underlying graph. It has been shown that the expressive power of standard GNNs is bounded by the Weisfeiler-Lehman (WL) graph isomorphism test, from which they inherit proven limitations such as the inability to detect and count graph substructures. On the other hand, there is significant empirical evidence, e.g. in network science and bioinformatics, that substructures are often informative for downstream tasks, suggesting that it is desirable to design GNNs capable of leveraging this important source of information. To this end, we propose a novel topologically-aware message passing scheme based on substructure encoding. We show that our architecture allows incorporating domain-specific inductive biases and that it is strictly more expressive than the WL test. Importantly, in contrast to recent works on the expressivity of GNNs, we do not attempt to adhere to the WL hierarchy; this allows us to retain multiple attractive properties of standard GNNs such as locality and linear network complexity, while being able to disambiguate even hard instances of graph isomorphism. We extensively evaluate our method on graph classification and regression tasks and show state-of-the-art results on multiple datasets including molecular graphs and social networks.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=LT0KSFnQDWF", "arxiv_id": "2006.09252", "paper_pdf": "papers/LT0KSFnQDWF.pdf", "paper_pdf_sha256": "0690ee61df32cdcc835a1086e5195188358cfdfe10df993c3843ebeda7ade2f9", "paper_pdf_bytes": 1494478, "paper_pdf_source": "openreview", "code_url": "https://github.com/gbouritsas/GSN", "code_repository": "gbouritsas/GSN", "code_commit": "6cce24a2c0f59c183c388f3016d33502f63e8175", "code_archive": "repos/LT0KSFnQDWF.zip", "code_archive_sha256": "2fa53aa332338f5c639095025bbf449ab1f5711597b48f96f66373f90a809a66", "code_archive_bytes": 1190892, "code_file_count": 39, "code_extensions": {".py": 38, ".sh": 1}, "github_disk_usage_kb": 1228, "github_languages": {"Python": 260611, "Shell": 596}, "github_archived": false, "github_pushed_at": "2021-06-15T14:59:02Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/improving-graph-neural-network-expressivity"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "S1e4Q6EtDH", "year": 2020, "status": "rejected", "title": "Tensorized Embedding Layers for Efficient Model Compression", "authors": ["Oleksii Hrinchuk", "Valentin Khrulkov", "Leyla Mirvakhabova", "Ivan Oseledets"], "authorids": ["oleksii.hrinchuk@skoltech.ru", "khrulkov.v@gmail.com", "leyla.mirvakhabova@skoltech.ru", "i.oseledets@skoltech.ru"], "authors_source": "OpenReview API", "abstract": "The embedding layers transforming input words into real vectors are the key components of deep neural networks used in natural language processing. However, when the vocabulary is large, the corresponding weight matrices can be enormous, which precludes their deployment in a limited resource setting. We introduce a novel way of parametrizing embedding layers based on the Tensor Train (TT) decomposition, which allows compressing the model significantly at the cost of a negligible drop or even a slight gain in performance.  We evaluate our method on a wide range of benchmarks in natural language processing and analyze the trade-off between performance and compression ratios for a wide range of architectures, from MLPs to LSTMs and Transformers.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=S1e4Q6EtDH", "arxiv_id": "1901.10787", "paper_pdf": "papers/S1e4Q6EtDH.pdf", "paper_pdf_sha256": "a0d2097f59847dd4dbcd2ce5edcc31f62883ca79f4a3e62b586b7ebf2d7f4412", "paper_pdf_bytes": 1314343, "paper_pdf_source": "openreview", "code_url": "https://github.com/KhrulkovV/tt-pytorch", "code_repository": "KhrulkovV/tt-pytorch", "code_commit": "3182e8d90fb4c4473b66105159e54486562b6e33", "code_archive": "repos/S1e4Q6EtDH.zip", "code_archive_sha256": "a089dd475459ae098edbf1e32bbbb5e1d643319ddd57c56fe7fdae9ab94b801c", "code_archive_bytes": 21528, "code_file_count": 14, "code_extensions": {".py": 13, ".sh": 1}, "github_disk_usage_kb": 8765, "github_languages": {"Python": 82029, "Shell": 759}, "github_archived": false, "github_pushed_at": "2020-07-06T07:54:01Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/tensorized-embedding-layers-for-efficient"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "xg2ccsYpZY", "year": 2026, "status": "rejected", "title": "SIM-Shapley: A Stable and Computationally Efficient Approach to Shapley Value Approximation", "authors": ["Wangxuan Fan", "Siqi Li", "Doudou Zhou", "Yohei Okada", "Chuan Hong", "Molei Liu", "Nan Liu"], "authorids": ["~Wangxuan_Fan1", "~Siqi_Li4", "~Doudou_Zhou1", "~Yohei_Okada1", "~Chuan_Hong1", "~Molei_Liu1", "~Nan_Liu3"], "authors_source": "OpenReview API", "abstract": "Explainable artificial intelligence (XAI) is essential for trustworthy  machine learning (ML), particularly in high-stakes domains such as healthcare and finance. Shapley value (SV) methods provide a principled framework for feature attribution in complex  models but incur high computational costs, limiting their scalability in high-dimensional settings. We propose Stochastic IterativeMomentum for Shapley Value Approximation (SIM-Shapley), a stable and efficient SV approximation method inspired by stochastic optimization. We analyze variance theoretically, prove linear $Q$-convergence, and demonstrate improved empirical stability and low bias in practice on real-world datasets.\nIn our numerical experiments, SIM-Shapley reduces computation time by up to 85\\% relative to state-of-the-art baselines while maintaining comparable feature attribution quality. Beyond feature attribution, our stochastic mini-batch iterative framework extends naturally to a broader class of sample average approximation problems, offering a new avenue for improving computational efficiency with stability guarantees. Code is publicly available at https://anonymous.4open.science/r/SIM-Shapley.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=xg2ccsYpZY", "arxiv_id": "2505.08198", "paper_pdf": "papers/xg2ccsYpZY.pdf", "paper_pdf_sha256": "7cc6c06f4b2bff81d3a8db1a24d1def094c0e311299c11f673782a8f739f4c72", "paper_pdf_bytes": 760017, "paper_pdf_source": "openreview", "code_url": "https://github.com/nliulab/SIM-Shapley", "code_repository": "nliulab/SIM-Shapley", "code_commit": "afcb1456c28bcec078c5188e33a37ea6ed0f3b0e", "code_archive": "repos/xg2ccsYpZY.zip", "code_archive_sha256": "4b58356cc5e43ee89ca72b25cb97f050b60a664ab0f145d4ed694cc64e5f01fb", "code_archive_bytes": 3600452, "code_file_count": 31, "code_extensions": {".py": 27, ".ipynb": 4}, "github_disk_usage_kb": 4119, "github_languages": {"Python": 188239, "AGS Script": 46939}, "github_archived": false, "github_pushed_at": "2026-04-02T14:35:33Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/sim-shapley-a-stable-and-computationally"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "GYwH71ugtC", "year": 2025, "status": "rejected", "title": "Retrieval Augmented Time Series Forecasting", "authors": ["Sungwon Han", "Seungeon Lee", "Meeyoung Cha", "Sercan O Arik", "Jinsung Yoon"], "authorids": ["~Sungwon_Han1", "~Seungeon_Lee1", "~Meeyoung_Cha2", "~Sercan_O_Arik1", "~Jinsung_Yoon1"], "authors_source": "OpenReview API", "abstract": "Time series forecasting uses historical data to predict future trends, leveraging the relationships between past observations and available features. In this paper, we propose, RAFT, a retrieval-augmented time series forecasting method to provide sufficient inductive biases and complement the model's learning capacity. When forecasting the subsequent time frames, we directly retrieve historical data candidates from the training dataset with patterns most similar to the input, and utilize the future values of these candidates alongside the inputs to obtain predictions. This simple approach augments the model's capacity by externally providing information about past patterns via retrieval modules. Our empirical evaluations on eight benchmark datasets show that RAFT consistently outperforms contemporary baselines, an average win ratio of 86% for multivariate forecasting and 80% for univariate forecasting tasks.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=GYwH71ugtC", "arxiv_id": "2411.08249", "paper_pdf": "papers/GYwH71ugtC.pdf", "paper_pdf_sha256": "2a1a9363dbe8d0ec13e633b7733f77976a1f3e5279681b5b2d2cfeab5d608a63", "paper_pdf_bytes": 1696089, "paper_pdf_source": "openreview", "code_url": "https://github.com/kutaytire/Retrieval-Augmented-Time-Series-Forecasting", "code_repository": "kutaytire/Retrieval-Augmented-Time-Series-Forecasting", "code_commit": "425e2af35797d0d32b63cfc554c7d97b3c54b390", "code_archive": "repos/GYwH71ugtC.zip", "code_archive_sha256": "f987205d1168765598ead579c0889258cc3fc1886efb3f06bddfc2d4e555bc31", "code_archive_bytes": 358038, "code_file_count": 10, "code_extensions": {".py": 10}, "github_disk_usage_kb": 284, "github_languages": {"Python": 109293}, "github_archived": false, "github_pushed_at": "2024-11-24T17:48:03Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/retrieval-augmented-time-series-forecasting"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "9SwObx9Jdn", "year": 2024, "status": "rejected", "title": "Generation of Geodesics with Actor-Critic Reinforcement Learning to Predict Midpoints", "authors": ["Kazumi Kasaura"], "authorids": ["~Kazumi_Kasaura1"], "authors_source": "OpenReview API", "abstract": "Various tasks in the real world, such as path planning, can be reduced to the generation of geodesics on manifolds. For reinforcement learning to generate geodesics sequentially, we need to define rewards appropriately. To generate geodesics without any adjustment of rewards, we propose to use a modified version of sub-goal trees, called midpoint trees. While sub-goal trees consist of arbitrary intermediate points, midpoint trees consist of midpoints. In addition, we propose an actor-critic method to learn to predict midpoints and theoretically prove that, under mild assumptions, when the learning converges at the limit of infinite tree depth, the resulting policy generates exact midpoints.\nWe show experimentally that our proposed method outperforms existing methods in a certain path planning task.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=9SwObx9Jdn", "arxiv_id": "2407.01991", "paper_pdf": "papers/9SwObx9Jdn.pdf", "paper_pdf_sha256": "554508fbfe0ce295f7304355cf21d166a8718369080e27faccc162f6dff4588f", "paper_pdf_bytes": 757217, "paper_pdf_source": "openreview", "code_url": "https://github.com/omron-sinicx/midpoint_learning", "code_repository": "omron-sinicx/midpoint_learning", "code_commit": "6f8857ba0f9542dfb87c8941d5beeada55a39329", "code_archive": "repos/9SwObx9Jdn.zip", "code_archive_sha256": "b7db1d07fd8cd57bbc4de7cfa96e9bcfc718377c71716673613f0a9fee80bb46", "code_archive_bytes": 268713, "code_file_count": 79, "code_extensions": {".py": 78, ".sh": 1}, "github_disk_usage_kb": 225, "github_languages": {"Python": 338931, "Shell": 7148, "Dockerfile": 871}, "github_archived": false, "github_pushed_at": "2026-01-05T00:30:33Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/generation-of-geodesics-with-actor-critic"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "u6KhE9fapjX", "year": 2023, "status": "rejected", "title": "CAT: Collaborative Adversarial Training", "authors": ["xingbin liu", "Huafeng Kuang", "Xianming Lin", "GUANNAN JIANG", "YONGJIAN WU", "Xi Wang", "Rongrong Ji"], "authorids": ["~xingbin_liu1", "~Huafeng_Kuang1", "~Xianming_Lin1", "~GUANNAN_JIANG1", "~YONGJIAN_WU2", "~Xi_Wang12", "~Rongrong_Ji5"], "authors_source": "OpenReview API", "abstract": "Adversarial training can improve the robustness of neural networks. Previous adversarial training methods focus on a single training strategy and do not consider the collaboration between different training strategies. In this paper, we find different adversarial training methods have distinct robustness for sample instances. For example, an instance can be correctly classified by a model trained using standard adversarial training (AT) but not by a model trained using TRADES, and vice versa.  Based on this phenomenon, we propose a collaborative adversarial training framework to improve the robustness of neural networks. Specifically, we simultaneously use different adversarial training methods to train two robust models from scratch. We input the adversarial examples generated by each network to the peer network and use the logit of the peer network to guide the training of its network. Collaborative Adversarial Training (CAT) can improve both robustness and accuracy. Finally, Extensive experiments on CIFAR-10 and CIFAR-100 validated the effectiveness of our method.\nCAT achieved new state-of-the-art robustness without using any additional data on CIFAR-10 under the Auto-Attack benchmark.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=u6KhE9fapjX", "arxiv_id": "2303.14922", "paper_pdf": "papers/u6KhE9fapjX.pdf", "paper_pdf_sha256": "b29f8f5c386abbef054c0014f4c797abb4d3f1fef594758b17a33be1e8b880bf", "paper_pdf_bytes": 324232, "paper_pdf_source": "openreview", "code_url": "https://github.com/liuxingbin/CAT", "code_repository": "liuxingbin/CAT", "code_commit": "2325fa16af644c1a6e83bf4761d4a8dece86763d", "code_archive": "repos/u6KhE9fapjX.zip", "code_archive_sha256": "9b1d3dab839509c451199e9a98032d8038a38a8333789c066b4f090662fdb949", "code_archive_bytes": 496510, "code_file_count": 12, "code_extensions": {".py": 10, ".sh": 2}, "github_disk_usage_kb": 486, "github_languages": {"Python": 53003, "Shell": 660}, "github_archived": false, "github_pushed_at": "2023-05-31T07:07:52Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/cat-collaborative-adversarial-training"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "57T1ctyxtP", "year": 2022, "status": "rejected", "title": "Structured Stochastic Gradient MCMC", "authors": ["Antonios Alexos", "Alex James Boyd", "Stephan Mandt"], "authorids": ["~Antonios_Alexos1", "~Alex_James_Boyd1", "~Stephan_Mandt1"], "authors_source": "OpenReview API", "abstract": "Stochastic gradient Markov Chain Monte Carlo (SGMCMC) is considered the gold standard for Bayesian inference in large-scale models, such as Bayesian neural networks. Since practitioners face speed versus accuracy tradeoffs in these models, variational inference (VI) is often the preferable option. Unfortunately, VI makes strong assumptions on both the factorization and functional form of the posterior. In this work, we propose a new non-parametric variational approximation that makes no assumptions about the approximate posterior's functional form and allows practitioners to specify the exact dependencies the algorithm should respect or break. The approach relies on a new Langevin-type algorithm that operates on a modified energy function, where parts of the latent variables are averaged over samples from earlier iterations of the Markov chain. This way, statistical dependencies can be broken in a controlled way, allowing the chain to mix faster. This scheme can be further modified in a ``dropout'' manner, leading to even more scalability. By implementing the scheme on a ResNet-20 architecture, we obtain better predictive likelihoods and faster mixing time than full SGMCMC.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=57T1ctyxtP", "arxiv_id": "2107.09028", "paper_pdf": "papers/57T1ctyxtP.pdf", "paper_pdf_sha256": "f367cdaf6ed89f25b32cedc1a2ba7849ea7357d3f8554a6888292eaf778ab326", "paper_pdf_bytes": 871196, "paper_pdf_source": "openreview", "code_url": "https://github.com/mandt-lab/Structured-SG-MCMC", "code_repository": "mandt-lab/Structured-SG-MCMC", "code_commit": "c60e10c1eec51e9a1ff2c3f5e812c7eb3e7a5ae4", "code_archive": "repos/57T1ctyxtP.zip", "code_archive_sha256": "a318448525c031514261d1c8539f601bd0409c825d4d13c7fc12a0871900f246", "code_archive_bytes": 1799010, "code_file_count": 9, "code_extensions": {".py": 7, ".ipynb": 2}, "github_disk_usage_kb": 1798, "github_languages": {"Jupyter Notebook": 2406870, "Python": 142113}, "github_archived": false, "github_pushed_at": "2023-09-10T19:32:38Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/structured-stochastic-gradient-mcmc"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "BygWRaVYwH", "year": 2020, "status": "rejected", "title": "Generalized Inner Loop Meta-Learning", "authors": ["Edward Grefenstette", "Brandon Amos", "Denis Yarats", "Phu Mon Htut", "Artem Molchanov", "Franziska Meier", "Douwe Kiela", "Kyunghyun Cho", "Soumith Chintala"], "authorids": ["egrefen@gmail.com", "brandon.amos.cs@gmail.com", "denisyarats@cs.nyu.edu", "pmh330@nyu.edu", "a.molchanov86@gmail.com", "fmeier@fb.com", "dkiela@fb.com", "kyunghyun.cho@nyu.edu", "soumith@gmail.com"], "authors_source": "OpenReview API", "abstract": "Many (but not all) approaches self-qualifying as \"meta-learning\" in deep learning and reinforcement learning fit a common pattern of approximating the solution to a nested optimization problem. In this paper, we give a formalization of this shared pattern, which we call GIMLI, prove its general requirements, and derive a general-purpose algorithm for implementing similar approaches. Based on this analysis and algorithm, we describe a library of our design, unnamedlib, which we share with the community to assist and enable future research into these kinds of meta-learning approaches. We end the paper by showcasing the practical applications of this framework and library through illustrative experiments and ablation studies which they facilitate.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=BygWRaVYwH", "arxiv_id": "1910.01727", "paper_pdf": "papers/BygWRaVYwH.pdf", "paper_pdf_sha256": "ab398d39418bf7d0666226ef670c09163533ca61e9f7f35b682001c2b2164acd", "paper_pdf_bytes": 2402275, "paper_pdf_source": "openreview", "code_url": "https://github.com/facebookresearch/higher", "code_repository": "facebookresearch/higher", "code_commit": "15a247ac06cac0d22601322677daff0dcfff062e", "code_archive": "repos/BygWRaVYwH.zip", "code_archive_sha256": "9342ae241bce7246b75df89a2807f6d1abc91d91644d3f642cc6dbc8e09afc53", "code_archive_bytes": 9067093, "code_file_count": 12, "code_extensions": {".py": 12}, "github_disk_usage_kb": 8888, "github_languages": {"Python": 132092}, "github_archived": true, "github_pushed_at": "2022-03-25T15:56:51Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/generalized-inner-loop-meta-learning"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "YQ5wH5C69t", "year": 2026, "status": "rejected", "title": "Dimple: Discrete Diffusion Multimodal Large Language Model with Parallel Decoding", "authors": ["Runpeng Yu", "Xinyin Ma", "Xinchao Wang"], "authorids": ["~Runpeng_Yu2", "~Xinyin_Ma1", "~Xinchao_Wang1"], "authors_source": "OpenReview API", "abstract": "In this work, we present Dimple and Dimple+, two Discrete Diffusion Multimodal Large Language Models (dMLLMs). Dimple is initialized from a discrete diffusion Large Language Model (dLLM) without multimodal understanding ability, and learns such ability through a hybrid training paradigm that first applies autoregressive training and then switches to discrete diffusion training. Dimple+ is initialized from an autoregressive Multimodal Large Language Models, and acquires parallel decoding capability through pure discrete diffusion training. Both models achieve performance comparable to their autoregressive baselines, and Dimple+ establishes new state-of-the-art results among dMLLMs. To enhance inference efficiency, we propose Confident Decoding, which dynamically adjusts the number of tokens generated per iteration. Experiments show that it accelerates decoding by 2×–6× with only minor performance degradation. We also demonstrate that the Prefilling technique, previously used in autoregressive models, can be effectively applied to dMLLMs with bidirectional attention, achieving nearly lossless speedups of 1.7×–7×. Finally, we introduce the Structure Prior method, enabling fine-grained control over response format and reasoning structure, which is difficult to realize in autoregressive models.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=YQ5wH5C69t", "arxiv_id": "2505.16990", "paper_pdf": "papers/YQ5wH5C69t.pdf", "paper_pdf_sha256": "c8c968bea90e010653c90bc3b15b859fcfefb26d262e43c486b459d71b4b74ba", "paper_pdf_bytes": 978374, "paper_pdf_source": "openreview", "code_url": "https://github.com/yu-rp/Dimple", "code_repository": "yu-rp/Dimple", "code_commit": "e0ec36582a9ef9d9f205e65e2bc91b3726aab908", "code_archive": "repos/YQ5wH5C69t.zip", "code_archive_sha256": "1ffc3fc0b6dc3d4b634f979e6c5eb34e472b6d5274bc6160ec32528002c1b587", "code_archive_bytes": 4802750, "code_file_count": 480, "code_extensions": {".py": 463, ".ipynb": 14, ".sh": 3}, "github_disk_usage_kb": 4168, "github_languages": {"Python": 3754846, "Jupyter Notebook": 255855, "Shell": 4982}, "github_archived": false, "github_pushed_at": "2026-08-23T13:19:26Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/dimple-discrete-diffusion-multimodal-large"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "baSU1eVLwS", "year": 2025, "status": "rejected", "title": "TimeBridge: Non-Stationarity Matters for Long-term Time Series Forecasting", "authors": ["Peiyuan Liu", "Beiliang Wu", "Yifan Hu", "Naiqi Li", "Tao Dai", "Jigang Bao", "Shu-Tao Xia"], "authorids": ["~Peiyuan_Liu1", "~Beiliang_Wu1", "~Yifan_Hu8", "~Naiqi_Li1", "~Tao_Dai3", "~Jigang_Bao1", "~Shu-Tao_Xia1"], "authors_source": "OpenReview API", "abstract": "Non-stationarity poses significant challenges for multivariate time series forecasting due to the inherent short-term fluctuations and long-term trends that can lead to spurious regressions or obscure essential long-term relationships. Most existing methods either eliminate or retain non-stationarity without adequately addressing its distinct impacts on short-term and long-term modeling. Eliminating non-stationarity is essential for avoiding spurious regressions and capturing local dependencies in short-term modeling, while preserving it is crucial for revealing long-term cointegration across variates. In this paper, we propose TimeBridge, a novel framework designed to bridge the gap between non-stationarity and dependency modeling in long-term time series forecasting. By segmenting input series into smaller patches, TimeBridge applies Integrated Attention to mitigate short-term non-stationarity and capture stable dependencies within each variate, while Cointegrated Attention preserves non-stationarity to model long-term cointegration across variates. Extensive experiments show that TimeBridge consistently achieves state-of-the-art performance in both short-term and long-term forecasting. Additionally, TimeBridge demonstrates exceptional performance in financial forecasting on the CSI 500 and S\\&P 500 indices, further validating its robustness and effectiveness. The code is available in the supplementary material.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=baSU1eVLwS", "arxiv_id": "2410.04442", "paper_pdf": "papers/baSU1eVLwS.pdf", "paper_pdf_sha256": "301de0175d6aa8efdf82709625cb618a7d0b266789cebbda0cdcd4bbde239f2f", "paper_pdf_bytes": 7923752, "paper_pdf_source": "openreview", "code_url": "https://github.com/Hank0626/TimeBridge", "code_repository": "Hank0626/TimeBridge", "code_commit": "0f9a83fbc3e1260c9ddd527c522dff0ce4b9554b", "code_archive": "repos/baSU1eVLwS.zip", "code_archive_sha256": "bb20739597b546e0291fe3dc6e5093e6e27c83975786a8dcfcab09d75cd214fe", "code_archive_bytes": 177652, "code_file_count": 16, "code_extensions": {".py": 15, ".sh": 1}, "github_disk_usage_kb": 291, "github_languages": {"Python": 87904, "Shell": 6642}, "github_archived": false, "github_pushed_at": "2025-05-16T04:33:15Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/timebridge-non-stationarity-matters-for-long"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "tZ3JmSDbJM", "year": 2024, "status": "rejected", "title": "GRAPES: Learning to Sample Graphs for Scalable Graph Neural Networks", "authors": ["Taraneh Younesian", "Thiviyan Thanapalasingam", "Emile van Krieken", "Daniel Daza", "Peter Bloem"], "authorids": ["~Taraneh_Younesian2", "~Thiviyan_Thanapalasingam1", "~Emile_van_Krieken1", "~Daniel_Daza1", "~Peter_Bloem1"], "authors_source": "OpenReview API", "abstract": "Graph neural networks (GNNs) learn the representation of nodes in a graph by aggregating the neighborhood information in various ways. As these networks grow in depth, their receptive field grows exponentially due to the increase in neighborhood sizes, resulting in high memory costs. Graph sampling solves memory issues in GNNs by sampling a small ratio of the nodes in the graph. This way, GNNs can scale to much larger graphs. Most sampling methods focus on fixed sampling heuristics, which may not generalize to different structures or tasks. We introduce GRAPES, an adaptive graph sampling method that learns to identify sets of influential nodes for training a GNN classifier.  GRAPES uses a GFlowNet to learn node sampling probabilities given the classification objectives. We evaluate GRAPES across several small- and large-scale graph benchmarks and demonstrate its effectiveness in accuracy and scalability. In contrast to existing sampling methods, GRAPES maintains high accuracy even with small sample sizes and, therefore, can scale to very large graphs. Our code is publicly available at https://anonymous.4open.science/r/GRAPES.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=tZ3JmSDbJM", "arxiv_id": "2310.03399", "paper_pdf": "papers/tZ3JmSDbJM.pdf", "paper_pdf_sha256": "0981b4c3da53e2fdb882056523e1c2da3f4ff279c577b2613e9c2194f674ac1e", "paper_pdf_bytes": 754802, "paper_pdf_source": "openreview", "code_url": "https://github.com/dfdazac/grapes", "code_repository": "dfdazac/grapes", "code_commit": "71ecebeaac896800aa4dd1d0f38c57ec222ef396", "code_archive": "repos/tZ3JmSDbJM.zip", "code_archive_sha256": "6f9294fe075e65e3ecf03d20216a47ae4cd820870912a80385c5cf730b9f5ca3", "code_archive_bytes": 79344, "code_file_count": 12, "code_extensions": {".py": 12}, "github_disk_usage_kb": 242, "github_languages": {"Python": 90863}, "github_archived": false, "github_pushed_at": "2025-04-01T12:36:20Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/grapes-learning-to-sample-graphs-for-scalable"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "e9rdb24Yzqx", "year": 2023, "status": "rejected", "title": "Empirical analysis of representation learning and exploration in neural kernel bandits", "authors": ["Michal Lisicki", "Arash Afkanpour", "Graham W. Taylor"], "authorids": ["~Michal_Lisicki1", "~Arash_Afkanpour1", "~Graham_W._Taylor1"], "authors_source": "OpenReview API", "abstract": "Neural bandits have been shown to provide an efficient solution to practical sequential decision tasks that have nonlinear reward functions. The main contributor to that success is approximate Bayesian inference, which enables neural network (NN) training with uncertainty estimates. However, Bayesian NNs often suffer from a prohibitive computational overhead or operate on a subset of parameters. Alternatively, certain classes of infinite neural networks were shown to directly correspond to Gausian processes (GP) with neural kernels (NK). NK-GPs provide accurate uncertainty estimates and can be trained faster than most Bayesian NNs. We propose to guide common bandit policies with NK distributions and show that NK bandits achieve state-of-the-art performance on nonlinear structured data. Moreover, we propose a framework for measuring independently the ability of a bandit algorithm to learn representations and explore, and use it to analyze the impact of NK distributions w.r.t. those two aspects. We consider policies based on a GP and a Student's t-process (TP). Furthermore, we study practical considerations, such as training frequency and model partitioning. We believe our work will help better understand the impact of utilizing NKs in applied settings.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=e9rdb24Yzqx", "arxiv_id": "2111.03543", "paper_pdf": "papers/e9rdb24Yzqx.pdf", "paper_pdf_sha256": "6c16f7be104c6f8c6d00db1ab9dd01ef7785246e682be95232bfc67678ec7d8b", "paper_pdf_bytes": 614603, "paper_pdf_source": "openreview", "code_url": "https://github.com/VectorInstitute/NeuralKernelBandits", "code_repository": "VectorInstitute/NeuralKernelBandits", "code_commit": "6ddbfa3a410b50b84d9fb72c75803347da61b4cb", "code_archive": "repos/e9rdb24Yzqx.zip", "code_archive_sha256": "3c0bece48111776c6dc7eced07121c279a0d6eec96bdce20e17d3d9a23123d25", "code_archive_bytes": 409604, "code_file_count": 27, "code_extensions": {".py": 27}, "github_disk_usage_kb": 513, "github_languages": {}, "github_archived": false, "github_pushed_at": "2021-12-09T15:15:07Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/an-empirical-study-of-neural-kernel-bandits"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "TD-5kgf13mH", "year": 2022, "status": "rejected", "title": "Sparse MoEs meet Efficient Ensembles", "authors": ["James Urquhart Allingham", "Florian Wenzel", "Zelda E Mariet", "Basil Mustafa", "Joan Puigcerver", "Neil Houlsby", "Ghassen Jerfel", "Vincent Fortuin", "Balaji Lakshminarayanan", "Jasper Snoek", "Dustin Tran", "Carlos Riquelme Ruiz", "Rodolphe Jenatton"], "authorids": ["~James_Urquhart_Allingham1", "~Florian_Wenzel1", "~Zelda_E_Mariet1", "~Basil_Mustafa1", "~Joan_Puigcerver1", "~Neil_Houlsby1", "~Ghassen_Jerfel1", "~Vincent_Fortuin1", "~Balaji_Lakshminarayanan1", "~Jasper_Snoek1", "~Dustin_Tran1", "~Carlos_Riquelme_Ruiz1", "~Rodolphe_Jenatton3"], "authors_source": "OpenReview API", "abstract": "Machine learning models based on the aggregated outputs of submodels, either at the activation or prediction levels, lead to strong performance. We study the interplay of two popular classes of such models: ensembles of neural networks and sparse mixture of experts (sparse MoEs). First, we show that these two approaches have complementary features whose combination is beneficial. Then, we present partitioned batch ensembles, an efficient ensemble of sparse MoEs that takes the best of both classes of models. Extensive experiments on fine-tuned vision transformers demonstrate the accuracy, log-likelihood, few-shot learning, robustness, and uncertainty calibration improvements of our approach over several challenging baselines. Partitioned batch ensembles not only scale to models with up to 2.7B parameters, but also provide larger performance gains for larger models. ", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=TD-5kgf13mH", "arxiv_id": "2110.03360", "paper_pdf": "papers/TD-5kgf13mH.pdf", "paper_pdf_sha256": "092a43edd66084b67086ec2e0cc0a2049f8e4c0bfa227e1cdd177bd0f09515d2", "paper_pdf_bytes": 1097965, "paper_pdf_source": "openreview", "code_url": "https://github.com/google-research/vmoe", "code_repository": "google-research/vmoe", "code_commit": "6bb693bc81f949d654a9c5b1fc062374c92464cb", "code_archive": "repos/TD-5kgf13mH.zip", "code_archive_sha256": "02915bf096dec1d3d93d0033497b5ef0c394aabc5747f2e90eeed58ed70713b0", "code_archive_bytes": 1538795, "code_file_count": 102, "code_extensions": {".py": 98, ".ipynb": 3, ".sh": 1}, "github_disk_usage_kb": 1834, "github_languages": {"Jupyter Notebook": 1764344, "Python": 776887, "Shell": 2077}, "github_archived": false, "github_pushed_at": "2026-09-07T15:11:47Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/sparse-moes-meet-efficient-ensembles"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Sva-fwURywB", "year": 2021, "status": "rejected", "title": "Efficiently Disentangle Causal Representations", "authors": ["Yuanpeng Li", "Joel Hestness", "Mohamed Elhoseiny", "Liang Zhao", "Kenneth Church"], "authorids": ["~Yuanpeng_Li2", "~Joel_Hestness2", "~Mohamed_Elhoseiny1", "~Liang_Zhao2", "~Kenneth_Church1"], "authors_source": "OpenReview API", "abstract": "In this paper, we propose a novel approach to efficiently learning disentangled representations with causal mechanisms, based on the difference of conditional probabilities in original and new distributions. We approximate the difference with model's generalization abilities so that it fits in standard machine learning framework and can be efficiently computed. In contrast to the state-of-the-art approach, which relies on learner's adaptation speed to new distribution, the proposed approach only requires evaluating the generalization ability of the model. We provide theoretical explanation for the advantage of the proposed method, and our experiments show that the proposed technique is 1.9-11.0x more sample efficient and 9.4-32.4x quicker than the previous method on various tasks.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=Sva-fwURywB", "arxiv_id": "2201.01942", "paper_pdf": "papers/Sva-fwURywB.pdf", "paper_pdf_sha256": "bf67f56ed84f6d2d9af097ce4ef221895e78a6c6bc21de2fc264b3adf0bfc476", "paper_pdf_bytes": 518121, "paper_pdf_source": "openreview", "code_url": "https://github.com/yuanpeng16/EDCR", "code_repository": "yuanpeng16/EDCR", "code_commit": "781194191a82fbc85d43084af63e6a20c59ae670", "code_archive": "repos/Sva-fwURywB.zip", "code_archive_sha256": "e64bd48f6e961e3cfdecfcef5a7643c22c1d0b7ba4f82f3271226995d46e4813", "code_archive_bytes": 1298441, "code_file_count": 59, "code_extensions": {".py": 53, ".ipynb": 6}, "github_disk_usage_kb": 1256, "github_languages": {"Jupyter Notebook": 279672, "Python": 166392}, "github_archived": false, "github_pushed_at": "2023-11-24T16:13:49Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/efficiently-disentangle-causal-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "SkglVlSFPS", "year": 2020, "status": "rejected", "title": "Uncertainty - sensitive learning and planning with ensembles", "authors": ["Piotr Miłoś", "Łukasz Kuciński", "Konrad Czechowski", "Piotr Kozakowski", "Maciej Klimek"], "authorids": ["pmilos@mimuw.edu.pl", "lukasz.kucinski@gmail.com", "konrad.czechowski@gmail.com", "p.kozakowski@mimuw.edu.pl", "maciej.klimek@gmail.com"], "authors_source": "OpenReview API", "abstract": "We propose a reinforcement learning framework for discrete environments in which an agent optimizes its behavior on two timescales. For the short one, it uses tree search methods to perform tactical decisions. The long strategic level is handled with an ensemble of value functions learned using $TD$-like backups. Combining these two techniques brings synergies. The planning module performs \\textit{what-if} analysis allowing to avoid short-term pitfalls and boost backups of the value function. Notably, our method performs well in environments with sparse rewards where standard $TD(1)$ backups fail. On the other hand, the value functions compensate for inherent short-sightedness of planning. Importantly, we use ensembles to measure the epistemic uncertainty of value functions. This serves two purposes: a) it stabilizes planning, b) it guides exploration. \n\nWe evaluate our methods on discrete environments with sparse rewards: the Deep sea chain environment, toy Montezuma's Revenge, and Sokoban. In all the cases, we obtain speed-up of learning and boost to the final performance.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=SkglVlSFPS", "arxiv_id": null, "paper_pdf": "papers/SkglVlSFPS.pdf", "paper_pdf_sha256": "27ed82277dff27e82e29165d57b6a1e4763e80bbf172d1e44eff01c704454099", "paper_pdf_bytes": 1406946, "paper_pdf_source": "openreview", "code_url": "https://github.com/learningandplanningICLR/learningandplanning", "code_repository": "learningandplanningICLR/learningandplanning", "code_commit": "10eac96831d9e61331c2879d33e12a117d259a2b", "code_archive": "repos/SkglVlSFPS.zip", "code_archive_sha256": "b81d7b2067cab901138e40f947b1a23f0c7e5aa9abfe3cbe4f287654457fd396", "code_archive_bytes": 8693125, "code_file_count": 291, "code_extensions": {".py": 284, ".ipynb": 4, ".cpp": 2, ".sh": 1}, "github_disk_usage_kb": 9515, "github_languages": {"Python": 355147, "Shell": 104}, "github_archived": false, "github_pushed_at": "2020-02-12T18:43:27Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/uncertainty-sensitive-learning-and-planning"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "FafH4GM4uc", "year": 2026, "status": "rejected", "title": "Universal Reasoner: A Single, Composable Plug-and-Play Reasoner for Frozen LLMs", "authors": ["Jaemin Kim", "Hangeol Chang", "Hyunmin Hwang", "Choonghan Kim", "Jong Chul Ye"], "authorids": ["~Jaemin_Kim2", "~Hangeol_Chang1", "~Hyunmin_Hwang1", "~Choonghan_Kim1", "~Jong_Chul_Ye1"], "authors_source": "OpenReview API", "abstract": "Large Language Models (LLMs) have demonstrated remarkable general capabilities, but enhancing skills such as reasoning often demands substantial computational resources and may compromise their generalization. While Parameter-Efficient Fine-Tuning (PEFT) methods offer a more resource-conscious alternative, they typically require retraining for each LLM backbone due to architectural dependencies. To address these challenges, here we propose Universal Reasoner (UniR) - a single, lightweight, composable, and plug-and-play reasoning module that can be used with larger frozen LLMs to endow it with specialized reasoning capabilities. Specifically, UniR decomposes the reward into a standalone reasoning module that is trained in a decoupled manner using predefined rewards, effectively translating trajectory-level signals into token-level guidance. Once trained, UniR can be combined with frozen LLMs at inference time by simply adding its output logits to those of the LLM backbones. This additive structure naturally enables modular composition: multiple UniR modules trained for different tasks can be jointly applied by summing their logits, enabling complex reasoning via composition. Experiments on mathematical reasoning and machine translation show that UniR surpasses existing baseline fine-tuning methods. Furthermore, UniR demonstrates weak-to-strong generalization: reasoning modules trained on smaller models effectively guide much larger LLMs. Beyond this, UniR generalizes across modalities such as in vision language models and domains such as medical reasoning. This makes UniR a cost-efficient, adaptable, and robust solution for enhancing reasoning in LLMs without compromising their core capabilities.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=FafH4GM4uc", "arxiv_id": "2505.19075", "paper_pdf": "papers/FafH4GM4uc.pdf", "paper_pdf_sha256": "7cdf1137b6779da375c9f77bbfdab694abb343a6cfa6c997640e22a5c04b707a", "paper_pdf_bytes": 3057012, "paper_pdf_source": "openreview", "code_url": "https://github.com/hangeol/UniR", "code_repository": "hangeol/UniR", "code_commit": "c262e784d40c226621949daad6a72203fe5f3b83", "code_archive": "repos/FafH4GM4uc.zip", "code_archive_sha256": "69b4a977714a2edf9b394fbc1faba67124487ca9ea3c06d24188a6b37f5925df", "code_archive_bytes": 4324357, "code_file_count": 19, "code_extensions": {".py": 16, ".sh": 3}, "github_disk_usage_kb": 4200, "github_languages": {"Python": 395487, "Shell": 8498}, "github_archived": false, "github_pushed_at": "2025-11-26T12:37:44Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/universal-reasoner-a-single-composable-plug"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "YUaQ6qrbRL", "year": 2025, "status": "rejected", "title": "Right Time to Learn: Promoting Generalization via Bio-inspired Spacing Effect in Knowledge Distillation", "authors": ["Guanglong Sun", "Hongwei Yan", "Liyuan Wang", "Qian Li", "Bo Lei", "Yi Zhong"], "authorids": ["~Guanglong_Sun1", "~Hongwei_Yan1", "~Liyuan_Wang1", "~Qian_Li3", "~Bo_Lei1", "~Yi_Zhong1"], "authors_source": "OpenReview API", "abstract": "Knowledge distillation (KD) is a powerful strategy for training deep neural networks (DNNs). While it was originally proposed to train a more compact “student” model from a large “teacher” model, many recent efforts have focused on adapting it as an effective way to promote generalization of the model itself, such as online KD and self KD. Here, we propose an easy-to-use and compatible strategy named Spaced KD to improve the effectiveness of both online KD and self KD, in which the student model distills knowledge from a teacher model trained with a space interval ahead. This strategy is inspired by a prominent theory named spacing effect in the field of biological learning and memory, positing that appropriate intervals between learning trials can significantly enhance learning performance. We provide an in-depth theoretical and empirical analysis showing that the benefits of the proposed spacing effect in KD stem from seeking a flat minima during stochastic gradient descent (SGD). We perform extensive experiments to demonstrate the effectiveness of our Spaced KD in improving the learning performance of DNNs (e.g., the additional performance gain is up to 2.31% and 3.34% on Tiny-ImageNet over online KD and self KD, respectively).", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=YUaQ6qrbRL", "arxiv_id": "2502.06192", "paper_pdf": "papers/YUaQ6qrbRL.pdf", "paper_pdf_sha256": "eff8106c1b9a890e1c0f9c069827106292f52e907e5ad1f09d0f100b547fd316", "paper_pdf_bytes": 2573459, "paper_pdf_source": "openreview", "code_url": "https://github.com/SunGL001/Spaced-KD", "code_repository": "SunGL001/Spaced-KD", "code_commit": "deecb4da498f1bc1ded450d3129c86fc46def40d", "code_archive": "repos/YUaQ6qrbRL.zip", "code_archive_sha256": "5ec914b86a2a663ddeb1711971c141e5169ef92352d716d6dc1672fbae9b5d11", "code_archive_bytes": 287559, "code_file_count": 77, "code_extensions": {".py": 75, ".sh": 2}, "github_disk_usage_kb": 300, "github_languages": {"Python": 460121, "Shell": 1157}, "github_archived": false, "github_pushed_at": "2025-05-20T05:12:01Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/right-time-to-learn-promoting-generalization"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "3mdCet7vVv", "year": 2024, "status": "rejected", "title": "Maestro: Uncovering Low-Rank Structures via Trainable Decomposition", "authors": ["Samuel Horváth", "Stefanos Laskaridis", "Shashank Rajput", "Hongyi Wang"], "authorids": ["~Samuel_Horváth1", "~Stefanos_Laskaridis1", "~Shashank_Rajput1", "~Hongyi_Wang1"], "authors_source": "OpenReview API", "abstract": "Deep Neural Networks (DNNs) have been a large drivers and enablers for AI breakthroughs in recent years. These models have been getting larger in their attempt to become more accurate and tackle new upcoming use-cases, including AR/VR and intelligent assistants. However, the training process of such large models is a costly and time-consuming process, which typically yields a single model to fit all targets. To mitigate this, various techniques have been proposed in the literature, including pruning, sparsification, or quantization of the model weights and updates. While able to achieve high compression rates, they often incur computational overheads or accuracy penalties. Alternatively, factorization methods have been leveraged to incorporate low-rank compression in the training process. Such techniques (e.g., SVD) also frequently rely on the computationally expensive decomposition of layers and are potentially sub-optimal for non-linear models, such as DNNs. In this work, we take a further step in designing efficient low-rank models and propose MAESTRO, a framework for trainable low-rank layers. Instead of regularly applying a priori decompositions such as SVD, the low-rank structure is built into the training process through a generalized variant of Ordered Dropout. This method imposes an importance ordering via sampling on the decomposed DNN structure. Our theoretical analysis demonstrates that our method recovers the SVD decomposition of linear mapping on uniformly distributed data and PCA for linear autoencoders. We further apply our technique on DNNs and empirically illustrate that MAESTRO enables the extraction of lower footprint models that preserve model performance while allowing for graceful accuracy-latency tradeoffs for the deployment to devices of different capabilities.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=3mdCet7vVv", "arxiv_id": "2308.14929", "paper_pdf": "papers/3mdCet7vVv.pdf", "paper_pdf_sha256": "39eea620e9f5a71b233a1140d659e16b28b60ba6990b229df7447172dcd6d455", "paper_pdf_bytes": 2326242, "paper_pdf_source": "openreview", "code_url": "https://github.com/SamuelHorvath/Maestro-LoD", "code_repository": "SamuelHorvath/Maestro-LoD", "code_commit": "0003014a94f78a63e48d971b773b40b00277a207", "code_archive": "repos/3mdCet7vVv.zip", "code_archive_sha256": "ad1521c7df926e61f3e2b2f22c590ea4ce5471bf2ab485010b1f92ffa37aabf7", "code_archive_bytes": 262308, "code_file_count": 26, "code_extensions": {".py": 23, ".sh": 2, ".ipynb": 1}, "github_disk_usage_kb": 247, "github_languages": {"Jupyter Notebook": 304265, "Python": 135895, "Shell": 1019}, "github_archived": false, "github_pushed_at": "2024-06-14T17:21:37Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/maestro-uncovering-low-rank-structures-via"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "rMbrVNxYuqZ", "year": 2023, "status": "rejected", "title": "EurNet: Efficient Multi-Range Relational Modeling of Spatial Multi-Relational Data", "authors": ["Minghao Xu", "Yuanfan Guo", "Yi Xu", "Jian Tang", "Xinlei Chen", "Yuandong Tian"], "authorids": ["~Minghao_Xu1", "~Yuanfan_Guo1", "~Yi_Xu2", "~Jian_Tang1", "~Xinlei_Chen1", "~Yuandong_Tian1"], "authors_source": "OpenReview API", "abstract": "Modeling spatial relationship in the data remains critical across many different tasks, such as image classification, semantic segmentation and protein structure understanding. Previous works often use a unified solution like relative positional encoding. However, there exists different kinds of spatial relations, including short-range, medium-range and long-range relations, and modeling them separately can better capture the focus of different tasks on the multi-range relations (e.g., short-range relations can be important in instance segmentation, while long-range relations should be upweighted for semantic segmentation). In this work, we introduce the EurNet for Efficient multi-range relational modeling. EurNet constructs the multi-relational graph, where each type of edge corresponds to short-, medium- or long-range spatial interactions. In the constructed graph, EurNet adopts a novel modeling layer, called gated relational message passing (GRMP), to propagate multi-relational information across the data. GRMP captures multiple relations within the data with little extra computational cost. We study EurNets in two important domains for image and protein structure modeling. Extensive experiments on ImageNet classification, COCO object detection and ADE20K semantic segmentation verify the gains of EurNet over the previous SoTA FocalNet. On the EC and GO protein function prediction benchmarks, EurNet consistently surpasses the previous SoTA GearNet. Our results demonstrate the strength of EurNets on modeling spatial multi-relational data from various domains. ", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=rMbrVNxYuqZ", "arxiv_id": "2211.12941", "paper_pdf": "papers/rMbrVNxYuqZ.pdf", "paper_pdf_sha256": "da3a8b2b93f2150d82241922068f81c805ec84c31ce08588077a43e86f72eb01", "paper_pdf_bytes": 1729572, "paper_pdf_source": "openreview", "code_url": "https://github.com/hirl-team/EurNet-Image", "code_repository": "hirl-team/EurNet-Image", "code_commit": "7770123cb289b763646ca65fe4cdc30494ea26ee", "code_archive": "repos/rMbrVNxYuqZ.zip", "code_archive_sha256": "f95928dae53ec47199060050ed23fc089e88cd5d6e585a9554a4d7f615712c4b", "code_archive_bytes": 414575, "code_file_count": 18, "code_extensions": {".py": 18}, "github_disk_usage_kb": 527, "github_languages": {"Python": 99800}, "github_archived": false, "github_pushed_at": "2022-11-24T01:26:02Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/eurnet-efficient-multi-range-relational"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "NP9T_pViXU", "year": 2022, "status": "rejected", "title": "VIMPAC: Video Pre-Training via Masked Token Prediction and Contrastive Learning", "authors": ["Hao Tan", "Jie Lei", "Thomas Wolf", "Mohit Bansal"], "authorids": ["~Hao_Tan1", "~Jie_Lei3", "~Thomas_Wolf1", "~Mohit_Bansal2"], "authors_source": "OpenReview API", "abstract": "Video understanding relies on perceiving the overall global content and modeling its internal connections (e.g., causality, movement, and spatio-temporal correspondence). To learn these interactions, we apply a mask-then-predict pre-training task on the discretized video tokens generated via VQ-VAE. Unlike language, where the text tokens are more independent, neighboring video tokens typically have strong correlations (e.g., consecutive video frames usually look very similar), and hence uniformly masking individual tokens will make the task too trivial to learn useful representations. To deal with this issue, we propose a block-wise masking strategy where we mask neighboring video tokens in both spatial and temporal domains. We also add an augmentation-free contrastive learning method to further capture the global content by predicting whether the video clips are sampled from the same video. We pre-train our model on uncurated videos and show that our pre-trained model can reach state-of-the-art results on several video understanding datasets (e.g., SSV2, Diving48). Lastly, we provide detailed analyses of the model scalability and pre-training method design. ", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=NP9T_pViXU", "arxiv_id": "2106.11250", "paper_pdf": "papers/NP9T_pViXU.pdf", "paper_pdf_sha256": "f95f9908f9ddfd86c9da5eee5c2d7bb3ab6ad5b4ab809d5a6b2d17892bb80ee6", "paper_pdf_bytes": 966814, "paper_pdf_source": "openreview", "code_url": "https://github.com/airsplay/vimpac", "code_repository": "airsplay/vimpac", "code_commit": "0760d19eecfc477bd39c448c3eec6226e07e7e4f", "code_archive": "repos/NP9T_pViXU.zip", "code_archive_sha256": "6ce028aaee79bf74d28eb976631270d0fed997e6d4c4c7f38570ead3bac525c3", "code_archive_bytes": 993474, "code_file_count": 33, "code_extensions": {".py": 22, ".sh": 11}, "github_disk_usage_kb": 1870, "github_languages": {"Python": 176892, "Shell": 12444}, "github_archived": false, "github_pushed_at": "2022-06-03T00:13:22Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/vimpac-video-pre-training-via-masked-token"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "_SKUm2AJpvN", "year": 2021, "status": "rejected", "title": "Decoupling Representation Learning from Reinforcement Learning", "authors": ["Adam Stooke", "Kimin Lee", "Pieter Abbeel", "Michael Laskin"], "authorids": ["~Adam_Stooke2", "~Kimin_Lee1", "~Pieter_Abbeel2", "mlaskin@berkeley.edu"], "authors_source": "OpenReview API", "abstract": "In an effort to overcome limitations of reward-driven feature learning in deep reinforcement learning (RL) from images, we propose decoupling representation learning from policy learning.  To this end, we introduce a new unsupervised learning (UL) task, called Augmented Temporal Contrast (ATC), which trains a convolutional encoder to associate pairs of observations separated by a short time difference, under image augmentations and using a contrastive loss.  In online RL experiments, we show that training the encoder exclusively using ATC matches or outperforms end-to-end RL in most environments.  Additionally, we benchmark several leading UL algorithms by pre-training encoders on expert demonstrations and using them, with weights frozen, in RL agents; we find that agents using ATC-trained encoders outperform all others.  We also train multi-task encoders on data from multiple environments and show generalization to different downstream RL tasks.  Finally, we ablate components of ATC, and introduce a new data augmentation to enable replay of (compressed) latent images from pre-trained encoders when RL requires augmentation.  Our experiments span visually diverse RL benchmarks in DeepMind Control, DeepMind Lab, and Atari, and our complete code is available at \\url{hidden url}.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=_SKUm2AJpvN", "arxiv_id": "2009.08319", "paper_pdf": "papers/_SKUm2AJpvN.pdf", "paper_pdf_sha256": "3b7defc2675a66ae7a8698766e02295d20593f3ed0f6a3ad5a479be40e82bb2d", "paper_pdf_bytes": 1957948, "paper_pdf_source": "openreview", "code_url": "https://github.com/astooke/rlpyt", "code_repository": "astooke/rlpyt", "code_commit": "f04f23db1eb7b5915d88401fca67869968a07a37", "code_archive": "repos/_SKUm2AJpvN.zip", "code_archive_sha256": "9d6653ab186008726d8ed589c0656516f5d6b55d4e20dbf9bfa53e312a7e95cf", "code_archive_bytes": 618818, "code_file_count": 410, "code_extensions": {".py": 410}, "github_disk_usage_kb": 1272, "github_languages": {"Python": 1387616}, "github_archived": false, "github_pushed_at": "2021-01-04T20:48:15Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/decoupling-representation-learning-from"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Skeh1krtvH", "year": 2020, "status": "rejected", "title": "WaveFlow: A Compact Flow-based Model for Raw Audio", "authors": ["Wei Ping", "Kainan Peng", "Kexin Zhao", "Zhao Song"], "authorids": ["weiping.thu@gmail.com"], "authors_source": "OpenReview API", "abstract": "In this work, we present WaveFlow, a small-footprint generative flow for raw audio, which is trained with maximum likelihood without complicated density distillation and auxiliary losses as used in Parallel WaveNet.  It provides a unified view of flow-based models for raw audio, including autoregressive flow (e.g., WaveNet) and bipartite flow (e.g., WaveGlow) as special cases. We systematically study these likelihood-based generative models for raw waveforms in terms of test likelihood and speech fidelity. We demonstrate that WaveFlow can synthesize high-fidelity speech and obtain comparable likelihood as WaveNet, while only requiring a few sequential steps to generate very long waveforms.  In particular, our small-footprint WaveFlow has only 5.91M parameters and can generate 22.05kHz speech 15.39 times faster than real-time on a GPU without customized inference kernels.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=Skeh1krtvH", "arxiv_id": "1912.01219", "paper_pdf": "papers/Skeh1krtvH.pdf", "paper_pdf_sha256": "782282378863686b403d582a694a2992d6258dceb46adba6ad1425fdb4b6831b", "paper_pdf_bytes": 353151, "paper_pdf_source": "openreview", "code_url": "https://github.com/PaddlePaddle/Parakeet", "code_repository": "PaddlePaddle/Parakeet", "code_commit": "8705a2a8405e3c63f2174d69880d2b5525a6c9fd", "code_archive": "repos/Skeh1krtvH.zip", "code_archive_sha256": "6231ae9f0dcb71998bdcdb5cd4c411856d00d56afaafb8193989ad4fbb0fb22e", "code_archive_bytes": 5527625, "code_file_count": 242, "code_extensions": {".py": 196, ".sh": 44, ".ipynb": 2}, "github_disk_usage_kb": 9775, "github_languages": {"Python": 725526, "Shell": 22908}, "github_archived": true, "github_pushed_at": "2021-11-19T02:21:01Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/waveflow-a-compact-flow-based-model-for-raw-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "UyiC52OPjn", "year": 2026, "status": "rejected", "title": "Characterizing Nonlinear Dynamics via Smooth Prototype Equivalences", "authors": ["Roy Friedman", "Noa Moriel", "Matthew Ricci", "Guy Pelc", "Yair Weiss", "Mor Nitzan"], "authorids": ["~Roy_Friedman2", "~Noa_Moriel1", "~Matthew_Ricci1", "~Guy_Pelc1", "~Yair_Weiss1", "~Mor_Nitzan1"], "authors_source": "OpenReview API", "abstract": "Characterizing dynamical systems given limited measurements is a common challenge throughout the physical and biological sciences. However, this task is challenging, especially due to transient variability in systems with equivalent long-term dynamics. We address this by introducing smooth prototype equivalences (SPE), a framework matches between sparse observations of phase space and prototypical behaviors using invertible neural networks. SPE enables classification by comparing the deformation loss of the observed sparse measurements to the prototype dynamics. Furthermore, our approach enables estimation of the invariant sets of the observed dynamics through the learned mapping from prototype space to data space. Our method outperforms existing techniques in the classification of oscillatory systems and can efficiently identify invariant structures like limit cycles and fixed points in an equation-free manner, even when only a small, noisy subset of the phase space is observed. Finally, we show how our method can be used for the detection of biological processes like the cell cycle trajectory from high-dimensional single-cell gene expression data.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=UyiC52OPjn", "arxiv_id": "2503.10336", "paper_pdf": "papers/UyiC52OPjn.pdf", "paper_pdf_sha256": "4541fb38110d373647f468a7540ce5d1a8b3e0e702b58b4188ec48e910304b7a", "paper_pdf_bytes": 5491647, "paper_pdf_source": "openreview", "code_url": "https://github.com/nitzanlab/prototype-equivalences", "code_repository": "nitzanlab/prototype-equivalences", "code_commit": "5291a487ce4a3befe034b91d7386e6872c7509b9", "code_archive": "repos/UyiC52OPjn.zip", "code_archive_sha256": "5c4793bc67fd42f8440426352310e3e4aa1c31e1ed7cff46884778e00941b0b8", "code_archive_bytes": 1649493, "code_file_count": 13, "code_extensions": {".py": 11, ".ipynb": 2}, "github_disk_usage_kb": 4272, "github_languages": {"Jupyter Notebook": 296751, "Python": 125709}, "github_archived": false, "github_pushed_at": "2026-08-05T08:43:05Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/characterizing-nonlinear-dynamics-via-smooth"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "S1OAqOtN5U", "year": 2025, "status": "rejected", "title": "Bayes Adaptive Monte Carlo Tree Search for Offline Model-based Reinforcement Learning", "authors": ["Jiayu Chen", "Wentse Chen", "Jeff Schneider"], "authorids": ["~Jiayu_Chen2", "~Wentse_Chen1", "~Jeff_Schneider1"], "authors_source": "OpenReview API", "abstract": "Offline reinforcement learning (RL) is a powerful approach for data-driven decision-making and control. Compared to model-free methods, offline model-based reinforcement learning (MBRL) explicitly learns world models from a static dataset and uses them as surrogate simulators, improving the data efficiency and enabling the learned policy to potentially generalize beyond the dataset support. However, there could be various MDPs that behave identically on the offline dataset and so dealing with the uncertainty about the true MDP can be challenging.  In this paper, we propose modeling offline MBRL as a Bayes Adaptive Markov Decision Process (BAMDP), which is a principled framework for addressing model uncertainty. We further introduce a novel Bayes Adaptive Monte-Carlo planning algorithm capable of solving BAMDPs in continuous state and action spaces with stochastic transitions. This planning process is based on Monte Carlo Tree Search and can be integrated into offline MBRL as a policy improvement operator in policy iteration. Our \"RL + Search\" framework follows in the footsteps of superhuman AIs like AlphaZero, improving on current offline MBRL methods by incorporating more computation input. The proposed algorithm significantly outperforms state-of-the-art model-based and model-free offline RL methods on twelve D4RL MuJoCo benchmark tasks and three target tracking tasks in a challenging, stochastic tokamak control simulator.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=S1OAqOtN5U", "arxiv_id": "2410.11234", "paper_pdf": "papers/S1OAqOtN5U.pdf", "paper_pdf_sha256": "12afde87eec63fec4b6a4eacc7d67bf41c246e020f4d8a5fd63098c16d439d36", "paper_pdf_bytes": 1755746, "paper_pdf_source": "openreview", "code_url": "https://github.com/LucasCJYSDL/Offline-RL-Kit", "code_repository": "LucasCJYSDL/Offline-RL-Kit", "code_commit": "0b15a36a7dad4ab9ed6b927c1368feca75670bca", "code_archive": "repos/S1OAqOtN5U.zip", "code_archive_sha256": "38a284549e567b6d69b1e32dcd833d9d156bd9afc96277356328ed9b49e724e8", "code_archive_bytes": 409140, "code_file_count": 56, "code_extensions": {".py": 51, ".cpp": 3, ".h": 2}, "github_disk_usage_kb": 305, "github_languages": {"Python": 225440, "C++": 23735, "Cython": 4004}, "github_archived": false, "github_pushed_at": "2025-03-19T22:18:09Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/bayes-adaptive-monte-carlo-tree-search-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "3b8CgMO5ix", "year": 2024, "status": "rejected", "title": "Model guidance via explanations turns image classifiers into segmentation models", "authors": ["Xiaoyan Yu", "Jannik Franzen", "Wojciech Samek", "Marina MC Höhne", "Dagmar Kainmueller"], "authorids": ["~Xiaoyan_Yu2", "~Jannik_Franzen1", "~Wojciech_Samek1", "~Marina_MC_Höhne1", "~Dagmar_Kainmueller2"], "authors_source": "OpenReview API", "abstract": "Heatmaps generated on inputs of image classification networks via explainable AI methods like Grad-CAM and LRP have been observed to resemble segmentations of input images in many cases. Consequently, heatmaps have also been leveraged for achieving weakly supervised segmentation with image-level supervision.\nOn the other hand, losses can be imposed on differentiable heatmaps, which has been shown to serve for (1) improving heatmaps to be more human-interpretable, (2) regularization of networks towards better generalization, (3) training diverse ensembles of networks, and (4) for explicitly ignoring confounding input features. Due to the latter use case, the paradigm of imposing losses on heatmaps is often referred to as \"Right for the right reasons\". \nWe unify these two lines of research by investigating semi-supervised segmentation as a novel use case for the Right for the Right Reasons paradigm. \nFirst, we show formal parallels between differentiable heatmap architectures and standard encoder-decoder architectures for image segmentation. \nSecond, we show that such differentiable heatmap architectures yield competitive results when trained with standard segmentation losses. \nThird, we show that such architectures allow for training with weak supervision in the form of image-level labels and small numbers of pixel-level labels, outperforming comparable encoder-decoder models.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=3b8CgMO5ix", "arxiv_id": "2407.03009", "paper_pdf": "papers/3b8CgMO5ix.pdf", "paper_pdf_sha256": "f5c1193a783f1239784eab92b25fc01bd7a8ea14bae458855b71d0658f8d7466", "paper_pdf_bytes": 12292941, "paper_pdf_source": "openreview", "code_url": "https://github.com/Kainmueller-Lab/TW-autoencoder", "code_repository": "Kainmueller-Lab/TW-autoencoder", "code_commit": "01d3932a173a4504f4bfe9638864cb46da607e63", "code_archive": "repos/3b8CgMO5ix.zip", "code_archive_sha256": "d1c0ac05f0c7f125536f2090c1c1f0f6ac512ee51d994e1d7623a4fe9000636a", "code_archive_bytes": 75742, "code_file_count": 46, "code_extensions": {".py": 45, ".sh": 1}, "github_disk_usage_kb": 252, "github_languages": {"Python": 222916, "Shell": 953}, "github_archived": false, "github_pushed_at": "2024-07-18T11:09:09Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/model-guidance-via-explanations-turns-image"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "deit1AdsFU", "year": 2023, "status": "rejected", "title": "Learning To Invert: Simple Adaptive Attacks for Gradient Inversion in Federated Learning", "authors": ["Ruihan Wu", "Xiangyu Chen", "Chuan Guo", "Kilian Q Weinberger"], "authorids": ["~Ruihan_Wu1", "~Xiangyu_Chen1", "~Chuan_Guo1", "~Kilian_Q_Weinberger1"], "authors_source": "OpenReview API", "abstract": "Gradient inversion attack enables recovery of training samples from model updates in federated learning (FL) and constitutes a serious threat to data privacy. To mitigate this vulnerability, prior work proposed both principled defenses based on differential privacy, as well as heuristic defenses based on gradient compression as countermeasures. These defenses have so far been very effective, in particular those based on gradient compression that allow the model to maintain high accuracy while greatly reducing the attack's effectiveness. In this work, we argue that such findings do not accurately reflect the privacy risk in FL, and show that existing defenses can be broken by a simple adaptive attack that trains a model using auxiliary data to learn how to invert gradients on both vision and language tasks.", "decision": "Reject", "meta_review": null, "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=deit1AdsFU", "arxiv_id": "2210.10880", "paper_pdf": "papers/deit1AdsFU.pdf", "paper_pdf_sha256": "a15867ba5dafaf263010a24527d5a2eb716e3be31c8b322c251d1e1668f71843", "paper_pdf_bytes": 8404531, "paper_pdf_source": "openreview", "code_url": "https://github.com/wrh14/Learning_to_Invert", "code_repository": "wrh14/Learning_to_Invert", "code_commit": "a3464c58166967ce12c917d368e03e8129fb60a6", "code_archive": "repos/deit1AdsFU.zip", "code_archive_sha256": "f5ec677a330f0bfdf83db62d47c72f36476d0bd8723714671400df5c0ba8385a", "code_archive_bytes": 637585, "code_file_count": 68, "code_extensions": {".py": 66, ".ipynb": 2}, "github_disk_usage_kb": 528, "github_languages": {"Python": 606402, "Jupyter Notebook": 10915}, "github_archived": false, "github_pushed_at": "2024-06-15T03:55:06Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/learning-to-invert-simple-adaptive-attacks"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "FFGDKzLasUa", "year": 2022, "status": "rejected", "title": "Stochastic Deep Networks with Linear Competing Units for Model-Agnostic Meta-Learning", "authors": ["Konstantinos Ι. Kalais", "Sotirios Chatzis"], "authorids": ["~Konstantinos_Ι._Kalais1", "~Sotirios_Chatzis1"], "authors_source": "OpenReview API", "abstract": "This work addresses meta-learning (ML) by considering deep networks with stochastic local winner-takes-all (LWTA) activations. This type of network units result in sparse representations from each model layer, as the units are organized into blocks where only one unit generates a non-zero output. The main operating principle of the introduced units lies on stochastic arguments, as the network performs posterior sampling over competing units to select the winner. Therefore, the proposed networks are explicitly designed to extract input data representations of sparse stochastic nature, as opposed to the currently standard deterministic representation paradigm. We posit that these modeling arguments, inspired from Bayesian statistics, allow for more robust modeling when uncertainty is high due to the limited availability of task-related training data; this is exactly the case with ML, which is the focus of this work. At training time, we rely on the reparameterization trick for Discrete distributions to perform reliable training via Monte-Carlo sampling. At inference time, we rely on Bayesian Model Averaging, which effectively averages over a number of sampled representations. As we experimentally show, our approach produces state-of-the-art predictive accuracy on standard few-shot image classification benchmarks; this is achieved without compromising computational efficiency.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=FFGDKzLasUa", "arxiv_id": "2208.01573", "paper_pdf": "papers/FFGDKzLasUa.pdf", "paper_pdf_sha256": "fc9fc13c23ca5b0cc33765ec221519988622468fd2d3069437fffd1c04c565ac", "paper_pdf_bytes": 1730465, "paper_pdf_source": "openreview", "code_url": "https://github.com/Kkalais/StochLWTA-ML", "code_repository": "Kkalais/StochLWTA-ML", "code_commit": "42d01bf70bf08612cd367e170cebe67fc9d1cfd1", "code_archive": "repos/FFGDKzLasUa.zip", "code_archive_sha256": "59a34a9bbe0eb248f2ff99e9cd2ee1e7ceba225c84a76977c755cdbe8452a03f", "code_archive_bytes": 2143291, "code_file_count": 26, "code_extensions": {".py": 23, ".ipynb": 3}, "github_disk_usage_kb": 2103, "github_languages": {"Jupyter Notebook": 574286, "Python": 186651}, "github_archived": false, "github_pushed_at": "2025-07-21T14:34:02Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/stochastic-deep-networks-with-linear-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "1NRMmEUyXMu", "year": 2021, "status": "rejected", "title": "World Model as a Graph: Learning Latent Landmarks for Planning", "authors": ["Lunjun Zhang", "Ge Yang", "Bradly C Stadie"], "authorids": ["~Lunjun_Zhang1", "~Ge_Yang1", "~Bradly_C_Stadie1"], "authors_source": "OpenReview API", "abstract": "Planning, the ability to analyze the structure of a problem in the large and decompose it into interrelated subproblems, is a hallmark of human intelligence. While deep reinforcement learning (RL) has shown great promise for solving relatively straightforward control tasks, it remains an open problem how to best incorporate planning into existing deep RL paradigms to handle increasingly complex environments. One prominent framework, Model-Based RL, learns a world model and plans using step-by-step virtual rollouts. This type of world model quickly diverges from reality when the planning horizon increases, thus struggling at long-horizon planning. How can we learn world models that endow agents with the ability to do temporally extended reasoning? In this work, we propose to learn graph-structured world models composed of sparse, multi-step transitions. We devise a novel algorithm to learn latent landmarks that are scattered (in terms of reachability) across the goal space as the nodes on the graph. In this same graph, the edges are the reachability estimates distilled from Q-functions. On a variety of high-dimensional continuous control tasks ranging from robotic manipulation to navigation, we demonstrate that our method, named L^{3}P, significantly outperforms prior work, and is oftentimes the only method capable of leveraging both the robustness of model-free RL and generalization of graph-search algorithms. We believe our work is an important step towards scalable planning in reinforcement learning. ", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=1NRMmEUyXMu", "arxiv_id": "2011.12491", "paper_pdf": "papers/1NRMmEUyXMu.pdf", "paper_pdf_sha256": "75a9eb65cc96c9c4bfa503ee1c414e68607d1a22c10add46d2c5201abc44ed17", "paper_pdf_bytes": 6599267, "paper_pdf_source": "openreview", "code_url": "https://github.com/LunjunZhang/world-model-as-a-graph", "code_repository": "LunjunZhang/world-model-as-a-graph", "code_commit": "b7d99b2efe15b9be162efe9a83cd99e9fdc88c33", "code_archive": "repos/1NRMmEUyXMu.zip", "code_archive_sha256": "f5de94aba1d2711cf3c4ffdafaa3f0dfeace888dc18d9255533c70a676448fd6", "code_archive_bytes": 1416445, "code_file_count": 69, "code_extensions": {".py": 64, ".sh": 5}, "github_disk_usage_kb": 1350, "github_languages": {"Python": 333076, "Shell": 4882}, "github_archived": false, "github_pushed_at": "2021-07-17T19:16:17Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/world-model-as-a-graph-learning-latent"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "MmYiosUoay", "year": 2026, "status": "rejected", "title": "Seeing is Understanding: Unlocking Causal Attention into Modality-Mutual Attention for Multimodal LLMs", "authors": ["Wei-Yao Wang", "Zhao Wang", "Helen Suzuki", "Yoshiyuki Kobayashi"], "authorids": ["~Wei-Yao_Wang1", "~Zhao_Wang6", "~Helen_Suzuki1", "~Yoshiyuki_Kobayashi2"], "authors_source": "OpenReview API", "abstract": "Recent Multimodal Large Language Models (MLLMs) have demonstrated significant progress in perceiving and reasoning over multimodal inquiries, ushering in a new research era for foundation models.\nHowever, vision-language misalignment in MLLMs has emerged as a critical challenge, where the textual responses generated by these models are not factually aligned with the given text-image inputs.\nExisting efforts to address vision-language misalignment have focused on developing specialized vision-language connectors or leveraging visual instruction tuning from diverse domains.\nIn this paper, we tackle this issue from a fundamental yet unexplored perspective by revisiting the core architecture of MLLMs.\nMost MLLMs are typically built on decoder-only LLMs consisting of a causal attention mechanism, which **limits the ability of the earlier modalities (e.g., images) to incorporate information from the latter modalities (e.g., text)**.\nTo address this problem a MLLM that unlocks causal attention into our proposed modality-mutual attention (MMA) to enable image tokens to attend to text tokens.\nThis simple yet effective design allows MMA to achieve state-of-the-art performance in 12 multimodal understanding benchmarks (**+5.5\\% on average across 2 LLMs backbones**) without introducing additional parameters.\nOur MMA design is intended to be generic, allowing for applications across various modalities, and scalable to accommodate diverse multimodal scenarios.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=MmYiosUoay", "arxiv_id": "2503.02597", "paper_pdf": "papers/MmYiosUoay.pdf", "paper_pdf_sha256": "ea0d4fe2dbc08473e6f1fe100785ae95e241c0211fe74034a2f9e87c9bb55d3a", "paper_pdf_bytes": 6370324, "paper_pdf_source": "openreview", "code_url": "https://github.com/sony/aki", "code_repository": "sony/aki", "code_commit": "c7642317a4c9078efe87a5533360ff07f1fdc228", "code_archive": "repos/MmYiosUoay.zip", "code_archive_sha256": "79f33205054cd0bd6bb53cbdc07ee487d00df769d4c828e3a1749accd7113619", "code_archive_bytes": 4505644, "code_file_count": 41, "code_extensions": {".py": 37, ".sh": 4}, "github_disk_usage_kb": 4404, "github_languages": {"Python": 256569, "Shell": 1850}, "github_archived": false, "github_pushed_at": "2026-05-11T07:11:01Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/seeing-is-understanding-unlocking-causal"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "qUJsX3XMBH", "year": 2025, "status": "rejected", "title": "Rethinking Data Selection at Scale: Random Selection is Almost All You Need", "authors": ["Tingyu Xia", "Bowen Yu", "Kai Dang", "An Yang", "Yuan Wu", "Yuan Tian", "Yi Chang", "Junyang Lin"], "authorids": ["~Tingyu_Xia1", "~Bowen_Yu3", "~Kai_Dang1", "~An_Yang1", "~Yuan_Wu2", "~Yuan_Tian6", "~Yi_Chang4", "~Junyang_Lin1"], "authors_source": "OpenReview API", "abstract": "Supervised fine-tuning (SFT) is crucial for aligning Large Language Models (LLMs) with human instructions. The primary goal during SFT is to select a small yet representative subset of training data from the larger pool, such that fine-tuning with this subset achieves results comparable to or even exceeding those obtained using the entire dataset. However, most existing data selection techniques are designed for small-scale data pools, which fail to meet the demands of real-world SFT scenarios. In this paper, we replicated several self-scoring methods—those that do not rely on external model assistance—on two million-scale datasets, and found that nearly all methods struggled to significantly outperform random selection when dealing with such large-scale data pools. Moreover, our comparisons suggest that, during SFT, diversity in data selection is more critical than simply focusing on high-quality data. We also analyzed the limitations of several current approaches, explaining why they perform poorly on large-scale datasets and why they are unsuitable for such contexts. Finally, we found that filtering data by token length offers a stable and efficient method for improving results. This approach, particularly when training on long-text data, proves highly beneficial for relatively weaker base models, such as Llama3.", "decision": "Reject", "meta_review": "", "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=qUJsX3XMBH", "arxiv_id": "2410.09335", "paper_pdf": "papers/qUJsX3XMBH.pdf", "paper_pdf_sha256": "65388a79d3f34feea031cce7c8ec4c34e132619e847fa6aa2cc6ef8330547d68", "paper_pdf_bytes": 418887, "paper_pdf_source": "openreview", "code_url": "https://github.com/xiatingyu/SFT-DataSelection-at-scale", "code_repository": "xiatingyu/SFT-DataSelection-at-scale", "code_commit": "5a635e42f84640f3c6b23a526fcfb1f251765f2b", "code_archive": "repos/qUJsX3XMBH.zip", "code_archive_sha256": "6e320586ae82a8f633e014d8d354d3bd5293a53650dc04c18c406c41e102c595", "code_archive_bytes": 375245, "code_file_count": 70, "code_extensions": {".py": 39, ".sh": 30, ".ipynb": 1}, "github_disk_usage_kb": 313, "github_languages": {"Jupyter Notebook": 285690, "Python": 227534, "Shell": 24823}, "github_archived": false, "github_pushed_at": "2025-02-09T03:05:52Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/rethinking-data-selection-at-scale-random"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "9Z0yB8rmQ2", "year": 2024, "status": "rejected", "title": "Lyra: Orchestrating Dual Correction in Automated Theorem Proving", "authors": ["Chuanyang Zheng", "Haiming Wang", "Enze Xie", "Zhengying Liu", "Jiankai Sun", "Huajian Xin", "Jianhao Shen", "Zhenguo Li", "Yu Li"], "authorids": ["~Chuanyang_Zheng3", "~Haiming_Wang1", "~Enze_Xie1", "~Zhengying_Liu2", "~Jiankai_Sun6", "~Huajian_Xin1", "~Jianhao_Shen1", "~Zhenguo_Li1", "~Yu_Li1"], "authors_source": "OpenReview API", "abstract": "Large Language Models (LLMs) present an intriguing avenue for exploration in the field of formal theorem proving. Nevertheless, their full potential, particularly concerning the mitigation of hallucinations and refinement through prover error messages, remains an area that has yet to be thoroughly investigated. To enhance the effectiveness of LLMs in the field, we introduce the Lyra, a new framework that employs two distinct correction mechanisms: Tool Correction (TC) and Conjecture Correction (CC). To implement Tool Correction in the post-processing of formal proofs, we leverage prior knowledge to utilize predefined prover tools (e.g., Sledgehammer) for guiding the replacement of incorrect tools. Tool Correction significantly contributes to mitigating hallucinations, thereby improving the overall accuracy of the proof. In addition, we introduce Conjecture Correction, an error feedback mechanism designed to interact with prover to refine formal proof conjectures with prover error messages. Compared to the previous refinement framework, the proposed Conjecture Correction refines generation with instruction but does not collect paired (generation, error & refinement) prompts. Our method has achieved state-of-the-art (SOTA) performance on both miniF2F validation (48.0% → 55.3%) and test (45.5% → 51.2%). We also present 3 IMO problems solved by Lyra. We believe Tool Correction (post-process for hallucination mitigation) and Conjecture Correction (subgoal adjustment from interaction with environment) could provide a promising avenue for future research in this field.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=9Z0yB8rmQ2", "arxiv_id": "2309.15806", "paper_pdf": "papers/9Z0yB8rmQ2.pdf", "paper_pdf_sha256": "4c9c2f59f6b73ce10222f6ea5dbf2fadd45a060352f84a5435485328a01da51f", "paper_pdf_bytes": 1024887, "paper_pdf_source": "openreview", "code_url": "https://github.com/chuanyang-Zheng/Lyra-theorem-prover", "code_repository": "chuanyang-Zheng/Lyra-theorem-prover", "code_commit": "4424ad734f1fb1ef9aaa13979cd273a54247acb7", "code_archive": "repos/9Z0yB8rmQ2.zip", "code_archive_sha256": "4463504a264ab7abfde1125b9259107b0b83ec3e150d030ba9214f3e5526aaa4", "code_archive_bytes": 286377, "code_file_count": 24, "code_extensions": {".py": 22, ".sh": 2}, "github_disk_usage_kb": 256, "github_languages": {"Python": 215424, "Isabelle": 154664, "Shell": 4382}, "github_archived": false, "github_pushed_at": "2024-07-02T06:22:56Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/lyra-orchestrating-dual-correction-in"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "zU2v47WF0Ku", "year": 2022, "status": "rejected", "title": "Implicit Bias of Linear Equivariant Networks", "authors": ["Hannah Lawrence", "Kristian Georgiev", "Andrew Dienes", "Bobak Kiani"], "authorids": ["~Hannah_Lawrence1", "~Kristian_Georgiev1", "adienes@mit.edu", "~Bobak_Kiani1"], "authors_source": "OpenReview API", "abstract": "Group equivariant convolutional neural networks (G-CNNs) are generalizations of convolutional neural networks (CNNs) which excel in a wide range of scientific and technical applications by explicitly encoding particular group symmetries, such as rotations and permutations, in their architectures. Although the success of G-CNNs is driven by the explicit symmetry bias of their convolutional architecture, a recent line of work has proposed that the implicit bias of training algorithms on a particular parameterization (or architecture) is key to understanding generalization for overparameterized neural nets. In this context, we show that $L$-layer full-width linear G-CNNs trained via gradient descent in a binary classification task converge to solutions with low-rank Fourier matrix coefficients, regularized by the $2/L$-Schatten matrix norm. Our work strictly generalizes previous analysis on the implicit bias of linear CNNs to linear G-CNNs over all finite groups, including the challenging setting of non-commutative symmetry groups (such as permutations). We validate our theorems via experiments on a variety of groups and empirically explore more realistic nonlinear networks, which locally capture similar regularization patterns. Finally, we provide intuitive interpretations of our Fourier-space implicit regularization results in real space via uncertainty principles.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=zU2v47WF0Ku", "arxiv_id": "2110.06084", "paper_pdf": "papers/zU2v47WF0Ku.pdf", "paper_pdf_sha256": "01991cbb43389206a334071fd0632d3079a17d68f5e21417c045c2c7720bfdac", "paper_pdf_bytes": 4811337, "paper_pdf_source": "openreview", "code_url": "https://github.com/kristian-georgiev/implicit-bias-of-linear-equivariant-networks", "code_repository": "kristian-georgiev/implicit-bias-of-linear-equivariant-networks", "code_commit": "ac5a5536ea98f06b3b33870b73e700a09b736777", "code_archive": "repos/zU2v47WF0Ku.zip", "code_archive_sha256": "81c6720db7cbbee21ca7e46d4ac747780980a55481fdeaf21cadfd7e5b7e5ee2", "code_archive_bytes": 2393383, "code_file_count": 15, "code_extensions": {".ipynb": 10, ".py": 5}, "github_disk_usage_kb": 2153, "github_languages": {"Jupyter Notebook": 1370211, "Python": 39216, "Sage": 3859, "GAP": 1969}, "github_archived": false, "github_pushed_at": "2022-06-17T05:08:58Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/implicit-bias-of-linear-equivariant-networks-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "cbtV7xGO9pS", "year": 2021, "status": "rejected", "title": "TEAC: Intergrating Trust Region and Max Entropy Actor Critic for Continuous Control", "authors": ["Hongyu Zang", "Xin Li", "Li Zhang", "Peiyao Zhao", "Mingzhong Wang"], "authorids": ["~Hongyu_Zang1", "~Xin_Li31", "~Li_Zhang18", "~Peiyao_Zhao1", "~Mingzhong_Wang1"], "authors_source": "OpenReview API", "abstract": "Trust region methods and maximum entropy methods are two state-of-the-art branches used in reinforcement learning (RL) for the benefits of stability and exploration in continuous environments, respectively. This paper proposes to integrate both branches in a unified framework, thus benefiting from both sides. We first transform the original RL objective to a constraint optimization problem and then proposes trust entropy actor-critic (TEAC), an off-policy algorithm to learn stable and sufficiently explored policies for continuous states and actions. TEAC trains the critic by minimizing the refined Bellman error and updates the actor by minimizing KL-divergence loss derived from the closed-form solution to the Lagrangian. \nWe prove that the policy evaluation and policy improvement in TEAC is guaranteed to converge.\nWe compare TEAC with 4 state-of-the-art solutions on 6 tasks in the MuJoCo environment. The results show that TEAC outperforms state-of-the-art solutions in terms of efficiency and effectiveness.\n", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=cbtV7xGO9pS", "arxiv_id": null, "paper_pdf": "papers/cbtV7xGO9pS.pdf", "paper_pdf_sha256": "7b08de7608f3ab5794d9a32f5bf2aab60c35465d5b0a2c719d78bed6f1f5fe48", "paper_pdf_bytes": 3088734, "paper_pdf_source": "openreview", "code_url": "https://github.com/ICLR2021papersub/TEAC", "code_repository": "ICLR2021papersub/TEAC", "code_commit": "bad3488749963daf8cc71bc5dcb870e71b97abdd", "code_archive": "repos/cbtV7xGO9pS.zip", "code_archive_sha256": "dd32d7a66f767c10d0e1b463eda84c0ad1fab4931403ef59ceb66b4468fe5650", "code_archive_bytes": 724386, "code_file_count": 27, "code_extensions": {".py": 27}, "github_disk_usage_kb": 1368, "github_languages": {"Python": 212088}, "github_archived": false, "github_pushed_at": "2020-10-30T06:11:35Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/teac-intergrating-trust-region-and-max"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "wyCnT4BUsT", "year": 2026, "status": "rejected", "title": "DeepCritic: Deliberate Critique with Large Language Models", "authors": ["Wenkai Yang", "Jingwen Chen", "Yankai Lin", "Ji-Rong Wen"], "authorids": ["~Wenkai_Yang1", "~Jingwen_Chen4", "~Yankai_Lin1", "~Ji-Rong_Wen1"], "authors_source": "OpenReview API", "abstract": "As Large Language Models (LLMs) are rapidly evolving, providing accurate feedback and scalable oversight on their outputs becomes an urgent and critical problem. Leveraging LLMs as critique models to achieve automated supervision is a promising solution. In this work, we focus on studying and enhancing the math critique ability of LLMs. Current LLM critics provide critiques that are too shallow and superficial on each step, leading to low judgment accuracy and struggling to offer sufficient feedback for the LLM generator to correct mistakes. To tackle this issue, we propose a novel and effective two-stage framework to develop LLM critics that are capable of deliberately critiquing on each reasoning step of math solutions. In the first stage, we carefully curate 4.5K long-form critiques as seed data for supervised fine-tuning. Each seed critique consists of deliberate step-wise critiques that includes multi-perspective verifications as well as in-depth critiques of initial critiques for each reasoning step. Then, we perform reinforcement learning on the fine-tuned model with either existing human-labeled data from PRM800K or our automatically annotated data obtained via Monte Carlo sampling-based correctness estimation, to further incentivize its critique ability. Our developed critique model built on Qwen2.5-7B-Instruct not only significantly outperforms existing LLM critics (including the same-sized DeepSeek-R1-Distill models and GPT-4o) on various error identification benchmarks, but also more effectively helps the LLM generator refine erroneous steps through more detailed feedback.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=wyCnT4BUsT", "arxiv_id": "2505.00662", "paper_pdf": "papers/wyCnT4BUsT.pdf", "paper_pdf_sha256": "a5bc0f8fdd2ebe2b03684b968e67f72315e68ebc056c2403d248c3647d2372af", "paper_pdf_bytes": 1271889, "paper_pdf_source": "openreview", "code_url": "https://github.com/RUCBM/DeepCritic", "code_repository": "RUCBM/DeepCritic", "code_commit": "53eaf5e048187162451ff165e823c27d976d0e09", "code_archive": "repos/wyCnT4BUsT.zip", "code_archive_sha256": "d2db3f057f7b21ec1c4c7d9ce0c1e37a2d9af3590a4b845441a60dd01ce51a20", "code_archive_bytes": 4651850, "code_file_count": 24, "code_extensions": {".py": 19, ".sh": 5}, "github_disk_usage_kb": 4535, "github_languages": {"Python": 125044, "Shell": 3698}, "github_archived": false, "github_pushed_at": "2025-06-24T12:50:24Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/deepcritic-deliberate-critique-with-large"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "DDxLsxiZR8", "year": 2025, "status": "rejected", "title": "CAT Pruning: Cluster-Aware Token Pruning For Text-to-Image Diffusion Models", "authors": ["Xinle Cheng", "Zhuoming Chen", "Zhihao Jia"], "authorids": ["~Xinle_Cheng1", "~Zhuoming_Chen1", "~Zhihao_Jia2"], "authors_source": "OpenReview API", "abstract": "Diffusion models have transformed generative tasks, particularly in text-to-image synthesis, but their iterative denoising process is computationally intensive. We present a novel acceleration strategy that combines token-level pruning with cache mechanisms to address this challenge. By utilizing Noise Relative Magnitude, we identify significant token changes across iterations. Additionally, we incorporate spatial clustering and distributional balance to enhance token selection. Our experiments demonstrate 50\\%-60\\% reduction in computational cost while maintaining model performance, offering a substantial improvement in the efficiency of diffusion models.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=DDxLsxiZR8", "arxiv_id": "2502.00433", "paper_pdf": "papers/DDxLsxiZR8.pdf", "paper_pdf_sha256": "b0e71fb65d37783c4203afae0f90ff1b3cff73ec508469de303a702aeecb89d7", "paper_pdf_bytes": 3146516, "paper_pdf_source": "openreview", "code_url": "https://github.com/ada-cheng/CAT-Pruning", "code_repository": "ada-cheng/CAT-Pruning", "code_commit": "3b88889e4deea8f1033b509efe3fd5bab90a7b43", "code_archive": "repos/DDxLsxiZR8.zip", "code_archive_sha256": "09f45b3d3140b619917c40786910176793d7e5d301296f6d975da1f82b5fd511", "code_archive_bytes": 334769, "code_file_count": 18, "code_extensions": {".py": 18}, "github_disk_usage_kb": 324, "github_languages": {"Python": 207497}, "github_archived": false, "github_pushed_at": "2025-07-26T06:11:14Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/cat-pruning-cluster-aware-token-pruning-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "tAmfM1sORP", "year": 2024, "status": "rejected", "title": "Large Language Models can Learn Rules", "authors": ["Zhaocheng Zhu", "Yuan Xue", "Xinyun Chen", "Denny Zhou", "Jian Tang", "Dale Schuurmans", "Hanjun Dai"], "authorids": ["~Zhaocheng_Zhu1", "~Yuan_Xue5", "~Xinyun_Chen1", "~Denny_Zhou1", "~Jian_Tang1", "~Dale_Schuurmans1", "~Hanjun_Dai1"], "authors_source": "OpenReview API", "abstract": "When prompted with a few examples and intermediate steps, large language models (LLMs) have demonstrated impressive performance in various reasoning tasks. However, prompting methods that rely on implicit knowledge in an LLM often hallucinate incorrect answers when the implicit knowledge is wrong or inconsistent with the task. To tackle this problem, we present Hypotheses-to-Theories (HtT), a framework that learns a rule library for reasoning with LLMs. HtT contains two stages, an induction stage and a deduction stage. In the induction stage, an LLM is first asked to generate and verify rules over a set of training examples. Rules that appear and lead to correct answers sufficiently often are collected to form a rule library. In the deduction stage, the LLM is then prompted to employ the learned rule library to perform reasoning to answer test questions. Experiments on both numerical reasoning and relational reasoning problems show that HtT improves existing prompting methods, with an absolute gain of 11-27% in accuracy. The learned rules are also transferable to different models and to different forms of the same problem.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=tAmfM1sORP", "arxiv_id": "2310.07064", "paper_pdf": "papers/tAmfM1sORP.pdf", "paper_pdf_sha256": "cbcab388152cd96c31c8f6ea2b79f1966bef41583f08ee6ff5b33cdb222345b9", "paper_pdf_bytes": 423040, "paper_pdf_source": "openreview", "code_url": "https://github.com/google-deepmind/llms_can_learn_rules", "code_repository": "google-deepmind/llms_can_learn_rules", "code_commit": "5c2c523690720a314932dcd8f77c321a3bd9c1f8", "code_archive": "repos/tAmfM1sORP.zip", "code_archive_sha256": "3eeb7d588e9298e9c126088986ea527b9ad2ba7c383a91a8a6ff16c204bbc08c", "code_archive_bytes": 265211, "code_file_count": 7, "code_extensions": {".py": 6, ".sh": 1}, "github_disk_usage_kb": 262, "github_languages": {"Python": 30326, "Shell": 1770}, "github_archived": false, "github_pushed_at": "2024-12-06T01:54:21Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/large-language-models-can-learn-rules"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "7PURWDjJCf3", "year": 2023, "status": "rejected", "title": "Slimmable Networks for Contrastive Self-supervised Learning", "authors": ["Shuai Zhao", "Xiaohan Wang", "Linchao Zhu", "Yi Yang"], "authorids": ["~Shuai_Zhao1", "~Xiaohan_Wang2", "~Linchao_Zhu1", "~Yi_Yang22"], "authors_source": "OpenReview API", "abstract": "Self-supervised learning makes great progress in large model pre-training but suffers in training small models. Previous solutions to this problem mainly rely on knowledge distillation and indeed have a two-stage learning procedure: first train a large teacher model, then distill it to improve the generalization ability of small ones. In this work, we present a new one-stage solution to obtain pre-trained small models without extra teachers: slimmable networks for contrastive self-supervised learning (SlimCLR). A slimmable network contains a full network and several weight-sharing sub-networks. We can pre-train for only one time and obtain various networks including small ones with low computation costs. However, in self-supervised cases, the interference between weight-sharing networks leads to severe performance degradation. One evidence of the interference is gradient imbalance: a small proportion of parameters produces dominant gradients during backpropagation, and the main parameters may not be fully optimized. The divergence in gradient directions of various networks may also cause interference between networks. To overcome these problems, we make the main parameters produce dominant gradients and provide consistent guidance for sub-networks via three techniques: slow start training of sub-networks, online distillation, and loss re-weighting according to model sizes. Besides, a switchable linear probe layer is applied during linear evaluation to avoid the interference of weight-sharing linear layers. We instantiate SlimCLR with typical contrastive learning frameworks and achieve better performance than previous arts with fewer parameters and FLOPs.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=7PURWDjJCf3", "arxiv_id": "2209.15525", "paper_pdf": "papers/7PURWDjJCf3.pdf", "paper_pdf_sha256": "768457d3d4c79d1c6cb0934192c3d524f8ad73eea3bf0d7fb95e7291351d8085", "paper_pdf_bytes": 2204918, "paper_pdf_source": "openreview", "code_url": "https://github.com/mzhaoshuai/SlimCLR", "code_repository": "mzhaoshuai/SlimCLR", "code_commit": "d975c2ab3fb1aaa0c3dbe2520a09ed449970d5cd", "code_archive": "repos/7PURWDjJCf3.zip", "code_archive_sha256": "966816b32815780d66e617f915c1af091753185541fb30c529b979f78f2a4696", "code_archive_bytes": 629787, "code_file_count": 155, "code_extensions": {".py": 146, ".sh": 9}, "github_disk_usage_kb": 582, "github_languages": {"Python": 657129, "Shell": 20640}, "github_archived": false, "github_pushed_at": "2025-11-18T04:07:40Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/slimmable-networks-for-contrastive-self"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Qu_XudmGajz", "year": 2022, "status": "rejected", "title": "Structured Uncertainty in the Observation Space of Variational Autoencoders", "authors": ["James Langley", "Miguel Monteiro", "Charles Jones", "Nick Pawlowski", "Ben Glocker"], "authorids": ["~James_Langley1", "~Miguel_Monteiro1", "~Charles_Jones4", "~Nick_Pawlowski2", "~Ben_Glocker1"], "authors_source": "OpenReview API", "abstract": "Variational autoencoders (VAEs) are a popular class of deep generative models with many variants and a wide range of applications. Improvements upon the standard VAE mostly focus on the modelling of the posterior distribution over the latent space and the properties of the neural network decoder. In contrast, improving the model for the observational distribution is rarely considered and typically defaults to a pixel-wise independent categorical or normal distribution. In image synthesis, sampling from such distributions produces spatially-incoherent results with uncorrelated pixel noise, resulting in only the sample mean being somewhat useful as an output prediction. In this paper, we aim to stay true to VAE theory by improving the samples from the observational distribution. We propose an alternative model for the observation space, encoding spatial dependencies via a low-rank parameterization. We demonstrate that this new observational distribution has the ability to capture relevant covariance between pixels, resulting in spatially-coherent samples. In contrast to pixel-wise independent distributions, our samples seem to contain semantically meaningful variations from the mean allowing the prediction of multiple plausible outputs with a single forward pass.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=Qu_XudmGajz", "arxiv_id": "2205.12533", "paper_pdf": "papers/Qu_XudmGajz.pdf", "paper_pdf_sha256": "baad970b9fe313794529d8fc7729e3b6b34752b6af86a988c1caa959bf8a5af1", "paper_pdf_bytes": 47036636, "paper_pdf_source": "openreview", "code_url": "https://github.com/biomedia-mira/sos-vae", "code_repository": "biomedia-mira/sos-vae", "code_commit": "cc02afb2551dee5b472741dce3d1e82e0996c83a", "code_archive": "repos/Qu_XudmGajz.zip", "code_archive_sha256": "12e59ab95dd7d30e72f07eb9a745e0b6c973405eb6c5d6fe0b3749a7207ce0c7", "code_archive_bytes": 2262927, "code_file_count": 27, "code_extensions": {".py": 27}, "github_disk_usage_kb": 2204, "github_languages": {"Python": 113931}, "github_archived": false, "github_pushed_at": "2022-10-28T22:11:51Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/structured-uncertainty-in-the-observation-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "lJgbDxGhJ4r", "year": 2021, "status": "rejected", "title": "OpenCoS: Contrastive Semi-supervised Learning for Handling Open-set Unlabeled Data", "authors": ["Jongjin Park", "Sukmin Yun", "Jongheon Jeong", "Jinwoo Shin"], "authorids": ["~Jongjin_Park1", "~Sukmin_Yun1", "~Jongheon_Jeong1", "~Jinwoo_Shin1"], "authors_source": "OpenReview API", "abstract": "Modern semi-supervised learning methods conventionally assume both labeled and unlabeled data have the same class distribution. However, unlabeled data may include out-of-class samples in practice; those that cannot have one-hot encoded labels from a closed-set of classes in label data, i.e., unlabeled data is an open-set. In this paper, we introduce OpenCoS, a method for handling this realistic semi-supervised learning scenario based on a recent framework of contrastive learning. One of our key findings is that out-of-class samples in the unlabeled dataset can be identified effectively via (unsupervised) contrastive learning. OpenCoS utilizes this information to overcome the failure modes in the existing state-of-the-art semi-supervised methods, e.g., ReMixMatch or FixMatch. In particular, we propose to assign soft-labels for out-of-class samples using the representation learned from contrastive learning. Our extensive experimental results show the effectiveness of OpenCoS, fixing the state-of-the-art semi-supervised methods to be suitable for diverse scenarios involving open-set unlabeled data.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=lJgbDxGhJ4r", "arxiv_id": "2107.08943", "paper_pdf": "papers/lJgbDxGhJ4r.pdf", "paper_pdf_sha256": "a4fc42ef14d533800188e561a531b632211d36ab862dd782ef0b7f02cf2148c6", "paper_pdf_bytes": 1422343, "paper_pdf_source": "openreview", "code_url": "https://github.com/alinlab/OpenCoS", "code_repository": "alinlab/OpenCoS", "code_commit": "59003724045f82cf1ca54b2d509da7abc5aefe96", "code_archive": "repos/lJgbDxGhJ4r.zip", "code_archive_sha256": "6b8950d86121ad9c904ac288f5d6972dc2312275058e69a3bfc8c63a58a6a078", "code_archive_bytes": 1502669, "code_file_count": 31, "code_extensions": {".py": 31}, "github_disk_usage_kb": 1425, "github_languages": {"Python": 281937}, "github_archived": false, "github_pushed_at": "2022-06-16T08:04:53Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/opencos-contrastive-semi-supervised-learning-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "6rm4ZC1nnP", "year": 2026, "status": "rejected", "title": "LDC-MTL: Balancing Multi-Task Learning through Scalable Loss Discrepancy Control", "authors": ["Peiyao Xiao", "Chaosheng Dong", "Shaofeng Zou", "Kaiyi Ji"], "authorids": ["~Peiyao_Xiao1", "~Chaosheng_Dong1", "~Shaofeng_Zou1", "~Kaiyi_Ji1"], "authors_source": "OpenReview API", "abstract": "Multi-task learning (MTL) has been widely adopted for its ability to simultaneously learn multiple tasks. While existing gradient manipulation methods often yield more balanced solutions than simple scalarization-based approaches, they typically incur a significant computational overhead of $\\mathcal{O}(K)$ in both time and memory, where $K$ is the number of tasks. In this paper, we propose LDC-MTL, a simple and scalable loss discrepancy control approach for MTL, formulated from a bilevel optimization perspective. Our method incorporates two key components: (i) a bilevel formulation for fine-grained loss discrepancy control, and (ii) a scalable first-order bilevel algorithm that requires only $\\mathcal{O}(1)$ time and memory. Theoretically, we prove that LDC-MTL guarantees convergence not only to a stationary point of the bilevel problem with loss discrepancy control but also to an $\\epsilon$-accurate Pareto stationary point for all $K$ loss functions under mild conditions. Extensive experiments on diverse multi-task datasets demonstrate the superior performance of LDC-MTL in both accuracy and efficiency.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=6rm4ZC1nnP", "arxiv_id": "2502.08585", "paper_pdf": "papers/6rm4ZC1nnP.pdf", "paper_pdf_sha256": "a1cb15bb99381f7a685fd643cfc7b43ccbb1c3393936e8e0fa4cf9b760c69fba", "paper_pdf_bytes": 1512519, "paper_pdf_source": "openreview", "code_url": "https://github.com/OptMN-Lab/LDC-MTL", "code_repository": "OptMN-Lab/LDC-MTL", "code_commit": "7c3946518689bf05be196c7bd51947b2cf885c9e", "code_archive": "repos/6rm4ZC1nnP.zip", "code_archive_sha256": "b659cca14ea0b2b5787299f6f00a17213201b26367d66fba2a84d7eb352b6b30", "code_archive_bytes": 3543370, "code_file_count": 24, "code_extensions": {".py": 20, ".sh": 4}, "github_disk_usage_kb": 4871, "github_languages": {"Python": 190554, "Shell": 806}, "github_archived": false, "github_pushed_at": "2025-05-15T16:15:10Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/scalable-bilevel-loss-balancing-for-multi"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ogmzNfeRl7", "year": 2025, "status": "rejected", "title": "Correlations Are Ruining Your Gradient Descent", "authors": ["Nasir Ahmad"], "authorids": ["~Nasir_Ahmad1"], "authors_source": "OpenReview API", "abstract": "Herein the topics of (natural) gradient descent, data decorrelation, and approximate methods for backpropagation are brought into a common discussion. Natural gradient descent illuminates how gradient vectors, pointing at directions of steepest descent, can be improved by considering the local curvature of loss landscapes. We extend this perspective and show that to fully solve the problem illuminated by natural gradients in neural networks, one must recognise that correlations in the data at any linear transformation, including node responses at every layer of a neural network, cause a non-orthonormal relationship between the model's parameters. To solve this requires a method for decorrelating inputs at each individual layer of a neural network. We describe a range of methods which have been proposed for decorrelation and whitening of node output, and expand on these to provide a novel method specifically useful for distributed computing and computational neuroscience. Implementing decorrelation within multi-layer neural networks, we can show that not only is training via backpropagation sped up significantly but also existing approximations of backpropagation, which have failed catastrophically in the past, benefit significantly in their accuracy and convergence speed. This has the potential to provide a route forward for approximate gradient descent methods which have previously been discarded, training approaches for analogue and neuromorphic hardware, and potentially insights as to the efficacy and utility of decorrelation processes in the brain.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=ogmzNfeRl7", "arxiv_id": "2407.10780", "paper_pdf": "papers/ogmzNfeRl7.pdf", "paper_pdf_sha256": "613e8814a56b2b3850c33e230833b73d32bae9ce91cf3e7c0e94c2af0e3cc46f", "paper_pdf_bytes": 1715785, "paper_pdf_source": "openreview", "code_url": "https://github.com/nasiryahm/CorrelationsRuinGD", "code_repository": "nasiryahm/CorrelationsRuinGD", "code_commit": "00cf47050bbd20e6a153e10bd86e1651524f9779", "code_archive": "repos/ogmzNfeRl7.zip", "code_archive_sha256": "880c79deed5c0065c46f1d09d3963ac9dda1ef9dcae8b87c9eccff103ba647ac", "code_archive_bytes": 80833, "code_file_count": 10, "code_extensions": {".py": 9, ".ipynb": 1}, "github_disk_usage_kb": 329, "github_languages": {"Jupyter Notebook": 98905, "Python": 53694}, "github_archived": false, "github_pushed_at": "2025-06-25T09:21:55Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/correlations-are-ruining-your-gradient"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Q8cVivO5k5", "year": 2024, "status": "rejected", "title": "Large-Batch, Iteration-Efficient Neural Bayesian Design Optimization", "authors": ["Navid Ansari", "Hans-peter Seidel", "Vahid Babaei"], "authorids": ["~Navid_Ansari1", "~Hans-peter_Seidel1", "~Vahid_Babaei1"], "authors_source": "OpenReview API", "abstract": "Bayesian optimization (BO) provides a powerful framework for optimizing black-box, expensive-to-evaluate functions. It is therefore an attractive tool for engineering design problems, typically involving multiple objectives. Thanks to the rapid advances in fabrication and measurement methods as well as parallel computing infrastructure, querying many design problems can be heavily parallelized. This class of problems challenges BO with an unprecedented setup where it has to deal with very large batches, shifting its focus from sample efficiency to iteration efficiency. We present a novel Bayesian optimization framework specifically tailored to address these limitations. Our key contribution is a highly scalable, sample-based acquisition function that performs a non-dominated sorting of not only the objectives but also their associated uncertainty. We show that our acquisition function in combination with different Bayesian neural network surrogates is effective in data-intensive environments with a minimal number of iterations. We demonstrate the superiority of our method by comparing it with state-of-the-art multi-objective optimizations. We perform our evaluation on two real-world problems - airfoil design and 3D printing -showcasing the applicability and efficiency of our approach. Our code is available at: https://github.com/an-on-ym-ous/lbn_mobo", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=Q8cVivO5k5", "arxiv_id": "2306.01095", "paper_pdf": "papers/Q8cVivO5k5.pdf", "paper_pdf_sha256": "f97e8cccb755588400aeb1d9819a025066426fca479a6f88ac461925b2467efe", "paper_pdf_bytes": 7788836, "paper_pdf_source": "openreview", "code_url": "https://github.com/AnsariNavid/lbn_mobo", "code_repository": "AnsariNavid/lbn_mobo", "code_commit": "da980750ecea1ed386d9ff54c16f6ae9c279e905", "code_archive": "repos/Q8cVivO5k5.zip", "code_archive_sha256": "04c4f3012720256a5464e793d4d5bcd545baf628f5cc8624e6bf4be97931e9cd", "code_archive_bytes": 211249, "code_file_count": 67, "code_extensions": {".py": 46, ".sh": 19, ".ipynb": 1, ".m": 1}, "github_disk_usage_kb": 264, "github_languages": {"Jupyter Notebook": 314060, "Python": 108599, "Shell": 4986, "MATLAB": 2605}, "github_archived": false, "github_pushed_at": "2023-05-22T14:44:37Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/large-batch-neural-multi-objective-bayesian"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "NHfSJAWhKTw", "year": 2023, "status": "rejected", "title": "A Closer Look at Self-supervised Lightweight Vision Transformers", "authors": ["Shaoru Wang", "Jin Gao", "Zeming Li", "Weiming Hu"], "authorids": ["~Shaoru_Wang1", "~Jin_Gao1", "~Zeming_Li2", "~Weiming_Hu1"], "authors_source": "OpenReview API", "abstract": "Self-supervised learning on large-scale Vision Transformers (ViTs) as pre-training methods has achieved promising downstream performance. Yet, how much these pre-training paradigms promote lightweight ViTs' performance is considerably less studied. In this work, we mainly develop and benchmark self-supervised pre-training methods, e.g., contrastive-learning-based MoCo-v3, masked-image-modeling-based MAE on image classification tasks, and some downstream dense prediction tasks. We surprisingly find that if proper pre-training is adopted, even vanilla lightweight ViTs show comparable performance on ImageNet to previous SOTA networks with delicate architecture design. We also point out some defects of such pre-training, \\eg, failing to benefit from large-scale pre-training data and showing inferior performance on data-insufficient downstream tasks. Furthermore, we analyze and clearly show the effect of such pre-training by analyzing the properties of the layer representation and attention maps for related models. Finally, based on the above analyses, a distillation strategy during pre-training is developed, which leads to further downstream performance improvement for MAE-based pre-training.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=NHfSJAWhKTw", "arxiv_id": "2205.14443", "paper_pdf": "papers/NHfSJAWhKTw.pdf", "paper_pdf_sha256": "677cdd497dadddf393f8d5f6e86987d727cf0b78100afedafe921868f58f4856", "paper_pdf_bytes": 1274963, "paper_pdf_source": "openreview", "code_url": "https://github.com/wangsr126/MAE-Lite", "code_repository": "wangsr126/MAE-Lite", "code_commit": "a5d1a74865b661c0e3a91e3ab1d5331ad1deac9c", "code_archive": "repos/NHfSJAWhKTw.zip", "code_archive_sha256": "66bb39272f229210ea715058df81c31d6795938d61cd96020aa8f8a6127c7910", "code_archive_bytes": 753174, "code_file_count": 238, "code_extensions": {".py": 229, ".sh": 8, ".ipynb": 1}, "github_disk_usage_kb": 611, "github_languages": {"Python": 1158942, "Shell": 9347, "Jupyter Notebook": 2581, "Makefile": 198}, "github_archived": false, "github_pushed_at": "2025-03-02T15:39:59Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/a-closer-look-at-self-supervised-lightweight"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ZzwfldvDLpC", "year": 2022, "status": "rejected", "title": "Let Your Heart Speak in its Mother Tongue: Multilingual Captioning of Cardiac Signals", "authors": ["Dani Kiyasseh", "Tingting Zhu", "David A. Clifton"], "authorids": ["~Dani_Kiyasseh1", "~Tingting_Zhu1", "~David_A._Clifton1"], "authors_source": "OpenReview API", "abstract": "Cardiac signals convey a significant amount of information about the health status of a patient. Upon recording these signals, cardiologists are expected to manually generate an accompanying report to share with physicians and patients. Generating these reports, however, can be time-consuming and error-prone, while also exhibiting a high degree of intra- and inter-physician variability. To address this, we design a neural, multilingual, cardiac signal captioning framework. In the process, we propose a discriminative multilingual representation learning method, RTLP, which randomly replaces tokens with those from a different language and tasks a network with identifying the language of all tokens. We show that RTLP performs on par with state-of-the-art pre-training methods such as MLM and MARGE, while generating more clinically accurate reports than MLM. We also show that, with RTLP, multilingual fine-tuning can be preferable to its monolingual counterpart, a phenomenon we refer to as the \\textit{blessing of multilinguality}.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=ZzwfldvDLpC", "arxiv_id": "2103.11011", "paper_pdf": "papers/ZzwfldvDLpC.pdf", "paper_pdf_sha256": "eabcf26638e62b32986478cf48f85ea5299213d4c7a74a3bc14133df4bbdcc72", "paper_pdf_bytes": 4227129, "paper_pdf_source": "openreview", "code_url": "https://github.com/danikiyasseh/RTLP", "code_repository": "danikiyasseh/RTLP", "code_commit": "e52f3e2488ba2d5eb6be0fa011daac002e736c16", "code_archive": "repos/ZzwfldvDLpC.zip", "code_archive_sha256": "5610078a8f05b39912a7cbbe2438d94353f900eb4f5f74920cf88e149e0f1a78", "code_archive_bytes": 1567052, "code_file_count": 15, "code_extensions": {".py": 15}, "github_disk_usage_kb": 2341, "github_languages": {"Python": 429536, "HTML": 1}, "github_archived": false, "github_pushed_at": "2022-08-25T16:43:17Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/let-your-heart-speak-in-its-mother-tongue"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "j5d9qacxdZa", "year": 2021, "status": "rejected", "title": "Energy-Based Models for Continual Learning", "authors": ["Shuang Li", "Yilun Du", "Gido Martijn van de Ven", "Antonio Torralba", "Igor Mordatch"], "authorids": ["~Shuang_Li5", "~Yilun_Du1", "~Gido_Martijn_van_de_Ven1", "~Antonio_Torralba1", "~Igor_Mordatch4"], "authors_source": "OpenReview API", "abstract": "We motivate Energy-Based Models (EBMs) as a promising model class for continual learning problems. Instead of tackling continual learning via the use of external memory, growing models, or regularization, EBMs have a natural way to support a dynamically-growing number of tasks and classes and less interference with old tasks. We show that EBMs are adaptable to a more general continual learning setting where the data distribution changes without the notion of explicitly delineated tasks. We also find that EBMs outperform the baseline methods by a large margin on several continual learning benchmarks. These observations point towards EBMs as a class of models naturally inclined towards the continual learning regime.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=j5d9qacxdZa", "arxiv_id": "2011.12216", "paper_pdf": "papers/j5d9qacxdZa.pdf", "paper_pdf_sha256": "d82cdf180c473dde5a7fcb2607cc9f5ea4dde8d4e4c3ab761c541718ae4fdea6", "paper_pdf_bytes": 4948780, "paper_pdf_source": "openreview", "code_url": "https://github.com/ShuangLI59/ebm-continual-learning", "code_repository": "ShuangLI59/ebm-continual-learning", "code_commit": "0450d69ac01625c1d227356e5367e002aaae65a4", "code_archive": "repos/j5d9qacxdZa.zip", "code_archive_sha256": "f12f2ed64e2c5a7d6e808a266336af2c0dcff2959bb426be6e70a50b426e013e", "code_archive_bytes": 1536377, "code_file_count": 41, "code_extensions": {".py": 25, ".sh": 16}, "github_disk_usage_kb": 1484, "github_languages": {"Python": 232622, "Shell": 3426}, "github_archived": false, "github_pushed_at": "2022-09-13T02:08:39Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/energy-based-models-for-continual-learning-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "FnwU7ogRzv", "year": 2026, "status": "rejected", "title": "CRAKEN: Cybersecurity LLM Agent with Knowledge-Based Execution", "authors": ["Minghao Shao", "Haoran Xi", "NANDA RANI", "Meet Udeshi", "Venkata Sai Charan Putrevu", "Kimberly Milner", "Brendan Dolan-Gavitt", "Sandeep K. Shukla", "Prashanth Krishnamurthy", "Farshad Khorrami", "Ramesh Karri", "Muhammad Shafique"], "authorids": ["~Minghao_Shao3", "~Haoran_Xi1", "~NANDA_RANI1", "~Meet_Udeshi1", "~Venkata_Sai_Charan_Putrevu1", "~Kimberly_Milner1", "~Brendan_Dolan-Gavitt1", "~Sandeep_K._Shukla1", "~Prashanth_Krishnamurthy1", "~Farshad_Khorrami1", "~Ramesh_Karri1", "~Muhammad_Shafique1"], "authors_source": "OpenReview API", "abstract": "Large Language Model (LLM) agents can automate cybersecurity tasks and can adapt to the evolving cybersecurity landscape without re-engineering. While LLM agents have demonstrated cybersecurity capabilities on Capture-The-Flag (CTF) competitions, they have two key limitations: accessing latest cybersecurity expertise beyond training data, and integrating new knowledge into complex task planning. Knowledge-based approaches that incorporate technical understanding into the task-solving automation can tackle these limitations. We present CRAKEN, a knowledge-based LLM agent framework that improves cybersecurity capability through three core mechanisms: contextual decomposition of task-critical information, iterative self-reflected knowledge retrieval, and solution injection that transforms insights into adaptive attack strategies. CRAKEN combines advanced retrieval algorithms with prompt injection-based integration. Comprehensive evaluations with different configurations show CRAKEN's effectiveness in multi-stage vulnerability detection and exploitation compared to previous approaches. Our extensible architecture establishes new methodologies for embedding new security knowledge into LLM-driven cybersecurity agentic systems. CRAKEN obtained an accuracy of 22% on NYU CTF Bench with our collected simple CTF write-up dataset and shows a 15.7% of the solution distribution difference, indicating the effectiveness of integrating knowledge into automated CTF solving for challenges that requires knowledge from specific domains.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=FnwU7ogRzv", "arxiv_id": "2505.17107", "paper_pdf": "papers/FnwU7ogRzv.pdf", "paper_pdf_sha256": "b9d83bb69e045abc6e4e42de67e57f736f97e56e54958164c761427d1dbee26d", "paper_pdf_bytes": 874852, "paper_pdf_source": "openreview", "code_url": "https://github.com/NYU-LLM-CTF/nyuctf_agents_craken", "code_repository": "NYU-LLM-CTF/nyuctf_agents_craken", "code_commit": "748bc8986bae2e02eb69dbebce07006a8b325a73", "code_archive": "repos/FnwU7ogRzv.zip", "code_archive_sha256": "e08a3112d197e51e8294a3ae805dad3198236a528c517e95087c00a4f0495604", "code_archive_bytes": 54704, "code_file_count": 19, "code_extensions": {".py": 18, ".sh": 1}, "github_disk_usage_kb": 5291, "github_languages": {"Python": 170827, "Shell": 58}, "github_archived": false, "github_pushed_at": "2025-07-13T23:23:15Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/craken-cybersecurity-llm-agent-with-knowledge"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "2bn7gayfz9", "year": 2025, "status": "rejected", "title": "CTBench: A Library and Benchmark for Certified Training", "authors": ["Yuhao Mao", "Stefan Balauca", "Martin Vechev"], "authorids": ["~Yuhao_Mao1", "~Stefan_Balauca1", "~Martin_Vechev1"], "authors_source": "OpenReview API", "abstract": "Training certifiably robust neural networks is an important but challenging task. While many algorithms for (deterministic) certified training have been proposed, they are often evaluated on different training schedules, certification methods, and systematically under-tuned hyperparameters, making it difficult to compare their performance. To address this challenge, we introduce CTBench, a unified library and a high-quality benchmark for certified training that evaluates all algorithms under fair settings and systematically tuned hyperparameters. We show that (1) almost all algorithms in CTBench surpass the corresponding reported performance in literature in the magnitude of algorithmic improvements, thus establishing new state-of-the-art, and (2) the claimed advantage of recent algorithms drops significantly when we enhance the outdated baselines with a fair training schedule, a fair certification method and well-tuned hyperparameters. Based on CTBench, we provide insights into the current state of certified training and suggest future research directions. We are confident that CTBench will serve as a benchmark and testbed for future research in certified training.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=2bn7gayfz9", "arxiv_id": "2406.04848", "paper_pdf": "papers/2bn7gayfz9.pdf", "paper_pdf_sha256": "f74b6a53ba22f84d39e82c30946768df6f2e691a42e7a2632fd250123dc0b8e7", "paper_pdf_bytes": 422592, "paper_pdf_source": "openreview", "code_url": "https://github.com/eth-sri/CTBench", "code_repository": "eth-sri/CTBench", "code_commit": "0f18162cba85c54d7fc28af62b2c5926efd3695d", "code_archive": "repos/2bn7gayfz9.zip", "code_archive_sha256": "ad1d962e99a18871f250d4d0fb4bb19106535b7c6348e0b8e1a676106214be26", "code_archive_bytes": 214501, "code_file_count": 93, "code_extensions": {".sh": 65, ".py": 28}, "github_disk_usage_kb": 335, "github_languages": {"Python": 578214, "Shell": 58890}, "github_archived": false, "github_pushed_at": "2026-04-25T17:55:36Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/ctbench-a-library-and-benchmark-for-certified"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "hJEMTDOwKx", "year": 2024, "status": "rejected", "title": "Language Models as Semantic Indexers", "authors": ["Bowen Jin", "Hansi Zeng", "Guoyin Wang", "Xiusi Chen", "Tianxin Wei", "Ruirui Li", "Zhengyang Wang", "Zheng Li", "Yang Li", "Hanqing Lu", "Suhang Wang", "Jiawei Han", "Xianfeng Tang"], "authorids": ["~Bowen_Jin1", "~Hansi_Zeng1", "~Guoyin_Wang1", "~Xiusi_Chen1", "~Tianxin_Wei1", "~Ruirui_Li3", "~Zhengyang_Wang1", "~Zheng_Li9", "~Yang_Li80", "~Hanqing_Lu3", "~Suhang_Wang1", "~Jiawei_Han1", "~Xianfeng_Tang1"], "authors_source": "OpenReview API", "abstract": "Semantic identifier (ID) is an important concept in information retrieval that aims to preserve the semantics of objects such as documents and items inside their IDs. Previous studies typically adopt a two-stage pipeline to learn semantic IDs by first procuring embeddings using off-the-shelf text encoders and then deriving IDs based on the embeddings. However, each step introduces potential information loss and there is usually an inherent mismatch between the distribution of embeddings within the latent space produced by text encoders and the anticipated distribution required for semantic indexing. Nevertheless, it is non-trivial to design a method that can learn the document’s semantic representations and its hierarchical structure simultaneously, given that semantic IDs are discrete and sequentially structured, and the semantic supervision is deficient. In this paper, we introduce LMINDEXER, a self-supervised framework to learn semantic IDs with a generative language model. We tackle the challenge of sequential discrete ID by introducing a semantic indexer capable of generating neural sequential discrete representations with progressive training and contrastive learning. In response to the semantic supervision deficiency, we propose to train the model with a self-supervised document reconstruction objective. The learned semantic indexer can facilitate various downstream tasks, such as recommendation and retrieval. We conduct experiments on three tasks including recommendation, product search, and document retrieval on five datasets from various domains, where LMINDEXER outperforms competitive baselines significantly and consistently. Code is available at https://anonymous.4open.science/r/ICLR24-submit-B2E7/.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=hJEMTDOwKx", "arxiv_id": "2310.07815", "paper_pdf": "papers/hJEMTDOwKx.pdf", "paper_pdf_sha256": "05ff3e71cb4c675b5bfbeef76d4e1b5b04bff322c831d71e98419cc129199f1a", "paper_pdf_bytes": 1254820, "paper_pdf_source": "openreview", "code_url": "https://github.com/PeterGriffinJin/LMIndexer", "code_repository": "PeterGriffinJin/LMIndexer", "code_commit": "e6dc4cc1ddf1f1615f643771c9047063b8e49528", "code_archive": "repos/hJEMTDOwKx.zip", "code_archive_sha256": "8ca1c376945ef5828c23a88ea6cc60771394f74e0a08f2932d6bc2426cc4cb63", "code_archive_bytes": 316165, "code_file_count": 97, "code_extensions": {".py": 51, ".sh": 45, ".ipynb": 1}, "github_disk_usage_kb": 267, "github_languages": {"Python": 630707, "Shell": 83923, "Jupyter Notebook": 19503}, "github_archived": false, "github_pushed_at": "2024-05-02T12:48:23Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/language-models-as-semantic-indexers"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "hUr6K4D9f7P", "year": 2022, "status": "rejected", "title": "Adversarial Weight Perturbation Improves Generalization in Graph Neural Networks", "authors": ["Yihan Wu", "Aleksandar Bojchevski", "Heng Huang"], "authorids": ["~Yihan_Wu1", "~Aleksandar_Bojchevski1", "~Heng_Huang1"], "authors_source": "OpenReview API", "abstract": "There is growing theoretical and empirical evidence that flatter local minima tend to improve generalization. An efficient and effective technique for finding such minima is Adversarial Weight Perturbation (AWP). The main idea is to minimize the loss w.r.t. a bounded worst-case perturbation of the model parameters by (approximately) solving an associated min-max problem. Intuitively, we favor local minima with a small loss in a neighborhood around them. The benefits of AWP, and more generally the connections between flatness and generalization, have been extensively studied for i.i.d. data such as images. In this paper we initiate the first study of this phenomenon for graph data. Along the way, we identify a vanishing-gradient issue with all existing formulations of AWP and we propose Weighted Truncated AWP (WT-AWP) to alleviate this issue. We show that regularizing graph neural networks with WT-AWP consistently improves both natural and robust generalization across many different graph learning tasks and models.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=hUr6K4D9f7P", "arxiv_id": "2212.04983", "paper_pdf": "papers/hUr6K4D9f7P.pdf", "paper_pdf_sha256": "c541437eda108e345a55204f7db67e35e187f3ef76edbf29fdc940ce104139f3", "paper_pdf_bytes": 1108766, "paper_pdf_source": "openreview", "code_url": "https://github.com/yihwu/WT-AWP", "code_repository": "yihwu/WT-AWP", "code_commit": "aa0dae0c64521d885fed6dce32c5a35a3a292f3e", "code_archive": "repos/hUr6K4D9f7P.zip", "code_archive_sha256": "aa802b68a1dfa8e2defeb045525eb97e555f6f6f3e6769c0a98bfec8c0f01402", "code_archive_bytes": 2581455, "code_file_count": 59, "code_extensions": {".py": 56, ".ipynb": 3}, "github_disk_usage_kb": 2349, "github_languages": {"Python": 2912518, "Jupyter Notebook": 341294}, "github_archived": false, "github_pushed_at": "2022-11-22T07:28:42Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/adversarial-weight-perturbation-improves-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "zbEupOtJFF", "year": 2021, "status": "rejected", "title": "On interaction between augmentations and corruptions in natural corruption robustness", "authors": ["Eric Mintun", "Alexander Kirillov", "Saining Xie"], "authorids": ["~Eric_Mintun1", "~Alexander_Kirillov1", "~Saining_Xie2"], "authors_source": "OpenReview API", "abstract": "Invariance to a broad array of image corruptions, such as warping, noise, or color shifts, is an important aspect of building robust models in computer vision. Recently, several new data augmentations have been proposed that significantly improve performance on ImageNet-C, a benchmark of such corruptions. However, there is still a lack of basic understanding on the relationship between data augmentations and test-time corruptions. To this end, we develop a feature space for image transforms, and then use a new measure in this space between augmentations and corruptions called the Minimal Sample Distance to demonstrate there is a strong correlation between similarity and performance. We then investigate recent data augmentations and observe a significant degradation in corruption robustness when the test-time corruptions are sampled to be perceptually dissimilar from ImageNet-C in this feature space. Our results suggest that test error can be improved by training on perceptually similar augmentations, and data augmentations may risk overfitting to the existing benchmark.  We hope our results and tools will allow for more robust progress towards improving robustness to image corruptions.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=zbEupOtJFF", "arxiv_id": "2102.11273", "paper_pdf": "papers/zbEupOtJFF.pdf", "paper_pdf_sha256": "c9a3fe7b9361c8a05078ba00cc2271f420f6adb682c789ef46803f3865ee4a6e", "paper_pdf_bytes": 6672450, "paper_pdf_source": "openreview", "code_url": "https://github.com/facebookresearch/augmentation-corruption", "code_repository": "facebookresearch/augmentation-corruption", "code_commit": "ba4d5a5e9132fe98fcd4be1ac90abedbbb188794", "code_archive": "repos/zbEupOtJFF.zip", "code_archive_sha256": "f536df67ed4b1ed909f05970784c4d9b43a83f28810bb2e7e8b032a72ca7092d", "code_archive_bytes": 1586667, "code_file_count": 74, "code_extensions": {".py": 55, ".sh": 18, ".ipynb": 1}, "github_disk_usage_kb": 1510, "github_languages": {"Python": 347384, "Shell": 48867, "Jupyter Notebook": 17469}, "github_archived": true, "github_pushed_at": "2022-11-06T23:27:04Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/on-interaction-between-augmentations-and"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Tj5rHP8yrQ", "year": 2026, "status": "rejected", "title": "CompeteSMoE - Statistically Guaranteed Mixture of Experts Training via Competition", "authors": ["Nam V. Nguyen", "Huy Nguyen", "Quang Pham", "Van Nguyen", "Savitha Ramasamy", "Nhat Ho"], "authorids": ["~Nam_V._Nguyen1", "~Huy_Nguyen5", "~Quang_Pham1", "~Van_Nguyen3", "~Savitha_Ramasamy1", "~Nhat_Ho1"], "authors_source": "OpenReview API", "abstract": "Sparse mixture of experts (SMoE) offers an appealing solution to scale up the model complexity beyond the mean of increasing the network's depth or width. However, we argue that effective SMoE training remains challenging because of the suboptimal routing process where experts that perform computation do not directly contribute to the routing process. In this work, we propose competition, a novel mechanism to route tokens to experts with the highest neural response. Theoretically, we show that the competition mechanism enjoys a better sample efficiency than the traditional softmax routing. Furthermore, we develop CompeteSMoE, a simple yet effective algorithm to train large language models by deploying a router to learn the competition policy, thus enjoying strong performances at a low training overhead. Our extensive empirical evaluations on both the visual instruction tuning and language pre-training tasks demonstrate the efficacy, robustness, and scalability of CompeteSMoE compared to state-of-the-art SMoE strategies. We will publish the implementation upon acceptance.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=Tj5rHP8yrQ", "arxiv_id": "2505.13380", "paper_pdf": "papers/Tj5rHP8yrQ.pdf", "paper_pdf_sha256": "2129ac6aa9e0482889ed7975d6b3b3c41e47f09f7c3b7b1f387fb05193c493d9", "paper_pdf_bytes": 832107, "paper_pdf_source": "openreview", "code_url": "https://github.com/Fsoft-AIC/CompeteSMoE", "code_repository": "Fsoft-AIC/CompeteSMoE", "code_commit": "ab48bb62aa7edb855328e375b9e089e9b922c9d7", "code_archive": "repos/Tj5rHP8yrQ.zip", "code_archive_sha256": "ff551a9b49cb66865030ec64c3f857b3aed36e3733cf604f13e094b3b6983101", "code_archive_bytes": 6254766, "code_file_count": 449, "code_extensions": {".py": 429, ".sh": 11, ".ipynb": 9}, "github_disk_usage_kb": 5686, "github_languages": {"Python": 3123102, "Jupyter Notebook": 151475, "Shell": 11757}, "github_archived": false, "github_pushed_at": "2025-08-23T08:02:14Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/competesmoe-statistically-guaranteed-mixture"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "GYk0thSY1M", "year": 2025, "status": "rejected", "title": "Recurrent Context Compression: Efficiently Expanding the Context Window of LLM", "authors": ["ChensenHuang", "Guibo Zhu", "Xuepeng Wang", "Dong Yi", "Yifei Luo", "Haoran Chen", "Guojing Ge", "Jinqiao Wang"], "authorids": ["~ChensenHuang1", "~Guibo_Zhu1", "~Xuepeng_Wang1", "~Dong_Yi2", "~Yifei_Luo2", "~Haoran_Chen1", "~Guojing_Ge2", "~Jinqiao_Wang1"], "authors_source": "OpenReview API", "abstract": "To extend the context length of Transformer-based large language models (LLMs) and improve comprehension capabilities, researchers often encounter constraints stemming from finite computational resources and bounded memory capacities. This work proposes a novel approach, termed Recurrent Context Compression (RCC), designed to efficiently expand the context window length of LLMs. Furthermore, we delve into the prevalent issue of degraded model performance when both instructional prompts and contextual information undergo compression for downstream tasks. To address this challenge, we propose a novel instruction reconstruction methodology aimed at mitigating the detrimental effects of this compression process. The effectiveness of our proposed approach was validated across multiple tasks while achieving an impressive context compression rate of at least 32x. On text reconstruction task, we maintain a BLEU-4 score close to 0.95. On passkey retrieval task, we achieve nearly 100% accuracy involving an extensive sequence length of 1 million tokens. On long-text question-answering task, we obtain comparable performance with the non-compressed LLM in F1 and Rouge scores. Our method also demonstrated competitive performance in long-text question-answering tasks compared to non-compressed methods, while significantly saving storage resources.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=GYk0thSY1M", "arxiv_id": "2406.06110", "paper_pdf": "papers/GYk0thSY1M.pdf", "paper_pdf_sha256": "b628086ad454173cf7779ed38cd7383a4eae84e8bc94934859d635547f9439d1", "paper_pdf_bytes": 587121, "paper_pdf_source": "openreview", "code_url": "https://github.com/WUHU-G/RCC_Transformer", "code_repository": "WUHU-G/RCC_Transformer", "code_commit": "b9c5486c49708cd284dbf1229fe6bc054d424453", "code_archive": "repos/GYk0thSY1M.zip", "code_archive_sha256": "d95d1c0e10e74490ebfe2dfb319de02fcf96746a3c7a4cc88b37dbbedff2a15e", "code_archive_bytes": 274055, "code_file_count": 3, "code_extensions": {".py": 3}, "github_disk_usage_kb": 339, "github_languages": {"Python": 40685}, "github_archived": false, "github_pushed_at": "2024-06-13T02:27:03Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/recurrent-context-compression-efficiently"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "U2ZIgcrg7Z", "year": 2024, "status": "rejected", "title": "ZOOPFL: EXPLORING BLACK-BOX FOUNDATION MODELS FOR PERSONALIZED FEDERATED LEARNING", "authors": ["Wang Lu", "Hao Yu", "Jindong Wang", "Damien Teney", "Haohan Wang", "Yiqiang Chen", "Qiang Yang", "Xing Xie", "Xiangyang Ji"], "authorids": ["~Wang_Lu2", "~Hao_Yu8", "~Jindong_Wang1", "~Damien_Teney1", "~Haohan_Wang1", "~Yiqiang_Chen1", "~Qiang_Yang1", "~Xing_Xie3", "~Xiangyang_Ji1"], "authors_source": "OpenReview API", "abstract": "When personalized federated learning (FL) meets large foundation models, new challenges arise from various limitations in resources. In addition to typical limitations such as data, computation, and communication costs, access to the models is also often limited. This paper endeavors to solve both the challenges of limited resources and personalization. i.e., distribution shifts between clients. To do so, we propose a method named ZOOPFL that uses Zeroth-Order Optimization for Personalized Federated Learning. ZOOPFL avoids direct interference with the foundation models and instead learns to adapt its inputs through zeroth-order optimization. In addition, we employ simple yet effective linear projections to remap its predictions for personalization. To reduce the computation costs and enhance personalization, we propose input surgery to incorporate an auto-encoder with low-dimensional and client-specific embeddings. We provide theoretical support for ZOOPFL to analyze its convergence. Extensive empirical experiments on computer vision and natural language processing tasks using popular foundation models demonstrate its effectiveness for FL on black-box foundation models.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=U2ZIgcrg7Z", "arxiv_id": "2310.05143", "paper_pdf": "papers/U2ZIgcrg7Z.pdf", "paper_pdf_sha256": "a51156d337e77165e8ec1552603dd54bedafe5d76c58c3bb6895b3feb3cb4225", "paper_pdf_bytes": 2140291, "paper_pdf_source": "openreview", "code_url": "https://github.com/microsoft/PersonalizedFL", "code_repository": "microsoft/PersonalizedFL", "code_commit": "441670edb5744d23af46867cce9a5fbec0f8ead3", "code_archive": "repos/U2ZIgcrg7Z.zip", "code_archive_sha256": "521bb8e68c430baf3e428468640c02aba5f11ce74194fbed0c9acd9ae583d203", "code_archive_bytes": 264542, "code_file_count": 22, "code_extensions": {".py": 21, ".sh": 1}, "github_disk_usage_kb": 279, "github_languages": {"Python": 75730, "Shell": 1210}, "github_archived": false, "github_pushed_at": "2023-10-04T01:05:44Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/zoopfl-exploring-black-box-foundation-models"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Gb2Rndy5595", "year": 2023, "status": "rejected", "title": "Context Autoencoder for Self-Supervised Representation Learning", "authors": ["Xiaokang Chen", "Mingyu Ding", "Xiaodi Wang", "Ying Xin", "Shentong Mo", "Yunhao Wang", "Shumin Han", "Ping Luo", "Gang Zeng", "Jingdong Wang"], "authorids": ["~Xiaokang_Chen1", "~Mingyu_Ding1", "~Xiaodi_Wang2", "~Ying_Xin1", "~Shentong_Mo1", "~Yunhao_Wang1", "~Shumin_Han1", "~Ping_Luo2", "~Gang_Zeng1", "~Jingdong_Wang1"], "authors_source": "OpenReview API", "abstract": "We present a novel masked image modeling (MIM) approach, context autoencoder (CAE), for self-supervised representation pretraining. The goal is to pretrain an encoder by solving the pretext task: estimate the masked patches from the visible patches in an image. Our approach first feeds the visible patches into the encoder, extracting the representations. Then, we make predictions from visible patches to masked patches in the encoded representation space. We introduce an alignment constraint, encouraging that the representations for masked patches, predicted from the encoded representations of visible patches, are aligned with the masked patch presentations computed from the encoder. In other words, the predicted representations are expected to lie in the encoded representation space, which empirically shows the benefit to representation learning. Last, the predicted masked patch representations are mapped to the targets of the pretext task through a decoder.\nOne additional characteristic is that our approach encourages the separation of the representation learning part (encoder), and the pretext task completion part that will be replaced by the downstream task part. In contrast, previous MIM methods (e.g., BEiT and MAE) couple the two parts, potentially limiting the representation learning quality. We demonstrate the effectiveness of our CAE through superior transfer performance in downstream tasks: semantic segmentation, and object detection and instance segmentation.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=Gb2Rndy5595", "arxiv_id": "2202.03026", "paper_pdf": "papers/Gb2Rndy5595.pdf", "paper_pdf_sha256": "4a8595a46e8939fb558f5ea4172977215b515901ea72201e30d8c4c1dbd6798a", "paper_pdf_bytes": 12318393, "paper_pdf_source": "openreview", "code_url": "https://github.com/Atten4Vis/CAE", "code_repository": "Atten4Vis/CAE", "code_commit": "a7fd1628176358e3e76b0b042b0f6e9f3cd7c76c", "code_archive": "repos/Gb2Rndy5595.zip", "code_archive_sha256": "a6976061e708b93dc0342f85a13ad873737c513d26ee862459164dd3fb9f7a79", "code_archive_bytes": 756589, "code_file_count": 236, "code_extensions": {".py": 223, ".sh": 13}, "github_disk_usage_kb": 628, "github_languages": {"Python": 1231842, "Shell": 23917}, "github_archived": false, "github_pushed_at": "2023-11-28T06:45:38Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/context-autoencoder-for-self-supervised"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "UFYYol-bRq", "year": 2022, "status": "rejected", "title": "ANCER: Anisotropic Certification  via Sample-wise Volume Maximization", "authors": ["Francisco Eiras", "Motasem Alfarra", "Philip Torr", "M. Pawan Kumar", "Puneet K. Dokania", "Bernard Ghanem", "Adel Bibi"], "authorids": ["~Francisco_Eiras1", "~Motasem_Alfarra1", "~Philip_Torr1", "~M._Pawan_Kumar1", "~Puneet_K._Dokania1", "~Bernard_Ghanem1", "~Adel_Bibi1"], "authors_source": "OpenReview API", "abstract": "Randomized smoothing has recently emerged as an effective tool that enables certification of deep neural network classifiers at scale. All prior art on randomized smoothing has focused on isotropic $\\ell_p$ certification, which has the advantage of yielding certificates that can be easily compared among isotropic methods via $\\ell_p$-norm radius. However, isotropic certification limits the region that can be certified around an input to worst-case adversaries, i.e., it cannot reason about other \"close\", potentially large, constant prediction safe regions. To alleviate this issue, (i) we theoretically extend the isotropic randomized smoothing $\\ell_1$ and $\\ell_2$ certificates to their generalized anisotropic counterparts following a simplified analysis. Moreover, (ii) we propose evaluation metrics allowing for the comparison of general certificates - a certificate is superior to another if it certifies a superset region - with the quantification of each certificate through the volume of the certified region. We introduce ANCER, a framework for obtaining anisotropic certificates for a given test set sample via volume maximization. We achieve it by generalizing memory-based certification of data-dependent classifiers. Our empirical results demonstrate that ANCER achieves state-of-the-art $\\ell_1$ and $\\ell_2$ certified accuracy on CIFAR-10 and ImageNet, while certifying larger regions in terms of volume, highlighting the benefits of moving away from isotropic analysis.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=UFYYol-bRq", "arxiv_id": "2107.04570", "paper_pdf": "papers/UFYYol-bRq.pdf", "paper_pdf_sha256": "be7e9b329ebbd8c439ef1f81e3e4f0137de709fa6817e36a2a855dc472af25d4", "paper_pdf_bytes": 5273045, "paper_pdf_source": "openreview", "code_url": "https://github.com/MotasemAlfarra/ANCER", "code_repository": "MotasemAlfarra/ANCER", "code_commit": "98f869e924c482ac246c2245ae3b247e08c7e04b", "code_archive": "repos/UFYYol-bRq.zip", "code_archive_sha256": "5c6a45368fd08960b3750c07b16eccb551b730fda282cf253222b049e909a1fb", "code_archive_bytes": 270942, "code_file_count": 8, "code_extensions": {".py": 8}, "github_disk_usage_kb": 2392, "github_languages": {"Python": 32651}, "github_archived": false, "github_pushed_at": "2022-09-09T08:45:29Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/ancer-anisotropic-certification-via-sample"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Q1aiM7sCi1", "year": 2021, "status": "rejected", "title": "Fuzzy c-Means Clustering for Persistence Diagrams", "authors": ["Thomas Davies", "Jack Aspinall", "Bryan Wilder", "Long Tran-Thanh"], "authorids": ["~Thomas_Davies1", "jack.aspinall@materials.ox.ac.uk", "~Bryan_Wilder1", "long.tran-thanh@warwick.ac.uk"], "authors_source": "OpenReview API", "abstract": "Persistence diagrams concisely represent the topology of a point cloud whilst having strong theoretical guarantees. Most current approaches to integrating topological information into machine learning implicitly map persistence diagrams to a Hilbert space, resulting in deformation of the underlying metric structure whilst also generally requiring prior knowledge about the true topology of the space. In this paper we give an algorithm for Fuzzy c-Means (FCM) clustering directly on the space of persistence diagrams, enabling unsupervised learning that automatically captures the topological structure of data, with no prior knowledge or additional processing of persistence diagrams. We prove the same convergence guarantees as traditional FCM clustering: every convergent subsequence of iterates tends to a local minimum or saddle point. We end by presenting experiments where our fuzzy topological clustering algorithm allows for unsupervised top-$k$ candidate selection in settings where (i) the properties of persistence diagrams make them the natural choice over geometric equivalents, and (ii) the probabilistic membership values let us rank candidates in settings where verifying candidate suitability is expensive: lattice structure classification in materials science and pre-trained model selection in machine learning.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=Q1aiM7sCi1", "arxiv_id": "2006.02796", "paper_pdf": "papers/Q1aiM7sCi1.pdf", "paper_pdf_sha256": "c91cae8d7bf52f22210719cd3a4a51a0415206c53211909163e45557211892b3", "paper_pdf_bytes": 1655816, "paper_pdf_source": "openreview", "code_url": "https://github.com/tomogwen/fpdcluster", "code_repository": "tomogwen/fpdcluster", "code_commit": "8bcd7c01e6120893a545aaaac1e73d793b6210f0", "code_archive": "repos/Q1aiM7sCi1.zip", "code_archive_sha256": "28e21f59c6dc8f3325aab16023e01ec5da4744b6049cffdda912cb48f27e2b18", "code_archive_bytes": 1023682, "code_file_count": 5, "code_extensions": {".py": 5}, "github_disk_usage_kb": 1534, "github_languages": {"D2": 311660, "Python": 27322}, "github_archived": false, "github_pushed_at": "2023-10-14T12:00:26Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/fuzzy-c-means-clustering-for-persistence"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "mAutPdnHIN", "year": 2026, "status": "rejected", "title": "ASyMOB: Algebraic Symbolic Mathematical Operations Benchmark", "authors": ["Michael Shalyt", "Rotem Elimelech", "Ido Kaminer"], "authorids": ["~Michael_Shalyt1", "~Rotem_Elimelech1", "~Ido_Kaminer1"], "authors_source": "OpenReview API", "abstract": "Large language models (LLMs) are increasingly applied to symbolic mathematics, yet existing evaluations often conflate pattern memorization with genuine reasoning. To address this gap, we present **ASyMOB**, a high-resolution dataset of **35,368** validated symbolic math problems spanning integration, limits, differential equations, series, and hypergeometrics. Unlike prior benchmarks, **ASyMOB** systematically perturbs each seed problem using symbolic, numeric, and equivalence-preserving transformations, enabling a fine-grained assessment of generalization and robustness.\nOur evaluation reveals three key findings: (1) most models’ performance collapses under minor perturbations, while frontier systems exhibit substantial robustness, suggesting an emerging *\"phase transition\"* from memorization to generalization; (2) integrated code tools stabilize performance, particularly for weaker models; and (3) we identify examples where Computer Algebra Systems (CAS) fail while LLMs succeed, as well as problems solved only via a hybrid LLM-CAS approach, highlighting a promising integration frontier.\n**ASyMOB** serves as a principled diagnostic tool for measuring and accelerating progress toward building verifiable, trustworthy AI for scientific discovery.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=mAutPdnHIN", "arxiv_id": "2505.23851", "paper_pdf": "papers/mAutPdnHIN.pdf", "paper_pdf_sha256": "7b02a17294644008602dd70a7d99e4fbc8aeebd30ec0bc30a6f01d455e56915b", "paper_pdf_bytes": 899298, "paper_pdf_source": "openreview", "code_url": "https://github.com/RamanujanMachine/ASyMOB", "code_repository": "RamanujanMachine/ASyMOB", "code_commit": "172712d0a12eed39a47cd1df82f690121c57520d", "code_archive": "repos/mAutPdnHIN.zip", "code_archive_sha256": "166001cc68ff36168a2231af4f327a372b7af0235df1a7b8f571c1093f5e3337", "code_archive_bytes": 1439949, "code_file_count": 17, "code_extensions": {".py": 16, ".sql": 1}, "github_disk_usage_kb": 5714, "github_languages": {"Python": 96250}, "github_archived": false, "github_pushed_at": "2026-06-08T23:49:01Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/asymob-algebraic-symbolic-mathematical"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "0fwJMANq9P", "year": 2025, "status": "rejected", "title": "Efficient Heuristics Generation for Solving Combinatorial Optimization Problems Using Large Language Models", "authors": ["Xuan Wu", "Di Wang", "Zhiguang Cao", "Chunguo Wu", "Lijie Wen", "Chunyan Miao", "Yubin Xiao", "You Zhou"], "authorids": ["~Xuan_Wu7", "~Di_Wang26", "~Zhiguang_Cao1", "~Chunguo_Wu1", "~Lijie_Wen1", "~Chunyan_Miao1", "~Yubin_Xiao1", "~You_Zhou5"], "authors_source": "OpenReview API", "abstract": "Recent studies exploited Large Language Models (LLMs) to autonomously generate heuristics for solving Combinatorial Optimization Problems (COPs), by prompting LLMs to first provide search directions and then derive heuristics accordingly. However, the absence of task-specific knowledge in prompts often leads LLMs to provide unspecific search directions, obstructing the derivation of well-performing heuristics. Moreover, evaluating the derived heuristics remains resource-intensive, especially for those semantically equivalent ones, often requiring unnecessary resource expenditure. To enable LLMs to provide specific search directions, we propose the Hercules algorithm, which leverages our designed Core Abstraction Prompting (CAP) method to abstract the core components from elite heuristics and incorporate them as prior knowledge in prompts. We theoretically prove the effectiveness of CAP in reducing unspecificity and provide empirical results in this work. To reduce the required computing resources for evaluating the derived heuristics, we propose few-shot Performance Prediction Prompting (PPP), a first-of-its-kind method for the Heuristic Generation (HG) task. PPP leverages LLMs to predict the fitness values of newly derived heuristics by analyzing their semantic similarity to previously evaluated ones. We further develop two tailored mechanisms for PPP to enhance predictive accuracy and determine unreliable predictions, respectively. The use of PPP makes Hercules more resource-efficient and we name this variant Hercules-P. Extensive experiments across various HG tasks, COPs, and LLMs demonstrate that Hercules outperforms the state-of-the-art LLM-based HG algorithms, while Hercules-P excels at minimizing computing resources. In addition, we illustrate the effectiveness of CAP, PPP, and the other proposed mechanisms by conducting relevant ablation studies.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=0fwJMANq9P", "arxiv_id": "2505.12627", "paper_pdf": "papers/0fwJMANq9P.pdf", "paper_pdf_sha256": "3e4d0c5fb444d440dde3c0e1e1c341f5225801123bc032eefed33ffe68a628bf", "paper_pdf_bytes": 668719, "paper_pdf_source": "openreview", "code_url": "https://github.com/wuuu110/Hercules", "code_repository": "wuuu110/Hercules", "code_commit": "bc532df7325373c564e1926916c8be6bbb645216", "code_archive": "repos/0fwJMANq9P.zip", "code_archive_sha256": "5bf3ec189239dda797d2257ef7c958deff95c6afd4e9dab7d9e365332db6b612", "code_archive_bytes": 200195, "code_file_count": 51, "code_extensions": {".py": 51}, "github_disk_usage_kb": 340, "github_languages": {"Python": 361661}, "github_archived": false, "github_pushed_at": "2025-05-20T05:59:36Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/efficient-heuristics-generation-for-solving"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "w8eCnnq57m", "year": 2024, "status": "rejected", "title": "LoraHub: Efficient Cross-Task Generalization via Dynamic LoRA Composition", "authors": ["Chengsong Huang", "Qian Liu", "Bill Yuchen Lin", "Chao Du", "Tianyu Pang", "Min Lin"], "authorids": ["~Chengsong_Huang1", "~Qian_Liu2", "~Bill_Yuchen_Lin1", "~Chao_Du1", "~Tianyu_Pang1", "~Min_Lin1"], "authors_source": "OpenReview API", "abstract": "Low-rank adaptations (LoRA) are often employed to fine-tune large language models (LLMs) for new tasks. This paper investigates LoRA composability for cross-task generalization and introduces LoraHub, a simple framework devised for the purposive assembly of LoRA modules trained on diverse given tasks, with the objective of achieving adaptable performance on unseen tasks. With just a few examples from a new task, LoraHub can fluidly combine multiple LoRA modules, eliminating the need for human expertise and assumptions.  Notably, the composition requires neither additional model parameters nor gradients.  Our empirical results, derived from the Big-Bench Hard benchmark, suggest that LoraHub can effectively mimic the performance of in-context learning in few-shot scenarios, excluding the necessity of in-context examples alongside each inference input.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=w8eCnnq57m", "arxiv_id": "2307.13269", "paper_pdf": "papers/w8eCnnq57m.pdf", "paper_pdf_sha256": "4ce8c9e6faf96627c129463d6db335743eb6ab6445e202e583d86b32c45a6a26", "paper_pdf_bytes": 557225, "paper_pdf_source": "openreview", "code_url": "https://github.com/sail-sg/lorahub", "code_repository": "sail-sg/lorahub", "code_commit": "df73afe5f38d9ff0fd1cd43774be51c79c581cc3", "code_archive": "repos/w8eCnnq57m.zip", "code_archive_sha256": "c4d5490c774ebc9ea13f7ca51137a75fa12b71836a3d7306b640aea101ba58b3", "code_archive_bytes": 268071, "code_file_count": 9, "code_extensions": {".py": 8, ".sh": 1}, "github_disk_usage_kb": 281, "github_languages": {"Python": 110046, "Shell": 1501}, "github_archived": false, "github_pushed_at": "2024-07-22T00:38:28Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/lorahub-efficient-cross-task-generalization"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "fUX3bszZSOw", "year": 2023, "status": "rejected", "title": "Do You Remember? Overcoming Catastrophic Forgetting for Fake Audio Detection", "authors": ["XiaoHui Zhang", "Jiangyan Yi", "Chenglong Wang", "Chu Yuan Zhang", "Jianhua Tao"], "authorids": ["~XiaoHui_Zhang4", "~Jiangyan_Yi1", "~Chenglong_Wang5", "~Chu_Yuan_Zhang1", "~Jianhua_Tao2"], "authors_source": "OpenReview API", "abstract": "Current fake audio detection algorithms achieve promising performances on most datasets. However, their performance may be significantly degraded when dealing with audio of a different dataset. The orthogonal weight modification to overcome catastrophic forgetting does not consider the similarity of some audio, including fake audio obtained by the same algorithm and genuine audio, on different datasets. To overcome this limitation, we propose a continual learning algorithm for fake audio detection to overcome catastrophic forgetting, called Regularized Adaptive Weight Modification (RAWM). Specifically, when fine-tuning a detection network, our approach adaptively computes the direction of weight modification according to the ratio of genuine utterances and fake utterances. The adaptive modification direction ensures the network can detect fake audio on the new dataset while preserving its knowledge of previous model, thus mitigating catastrophic forgetting. In addition, orthogonal weight modification of fake audios in the new dataset will skew the distribution of inferences on audio in the previous dataset with similar acoustic characteristics, so we introduce a regularization constraint to force the network to remember this distribution. We evaluate our approach across multiple datasets and obtain a significant performance improvement on cross-dataset experiments.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=fUX3bszZSOw", "arxiv_id": "2308.03300", "paper_pdf": "papers/fUX3bszZSOw.pdf", "paper_pdf_sha256": "4e3e72b1c89b0085d90f2aae4cf92eacd32aee980d6e69703c3961fd01e091b5", "paper_pdf_bytes": 1633379, "paper_pdf_source": "openreview", "code_url": "https://github.com/Cecile-hi/Regularized-Adaptive-Weight-Modification", "code_repository": "Cecile-hi/Regularized-Adaptive-Weight-Modification", "code_commit": "bb663ee083d8d9f8309701f2255b87916eff55c6", "code_archive": "repos/fUX3bszZSOw.zip", "code_archive_sha256": "06be15f23719a8053871703285a433ab3dbed68ee8d4c151d00b8a840d1ff4a3", "code_archive_bytes": 714128, "code_file_count": 156, "code_extensions": {".py": 154, ".sh": 2}, "github_disk_usage_kb": 647, "github_languages": {"Python": 1307444, "Shell": 363}, "github_archived": false, "github_pushed_at": "2024-09-26T08:14:30Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/do-you-remember-overcoming-catastrophic"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Yn4CPz_LRKO", "year": 2022, "status": "rejected", "title": "Conditional GANs with Auxiliary Discriminative Classifier", "authors": ["Liang Hou", "Qi Cao", "Huawei Shen", "Xueqi Cheng"], "authorids": ["~Liang_Hou1", "~Qi_Cao1", "~Huawei_Shen1", "~Xueqi_Cheng1"], "authors_source": "OpenReview API", "abstract": "Conditional generative models aim to learn the underlying joint distribution of data and labels, and thus realize conditional generation. Among them, auxiliary classifier generative adversarial networks (AC-GAN) have been widely used, but suffer from the problem of low intra-class diversity on generated samples. In this paper, we point out that the fundamental reason is that the classifier of AC-GAN is generator-agnostic, and therefore cannot provide informative guidance to the generator to approximate the target distribution, resulting in minimization of conditional entropy that decreases the intra-class diversity. Motivated by this observation, we propose a novel conditional GAN with auxiliary \\textit{discriminative} classifier (ADC-GAN) to resolve the problem of AC-GAN. Specifically, the proposed auxiliary \\textit{discriminative} classifier becomes generator-aware by recognizing the labels of the real data and the generated data \\textit{discriminatively}. Our theoretical analysis reveals that the generator can faithfully replicate the target distribution even without the original discriminator, making the proposed ADC-GAN robust to the hyper-parameter and stable during the training process. Extensive experimental results on synthetic and real-world datasets demonstrate the superiority of ADC-GAN on conditional generative modeling compared to competing methods.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=Yn4CPz_LRKO", "arxiv_id": "2107.10060", "paper_pdf": "papers/Yn4CPz_LRKO.pdf", "paper_pdf_sha256": "6082ae521f88fe5a135412119572edf86d9fa40aad32849c5cc27253d2f8db6c", "paper_pdf_bytes": 1694234, "paper_pdf_source": "openreview", "code_url": "https://github.com/liang-hou/adcgan", "code_repository": "liang-hou/adcgan", "code_commit": "29ecfa74dff78286e5966035a2ec8c6e96e6b4e1", "code_archive": "repos/Yn4CPz_LRKO.zip", "code_archive_sha256": "e55345ca538a7bd1d6e712da66a5db6798ff9092116aaafdc7a96f020a368203", "code_archive_bytes": 2634450, "code_file_count": 48, "code_extensions": {".sh": 25, ".py": 23}, "github_disk_usage_kb": 2613, "github_languages": {"Python": 276530, "Shell": 14184}, "github_archived": false, "github_pushed_at": "2023-06-11T17:18:28Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/cgans-with-auxiliary-discriminative"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "VbCVU10R7K", "year": 2021, "status": "rejected", "title": "Offline policy selection under Uncertainty", "authors": ["Mengjiao Yang", "Bo Dai", "Ofir Nachum", "George Tucker", "Dale Schuurmans"], "authorids": ["~Mengjiao_Yang1", "~Bo_Dai1", "~Ofir_Nachum1", "~George_Tucker1", "~Dale_Schuurmans1"], "authors_source": "OpenReview API", "abstract": "The presence of uncertainty in policy evaluation  significantly complicates the process of policy ranking and selection in real-world settings. We formally consider offline policy selection as learning preferences over a set of policy prospects given a fixed experience dataset. While one can select or rank policies based on point estimates of their policy values or high-confidence intervals, access to the full distribution over one's belief of the policy value enables more flexible selection algorithms under a wider range of downstream evaluation metrics. We propose BayesDICE for estimating this belief distribution in terms of posteriors of distribution correction ratios derived from stochastic constraints (as opposed to explicit likelihood, which is not available). Empirically, BayesDICE is highly competitive to existing state-of-the-art approaches in confidence interval estimation. More importantly, we show how the belief distribution estimated by BayesDICE may be used to rank policies with respect to any arbitrary downstream policy selection metric, and we empirically demonstrate that this selection procedure significantly outperforms existing approaches, such as ranking policies according to mean or high-confidence lower bound value estimates.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=VbCVU10R7K", "arxiv_id": "2012.06919", "paper_pdf": "papers/VbCVU10R7K.pdf", "paper_pdf_sha256": "9e5ff9eba10b12dc490ce3399498f90b99c00ec05c8a07b5088c432ba98a1cfd", "paper_pdf_bytes": 5387397, "paper_pdf_source": "openreview", "code_url": "https://github.com/google-research/dice_rl", "code_repository": "google-research/dice_rl", "code_commit": "5ee67f7f2145d295bdc4759b62e5c6193af31acf", "code_archive": "repos/VbCVU10R7K.zip", "code_archive_sha256": "cc05b5fa6d97faaf799830bbfde82cd9de5afdaa9d595e8ff33781f76b60acb5", "code_archive_bytes": 1415580, "code_file_count": 98, "code_extensions": {".py": 96, ".sh": 2}, "github_disk_usage_kb": 1569, "github_languages": {"Python": 582029, "Shell": 1366}, "github_archived": false, "github_pushed_at": "2026-07-30T00:11:12Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/offline-policy-selection-under-uncertainty-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "KuhCUX2oIt", "year": 2026, "status": "rejected", "title": "LEAD: Large Foundation Model for EEG-Based Alzheimer’s Disease Detection", "authors": ["Yihe Wang", "Nan Huang", "Nadia Mammone", "Marco Cecchi", "Xiang Zhang"], "authorids": ["~Yihe_Wang2", "~Nan_Huang2", "~Nadia_Mammone1", "~Marco_Cecchi1", "~Xiang_Zhang10"], "authors_source": "OpenReview API", "abstract": "Electroencephalography (EEG) provides a non-invasive, highly accessible, and cost-effective approach for detecting Alzheimer’s disease (AD). However, existing methods, whether based on handcrafted feature engineering or standard deep learning, face two major challenges: 1) the lack of large-scale EEG-AD datasets for robust representation learning, and 2) the absence of a dedicated deep learning pipeline for subject-level detection, which is more clinically meaningful than the commonly used sample-level detection. To address these gaps, we have curated the world's largest EEG-AD corpus to date, comprising 2,255 subjects. Leveraging this unique data corpus, we propose LEAD, the first large-scale foundation model for EEG analysis in dementia. Our approach provides an innovative framework for subject-level AD detection, including: 1) a comprehensive preprocessing pipeline such as artifact removal, resampling, and filtering, and a newly proposed multi-scale segmentation strategy, 2) a subject-regularized spatio–temporal transformer trained with a novel subject-level cross-entropy loss and an indices group-shuffling algorithm, and 3) AD-guided contrastive pre-training. We pre-train on 12 datasets (3 AD-related and 9 non-AD) and fine-tune/test on 4 AD datasets. Compared with 10 baselines, LEAD consistently obtains superior subject-level detection performance under the challenging subject-independent cross-validation protocol. On the benchmark ADFTD dataset, our model achieves an impressive subject-level Sensitivity of 90.91\\% under the leave-one-subject-out (LOSO) setting. These results strongly validate the effectiveness of our method for real-world EEG-based AD detection. Source code: \\url{https://anonymous.4open.science/r/LEAD-3B51}", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=KuhCUX2oIt", "arxiv_id": "2502.01678", "paper_pdf": "papers/KuhCUX2oIt.pdf", "paper_pdf_sha256": "28d845a6a3c2bf6aaad13d787b5a3f87b6d8778abed3b42124a598a427286cdf", "paper_pdf_bytes": 2148190, "paper_pdf_source": "openreview", "code_url": "https://github.com/DL4mHealth/LEAD", "code_repository": "DL4mHealth/LEAD", "code_commit": "ec35aadb1bc068fcacdcd8d036964514ca2a708f", "code_archive": "repos/KuhCUX2oIt.zip", "code_archive_sha256": "1b5c8b68a7b4f426f3e86a2fe6af37791322193010525516be45fd384bdf1a08", "code_archive_bytes": 3205704, "code_file_count": 114, "code_extensions": {".py": 68, ".sh": 26, ".ipynb": 20}, "github_disk_usage_kb": 5765, "github_languages": {"Jupyter Notebook": 22251065, "Python": 451390, "Shell": 52879}, "github_archived": false, "github_pushed_at": "2026-04-01T07:40:39Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/lead-large-foundation-model-for-eeg-based"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "3Wuvqc4xoy", "year": 2025, "status": "rejected", "title": "Learning Efficient Representations of Neutrino Telescope Events", "authors": ["Felix J. Yu", "Nicholas Kamp", "Carlos A. Argüelles"], "authorids": ["~Felix_J._Yu1", "~Nicholas_Kamp1", "~Carlos_A._Argüelles1"], "authors_source": "OpenReview API", "abstract": "Neutrino telescopes detect rare interactions of particles produced in some of the most extreme environments in the Universe. This is accomplished by instrumenting a cubic-kilometer volume of naturally occurring transparent medium with light sensors. Given their substantial size and the high frequency of background interactions, these telescopes amass an enormous quantity of large variance, high-dimensional data. These attributes create substantial challenges for analyzing and reconstructing interactions, particularly when utilizing machine learning (ML) techniques. In this paper, we present a novel approach, called om2vec, that employs transformer-based variational autoencoders to efficiently represent neutrino telescope events by learning compact and descriptive latent representations. We demonstrate that these latent representations offer enhanced flexibility and improved computational efficiency, thereby facilitating downstream tasks in data analysis.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=3Wuvqc4xoy", "arxiv_id": "2410.13148", "paper_pdf": "papers/3Wuvqc4xoy.pdf", "paper_pdf_sha256": "b4a75aab91c7fe43593e28061037dd529d35043e8951855b5cdac92af7dfb66e", "paper_pdf_bytes": 3815067, "paper_pdf_source": "openreview", "code_url": "https://github.com/felixyu7/om2vec", "code_repository": "felixyu7/om2vec", "code_commit": "60394e1dc5293126e32e4dab8510eca1f4fcbd12", "code_archive": "repos/3Wuvqc4xoy.zip", "code_archive_sha256": "0f36d3f6efe9befbc2d6b29ea75b5b6e6eeb9c482c1125084d1c1d8da82316a8", "code_archive_bytes": 14083, "code_file_count": 7, "code_extensions": {".py": 7}, "github_disk_usage_kb": 344, "github_languages": {"Python": 31260}, "github_archived": false, "github_pushed_at": "2025-07-07T19:26:05Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/learning-efficient-representations-of"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "UTGv8CayNt", "year": 2024, "status": "rejected", "title": "Chain-of-Thought Predictive Control", "authors": ["Zhiwei Jia", "Vineet Thumuluri", "Fangchen Liu", "Linghao Chen", "Zhiao Huang", "Hao Su"], "authorids": ["~Zhiwei_Jia1", "~Vineet_Thumuluri1", "~Fangchen_Liu2", "~Linghao_Chen2", "~Zhiao_Huang1", "~Hao_Su1"], "authors_source": "OpenReview API", "abstract": "We study generalizable policy learning from demonstrations for complex low-level control tasks (e.g., contact-rich object manipulations). We propose a novel hierarchical imitation learning method that utilizes scalable, albeit sub-optimal, demonstrations. Firstly, we propose an observation space-agnostic approach that efficiently discovers the multi-step subgoal decomposition (sequences of key observations) of the demos in an unsupervised manner. By grouping temporarily close and functionally similar actions into subskill-level segments, the discovered breakpoints (the segment boundaries) constitute a chain of planning steps (i.e., the chain-of-thought) to complete the task. Next, we propose a Transformer-based design that effectively learns to predict the chain-of-thought (CoT) as the high-level guidance for low-level action. We couple action and CoT predictions via prompt tokens and a hybrid masking strategy, which enable dynamically updated CoT guidance at test time and improve feature representation of the trajectory for generalizable policy learning. Our method, named Chain-of-Thought Predictive Control (CoTPC), consistently surpasses existing strong baselines on a wide range of challenging low-level manipulation tasks with scalable yet sub-optimal demos.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=UTGv8CayNt", "arxiv_id": "2304.00776", "paper_pdf": "papers/UTGv8CayNt.pdf", "paper_pdf_sha256": "a8356f018ce2525dd1ac4f5d8becf5410d6861b0a751cfd9439fe05d08a85fef", "paper_pdf_bytes": 2263379, "paper_pdf_source": "openreview", "code_url": "https://github.com/SeanJia/CoTPC", "code_repository": "SeanJia/CoTPC", "code_commit": "1c971be6cd5bdbfd59e0ad5607be0d5c149b9b5b", "code_archive": "repos/UTGv8CayNt.zip", "code_archive_sha256": "fb46e2bfca37887a5954b748d773d02cce036b29201d5fe3f048218dbf28f824", "code_archive_bytes": 215039, "code_file_count": 13, "code_extensions": {".py": 10, ".sh": 3}, "github_disk_usage_kb": 297, "github_languages": {"Python": 67575, "Shell": 1818}, "github_archived": false, "github_pushed_at": "2023-05-01T00:43:24Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/chain-of-thought-predictive-control"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "alaQzRbCY9w", "year": 2022, "status": "rejected", "title": "Bolstering Stochastic Gradient Descent with Model Building", "authors": ["Ilker Birbil", "Özgür Martin", "Gönenc Onay", "Figen Öztoprak"], "authorids": ["~Ilker_Birbil1", "~Özgür_Martin1", "~Gönenc_Onay1", "~Figen_Öztoprak1"], "authors_source": "OpenReview API", "abstract": "Stochastic gradient descent method and its variants constitute the core optimization algorithms that achieve good convergence rates for solving machine learning problems. These rates are obtained especially when these algorithms are fine-tuned for the application at hand. Although this tuning process can require large computational costs, recent work has shown that these costs can be reduced by line search methods that iteratively adjust the stepsize. We propose an alternative approach to stochastic line search by using a new algorithm based on forward step model building. This model building step incorporates a second-order information that allows adjusting not only the stepsize but also the search direction. Noting that deep learning model parameters come in groups (layers of tensors), our method builds its model and calculates a new step for each parameter group.  This novel diagonalization approach makes the selected step lengths adaptive. We provide convergence rate analysis, and experimentally show that the proposed algorithm achieves faster convergence and better generalization in most problems. Moreover, our experiments show that the proposed method is quite robust as it converges for a wide range of initial stepsizes.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=alaQzRbCY9w", "arxiv_id": "2111.07058", "paper_pdf": "papers/alaQzRbCY9w.pdf", "paper_pdf_sha256": "c8640f5e63b98f80e48a730ebf87250d83fe93a6f5750bca93e54fd843963406", "paper_pdf_bytes": 1577973, "paper_pdf_source": "openreview", "code_url": "https://github.com/sibirbil/SMB", "code_repository": "sibirbil/SMB", "code_commit": "17fb8ba4f440a5e36e66d701385ecdef6cb05723", "code_archive": "repos/alaQzRbCY9w.zip", "code_archive_sha256": "fe6636bb802a0fefd225b034c8c5aaf15cc298ffaed48f9d5f501c79b7797f27", "code_archive_bytes": 86928, "code_file_count": 15, "code_extensions": {".py": 15}, "github_disk_usage_kb": 2671, "github_languages": {"Python": 48703}, "github_archived": false, "github_pushed_at": "2023-02-15T16:44:03Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/bolstering-stochastic-gradient-descent-with-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "XEw5Onu69uu", "year": 2021, "status": "rejected", "title": "Self-Labeling of Fully Mediating Representations by Graph Alignment", "authors": ["Martijn Oldenhof", "Adam Arany", "Yves Moreau", "Jaak Simm"], "authorids": ["~Martijn_Oldenhof1", "~Adam_Arany1", "~Yves_Moreau2", "~Jaak_Simm1"], "authors_source": "OpenReview API", "abstract": "To be able to predict a molecular graph structure ($W$) given a 2D image of a chemical compound ($U$)  is a challenging problem in machine learning. We are interested to learn $f: U \\rightarrow W$ where we have a fully mediating representation $V$ such that $f$ factors into $U \\rightarrow V \\rightarrow W$. However, observing V requires detailed and expensive labels. We propose \\textbf{graph aligning} approach that generates rich or detailed labels given normal labels $W$. In this paper we investigate the scenario of domain adaptation from the source domain where we have access to the expensive labels $V$ to the target domain where only normal labels W are available. Focusing on the problem of predicting chemical compound graphs from 2D images the fully mediating layer is represented using the planar embedding of the chemical graph structure we are predicting. The use of a fully mediating layer implies some assumptions on the mechanism of the underlying process. However if the assumptions are correct it should allow the machine learning model to be more interpretable, generalize better and be more  data efficient at training time.\nThe empirical results show that, using only 4000 data points, we obtain up to 4x improvement of performance after domain adaptation to target domain compared to pretrained model only on the source domain. After domain adaptation, the model is even able to detect atom types that were never seen in the original source domain. Finally, on the Maybridge data set the proposed self-labeling approach reached higher performance than the current state of the art.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=XEw5Onu69uu", "arxiv_id": "2103.14133", "paper_pdf": "papers/XEw5Onu69uu.pdf", "paper_pdf_sha256": "11b887ee3ea1e57b05d84ba267dfda746ee2aaede94e6d5aa21e41f344f241ee", "paper_pdf_bytes": 1208307, "paper_pdf_source": "openreview", "code_url": "https://github.com/biolearning-stadius/chemgrapher-self-rich-labeling", "code_repository": "biolearning-stadius/chemgrapher-self-rich-labeling", "code_commit": "7018a7d2ea0e9288dc94cdce5fb7024d4c14cd55", "code_archive": "repos/XEw5Onu69uu.zip", "code_archive_sha256": "a5fd427ca5577951a8b26772e7ccb79c1c0f3e311811b6a5a21b412f8fa5f1f1", "code_archive_bytes": 1553802, "code_file_count": 23, "code_extensions": {".py": 22, ".ipynb": 1}, "github_disk_usage_kb": 1635, "github_languages": {"Jupyter Notebook": 185146, "Python": 122012}, "github_archived": false, "github_pushed_at": "2021-09-02T09:39:42Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/self-labeling-of-fully-mediating-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "qiOIqQ9CwA", "year": 2026, "status": "rejected", "title": "Optimal Stepsize for Diffusion Sampling", "authors": ["Jianning Pei", "Han Hu", "Shuyang Gu"], "authorids": ["~Jianning_Pei1", "~Han_Hu1", "~Shuyang_Gu1"], "authors_source": "OpenReview API", "abstract": "Diffusion models achieve remarkable generation quality but suffer from computational intensive sampling due to suboptimal step discretization. While existing works focus on optimizing denoising directions, we address the principled design of stepsize schedules. This paper proposes Optimal Stepsize Distillation, a dynamic programming framework that extracts theoretically optimal schedules by distilling knowledge from reference trajectories. By reformulating stepsize optimization as recursive error minimization, our method guarantees global discretization bounds through optimal substructure exploitation. Crucially, the distilled schedules demonstrate strong robustness across architectures, ODE solvers, and noise schedules. Experiments show 10x accelerated text-to-image generation while preserving 99.4% performance on GenEval.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=qiOIqQ9CwA", "arxiv_id": "2503.21774", "paper_pdf": "papers/qiOIqQ9CwA.pdf", "paper_pdf_sha256": "ade4e5a2ed53aff1db1fe45f258548b09292d60c941a17e739fde0381e5ee9dd", "paper_pdf_bytes": 43910505, "paper_pdf_source": "openreview", "code_url": "https://github.com/bebebe666/OptimalSteps", "code_repository": "bebebe666/OptimalSteps", "code_commit": "ee350436c86e29088c0cf550300a2e6a5911a7ee", "code_archive": "repos/qiOIqQ9CwA.zip", "code_archive_sha256": "8a6e85af2d138a34cf05c7a4bc3fa53b1b4ac03a3c89140fd74e90305c7cfedc", "code_archive_bytes": 6127121, "code_file_count": 15, "code_extensions": {".py": 11, ".sh": 4}, "github_disk_usage_kb": 5993, "github_languages": {"Python": 17789, "Shell": 3320}, "github_archived": false, "github_pushed_at": "2025-04-13T05:54:03Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/optimal-stepsize-for-diffusion-sampling"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "PIHPmNNp7w", "year": 2025, "status": "rejected", "title": "Retrieval-Augmented Decision Transformer: External Memory for In-context RL", "authors": ["Thomas Schmied", "Fabian Paischer", "Vihang Prakash Patil", "Markus Hofmarcher", "Razvan Pascanu", "Sepp Hochreiter"], "authorids": ["~Thomas_Schmied1", "~Fabian_Paischer1", "~Vihang_Prakash_Patil1", "~Markus_Hofmarcher1", "~Razvan_Pascanu1", "~Sepp_Hochreiter1"], "authors_source": "OpenReview API", "abstract": "In-context learning (ICL) is the ability of a model to learn a new task by observing a few exemplars in its context. While prevalent in NLP, this capability has recently also been observed in Reinforcement Learning (RL) settings. Prior in-context RL methods, however, require entire episodes in the agent's context. Given that complex environments typically lead to long episodes with sparse rewards, these methods are constrained to simple environments with short episodes. To address these challenges, we introduce Retrieval-Augmented Decision Transformer (RA-DT). RA-DT employs an external memory mechanism to store past experiences from which it retrieves only sub-trajectories relevant for the current situation. The retrieval component in RA-DT does not require training and can be entirely domain-agnostic. We evaluate the capabilities of RA-DT on grid-world environments, robotics simulations, and procedurally-generated video games. On grid worlds, RA-DT outperforms baselines, while using only a fraction of their context length. Furthermore, we illuminate the limitations of current in-context RL methods on complex environments and discuss future directions. To facilitate future research, we release datasets for four of the considered environments.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=PIHPmNNp7w", "arxiv_id": "2410.07071", "paper_pdf": "papers/PIHPmNNp7w.pdf", "paper_pdf_sha256": "1871449b61b208290f9be73ca01277721718faa459a8e1d3656a58d345234951", "paper_pdf_bytes": 11811785, "paper_pdf_source": "openreview", "code_url": "https://github.com/ml-jku/RA-DT", "code_repository": "ml-jku/RA-DT", "code_commit": "40adec5cc4a8f3aeef1e84c5a203eb55ebd9d481", "code_archive": "repos/PIHPmNNp7w.zip", "code_archive_sha256": "e0b5ab37a38cbba26b08aaaa8b1019a34a49501f0f2bd6e2fe9ec95e7bb84ac1", "code_archive_bytes": 384060, "code_file_count": 75, "code_extensions": {".py": 75}, "github_disk_usage_kb": 349, "github_languages": {"Python": 849387}, "github_archived": false, "github_pushed_at": "2024-10-27T16:10:03Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/retrieval-augmented-decision-transformer"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "GY1fKFXG5i", "year": 2024, "status": "rejected", "title": "Non-Vacuous Generalization Bounds for Large Language Models", "authors": ["Sanae Lotfi", "Marc Anton Finzi", "Yilun Kuang", "Tim G. J. Rudner", "Micah Goldblum", "Andrew Gordon Wilson"], "authorids": ["~Sanae_Lotfi1", "~Marc_Anton_Finzi1", "~Yilun_Kuang1", "~Tim_G._J._Rudner2", "~Micah_Goldblum1", "~Andrew_Gordon_Wilson1"], "authors_source": "OpenReview API", "abstract": "Modern language models can contain billions of parameters, raising the question of whether they can generalize beyond the training data or simply regurgitate their training corpora. We provide the first non-vacuous generalization bounds for pretrained large language models (LLMs), indicating that language models are capable of discovering regularities that generalize to unseen data. In particular, we derive a compression bound that is valid for the unbounded log-likelihood loss using prediction smoothing, and we extend the bound to handle subsampling, accelerating bound computation on massive datasets. To achieve the extreme level of compression required for non-vacuous generalization bounds, we devise SubLoRA, a low-dimensional non-linear parameterization. Using this approach, we find that larger models have better generalization bounds and are more compressible than smaller models.", "decision": "Reject", "meta_review": "", "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=GY1fKFXG5i", "arxiv_id": "2312.17173", "paper_pdf": "papers/GY1fKFXG5i.pdf", "paper_pdf_sha256": "04d76647540f51fc646fc5a187ee916012d89d1963d8f7f1dcb99c89fc3c20e7", "paper_pdf_bytes": 362216, "paper_pdf_source": "openreview", "code_url": "https://github.com/Sanaelotfi/sublora-bounds-for-llms", "code_repository": "Sanaelotfi/sublora-bounds-for-llms", "code_commit": "c606ea664e54fe60cfd870167bf6d1c1183fedf1", "code_archive": "repos/GY1fKFXG5i.zip", "code_archive_sha256": "9e3f91b35553c7183df48379e9e9122da9c8283abe392b5447ee5eeca52c964f", "code_archive_bytes": 310471, "code_file_count": 16, "code_extensions": {".py": 16}, "github_disk_usage_kb": 300, "github_languages": {"Python": 197077}, "github_archived": false, "github_pushed_at": "2024-06-04T01:38:45Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/non-vacuous-generalization-bounds-for-large"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "cwiFbXPW4G0", "year": 2023, "status": "rejected", "title": "Learning for Edge-Weighted Online Bipartite Matching with Robustness Guarantees", "authors": ["Pengfei Li", "Jianyi Yang", "Shaolei Ren"], "authorids": ["~Pengfei_Li2", "~Jianyi_Yang1", "~Shaolei_Ren1"], "authors_source": "OpenReview API", "abstract": "Many real-world problems, such as online ad display, can be formulated as online bipartite matching. The crucial challenge lies in the nature of sequentially-revealed online item information, based on which we make irreversible matching decisions at each step. While numerous expert online algorithms have been proposed with bounded worst-case competitive ratios, they may not offer satisfactory performance in average cases. On the other hand, reinforcement learning (RL) has been applied to improve the average performance, but they lack robustness and can perform arbitrarily badly. In this paper, we propose a novel RL-based approach to edge-weighted online bipartite matching with robustness guarantees (LOMAR), achieving both good average-case and good worst-case performance. The key novelty of LOMAR is a new online switching operation which, based on a judiciously-designed condition to hedge against future uncertainties, decides whether to follow the expert's decision or the RL decision for each online item arrival. We prove that for any $\\rho \\in [0,1]$, LOMAR is $\\rho$-competitive against any given expert online algorithm. To improve the average performance, we  train the RL policy by explicitly considering the online switching operation. Finally, we run empirical experiments to demonstrate the advantages of LOMAR compared to existing baselines.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=cwiFbXPW4G0", "arxiv_id": "2306.00172", "paper_pdf": "papers/cwiFbXPW4G0.pdf", "paper_pdf_sha256": "c22fb89087ec769907de6de169c2f567782c149838560417b3eec7e181246aea", "paper_pdf_bytes": 479149, "paper_pdf_source": "openreview", "code_url": "https://github.com/Ren-Research/LOMAR", "code_repository": "Ren-Research/LOMAR", "code_commit": "7710662a31b55fdafe1c542da1842d8ed9f2e7a9", "code_archive": "repos/cwiFbXPW4G0.zip", "code_archive_sha256": "705514101b0a81274b01b361c9136fb3914c41542aeb17c326c845caa01a2874", "code_archive_bytes": 704788, "code_file_count": 40, "code_extensions": {".py": 40}, "github_disk_usage_kb": 682, "github_languages": {"Python": 323617}, "github_archived": false, "github_pushed_at": "2023-08-09T02:45:38Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/learning-for-edge-weighted-online-bipartite"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "8QE3pwEVc8P", "year": 2022, "status": "rejected", "title": "Zero-Cost Operation Scoring in Differentiable Architecture Search", "authors": ["Lichuan Xiang", "Łukasz Dudziak", "Mohamed S Abdelfattah", "Thomas Chun Pong Chau", "Nicholas Donald Lane", "Hongkai Wen"], "authorids": ["~Lichuan_Xiang1", "~Łukasz_Dudziak1", "~Mohamed_S_Abdelfattah1", "~Thomas_Chun_Pong_Chau1", "~Nicholas_Donald_Lane1", "~Hongkai_Wen1"], "authors_source": "OpenReview API", "abstract": "Differentiable neural architecture search (NAS) has attracted significant attention in recent years due to its ability to quickly discover promising architectures of deep neural networks even in very large search spaces. Despite its success, many differentiable NAS methods lack robustness and may degenerate to trivial architectures with excessive parameter-free operations such as skip connections thus leading to inferior performance. In fact, selecting operations based on the magnitude of architectural parameters was recently proven to be fundamentally wrong, showcasing the need to rethink how operation scoring and selection occurs in differentiable NAS. To this end, we formalize and analyze a fundamental component of differentiable NAS: local \"operation scoring\" that occurs at each choice of operation.\nWhen comparing existing operation scoring functions, we find that existing methods can be viewed as inexact proxies for accuracy.\nWe also find that existing methods perform poorly when analyzed empirically on NAS benchmarks. From this perspective, we introduce new training-free proxies to the context of differentiable NAS, and show that we can significantly speed up the search process while improving accuracy on multiple search spaces. We take inspiration from zero-cost proxies that were recently studied in the context of sample-based NAS but shown to degrade significantly for larger search spaces like DARTS. Our novel \"perturbation-based zero-cost operation scoring\" (Zero-Cost-PT) improves searching time and accuracy compared to the best available differentiable architecture search for many search space sizes, including very large ones. Specifically, we are able improve accuracy compared to the best current method (DARTS-PT) on the DARTS CNN search space while being over 40x faster (total searching time 25 minutes on a single GPU). Our code is available at: https://github.com/avail-upon-acceptance.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=8QE3pwEVc8P", "arxiv_id": "2106.06799", "paper_pdf": "papers/8QE3pwEVc8P.pdf", "paper_pdf_sha256": "aa5dd9bcb28ebb7ec261ff973d60e9a01ecf2012c7287464d33f77e5417e77d0", "paper_pdf_bytes": 1426415, "paper_pdf_source": "openreview", "code_url": "https://github.com/visionbasicagent/zerocost_operation_score", "code_repository": "visionbasicagent/zerocost_operation_score", "code_commit": "55fc52b29a8d1be086937a7e8045d155ec0cf63f", "code_archive": "repos/8QE3pwEVc8P.zip", "code_archive_sha256": "16e9163ccc8694ced2039ff210c27c2528c06074bd659e816b9498c1f2c8480c", "code_archive_bytes": 4029109, "code_file_count": 230, "code_extensions": {".py": 123, ".sh": 94, ".ipynb": 13}, "github_disk_usage_kb": 3010, "github_languages": {"Jupyter Notebook": 15717226, "Python": 877775, "Shell": 166619}, "github_archived": false, "github_pushed_at": "2022-12-01T19:51:51Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/zero-cost-proxies-meet-differentiable"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "fycxGdpCCmW", "year": 2021, "status": "rejected", "title": "Hybrid Discriminative-Generative Training via Contrastive Learning", "authors": ["Hao Liu", "Pieter Abbeel"], "authorids": ["~Hao_Liu1", "~Pieter_Abbeel2"], "authors_source": "OpenReview API", "abstract": "Contrastive learning and supervised learning have both seen significant progress and success. However, thus far they have largely been treated as two separate objectives, brought together only by having a shared neural network. In this paper we show that through the perspective of hybrid discriminative-generative training of energy-based models we can make a direct connection between contrastive learning and supervised learning. Beyond presenting this unified view, we show our specific choice of approximation of the energy-based loss significantly improves energy-based models and contrastive learning based methods in confidence-calibration, out-of-distribution detection, adversarial robustness, generative modeling, and image classification tasks. In addition to significantly improved performance, our method also gets rid of SGLD training and does not suffer from training instability. Our evaluations also demonstrate that our method performs better than or on par with state-of-the-art hand-tailored methods in each task. ", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=fycxGdpCCmW", "arxiv_id": "2007.09070", "paper_pdf": "papers/fycxGdpCCmW.pdf", "paper_pdf_sha256": "b1f7de6d844c0fe1fa671c3b1531b6904291739e485a7eeb3a86caf75fd844c0", "paper_pdf_bytes": 716991, "paper_pdf_source": "openreview", "code_url": "https://github.com/haoliuhl/hybrid-discriminative-generative", "code_repository": "haoliuhl/hybrid-discriminative-generative", "code_commit": "9f12b3e8f53dcc20bdc47359ffba1894748f1c8f", "code_archive": "repos/fycxGdpCCmW.zip", "code_archive_sha256": "bba232cd59a576a245bdac2300dd0e375e25980095a77ae97100f4aed9d28cf1", "code_archive_bytes": 1483099, "code_file_count": 16, "code_extensions": {".py": 11, ".js": 5}, "github_disk_usage_kb": 1648, "github_languages": {"Python": 153412, "JavaScript": 28156, "HTML": 23946, "CSS": 8270}, "github_archived": false, "github_pushed_at": "2023-05-01T20:42:33Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/hybrid-discriminative-generative-training-via"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "4eYSSSDle6", "year": 2026, "status": "rejected", "title": "PRL: Prompts from Reinforcement Learning", "authors": ["Paweł Batorski", "Adrian Kosmala", "Paul Swoboda"], "authorids": ["~Paweł_Batorski2", "~Adrian_Kosmala1", "~Paul_Swoboda1"], "authors_source": "OpenReview API", "abstract": "Effective prompt engineering remains a central challenge in fully harnessing the\ncapabilities of LLMs. While well-designed prompts can dramatically enhance\nperformance, crafting them typically demands expert intuition and a nuanced understanding of the task. Moreover, the most impactful prompts often hinge on\nsubtle semantic cues, ones that may elude human perception but are crucial for\nguiding LLM behavior. In this paper, we introduce PRL (Prompts from Reinforcement Learning), a novel RL-based approach for automatic prompt generation.\nUnlike previous methods, PRL can produce novel few-shot examples that were not\nseen during training. Our approach achieves state-of-the-art performance across a\nrange of benchmarks, including text classification, simplification, summarization,\nand reasoning. On the classification task, it surpasses prior methods by 2.58%\nover APE and 1.00% over EvoPrompt. Additionally, it improves the average\nROUGE scores on the summarization task by 4.32 over APE and by 2.12 over\nEvoPrompt and the SARI score on simplification by 6.93 over APE and by 6.01\nover EvoPrompt. On the GSM8K mathematical reasoning benchmark, PRL further\nimproves accuracy by 2.72% over APE and by 4.53% over EvoPrompt. We will\nmake our implementation publicly available upon acceptance.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=4eYSSSDle6", "arxiv_id": "2505.14412", "paper_pdf": "papers/4eYSSSDle6.pdf", "paper_pdf_sha256": "f1566a76074a2d37315d2af3a19028f167133afec792b8cdef1ff017cabaac61", "paper_pdf_bytes": 380024, "paper_pdf_source": "openreview", "code_url": "https://github.com/Batorskq/PRL-Prompts-from-Reinforcement-Learning", "code_repository": "Batorskq/PRL-Prompts-from-Reinforcement-Learning", "code_commit": "f9797cb987f5fd9f2ce5aa7b9e1e15b12f6ea328", "code_archive": "repos/4eYSSSDle6.zip", "code_archive_sha256": "b1c37c1a30cb5a61d90d3ade1bb1af8207ed3e00720d5aacbd20ea2ff0bbf95f", "code_archive_bytes": 7940967, "code_file_count": 276, "code_extensions": {".py": 267, ".sh": 9}, "github_disk_usage_kb": 6953, "github_languages": {"Python": 1790258, "Shell": 11501}, "github_archived": false, "github_pushed_at": "2025-06-05T11:48:47Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/prl-prompts-from-reinforcement-learning"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "O6W9SJRZRA", "year": 2025, "status": "rejected", "title": "Right on Time: Revising Time Series Models by Constraining their Explanations", "authors": ["Maurice Kraus", "David Steinmann", "Antonia Wüst", "Andre Kokozinski", "Kristian Kersting"], "authorids": ["~Maurice_Kraus1", "~David_Steinmann1", "~Antonia_Wüst1", "~Andre_Kokozinski1", "~Kristian_Kersting1"], "authors_source": "OpenReview API", "abstract": "The reliability of deep time series models is often compromised by their tendency to rely on confounding factors, which may lead to incorrect outputs. Our newly recorded, naturally confounded dataset named P2S from a real mechanical production line emphasizes this. To avoid “Clever-Hans” moments in time series, i.e., to mitigate confounders, we introduce the method Right on Time (RioT). RioT enables, for the first time interactions with model explanations across both the time and frequency domain. Feedback on explanations in both domains is then used to constrain the model, steering it away from the annotated confounding factors. The dual-domain interaction strategy is crucial for effectively addressing confounders in time series datasets. We empirically demonstrate that RioT can effectively guide models away from the wrong reasons in P2S as well as popular time series classification and forecasting datasets.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=O6W9SJRZRA", "arxiv_id": "2402.12921", "paper_pdf": "papers/O6W9SJRZRA.pdf", "paper_pdf_sha256": "6bfb009dd3dc87b73bf435d1ed7fb5263e4730490588c417ad8ed4a8df0e50f4", "paper_pdf_bytes": 2753655, "paper_pdf_source": "openreview", "code_url": "https://github.com/ml-research/RioT", "code_repository": "ml-research/RioT", "code_commit": "0480158d53ec9b2218eec89d7b3f3f8f02277b09", "code_archive": "repos/O6W9SJRZRA.zip", "code_archive_sha256": "27d523b40012ccb9d5e532118e91723b54300dce2c644ac1c73196264aec2547", "code_archive_bytes": 391126, "code_file_count": 86, "code_extensions": {".py": 84, ".ipynb": 2}, "github_disk_usage_kb": 354, "github_languages": {"Jupyter Notebook": 373269, "Python": 291429, "Dockerfile": 811}, "github_archived": false, "github_pushed_at": "2024-06-18T15:51:51Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/right-on-time-revising-time-series-models-by"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "VmqTuFMk68", "year": 2024, "status": "rejected", "title": "Trainable Transformer in Transformer", "authors": ["Abhishek Panigrahi", "Sadhika Malladi", "Mengzhou Xia", "Sanjeev Arora"], "authorids": ["~Abhishek_Panigrahi1", "~Sadhika_Malladi2", "~Mengzhou_Xia1", "~Sanjeev_Arora1"], "authors_source": "OpenReview API", "abstract": "Recent works attribute the capability of in-context learning (ICL) in large pre-trained language models to implicitly simulating and fine-tuning an internal model (e.g., linear or 2-layer MLP) during inference. However, such constructions require large memory overhead, which makes simulation of more sophisticated internal models intractable. In this work, we propose a new efficient construction, Transformer in Transformer (in short, TINT), that allows a transformer to simulate and fine-tune more complex models during inference (e.g., pre-trained language models). In particular, we introduce innovative approximation techniques that allow a TINT model with less than 2 billion parameters to simulate and fine-tune a 125 million parameter transformer model within a single forward pass. TINT accommodates many common transformer variants and its design ideas also improve the efficiency of past instantiations of simple models inside transformers. We conduct end-to-end experiments to validate the internal fine-tuning procedure of TINT on various language modeling and downstream tasks. For example, even with a limited one-step budget, we observe TINT for a OPT-125M model improves performance by 4 − 16% absolute on average compared to OPT-125M. These findings suggest that large pre-trained language models are capable of performing intricate subroutines. To facilitate further work, a modular and extensible codebase for TINT is included.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=VmqTuFMk68", "arxiv_id": "2307.01189", "paper_pdf": "papers/VmqTuFMk68.pdf", "paper_pdf_sha256": "98ec028b07033358362f20d42a8140dfdfa00362f3773c28f8de52652c3832c1", "paper_pdf_bytes": 3692509, "paper_pdf_source": "openreview", "code_url": "https://github.com/abhishekpanigrahi1996/transformer_in_transformer", "code_repository": "abhishekpanigrahi1996/transformer_in_transformer", "code_commit": "1722081db59c1a9459941ec9e2c2f039fba9f331", "code_archive": "repos/VmqTuFMk68.zip", "code_archive_sha256": "5121d3b818a7748a5094d969bb0fc6da3d10596fd5a03d42b0ce550559403ea6", "code_archive_bytes": 321987, "code_file_count": 40, "code_extensions": {".py": 39, ".sh": 1}, "github_disk_usage_kb": 315, "github_languages": {"Python": 500283, "Shell": 3505}, "github_archived": false, "github_pushed_at": "2023-10-11T17:55:19Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/trainable-transformer-in-transformer"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "QEmn_Hvh7j8", "year": 2023, "status": "rejected", "title": "Private GANs, Revisited", "authors": ["Alex Bie", "Gautam Kamath", "Guojun Zhang"], "authorids": ["~Alex_Bie1", "~Gautam_Kamath1", "~Guojun_Zhang1"], "authors_source": "OpenReview API", "abstract": "We show that with improved training, the standard approach for differentially private GANs -- updating the discriminator with noisy gradients -- achieves or competes with state-of-the-art results for private image synthesis. Existing instantiations of this approach neglect to consider how adding noise only to discriminator updates disrupts the careful balance between generator and discriminator necessary for successful GAN training. We show that a simple fix restores parity: taking more discriminator steps between generator steps. Finally, with the goal of restoring parity between generator and discriminator, we experiment with further modifications to improve discriminator training and see further improvements. For MNIST at $\\eps=10$, our private GANs improve the record FID from 48.4 to 13.0, as well as downstream classifier accuracy from 83.2\\% to 95.0\\%.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=QEmn_Hvh7j8", "arxiv_id": "2302.02936", "paper_pdf": "papers/QEmn_Hvh7j8.pdf", "paper_pdf_sha256": "ea524c63ea455d0332e874d3b67c64a46d65b941288c9ee806dd1e5f1e177396", "paper_pdf_bytes": 3233598, "paper_pdf_source": "openreview", "code_url": "https://github.com/alexbie98/dpgan-revisit", "code_repository": "alexbie98/dpgan-revisit", "code_commit": "2233d9bc7f787e19750fc83c3a5944b43e25a9df", "code_archive": "repos/QEmn_Hvh7j8.zip", "code_archive_sha256": "3b3a4d5f98f2bb52c32b2b3accc0380ba68ccc652ae1997de502dd68da3b9f46", "code_archive_bytes": 489139, "code_file_count": 29, "code_extensions": {".py": 29}, "github_disk_usage_kb": 710, "github_languages": {"Python": 117435}, "github_archived": false, "github_pushed_at": "2023-10-05T14:00:31Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/private-gans-revisited"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "2PSrjVtj6gU", "year": 2022, "status": "rejected", "title": "Graph Attention Multi-layer Perceptron", "authors": ["Wentao Zhang", "Ziqi Yin", "Zeang Sheng", "Yang Li", "Wen Ouyang", "Xiaosen Li", "Yangyu Tao", "Zhi Yang", "Bin CUI"], "authorids": ["~Wentao_Zhang1", "~Ziqi_Yin1", "~Zeang_Sheng1", "~Yang_Li36", "~Wen_Ouyang1", "~Xiaosen_Li1", "~Yangyu_Tao2", "~Zhi_Yang4", "~Bin_CUI2"], "authors_source": "OpenReview API", "abstract": "Recently, graph neural networks (GNNs) have achieved a stride of success in many graph-based applications. However, most GNNs suffer from a critical issue: representation learned is constructed based on a fixed k-hop neighborhood and insensitive to individual needs for each node, which greatly hampers the performance of GNNs. To satisfy the unique needs of each node, we propose a new architecture -- Graph Attention Multi-Layer Perceptron (GAMLP). This architecture combines multi-scale knowledge and learns to capture the underlying correlations between different scales of knowledge with two novel attention mechanisms: Recursive attention and Jumping Knowledge (JK) attention. Instead of using node feature only, the knowledge within node labels is also exploited to reinforce the performance of GAMLP. Extensive experiments on 12 real-world datasets demonstrate that GAMLP achieves state-of-the-art performance while enjoying high scalability and efficiency.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=2PSrjVtj6gU", "arxiv_id": "2206.04355", "paper_pdf": "papers/2PSrjVtj6gU.pdf", "paper_pdf_sha256": "f0060f3ef66ee90a5ddf5e29fd6c06411e50d4e370c739f7482748e6af33faa0", "paper_pdf_bytes": 700295, "paper_pdf_source": "openreview", "code_url": "https://github.com/PKU-DAIR/GAMLP", "code_repository": "PKU-DAIR/GAMLP", "code_commit": "83dc5d4a414c1829ed4a793410d3bd6037ee89ee", "code_archive": "repos/2PSrjVtj6gU.zip", "code_archive_sha256": "cc28d0919c20311ff05a335c267ee9523e5b8a83df9ea94be48f684d4f5e30be", "code_archive_bytes": 1183856, "code_file_count": 9, "code_extensions": {".py": 8, ".sh": 1}, "github_disk_usage_kb": 3027, "github_languages": {"Python": 59499, "Shell": 1028}, "github_archived": false, "github_pushed_at": "2022-06-25T09:04:23Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/graph-attention-multi-layer-perceptron-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "nCY83KxoehA", "year": 2021, "status": "rejected", "title": "Automated Concatenation of Embeddings for Structured Prediction", "authors": ["Xinyu Wang", "Yong Jiang", "Nguyen Bach", "Tao Wang", "Zhongqiang Huang", "Fei Huang", "Kewei Tu"], "authorids": ["~Xinyu_Wang3", "~Yong_Jiang1", "~Nguyen_Bach1", "~Tao_Wang4", "~Zhongqiang_Huang1", "~Fei_Huang2", "~Kewei_Tu1"], "authors_source": "OpenReview API", "abstract": "Pretrained contextualized embeddings are powerful word representations for structured prediction tasks. Recent work found that better word representations can be obtained by concatenating different types of embeddings. However, the selection of embeddings to form the best concatenated representation usually varies depending on the task and the collection of candidate embeddings, and the ever-increasing number of embedding types makes it a more difficult problem. In this paper, we propose Automated Concatenation of Embeddings (ACE) to automate the process of finding better concatenations of embeddings for structured prediction tasks, based on a formulation inspired by recent progress on neural architecture search. Specifically, a controller alternately samples a concatenation of embeddings, according to its current belief of the effectiveness of individual embedding types in consideration for a task, and updates the belief based on a reward. We follow strategies in reinforcement learning to optimize the parameters of the controller and compute the reward based on the accuracy of a task model, which is fed with the sampled concatenation as input and trained on a task dataset. Empirical results on 6 tasks and 21 datasets show that our approach outperforms strong baselines and achieves state-of-the-art performance with fine-tuned embeddings in the vast majority of evaluations.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=nCY83KxoehA", "arxiv_id": "2010.05006", "paper_pdf": "papers/nCY83KxoehA.pdf", "paper_pdf_sha256": "a3c69d41fa029be962e56d93214e2fc09181deabb5f0be33fc1c7bfb04ffd848", "paper_pdf_bytes": 401053, "paper_pdf_source": "openreview", "code_url": "https://github.com/Alibaba-NLP/ACE", "code_repository": "Alibaba-NLP/ACE", "code_commit": "cf50440b5d4ab44f74feb4214733224265428db4", "code_archive": "repos/nCY83KxoehA.zip", "code_archive_sha256": "91055b0b5bac3f17a3e0649024ed0f846fe80507facaa1a22a9bff8164a4e169", "code_archive_bytes": 1456656, "code_file_count": 123, "code_extensions": {".py": 123}, "github_disk_usage_kb": 1758, "github_languages": {"Python": 1615035, "Perl": 12743}, "github_archived": false, "github_pushed_at": "2022-12-02T09:13:18Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/automated-concatenation-of-embeddings-for-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "AEgyitdRWf", "year": 2026, "status": "rejected", "title": "ML-Agent: Reinforcing LLM Agents for Autonomous Machine Learning Engineering", "authors": ["Zexi Liu", "Jingyi Chai", "Xinyu Zhu", "Shuo Tang", "Rui Ye", "Weiyu Ma", "Bo Zhang", "LEI BAI", "Siheng Chen"], "authorids": ["~Zexi_Liu1", "~Jingyi_Chai1", "~Xinyu_Zhu5", "~Shuo_Tang2", "~Rui_Ye1", "~Weiyu_Ma1", "~Bo_Zhang17", "~LEI_BAI1", "~Siheng_Chen1"], "authors_source": "OpenReview API", "abstract": "The emergence of large language model (LLM)-based agents has significantly advanced the development of autonomous machine learning (ML) engineering. However, the dominant prompt-based paradigm exhibits limitations: smaller models lack the capacity to learn from execution trajectories for generalization, while large proprietary models incur high computational overhead, restricting accessibility and scalability.\nFocusing on this, for the first time, we explore the paradigm of learning-based agentic ML, where an LLM agent learns through interactive experimentation on ML tasks using online reinforcement learning (RL). To realize this, we propose a novel agentic ML training framework with three key components:\n(1) exploration-enriched fine-tuning, which enables LLM agents to generate diverse actions for enhanced RL exploration;\n(2) step-wise RL, which enables training on a single action step, accelerating experience collection and improving training efficiency;\n(3) an agentic ML-specific reward module, which unifies varied ML feedback signals into consistent rewards for RL optimization. \nLeveraging this framework, we train ML-Agent, driven by a 7B-sized Qwen-2.5 LLM for autonomous ML.\nDespite training on only 9 ML tasks, our 7B-sized ML-Agent achieves comparable performance to agents using much larger proprietary LLMs (e.g., GPT-5) but at significantly lower computational cost, demonstrating strong performance and cross-task generalization.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=AEgyitdRWf", "arxiv_id": "2505.23723", "paper_pdf": "papers/AEgyitdRWf.pdf", "paper_pdf_sha256": "860c4bcd5f977fd96dee280d1b269e40187a1e98d91ba4605c4080df351f5137", "paper_pdf_bytes": 2392506, "paper_pdf_source": "openreview", "code_url": "https://github.com/MASWorks/ML-Agent", "code_repository": "MASWorks/ML-Agent", "code_commit": "15932e7525deb99d59f7416bbe8c75077cff3690", "code_archive": "repos/AEgyitdRWf.zip", "code_archive_sha256": "c0207c1ddb4cf5870bdda0284dfbfcac5ec13fa7a44cc5c4e0ed04a15d175414", "code_archive_bytes": 7311896, "code_file_count": 156, "code_extensions": {".py": 150, ".sh": 6}, "github_disk_usage_kb": 7394, "github_languages": {"Python": 1089531, "Shell": 2413, "Makefile": 391}, "github_archived": false, "github_pushed_at": "2025-06-21T18:17:12Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/ml-agent-reinforcing-llm-agents-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "f4mQ2SU5tp", "year": 2025, "status": "rejected", "title": "IntLoRA: Integral Low-rank Adaptation of Quantized Diffusion Models", "authors": ["Hang Guo", "Yawei Li", "Tao Dai", "Shu-Tao Xia", "Luca Benini"], "authorids": ["~Hang_Guo3", "~Yawei_Li1", "~Tao_Dai3", "~Shu-Tao_Xia1", "~Luca_Benini2"], "authors_source": "OpenReview API", "abstract": "Fine-tuning large-scale text-to-image diffusion models for various downstream tasks has yielded impressive results. However, the heavy computational burdens of tuning large models prevent personal customization. Recent advances have attempted to employ parameter-efficient fine-tuning (PEFT) techniques to adapt the floating-point (FP) or quantized pre-trained weights. Nonetheless, the adaptation parameters in existing works are still restricted to FP arithmetic, hindering hardware-friendly acceleration. In this work, we propose IntLoRA, to further push the efficiency limits by using integer type (INT) low-rank parameters to adapt the quantized diffusion models. By working in the integer arithmetic, our IntLoRA offers three key advantages: (i) for fine-tuning, the pre-trained weights are quantized, reducing memory usage; (ii) for storage, both pre-trained and low-rank weights are in INT which consumes less disk space; (iii) for inference, IntLoRA weights can be naturally merged into quantized pre-trained weights through efficient integer multiplication or bit-shifting, eliminating additional post-training quantization. Extensive experiments demonstrate that IntLoRA can achieve performance on par with or even superior to the vanilla LoRA, accompanied by significant efficiency improvements.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=f4mQ2SU5tp", "arxiv_id": "2410.21759", "paper_pdf": "papers/f4mQ2SU5tp.pdf", "paper_pdf_sha256": "492452adf5049baefb33b9651cbcacfebb16a8a6dfe5b92d6b8ab6c43ff447f0", "paper_pdf_bytes": 4691337, "paper_pdf_source": "openreview", "code_url": "https://github.com/csguoh/IntLoRA", "code_repository": "csguoh/IntLoRA", "code_commit": "65a8257a4311e0feab9b33477c9748d13d8ce17b", "code_archive": "repos/f4mQ2SU5tp.zip", "code_archive_sha256": "10ce454be8b0b948d631cf377c4b144f9adeca2d32df503b960660c75ac62578", "code_archive_bytes": 367037, "code_file_count": 9, "code_extensions": {".py": 8, ".sh": 1}, "github_disk_usage_kb": 358, "github_languages": {"Python": 99330, "Shell": 8177}, "github_archived": false, "github_pushed_at": "2024-11-25T04:23:17Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/intlora-integral-low-rank-adaptation-of"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "17ZbByq95E", "year": 2024, "status": "rejected", "title": "Memory-Efficient Backpropagation through Large Linear Layers", "authors": ["Daniel Bershatsky", "Aleksandr Mikhalev", "Aleksandr Katrutsa", "Julia Gusak", "Daniil Merkulov", "Ivan Oseledets"], "authorids": ["~Daniel_Bershatsky1", "~Aleksandr_Mikhalev1", "~Aleksandr_Katrutsa1", "~Julia_Gusak1", "~Daniil_Merkulov1", "~Ivan_Oseledets1"], "authors_source": "OpenReview API", "abstract": "In modern neural networks like Transformers, linear layers require significant memory to store activations during backward pass. This study proposes a memory reduction approach to perform backpropagation through linear layers. Since the gradients of linear layers are computed by matrix multiplications, we consider methods for randomized matrix multiplications and demonstrate that they require less memory with a moderate decrease of the test accuracy. Also, we investigate the variance of the gradient estimate induced by the randomized matrix multiplication. We compare this variance with the variance coming from gradient estimation based on the batch of samples. We demonstrate the benefits of the proposed method on the fine-tuning of the pretrained RoBERTa model on GLUE tasks.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=17ZbByq95E", "arxiv_id": "2201.13195", "paper_pdf": "papers/17ZbByq95E.pdf", "paper_pdf_sha256": "74f1ef669098cb715ce131858a8bbe5267936d9ea86d47a41cbcd8b4fac2f10a", "paper_pdf_bytes": 474137, "paper_pdf_source": "openreview", "code_url": "https://github.com/skolai/fewbit", "code_repository": "skolai/fewbit", "code_commit": "940706bbbc40b11ade61d42253f80dc960d35793", "code_archive": "repos/17ZbByq95E.zip", "code_archive_sha256": "330303eaa470a820abfde80e7eb67e3d37feaa9b41e71eef065f95da6ef2d0e8", "code_archive_bytes": 310980, "code_file_count": 53, "code_extensions": {".py": 33, ".cc": 5, ".h": 5, ".sh": 4, ".ipynb": 4, ".cu": 2}, "github_disk_usage_kb": 321, "github_languages": {"Python": 141112, "C++": 36054, "Cuda": 30862, "Shell": 5429, "CMake": 3968, "Dockerfile": 1542}, "github_archived": false, "github_pushed_at": "2023-07-26T04:42:05Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/memory-efficient-backpropagation-through"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "zHSaBQtj-l", "year": 2023, "status": "rejected", "title": "Differentiable Rendering with Reparameterized Volume Sampling", "authors": ["Kirill Struminsky", "Oleg Desheulin"], "authorids": ["~Kirill_Struminsky1", "~Oleg_Desheulin2"], "authors_source": "OpenReview API", "abstract": "We propose an alternative rendering algorithm for neural radiance fields based on importance sampling. In view synthesis, a neural radiance field approximates underlying density and radiance fields based on a sparse set of views of a scene. To generate a pixel of a novel view, it marches a ray through the pixel and computes a weighted sum of radiance emitted from a dense set of ray points. This rendering algorithm is fully differentiable and facilitates gradient-based optimization of the fields. However, in practice, only a tiny opaque portion of the ray contributes most of the radiance to the sum. Therefore, we can avoid computing radiance in the rest part. In this work, we use importance sampling to pick non-transparent points on the ray. Specifically, we generate samples according to the probability distribution induced by the density field. Our main contribution is the reparameterization of the sampling algorithm. It allows end-to-end learning with gradient descent as in the original rendering algorithm. With our approach, we can optimize a neural radiance field with just a few radiance field evaluations per ray. As a result, we alleviate the costs associated with the color component of the neural radiance field.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=zHSaBQtj-l", "arxiv_id": "2302.10970", "paper_pdf": "papers/zHSaBQtj-l.pdf", "paper_pdf_sha256": "d297a60270bccf9ea8cfdc0f012983c1d59f68dcb49de45bc14c247569dcf531", "paper_pdf_bytes": 4336943, "paper_pdf_source": "openreview", "code_url": "https://github.com/GreatDrake/reparameterized-volume-sampling", "code_repository": "GreatDrake/reparameterized-volume-sampling", "code_commit": "1fc6979f465e0e556786eee36648c7231c4006a9", "code_archive": "repos/zHSaBQtj-l.zip", "code_archive_sha256": "ead7086e7d32d43e88d4251c3bb2c08321eb3b1f3a18d2840208909b03c444f9", "code_archive_bytes": 703419, "code_file_count": 9, "code_extensions": {".py": 8, ".ipynb": 1}, "github_disk_usage_kb": 713, "github_languages": {"Jupyter Notebook": 603423, "Python": 77328}, "github_archived": false, "github_pushed_at": "2024-04-21T13:44:00Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/differentiable-rendering-with-reparameterized"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "8nXkyH2_s6", "year": 2021, "status": "rejected", "title": "Neural networks behave as hash encoders: An empirical study", "authors": ["Fengxiang He", "Shiye Lei", "Jianmin Ji", "Dacheng Tao"], "authorids": ["~Fengxiang_He1", "leishiye@gmail.com", "jianmin@ustc.edu.cn", "~Dacheng_Tao1"], "authors_source": "OpenReview API", "abstract": "The input space of a neural network with ReLU-like activations is partitioned into multiple linear regions, each corresponding to a specific activation pattern of the included ReLU-like activations. We demonstrate that this partition exhibits the following encoding properties across a variety of deep learning models: (1) {\\it determinism}: almost every linear region contains at most one training example. We can therefore represent almost every training example by a unique activation pattern, which is parameterized by a {\\it neural code}; and (2) {\\it categorization}: according to the neural code, simple algorithms, such as $K$-Means, $K$-NN, and logistic regression, can achieve fairly good performance on both training and test data. These encoding properties surprisingly suggest that {\\it normal neural networks well-trained for classification behave as hash encoders without any extra efforts.} In addition, the encoding properties exhibit variability in different scenarios. {Further experiments demonstrate that {\\it model size}, {\\it training time}, {\\it training sample size}, {\\it regularization}, and {\\it label noise} contribute in shaping the encoding properties, while the impacts of the first three are dominant.} We then define an {\\it activation hash phase chart} to represent the space expanded by {model size}, training time, training sample size, and the encoding properties, which is divided into three canonical regions: {\\it under-expressive regime}, {\\it critically-expressive regime}, and {\\it sufficiently-expressive regime}.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=8nXkyH2_s6", "arxiv_id": "2101.05490", "paper_pdf": "papers/8nXkyH2_s6.pdf", "paper_pdf_sha256": "e07a0c4bf55da59616a0fcf40bc32c09a0b9cb779bb1670aeaa8ba51badf72cb", "paper_pdf_bytes": 1089361, "paper_pdf_source": "openreview", "code_url": "https://github.com/LeavesLei/activation-code", "code_repository": "LeavesLei/activation-code", "code_commit": "6be962f5e6c081d0bbefd6a30f31b13728da7752", "code_archive": "repos/8nXkyH2_s6.zip", "code_archive_sha256": "a4bc5085e3e930bca96833b326175109900c7f7e546e7aebd615e73d37bdf373", "code_archive_bytes": 810980, "code_file_count": 136, "code_extensions": {".sh": 69, ".py": 67}, "github_disk_usage_kb": 1795, "github_languages": {"Python": 323069, "Shell": 23651}, "github_archived": false, "github_pushed_at": "2021-05-28T08:55:34Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/neural-networks-behave-as-hash-encoders-an-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Xq4NdAodoA", "year": 2026, "status": "rejected", "title": "Generative Modeling with Bayesian Sample Inference", "authors": ["Marten Lienen", "Marcel Kollovieh", "Stephan Günnemann"], "authorids": ["~Marten_Lienen1", "~Marcel_Kollovieh1", "~Stephan_Günnemann1"], "authors_source": "OpenReview API", "abstract": "We derive a novel generative model from iterative Gaussian posterior inference. By treating the generated sample as an unknown variable, we can formulate the sampling process in the language of Bayesian probability. Our model uses a sequence of prediction and posterior update steps to iteratively narrow down the unknown sample starting from a broad initial belief. In addition to a rigorous theoretical analysis, we establish a connection between our model and diffusion models and show that it includes Bayesian Flow Networks (BFNs) as a special case. In our experiments, we demonstrate that our model improves sample quality on ImageNet32 over both BFNs and the closely related Variational Diffusion Models, while achieving equivalent log-likelihoods on ImageNet32 and CIFAR10.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=Xq4NdAodoA", "arxiv_id": "2502.07580", "paper_pdf": "papers/Xq4NdAodoA.pdf", "paper_pdf_sha256": "45bc0fe5d84ca629ba40629a7afd658a34a9e54b0740dedb4aeee71a1b489ce8", "paper_pdf_bytes": 1576494, "paper_pdf_source": "openreview", "code_url": "https://github.com/martenlienen/bsi", "code_repository": "martenlienen/bsi", "code_commit": "5b8a6acf17f2c7106d6fa31a474738e3b412c9c1", "code_archive": "repos/Xq4NdAodoA.zip", "code_archive_sha256": "00094761be0e38c281c9f17c770f2ff23a01f14944dbb71a63401fb7d9083827", "code_archive_bytes": 7537638, "code_file_count": 49, "code_extensions": {".py": 48, ".ipynb": 1}, "github_disk_usage_kb": 7457, "github_languages": {"Python": 197117, "Jupyter Notebook": 179113}, "github_archived": false, "github_pushed_at": "2026-08-14T05:54:45Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/generative-modeling-with-bayesian-sample"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "e3odKmatZr", "year": 2025, "status": "rejected", "title": "Critique-out-Loud Reward Models", "authors": ["Zachary Ankner", "Mansheej Paul", "Brandon Cui", "Jonathan Daniel Chang", "Prithviraj Ammanabrolu"], "authorids": ["~Zachary_Ankner1", "~Mansheej_Paul1", "~Brandon_Cui1", "~Jonathan_Daniel_Chang1", "~Prithviraj_Ammanabrolu1"], "authors_source": "OpenReview API", "abstract": "Traditionally, reward models used for reinforcement learning from human feedback (RLHF) are trained to directly predict preference scores without leveraging the generation capabilities of the underlying large language model (LLM). This limits the capabilities of reward models as they must reason implicitly about the quality of a response, i.e., preference modeling must be performed in a single forward pass through the model. To enable reward models to reason explicitly about the quality of a response, we introduce Critique-out-Loud (CLoud) reward models. CLoud reward models operate by first generating a natural language critique of the assistant's response that is then used to predict a scalar reward for the quality of the response. We demonstrate the success of CLoud reward models for both Llama-3-8B and 70B base models: compared to classic reward models CLoud reward models improve pairwise preference classification accuracy on RewardBench by 4.65 and 5.84 percentage points for the 8B and 70B base models respectively. Furthermore, CLoud reward models lead to a Pareto improvement for win rate on ArenaHard when used as the scoring model for Best-of-N. Finally, we explore how to exploit the dynamic inference compute capabilities of CLoud reward models by performing self-consistency decoding for reward prediction.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=e3odKmatZr", "arxiv_id": "2408.11791", "paper_pdf": "papers/e3odKmatZr.pdf", "paper_pdf_sha256": "7a128b93c833ac7cfc77cf2cbba0d1770f1c289609ec2b549cfc84625fe53bc3", "paper_pdf_bytes": 1763330, "paper_pdf_source": "openreview", "code_url": "https://github.com/zankner/CLoud", "code_repository": "zankner/CLoud", "code_commit": "fac417e2c0f45fab9083bf9b77066de490e8288c", "code_archive": "repos/e3odKmatZr.zip", "code_archive_sha256": "da1b3911374b086d2d2cb64f352ce2ca1fa5847fb6d598cfb3d7e113429a8e24", "code_archive_bytes": 378121, "code_file_count": 16, "code_extensions": {".py": 16}, "github_disk_usage_kb": 362, "github_languages": {"Python": 95335}, "github_archived": false, "github_pushed_at": "2024-10-18T19:38:52Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/critique-out-loud-reward-models"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "01LMSeReNvY", "year": 2023, "status": "rejected", "title": "PromptBoosting: Black-Box Text Classification with Ten Forward Passes", "authors": ["Bairu Hou", "Joe O'Connor", "Jacob Andreas", "Shiyu Chang", "Yang Zhang"], "authorids": ["~Bairu_Hou2", "~Joe_O'Connor1", "~Jacob_Andreas1", "~Shiyu_Chang2", "~Yang_Zhang3"], "authors_source": "OpenReview API", "abstract": "We describe PromptBoosting, a query-efficient procedure for building a text classifier from a neural language model (LM) without access to the LM’s parameters, gradients, or hidden representations. This form of “black-box” classifier training has become increasingly important as the cost of training and inference in large-scale LMs grows. But existing black-box LM classifier learning approaches are themselves computationally inefficient, typically specializing LMs to the target task by searching in a large space of (discrete or continuous) prompts using zeroth-order optimization methods. Instead of directly optimizing in prompt space, PromptBoosting obtains a small pool of prompts via a gradient-free approach and then constructs a large pool of weak learners by pairing these prompts with different elements of the LM’s output distribution. These weak learners are then ensembled using the AdaBoost algorithm. The entire learning process requires only a small number of forward passes and no backward pass. Experiments show that PromptBoosting achieves state-of-the-art performance in multiple black-box few-shot classification tasks, and matches or outperforms full fine-tuning in both few-shot and standard learning paradigms, while training 10x faster than existing black-box methods.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=01LMSeReNvY", "arxiv_id": "2212.09257", "paper_pdf": "papers/01LMSeReNvY.pdf", "paper_pdf_sha256": "a7492816417762dbfd110f87eb6b6cbf4b16e9d7c1806461fa8e949b8ed458be", "paper_pdf_bytes": 700916, "paper_pdf_source": "openreview", "code_url": "https://github.com/UCSB-NLP-Chang/PromptBoosting", "code_repository": "UCSB-NLP-Chang/PromptBoosting", "code_commit": "6843af7da67a27f21f1de76950a7f0c0f3fb7785", "code_archive": "repos/01LMSeReNvY.zip", "code_archive_sha256": "df0b24f71dc2ce1c47abf0ac5eb5f215ae73dbdd41bb8600db3ebb40c36f3a11", "code_archive_bytes": 1088188, "code_file_count": 23, "code_extensions": {".py": 23}, "github_disk_usage_kb": 747, "github_languages": {"Python": 193104}, "github_archived": false, "github_pushed_at": "2023-09-05T19:08:49Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/promptboosting-black-box-text-classification"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "FOR2VqgJXb", "year": 2021, "status": "rejected", "title": "Evaluating representations by the complexity of learning low-loss predictors", "authors": ["William F Whitney", "Min Jae Song", "David Brandfonbrener", "Jaan Altosaar", "Kyunghyun Cho"], "authorids": ["~William_F_Whitney1", "~Min_Jae_Song1", "~David_Brandfonbrener1", "~Jaan_Altosaar1", "~Kyunghyun_Cho1"], "authors_source": "OpenReview API", "abstract": "We consider the problem of evaluating representations of data for use in solving a downstream task. We propose to measure the quality of a representation by the complexity of learning a predictor on top of the representation that achieves low loss on a task of interest. To this end, we introduce two measures: surplus description length (SDL) and $\\varepsilon$ sample complexity ($\\varepsilon$SC). To compare our methods to prior work, we also present a framework based on plotting the validation loss versus dataset size (the \"loss-data\" curve). Existing measures, such as mutual information and minimum description length, correspond to slices and integrals along the data-axis of the loss-data curve, while ours correspond to slices and integrals along the loss-axis. This analysis shows that prior methods measure properties of an evaluation dataset of a specified size, whereas our methods measure properties of a predictor with a specified loss. We conclude with experiments on real data to compare the behavior of these methods over datasets of varying size.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=FOR2VqgJXb", "arxiv_id": "2009.07368", "paper_pdf": "papers/FOR2VqgJXb.pdf", "paper_pdf_sha256": "d19be31d60dc849c00836981f1d287e0048530d717db528739d0a9f8518ed9e4", "paper_pdf_bytes": 354076, "paper_pdf_source": "openreview", "code_url": "https://github.com/willwhitney/reprieve", "code_repository": "willwhitney/reprieve", "code_commit": "180a82092c973fca572fad35d4c21f0075764f4a", "code_archive": "repos/FOR2VqgJXb.zip", "code_archive_sha256": "78d929b83bd50dee5b37d5b728d4e52708e901e3f56579a24192ce47c6258298", "code_archive_bytes": 389534, "code_file_count": 16, "code_extensions": {".py": 15, ".ipynb": 1}, "github_disk_usage_kb": 1869, "github_languages": {"Python": 46532}, "github_archived": false, "github_pushed_at": "2021-11-21T15:30:17Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/evaluating-representations-by-the-complexity"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "vnlsxFbWSB", "year": 2026, "status": "rejected", "title": "Bench-CoE: A Framework for Collaboration of Experts from Benchmark", "authors": ["Yuanshuai Wang", "Xingjian Zhang", "Jinkun Zhao", "Siwei Wen", "Peilin Feng", "Shuhao Liao", "Lei Huang", "wenjun wu"], "authorids": ["~Yuanshuai_Wang1", "~Xingjian_Zhang2", "~Jinkun_Zhao1", "~Siwei_Wen5", "~Peilin_Feng1", "~Shuhao_Liao2", "~Lei_Huang1", "~wenjun_wu3"], "authors_source": "OpenReview API", "abstract": "Large Language Models (LLMs) are key technologies that drive intelligent systems to handle multiple tasks. To meet the demands of various tasks, an increasing number of LLMs-driven experts with diverse capabilities have been developed, spreading from language to visual understanding and generalization, accompanied by corresponding benchmarks to evaluate their performance. This paper proposes the Bench-CoE framework, which enables Collaboration of Experts (CoE) by effectively leveraging benchmark evaluations to achieve optimal performance across various tasks. Bench-CoE consists of a set of specialized expert models, a router for assigning tasks to corresponding experts, and a benchmark dataset for training the router. Based on this framework, we first formulate Query-Level Bench-CoE that is an abstraction of existing CoE methods exploiting the benchmark dataset. We further propose Subject-Level Bench-CoE, a new method that effectively addresses the potential issues of Query-Level Bench-CoE in poor generalization and labeling costs during training the router. Experiments show that the Query-Level Bench-CoE excels in in-distribution tasks, while the Subject-Level Bench-CoE demonstrates stronger out-of-distribution generalization. Our proposed Bench-CoE achieves efficient expert collaboration with minimal training label costs, improving adaptability in multi-task and cross-domain scenarios.", "decision": "Reject", "meta_review": "", "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=vnlsxFbWSB", "arxiv_id": "2412.04167", "paper_pdf": "papers/vnlsxFbWSB.pdf", "paper_pdf_sha256": "6f392a8c90df25218c53f6852fb37d0ced17942cf7d9b0f9f2c6b2b1887eb19b", "paper_pdf_bytes": 1047702, "paper_pdf_source": "openreview", "code_url": "https://github.com/ZhangXJ199/Bench-CoE", "code_repository": "ZhangXJ199/Bench-CoE", "code_commit": "fab309d6c397e92d60907ec345a8989043afa04c", "code_archive": "repos/vnlsxFbWSB.zip", "code_archive_sha256": "e28ddcbfbb200506ef93fa136458dc675b2b54a97e0d069b0fb483806dc6bab4", "code_archive_bytes": 7402808, "code_file_count": 127, "code_extensions": {".py": 125, ".sh": 2}, "github_disk_usage_kb": 7522, "github_languages": {"Python": 1060126, "Shell": 1497}, "github_archived": false, "github_pushed_at": "2025-04-27T05:44:00Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/bench-coe-a-framework-for-collaboration-of"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "j9wBgcxa7N", "year": 2025, "status": "rejected", "title": "MAgICoRe: Multi-Agent, Iterative, Coarse-to-Fine Refinement for Reasoning", "authors": ["Justin Chen", "Archiki Prasad", "Swarnadeep Saha", "Elias Stengel-Eskin", "Mohit Bansal"], "authorids": ["~Justin_Chen1", "~Archiki_Prasad1", "~Swarnadeep_Saha2", "~Elias_Stengel-Eskin1", "~Mohit_Bansal2"], "authors_source": "OpenReview API", "abstract": "Large Language Models' (LLM) reasoning can be improved using test-time aggregation strategies, i.e., generating multiple samples for each problem and aggregating over them to find a better answer. While these improve performance, they often reach a saturation point beyond which additional samples provide no return. Refinement offers an alternative by using model-generated feedback to improve answer quality. However, refinement introduces three key challenges: (1) Excessive refinement: Uniformly refining all instances can cause over-correction and reduce overall performance. (2) Inability to localize and address errors: LLMs have a limited ability to self-correct and struggle to identify and correct their own mistakes in a targeted way. (3) Insufficient refinement: Deciding how many iterations of refinement are needed is non-trivial, and stopping too soon could leave errors unaddressed. To tackle these issues, we propose MAgICoRe, a framework for Multi-Agent Iteration for Coarse-to-fine Refinement. MAgICoRe aims to avoid excessive refinement by categorizing problems as easy or hard, solving easy problems with coarse-grained aggregation, and solving hard ones with fine-grained and iterative multi-agent refinement. To enable more granular error localization, we incorporate external step-wise reward model (RM) scores. To ensure effective refinement, we employ a multi-agent loop with three agents: the Solver, the Reviewer (which generates targeted feedback based on step-wise RM scores) and the Refiner (which incorporates feedback and generates new solutions). To ensure sufficient refinement, we re-evaluate updated solutions, iteratively initiating further rounds of multi-agent refinement. We evaluate MAgICoRe on Llama-3-8B and GPT-3.5 and show its effectiveness across five math reasoning datasets, with consistent gains for all datasets and models. Even one iteration of MAgICoRe beats Self-Consistency by 3.4%, Best-of-k by 3.2%, and Self-Refine by 4.0% while using less than 50% of the samples. Unlike iterative refinement with baseline methods, MAgICoRe continues to improve with more iterations. Finally, our ablations highlight the importance of MAgICoRe's use of RMs and multi-agent communication.", "decision": "Reject", "meta_review": "", "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=j9wBgcxa7N", "arxiv_id": "2409.12147", "paper_pdf": "papers/j9wBgcxa7N.pdf", "paper_pdf_sha256": "9845342614e4bfeb7d8cd67c48a4fea6f5b086f81f5505a26cf94c973efb14bf", "paper_pdf_bytes": 929378, "paper_pdf_source": "openreview", "code_url": "https://github.com/dinobby/MAgICoRE", "code_repository": "dinobby/MAgICoRE", "code_commit": "1388006b309663da69b142102d819d3be4d34890", "code_archive": "repos/j9wBgcxa7N.zip", "code_archive_sha256": "de87d379bd0bb1ca31f895558bf5c3450f82737d25a5de7242f137690d3b266e", "code_archive_bytes": 306128, "code_file_count": 8, "code_extensions": {".py": 8}, "github_disk_usage_kb": 365, "github_languages": {"Python": 48065}, "github_archived": false, "github_pushed_at": "2024-09-19T06:17:44Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/magicore-multi-agent-iterative-coarse-to-fine"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "0aEUd9UtiA", "year": 2024, "status": "rejected", "title": "DiffCPS: Diffusion Model based Constrained Policy Search for Offline Reinforcement Learning", "authors": ["Longxiang He", "Linrui Zhang", "Junbo Tan", "Xueqian Wang"], "authorids": ["~Longxiang_He2", "~Linrui_Zhang1", "~Junbo_Tan1", "~Xueqian_Wang1"], "authors_source": "OpenReview API", "abstract": "Constrained policy search (CPS) is a fundamental problem in offline reinforcement learning, which is generally solved by advantage weighted regression (AWR). However, previous methods may still encounter out-of-distribution actions due to the limited expressivity of Gaussian-based policies.  On the other hand, directly applying the state-of-the-art models with distribution expression capabilities (i.e., diffusion models) in the AWR framework is insufficient since AWR requires exact policy probability densities, which is intractable in diffusion models. In this paper, we propose a novel approach called $\\textbf{Diffusion Model based Constrained Policy Search (DiffCPS)}$, which tackles the diffusion-based constrained policy search without resorting to AWR.  The theoretical analysis reveals our key insights by leveraging the action distribution of the diffusion model to eliminate the policy distribution constraint in the CPS and then utilizing the Evidence Lower Bound (ELBO) of diffusion-based policy to approximate the KL constraint. Consequently, DiffCPS admits the high expressivity of diffusion models while circumventing the cumbersome density calculation brought by AWR. Extensive experimental results based on the D4RL benchmark demonstrate the efficacy of our approach. We empirically show that DiffCPS achieves better or at least competitive performance compared to traditional AWR-based baselines as well as recent diffusion-based offline RL methods. Code will be made publicly available upon acceptance.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=0aEUd9UtiA", "arxiv_id": "2310.05333", "paper_pdf": "papers/0aEUd9UtiA.pdf", "paper_pdf_sha256": "22c566f0eb1b3e96ce15bf9a5f87e9c6540f3de10d854215ef528987d829f16a", "paper_pdf_bytes": 2207264, "paper_pdf_source": "openreview", "code_url": "https://github.com/felix-thu/DiffCPS", "code_repository": "felix-thu/DiffCPS", "code_commit": "2d8de2878bdc2bb2f59bbb9f7ce929897cf7cb15", "code_archive": "repos/0aEUd9UtiA.zip", "code_archive_sha256": "3f7a18f629a0c2ad024c463a3080b848785bffd3208d275cdfd4a5f3d19db200", "code_archive_bytes": 332031, "code_file_count": 11, "code_extensions": {".py": 11}, "github_disk_usage_kb": 329, "github_languages": {"Python": 89657}, "github_archived": false, "github_pushed_at": "2024-09-09T16:28:24Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/diffcps-diffusion-model-based-constrained"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "YjKqWExiy6s", "year": 2023, "status": "rejected", "title": "Eliminating Catastrophic Overfitting Via Abnormal Adversarial Examples Regularization", "authors": ["Runqi Lin", "Chaojian Yu", "Tongliang Liu"], "authorids": ["~Runqi_Lin1", "~Chaojian_Yu1", "~Tongliang_Liu1"], "authors_source": "OpenReview API", "abstract": "Single-step adversarial training (SSAT) is shown to be able to defend against iterative-step adversarial attacks to achieve both efficiency and robustness. However, SSAT suffers from catastrophic overfitting (CO) with strong adversaries, showing that the classifier decision boundaries are highly distorted and robust accuracy against iterative-step adversarial attacks suddenly drops from peak to nearly 0% in a few epochs. In this work, we find that some adversarial examples generated on the network trained by SSAT exhibit anomalous behaviour, that is, although the training data is generated by the inner maximization process, the loss of some adversarial examples decreases instead, which we called abnormal adversarial examples. Furthermore, network optimization on these abnormal adversarial examples will further accelerate the model decision boundaries distortion, and correspondingly, the number of abnormal adversarial examples will sharply increase with CO. These observations motivate us to prevent CO by hindering the generation of abnormal adversarial examples. Specifically, we design a novel method, Abnormal Adversarial Examples Regularization (AAER), which explicitly regularizes the number and logits variation of abnormal adversarial examples to hinder the model from generating abnormal adversarial examples. Extensive experiments demonstrate that our method can prevent CO and further boost adversarial robustness with strong adversaries.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=YjKqWExiy6s", "arxiv_id": "2404.08154", "paper_pdf": "papers/YjKqWExiy6s.pdf", "paper_pdf_sha256": "75560f0a0429313afd4fdb7a50f72d05a1d31dbc36334b364aa475a022705a17", "paper_pdf_bytes": 2198026, "paper_pdf_source": "openreview", "code_url": "https://github.com/tmllab/2023_NeurIPS_AAER", "code_repository": "tmllab/2023_NeurIPS_AAER", "code_commit": "5ad2591c1f9be9e797420f713aa45a3e288da269", "code_archive": "repos/YjKqWExiy6s.zip", "code_archive_sha256": "91cd60175e8c2823362921813b3f1faa6fcdd70e39e4102b1cd25c4e56074800", "code_archive_bytes": 773838, "code_file_count": 21, "code_extensions": {".py": 21}, "github_disk_usage_kb": 767, "github_languages": {"Python": 112075}, "github_archived": false, "github_pushed_at": "2025-02-15T07:05:41Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/eliminating-catastrophic-overfitting-via-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "9Cwxjd6nRh", "year": 2022, "status": "rejected", "title": "High Fidelity Visualization of What Your Self-Supervised Representation Knows About", "authors": ["Florian Bordes", "Randall Balestriero", "Pascal Vincent"], "authorids": ["~Florian_Bordes1", "~Randall_Balestriero1", "~Pascal_Vincent1"], "authors_source": "OpenReview API", "abstract": "Discovering what is learned by neural networks remains a challenge. In self-supervised learning, classification is the most common task used to evaluate how good a representation is. However, relying only on such downstream task can limit our understanding of how much information is contained in the representation of a given input. In this work, we study how to visualize representations learned with self-supervised models. We investigate a simple gradient descend based method to match a target representation and show the limitations of such techniques. We overcome these limitations by developing a representation-conditioned diffusion model (RCDM) that is able to generate high-quality inputs that share commonalities with a given representation. We further demonstrate how our model's generation quality is on par with state-of-the-art generative models and how the representation conditioning brings new avenues to analyze and improve self-supervised models.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=9Cwxjd6nRh", "arxiv_id": "2112.09164", "paper_pdf": "papers/9Cwxjd6nRh.pdf", "paper_pdf_sha256": "e2819039495fa338837ee48e0f35ca1bb67dbdb02f8d45d0cca8106991b64b60", "paper_pdf_bytes": 44596143, "paper_pdf_source": "openreview", "code_url": "https://github.com/facebookresearch/RCDM", "code_repository": "facebookresearch/RCDM", "code_commit": "71daaf10a73bb2012864f0827c68d209fc92b0a5", "code_archive": "repos/9Cwxjd6nRh.zip", "code_archive_sha256": "4778ee2854148947cb035b8eb56a6e92c898e8db70648271145fb82e02e8b9c2", "code_archive_bytes": 3425002, "code_file_count": 28, "code_extensions": {".py": 28}, "github_disk_usage_kb": 3345, "github_languages": {"Python": 255010}, "github_archived": true, "github_pushed_at": "2023-05-03T20:06:06Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/high-fidelity-visualization-of-what-your-self-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ohdw3t-8VCY", "year": 2021, "status": "rejected", "title": "CTRLsum: Towards Generic Controllable Text Summarization", "authors": ["Junxian He", "Wojciech Maciej Kryscinski", "Bryan McCann", "Nazneen Rajani", "Caiming Xiong"], "authorids": ["~Junxian_He1", "~Wojciech_Maciej_Kryscinski1", "~Bryan_McCann1", "~Nazneen_Rajani1", "~Caiming_Xiong1"], "authors_source": "OpenReview API", "abstract": "Current summarization systems yield generic summaries that are disconnected from users' preferences and expectations. To address this limitation, we present CTRLsum, a novel framework for controllable summarization. Our approach enables users to control multiple aspects of generated summaries by interacting with the summarization system through textual input in the form of a set of keywords or descriptive prompts. Using a single unified model, CTRLsum is able to achieve a broad scope of summary manipulation at inference time without requiring additional human annotations or pre-defining a set of control aspects during training. We quantitatively demonstrate the effectiveness of our approach on three domains of summarization datasets and five control aspects: 1) entity-centric and 2) length-controllable summarization, 3) contribution summarization on scientific papers, 4) invention purpose summarization on patent filings, and 5) question-guided summarization on news articles in a reading comprehension setting. Moreover, when used in a standard, uncontrolled summarization setting, CTRLsum achieves state-of-the-art results on the CNN/DailyMail dataset.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=ohdw3t-8VCY", "arxiv_id": "2012.04281", "paper_pdf": "papers/ohdw3t-8VCY.pdf", "paper_pdf_sha256": "7df38c89ff91288cef0ac6eb33396b248fb88620209b215048792980dddb3d44", "paper_pdf_bytes": 387961, "paper_pdf_source": "openreview", "code_url": "https://github.com/salesforce/ctrl-sum", "code_repository": "salesforce/ctrl-sum", "code_commit": "01c1a6cf0e286346321829ea88d30c85fbede0ca", "code_archive": "repos/ohdw3t-8VCY.zip", "code_archive_sha256": "32126739d7e095fbd94c9be1414ee068b5334bfc5025cecf6029354994e8d363", "code_archive_bytes": 1831724, "code_file_count": 23, "code_extensions": {".py": 15, ".sh": 8}, "github_disk_usage_kb": 1877, "github_languages": {"Python": 174890, "Shell": 15348}, "github_archived": true, "github_pushed_at": "2025-05-01T17:29:00Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/ctrlsum-towards-generic-controllable-text-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "YBgjDBYPzz", "year": 2026, "status": "rejected", "title": "Top-Theta Attention: Sparsifying Transformers by Compensated Thresholding", "authors": ["Konstantin Berestizshevsky", "Renzo Andri", "Lukas Cavigelli"], "authorids": ["~Konstantin_Berestizshevsky1", "~Renzo_Andri1", "~Lukas_Cavigelli1"], "authors_source": "OpenReview API", "abstract": "We present Top-Theta (Top-θ) Attention, a training-free method for sparsifying transformer attention during inference. Our key insight is that static, per-head thresholds can be calibrated to retain the desired constant number of significant elements per attention row. This approach enables content-based sparsity without retraining, and it remains robust across data domains. We further introduce compensation techniques to preserve accuracy under aggressive sparsification, establishing attention thresholding as a practical and principled alternative to top-k attention. We provide extensive evaluation on natural language processing tasks, showing that Top-θ achieves 3-10x reduction in $V$-cache usage and up to 10x fewer attention elements during inference while degrading no more than 1% in accuracy.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=YBgjDBYPzz", "arxiv_id": "2502.08363", "paper_pdf": "papers/YBgjDBYPzz.pdf", "paper_pdf_sha256": "26fdc80d8cfeb1fc700da2f9aaffb44a0c5a4f84628ac1acf7ea523970f00ebf", "paper_pdf_bytes": 4913761, "paper_pdf_source": "openreview", "code_url": "https://github.com/huawei-csl/top-theta-attention", "code_repository": "huawei-csl/top-theta-attention", "code_commit": "fa3ab6889137994cd393162ff8236993f3391053", "code_archive": "repos/YBgjDBYPzz.zip", "code_archive_sha256": "0debb1e23d3363604f7f959602e361dffd9739123b68c86f148e94f70266926d", "code_archive_bytes": 5863761, "code_file_count": 25, "code_extensions": {".py": 17, ".ipynb": 6, ".sh": 2}, "github_disk_usage_kb": 7596, "github_languages": {"Jupyter Notebook": 6079920, "Python": 351171, "Shell": 3266}, "github_archived": false, "github_pushed_at": "2026-06-04T11:37:22Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/top-theta-attention-sparsifying-transformers"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "OANUpvmnuf", "year": 2025, "status": "rejected", "title": "Choices are More Important than Efforts: LLM Enables Efficient Multi-Agent Exploration", "authors": ["Yun Qu", "Boyuan Wang", "Yuhang Jiang", "Jianzhun Shao", "Yixiu Mao", "Chang Liu", "Cheems Wang", "Xiangyang Ji"], "authorids": ["~Yun_Qu2", "~Boyuan_Wang1", "~Yuhang_Jiang3", "~Jianzhun_Shao1", "~Yixiu_Mao2", "~Chang_Liu9", "~Cheems_Wang1", "~Xiangyang_Ji1"], "authors_source": "OpenReview API", "abstract": "With expansive state-action spaces, efficient multi-agent exploration remains a longstanding challenge in reinforcement learning.\nAlthough pursuing novelty, diversity, or uncertainty attracts increasing attention, redundant efforts brought by exploration without proper guidance choices poses a practical issue for the community.\nThis paper introduces a systematic approach, termed LEMAE, choosing to channel informative task-relevant guidance from a knowledgeable Large Language Model (LLM) for Efficient Multi-Agent Exploration. \nSpecifically, we ground linguistic knowledge from LLM into symbolic key states, that are critical for task fulfillment, in a discriminative manner at low LLM inference costs. \nTo unleash the power of key states, \nwe design Subspace-based Hindsight Intrinsic Reward (SHIR) to guide agents toward key states by increasing reward density.  Additionally, we build the Key State Memory Tree (KSMT) to track transitions between key states in a specific task for organized exploration. Benefiting from diminishing redundant explorations, LEMAE outperforms existing SOTA approaches on the challenging benchmarks (e.g., SMAC and MPE) by a large margin, achieving a 10x acceleration in certain scenarios.\nOur code is available at https://anonymous.4open.science/r/LEMAE.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=OANUpvmnuf", "arxiv_id": "2410.02511", "paper_pdf": "papers/OANUpvmnuf.pdf", "paper_pdf_sha256": "cc3a22bb34cd614ab808aa296535ef3e270687752fd0c153436cad4d978bea5b", "paper_pdf_bytes": 6752884, "paper_pdf_source": "openreview", "code_url": "https://github.com/hijkzzz/pymarl2", "code_repository": "hijkzzz/pymarl2", "code_commit": "8ccac7c5aa134422a2e3009be735d23cdf8ce2f8", "code_archive": "repos/OANUpvmnuf.zip", "code_archive_sha256": "240e13e2e93cc3360a458d7589b872e6143e2424c3160892b3bcb2288ac8057a", "code_archive_bytes": 350311, "code_file_count": 81, "code_extensions": {".py": 76, ".sh": 5}, "github_disk_usage_kb": 373, "github_languages": {"Python": 374490, "Shell": 3103}, "github_archived": false, "github_pushed_at": "2024-05-18T03:22:54Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/choices-are-more-important-than-efforts-llm"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "q38SZkUmUh", "year": 2024, "status": "rejected", "title": "FreshLLMs: Refreshing Large Language Models with Search Engine Augmentation", "authors": ["Tu Vu", "Mohit Iyyer", "Xuezhi Wang", "Noah Constant", "Jerry Wei", "Jason Wei", "Chris Tar", "Yun-Hsuan Sung", "Denny Zhou", "Quoc V Le", "Thang Luong"], "authorids": ["~Tu_Vu1", "~Mohit_Iyyer1", "~Xuezhi_Wang3", "~Noah_Constant1", "~Jerry_Wei1", "~Jason_Wei1", "ctar@google.com", "~Yun-Hsuan_Sung1", "~Denny_Zhou1", "~Quoc_V_Le1", "~Thang_Luong1"], "authors_source": "OpenReview API", "abstract": "Most large language models (LLMs) are trained once and never updated; thus, they lack the ability to dynamically adapt to our ever-changing world.  In this work, we perform a detailed study of the factuality of LLM-generated text in the context of answering questions that test current world knowledge. Specifically, we introduce FreshQA, a novel dynamic QA benchmark encompassing a diverse range of question and answer types, including questions that require fast-changing world knowledge as well as questions with false premises that need to be debunked. We benchmark a diverse array of both closed and open-source LLMs under a two-mode evaluation procedure that allows us to measure both correctness and hallucination. Through human evaluations involving more than 50K judgments, we shed light on limitations of these models and demonstrate significant room for improvement: for instance, all models (regardless of model size) struggle on questions that involve fast-changing knowledge and false premises. Motivated by these results, we present FreshPrompt, a simple few-shot prompting method that substantially boosts the performance of an LLM on FreshQA by incorporating relevant and up-to-date information retrieved from a search engine into the prompt.  Our experiments show that FreshPrompt outperforms both competing search engine-augmented prompting methods such as Self-Ask (Press et al., 2022) as well as commercial systems such as Perplexity AI. Further analysis of FreshPrompt reveals that both the number of retrieved evidences and their order play a key role in influencing the correctness of LLM-generated answers.  Additionally, instructing the LLM to generate concise and direct answers helps reduce hallucination compared to encouraging more verbose answers. To facilitate future research, we will release FreshQA after blind review and commit to updating it at regular intervals.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=q38SZkUmUh", "arxiv_id": "2310.03214", "paper_pdf": "papers/q38SZkUmUh.pdf", "paper_pdf_sha256": "65c3af0884306b6f5d750991b09bb4f6c1b6f6846917abf32af4009193544643", "paper_pdf_bytes": 2172992, "paper_pdf_source": "openreview", "code_url": "https://github.com/freshllms/freshqa", "code_repository": "freshllms/freshqa", "code_commit": "7d2d3683991916f3633e480548a6aa5c9a62e3db", "code_archive": "repos/q38SZkUmUh.zip", "code_archive_sha256": "7eb5ace9d7d9af0a39c1b3e4cf3e90c7f87bee1c39c1b1e61543e54c7ec03bee", "code_archive_bytes": 39840, "code_file_count": 3, "code_extensions": {".ipynb": 3}, "github_disk_usage_kb": 343, "github_languages": {"Jupyter Notebook": 146459}, "github_archived": false, "github_pushed_at": "2026-05-01T22:43:08Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/freshllms-refreshing-large-language-models"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "qhplAU1BOZW", "year": 2023, "status": "rejected", "title": "Lottery Aware Sparsity Hunting: Enabling Federated Learning on Resource-Limited Edge", "authors": ["Sara Babakniya", "Souvik Kundu", "Saurav Prakash", "Yue Niu", "Salman Avestimehr"], "authorids": ["~Sara_Babakniya1", "~Souvik_Kundu2", "~Saurav_Prakash1", "~Yue_Niu1", "~Salman_Avestimehr1"], "authors_source": "OpenReview API", "abstract": "Limited computation and communication capabilities of clients pose significant challenges in federated learning (FL) over resource-limited edge nodes. A potential solution to this problem is to deploy off-the-shelf sparse learning algorithms that train a binary sparse mask on each client with the expectation of training a consistent sparse server mask yielding sparse weight tensors. However, as we investigate in this paper, such naive deployments result in a significant drop in accuracy compared to FL with dense models, especially for clients with limited resource budgets. In particular, our investigations reveal a serious lack of consensus among the trained sparsity masks on clients, which prevents convergence for the server mask and potentially leads to a substantial drop in model performance. Based on such key observations, we propose federated lottery aware sparsity hunting (FLASH), a unified sparse learning framework to make the server win a lottery in terms of yielding a sparse sub-model, able to maintain classification performance under highly resource-limited client settings. Moreover, to support FL on different devices requiring different parameter density, we leverage our findings to present hetero-FLASH, where clients can have different target sparsity budgets based on their device resource limits. Experimental evaluations with multiple models on various datasets (both IID and non-IID) show superiority of our models in closing the gap with unpruned baseline while yielding up to ∼10.1% improved accuracy with ∼10.26x fewer communication costs, compared to existing alternatives, at similar hyperparameter settings.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=qhplAU1BOZW", "arxiv_id": "2208.13092", "paper_pdf": "papers/qhplAU1BOZW.pdf", "paper_pdf_sha256": "88534b715314720914f09d213eea012d2e13dde7fbf7f5c6beff0521bf7d95b9", "paper_pdf_bytes": 16715668, "paper_pdf_source": "openreview", "code_url": "https://github.com/SaraBabakN/flash_fl", "code_repository": "SaraBabakN/flash_fl", "code_commit": "fb883133c48490e7d4e3f21cce4212a27f618a22", "code_archive": "repos/qhplAU1BOZW.zip", "code_archive_sha256": "976ce2eb95014d02e09f63ed5d9f3c4a4b3a58971bc6ffb89ddc176e701b4e7a", "code_archive_bytes": 219002, "code_file_count": 13, "code_extensions": {".py": 13}, "github_disk_usage_kb": 824, "github_languages": {"Python": 66195}, "github_archived": false, "github_pushed_at": "2023-11-04T07:06:51Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/federated-sparse-training-lottery-aware-model"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "u2JeVfXIQa", "year": 2022, "status": "rejected", "title": "Adaptive Cross-Layer Attention for Image Restoration", "authors": ["Yancheng Wang", "Yingzhen Yang", "Chong Chen", "Ning Xu"], "authorids": ["~Yancheng_Wang2", "~Yingzhen_Yang1", "chongchen@kuaishou.com", "~Ning_Xu3"], "authors_source": "OpenReview API", "abstract": "Non-local attention module has been proven to be crucial for image restoration. Conventional non-local attention processes features of each layer separately, so it risks missing correlation between features among different layers. To address this problem, we propose Cross-Layer Attention (CLA) module in this paper. Instead of finding correlated key pixels within the same layer, each query pixel is allowed to attend to key pixels at previous layers of the network. In order to mitigate the expensive computational cost of such hierarchical attention design, only a small fixed number of keys can be selected for each query from a previous layer. We further propose a variant of CLA termed Adaptive Cross-Layer Attention (ACLA). In ACLA, the number of keys to be aggregated for each query is dynamically selected. A neural architecture search method is used to find the insert positions of ACLA modules to render a compact neural network with compelling performance. Extensive experiments on image restoration tasks including single image super-resolution, image denoising, image demosaicing, and image compression artifacts reduction validate the effectiveness and efficiency of ACLA.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=u2JeVfXIQa", "arxiv_id": "2203.03619", "paper_pdf": "papers/u2JeVfXIQa.pdf", "paper_pdf_sha256": "f0a2f5a94fab68e20d1b7273dd910d759ca4cf17af372ffa9be9d2d07658ded8", "paper_pdf_bytes": 1851316, "paper_pdf_source": "openreview", "code_url": "https://github.com/Statistical-Deep-Learning/ACLA-IKS", "code_repository": "Statistical-Deep-Learning/ACLA-IKS", "code_commit": "aaefb77dcc44ba81d06c13b3d12db2b641c15273", "code_archive": "repos/u2JeVfXIQa.zip", "code_archive_sha256": "3978e80dada063057fd8071aaf83a086103c758d428e01b6cda787227e2a7847", "code_archive_bytes": 3785892, "code_file_count": 160, "code_extensions": {".py": 127, ".sh": 18, ".ipynb": 6, ".h": 3, ".cpp": 2, ".m": 2, ".cu": 1, ".cuh": 1}, "github_disk_usage_kb": 3434, "github_languages": {"Python": 490756, "Cuda": 62010, "Jupyter Notebook": 52308, "MATLAB": 17294, "Shell": 15601, "C++": 6172}, "github_archived": false, "github_pushed_at": "2025-11-06T23:50:10Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/adaptive-cross-layer-attention-for-image-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "nPVlVsBTiJ", "year": 2021, "status": "rejected", "title": "Adversarial Boot Camp: label free certified robustness in one epoch", "authors": ["Ryan Campbell", "Chris Finlay", "Adam M Oberman"], "authorids": ["~Ryan_Campbell2", "~Chris_Finlay1", "~Adam_M_Oberman1"], "authors_source": "OpenReview API", "abstract": "Machine learning models are vulnerable to adversarial attacks.  One approach to addressing this vulnerability is certification, which focuses on models that are guaranteed to be robust for a given perturbation size.  A drawback of recent certified models is that they are stochastic: they require multiple computationally expensive model evaluations with random noise added to a given image. In our work, we present a deterministic certification approach which results in a certifiably robust model. This approach is based on an equivalence between training with a particular regularized loss, and the expected values of Gaussian averages. We achieve certified models on ImageNet-1k by retraining a model with this loss for one epoch without the use of label information.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=nPVlVsBTiJ", "arxiv_id": "2010.02508", "paper_pdf": "papers/nPVlVsBTiJ.pdf", "paper_pdf_sha256": "f9ca0805c0c46466ecec5c73facc7a969fec08078581600b7d57376d44843cf8", "paper_pdf_bytes": 785534, "paper_pdf_source": "openreview", "code_url": "https://github.com/ryancampbell514/HeatSmoothing", "code_repository": "ryancampbell514/HeatSmoothing", "code_commit": "a67b3d857ea6bb37c34c6ce591f7626052278035", "code_archive": "repos/nPVlVsBTiJ.zip", "code_archive_sha256": "f03dd8d49c3585c0ceef82f4cddec9c186a02444c496b4e312a3d494c5aeb38c", "code_archive_bytes": 562492, "code_file_count": 52, "code_extensions": {".py": 47, ".sh": 3, ".ipynb": 2}, "github_disk_usage_kb": 1988, "github_languages": {"Python": 286446, "Jupyter Notebook": 114568, "Shell": 4949}, "github_archived": false, "github_pushed_at": "2020-10-08T17:32:43Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/adversarial-boot-camp-label-free-certified-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "RYwtJyOP3k", "year": 2026, "status": "rejected", "title": "Probably Approximately Correct Labels", "authors": ["Emmanuel Candes", "Andrew Ilyas", "Tijana Zrnic"], "authorids": ["~Emmanuel_Candes1", "~Andrew_Ilyas1", "~Tijana_Zrnic1"], "authors_source": "OpenReview API", "abstract": "Obtaining high-quality labeled datasets is often costly, requiring either\nhuman annotation or expensive experiments. \nIn theory, powerful pre-trained AI models provide an opportunity to \nautomatically label datasets and save costs. \nUnfortunately, these models provide no guarantees on their accuracy, \nmaking wholesale replacement of manual labeling impractical.\nIn this work, we propose a method for leveraging pre-trained AI models to curate \ncost-effective and high-quality datasets.\nIn particular, our approach results in\n*probably approximately correct labels*: with high probability, the overall\nlabeling error is small. \nOur method is nonasymptotically valid under minimal assumptions on the dataset or\nthe AI model being studied, and thus enables rigorous yet efficient dataset\ncuration using modern AI models. We demonstrate the benefits of the methodology\nthrough text annotation with large language models, image labeling with\npre-trained vision models, and protein folding analysis with AlphaFold.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=RYwtJyOP3k", "arxiv_id": "2506.10908", "paper_pdf": "papers/RYwtJyOP3k.pdf", "paper_pdf_sha256": "f0e07a02481cec4b5b435452eb445f152afc99067218518f8d89203bffc1fbb0", "paper_pdf_bytes": 1030000, "paper_pdf_source": "openreview", "code_url": "https://github.com/tijana-zrnic/pac-labels", "code_repository": "tijana-zrnic/pac-labels", "code_commit": "b415b58756b14b384529ac9cf146bd5d4c8139aa", "code_archive": "repos/RYwtJyOP3k.zip", "code_archive_sha256": "38020d18668956f268bd2520397262c99c7ed7858bec0e9a31160cbcbf512dbc", "code_archive_bytes": 12267016, "code_file_count": 6, "code_extensions": {".ipynb": 3, ".py": 3}, "github_disk_usage_kb": 7615, "github_languages": {"Jupyter Notebook": 28007, "Python": 20683}, "github_archived": false, "github_pushed_at": "2025-06-13T06:30:15Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/probably-approximately-correct-labels"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "leSbzBtofH", "year": 2025, "status": "rejected", "title": "AutoAdvExBench: Benchmarking Autonomous Exploitation of Adversarial Example Defenses", "authors": ["Nicholas Carlini", "Edoardo Debenedetti", "Javier Rando", "Milad Nasr", "Florian Tramèr"], "authorids": ["~Nicholas_Carlini1", "~Edoardo_Debenedetti1", "~Javier_Rando2", "~Milad_Nasr2", "~Florian_Tramèr1"], "authors_source": "OpenReview API", "abstract": "We introduce AutoAdvExBench, a benchmark to evaluate if large language models (LLMs)\ncan autonomously exploit defenses to adversarial examples.\nWe believe our benchmark will be valuable to several distinct audiences. \nFirst, it measures if models can match the abilities of expert adversarial machine learning researchers.\nSecond, it serves as a challenging evaluation for reasoning capabilities that\ncan measure LLMs' ability to understand and interact with sophisticated codebases. \nAnd third, \nsince many adversarial examples defenses have been broken in the past,\nthis benchmark allows for evaluating the ability of LLMs to reproduce\nprior research results automatically.\nWe then benchmark the ability of current LLMs to solve this benchmark,\nand find most are unable to succeed.\nOur strongest agent, with a human-guided prompt,\nis only able to successfully generate adversarial examples on 6 of the 51 defenses in our benchmark.\nThis benchmark is publicly accessible at redacted for review.", "decision": "Reject", "meta_review": "", "num_reviews": 6, "openreview_url": "https://openreview.net/forum?id=leSbzBtofH", "arxiv_id": "2503.01811", "paper_pdf": "papers/leSbzBtofH.pdf", "paper_pdf_sha256": "3dbbb97e83361880472f876f10dd59a6ada4d4c4f1e0068eab4276bd9ec1dea3", "paper_pdf_bytes": 250543, "paper_pdf_source": "openreview", "code_url": "https://github.com/ethz-spylab/autoadvexbench", "code_repository": "ethz-spylab/autoadvexbench", "code_commit": "8da21de3aa3442b339b519a0f5c82b54905efbf1", "code_archive": "repos/leSbzBtofH.zip", "code_archive_sha256": "79ec0284af77407ac1cdfda75e7e82e5b0743baafc953763d99a853a743c6e06", "code_archive_bytes": 397012, "code_file_count": 32, "code_extensions": {".py": 31, ".js": 1}, "github_disk_usage_kb": 374, "github_languages": {"Python": 93972, "HTML": 21117}, "github_archived": false, "github_pushed_at": "2025-05-21T18:36:47Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/2503-01811"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "XJiN1VkgA0", "year": 2024, "status": "rejected", "title": "Generating with Confidence: Uncertainty Quantification for Black-box Large Language Models", "authors": ["Zhen Lin", "Shubhendu Trivedi", "Jimeng Sun"], "authorids": ["~Zhen_Lin2", "~Shubhendu_Trivedi2", "~Jimeng_Sun3"], "authors_source": "OpenReview API", "abstract": "Large language models (LLMs) specializing in natural language generation (NLG) have recently started exhibiting promising capabilities across a variety of domains. However, gauging the trustworthiness of responses generated by LLMs remains an open challenge, with limited research on uncertainty quantification (UQ) for NLG. Furthermore, existing literature typically assumes white-box access to language models, which is becoming unrealistic either due to the closed-source nature of the latest LLMs or computational constraints. In this work, we investigate UQ in NLG for *black-box* LLMs. We first differentiate *uncertainty* vs *confidence*: the former refers to the “dispersion” of the potential predictions for a fixed input, and the latter refers to the confidence on a particular prediction/generation. We then propose and compare several confidence/uncertainty metrics, applying them to *selective* NLG where unreliable results could either be ignored or yielded for further assessment. Experiments were carried out with several popular LLMs on question-answering datasets (for evaluation purposes). Results reveal that a simple metric for the semantic dispersion can be a reliable predictor of the quality of LLM responses, providing valuable insights for practitioners on uncertainty management when adopting LLMs.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=XJiN1VkgA0", "arxiv_id": "2305.19187", "paper_pdf": "papers/XJiN1VkgA0.pdf", "paper_pdf_sha256": "6fe887b68b1544c5ac4a5f576b66d4939c7ec6e27360566492527c8d9a52fcb0", "paper_pdf_bytes": 711403, "paper_pdf_source": "openreview", "code_url": "https://github.com/zlin7/UQ-NLG", "code_repository": "zlin7/UQ-NLG", "code_commit": "ecaea91741a6c1076e9069b56d32a9929eea8ff9", "code_archive": "repos/XJiN1VkgA0.zip", "code_archive_sha256": "b8914f75d45f8be4a7828b14affe9f54ec8a93cf61ec67c5f864569714605a56", "code_archive_bytes": 346125, "code_file_count": 18, "code_extensions": {".py": 16, ".ipynb": 2}, "github_disk_usage_kb": 345, "github_languages": {"Jupyter Notebook": 262628, "Python": 97829}, "github_archived": false, "github_pushed_at": "2024-06-30T21:12:23Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/generating-with-confidence-uncertainty"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "pcBJT4bgbpH", "year": 2023, "status": "rejected", "title": "Attention Flows for General Transformers", "authors": ["Niklas Metzger", "Christopher Hahn", "Julian Siber", "Frederik Schmitt", "Bernd Finkbeiner"], "authorids": ["~Niklas_Metzger1", "~Christopher_Hahn1", "~Julian_Siber1", "~Frederik_Schmitt1", "~Bernd_Finkbeiner1"], "authors_source": "OpenReview API", "abstract": "In this paper, we study the computation of how much an input token in a Transformer model influences its prediction. We formalize a method to construct a flow network out of the attention values of encoder-only Transformer models and extend it to general Transformer architectures, including an auto-regressive decoder. We show that running a maxflow algorithm on the flow network construction yields Shapley values, which determine a player's impact in cooperative game theory. By interpreting the input tokens in the flow network as players, we can compute their influence on the total attention flow leading to the decoder's decision. Additionally, we provide a library that computes and visualizes the attention flow of arbitrary Transformer models. We show the usefulness of our implementation on various models trained on natural language processing and reasoning tasks.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=pcBJT4bgbpH", "arxiv_id": "2205.15389", "paper_pdf": "papers/pcBJT4bgbpH.pdf", "paper_pdf_sha256": "8e3695108d43694082a958ca8e0c65057752efe072b63ed8bf478366b2821839", "paper_pdf_bytes": 8711270, "paper_pdf_source": "openreview", "code_url": "https://github.com/reactive-systems/ml2", "code_repository": "reactive-systems/ml2", "code_commit": "33d9696c94de6d27aa836ae8118118a7277ff35c", "code_archive": "repos/pcBJT4bgbpH.zip", "code_archive_sha256": "0c7557ed1e26ad1408b973c8a1f0e8f05314334a3d659f408e52f6f83f174a4a", "code_archive_bytes": 556930, "code_file_count": 436, "code_extensions": {".py": 411, ".ipynb": 23, ".sh": 2}, "github_disk_usage_kb": 833, "github_languages": {"Python": 1368065, "Jupyter Notebook": 116581, "Dockerfile": 15364, "Shell": 2114}, "github_archived": false, "github_pushed_at": "2025-04-01T14:13:17Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/attention-flows-for-general-transformers"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "TQ75Md-FqQp", "year": 2022, "status": "rejected", "title": "Efficient and Modular Implicit Differentiation", "authors": ["Mathieu Blondel", "Quentin Berthet", "marco cuturi", "Roy Frostig", "Stephan Hoyer", "Felipe Llinares-López", "Fabian Pedregosa", "Jean-Philippe Vert"], "authorids": ["~Mathieu_Blondel1", "~Quentin_Berthet2", "~marco_cuturi2", "~Roy_Frostig1", "~Stephan_Hoyer1", "~Felipe_Llinares-López1", "~Fabian_Pedregosa1", "~Jean-Philippe_Vert1"], "authors_source": "OpenReview API", "abstract": "Automatic differentiation (autodiff) has revolutionized machine learning.  It allows expressing complex computations by composing elementary ones in creative ways and removes the tedious burden of computing their derivatives by hand. More recently, differentiation of optimization problem solutions has attracted a great deal of research, with applications as a layer in a neural network, and in bi-level optimization, including hyper-parameter optimization. However, the formulae for these derivatives often involves a tedious manual derivation and implementation. In this paper, we propose a unified, efficient and modular approach for implicit differentiation of optimization problems. In our approach, the user defines directly in Python a function $F$ capturing the optimality conditions of the problem to be differentiated. Once this is done, we leverage autodiff of $F$ to automatically differentiate the optimization problem. This way, our approach combines the benefits of implicit differentiation and autodiff.  We show that seemingly simple principles allow to recover all recently proposed implicit differentiation methods and create new ones easily. We describe in details a JAX implementation of our framework and demonstrate the ease of differentiating through optimization problems thanks to it on four diverse tasks: hyperparameter optimization of multiclass SVMs, dataset distillation, task-driven dictionary learning and sensitivity analysis of molecular dynamics.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=TQ75Md-FqQp", "arxiv_id": "2105.15183", "paper_pdf": "papers/TQ75Md-FqQp.pdf", "paper_pdf_sha256": "c80fc1805f72a0ce8f7321dc71ace7191bda8ab64b2d513f37abfdf1e27a52f5", "paper_pdf_bytes": 2202952, "paper_pdf_source": "openreview", "code_url": "https://github.com/google/jaxopt", "code_repository": "google/jaxopt", "code_commit": "176b13830bc552236b32a4c3c4654f196c8b4cd6", "code_archive": "repos/TQ75Md-FqQp.zip", "code_archive_sha256": "97c9a4807ae55bc66f2027544a5f2d6e295aa18a7b5d0e49c9d7edf12986feb2", "code_archive_bytes": 2247572, "code_file_count": 126, "code_extensions": {".py": 118, ".ipynb": 8}, "github_disk_usage_kb": 3535, "github_languages": {"Python": 922179}, "github_archived": false, "github_pushed_at": "2026-09-07T21:43:54Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/efficient-and-modular-implicit"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "5wmNjjvGOXh", "year": 2021, "status": "rejected", "title": "Selfish Sparse RNN Training", "authors": ["Shiwei Liu", "Decebal Constantin Mocanu", "Yulong Pei", "Mykola Pechenizkiy"], "authorids": ["~Shiwei_Liu2", "~Decebal_Constantin_Mocanu1", "~Yulong_Pei1", "~Mykola_Pechenizkiy1"], "authors_source": "OpenReview API", "abstract": "Sparse neural networks have been widely applied to reduce the necessary resource requirements to train and deploy over-parameterized deep neural networks. For inference acceleration, methods that induce sparsity from a pre-trained dense network (dense-to-sparse) work effectively. Recently, dynamic sparse training (DST) has been proposed to train sparse neural networks without pre-training a large and dense network (sparse-to-sparse), so that the training process can also be accelerated. However, previous sparse-to-sparse methods mainly focus on Multilayer Perceptron Networks (MLPs) and Convolutional Neural Networks (CNNs), failing to match the performance of dense-to-sparse methods in Recurrent Neural Networks (RNNs) setting. In this paper, we propose an approach to train sparse RNNs with a fixed parameter count in one single run, without compromising performance. During training, we allow RNN layers to have a non-uniform redistribution across cell weights for a better regularization. Further, we introduce SNT-ASGD, a variant of the averaged stochastic gradient optimizer, which significantly improves the performance of all sparse training methods for RNNs. Using these strategies, we achieve state-of-the-art sparse training results, even better than dense model results, with various types of RNNs on Penn TreeBank and Wikitext-2 datasets.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=5wmNjjvGOXh", "arxiv_id": "2101.09048", "paper_pdf": "papers/5wmNjjvGOXh.pdf", "paper_pdf_sha256": "8d88788cb8f686585bc754a7b77b04ccd1adf48f5b65a4e949d3d0417c3ad671", "paper_pdf_bytes": 1853573, "paper_pdf_source": "openreview", "code_url": "https://github.com/Shiweiliuiiiiiii/Selfish-RNN", "code_repository": "Shiweiliuiiiiiii/Selfish-RNN", "code_commit": "06b03db04bfe93f20f8fc703e9a78a83cea777a7", "code_archive": "repos/5wmNjjvGOXh.zip", "code_archive_sha256": "0504ce22b9e498a9509623020f599bba977ede7efdc5b9201b8fde7f52123e55", "code_archive_bytes": 2046934, "code_file_count": 13, "code_extensions": {".py": 13}, "github_disk_usage_kb": 2101, "github_languages": {"Python": 104692}, "github_archived": false, "github_pushed_at": "2021-10-08T03:52:04Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/selfish-sparse-rnn-training-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "uwoA5iyTC6", "year": 2026, "status": "rejected", "title": "Time Series Forecasting as Reasoning: A Slow-Thinking Approach with Reinforced LLMs", "authors": ["Yucong Luo", "Yitong Zhou", "Mingyue Cheng", "Jiahao Wang", "Daoyu Wang"], "authorids": ["~Yucong_Luo1", "~Yitong_Zhou2", "~Mingyue_Cheng1", "~Jiahao_Wang25", "~Daoyu_Wang1"], "authors_source": "OpenReview API", "abstract": "To advance time series forecasting (TSF), various methods have been proposed to improve prediction accuracy, evolving from statistical techniques to data-driven deep learning architectures. Despite their effectiveness, most existing methods still adhere to a fast thinking paradigm - relying on pattern recognition and trend prediction as their core modeling philosophy, lacking an explicit \"thinking process\" that incorporates intermediate time series reasoning. Meanwhile, emerging slow-thinking LLMs (e.g., ChatGPT-o1) have shown remarkable multi-step reasoning capabilities, offering an alternative way to overcome these issues. However, prompt engineering alone presents several limitations—including high computational cost, privacy risks, and limited capacity for in-depth domain-specific time series reasoning. To address these limitations, a more promising approach is to train LLMs to develop slow-thinking capabilities and acquire strong time series reasoning skills. To acquire such slow thinking reasoning capabilities, we propose Time-R1, a two-stage reinforcement fine-tuning framework designed to enhance multi-step reasoning ability of LLMs for time series forecasting. Specifically, the first stage conducts supervised fine-tuning for warmup adaptation, while the second stage employs reinforcement learning to improve the model's generalization ability. Particularly, we introduce GRIP (group-based relative importance for policy optimization), which utilizes non-uniform sampling along with a fine-grained multi-objective reward specifically designed for time series forecasting to further encourage and optimize the model's exploration of effective reasoning paths. Experiments demonstrate that Time-R1 significantly improves forecast performance across diverse datasets. Source code is available https://anonymous.4open.science/r/Time-R1-NeurIPS-2025/.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=uwoA5iyTC6", "arxiv_id": "2506.10630", "paper_pdf": "papers/uwoA5iyTC6.pdf", "paper_pdf_sha256": "e71338135a187cf3ecf4d2440013466d2798916226d2f56fec55e3c53ad8c1ad", "paper_pdf_bytes": 6664272, "paper_pdf_source": "openreview", "code_url": "https://github.com/ustc-time-series/Time-R1", "code_repository": "ustc-time-series/Time-R1", "code_commit": "2cd666eaddac37734b45103d654b6d2d22bf23ed", "code_archive": "repos/uwoA5iyTC6.zip", "code_archive_sha256": "b16dc8a9161bbb16427a6e4d162446b69d6f827a345bba5fd99e1b53a634b135", "code_archive_bytes": 7496202, "code_file_count": 259, "code_extensions": {".py": 203, ".sh": 54, ".ipynb": 2}, "github_disk_usage_kb": 7771, "github_languages": {"Python": 1502870, "Shell": 23013}, "github_archived": false, "github_pushed_at": "2026-04-14T08:32:23Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/time-series-forecasting-as-reasoning-a-slow"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "pVL4bYKOGM", "year": 2025, "status": "rejected", "title": "Conformal prediction for causal effects of continuous treatments", "authors": ["Maresa Schröder", "Dennis Frauen", "Jonas Schweisthal", "Konstantin Hess", "Valentyn Melnychuk", "Stefan Feuerriegel"], "authorids": ["~Maresa_Schröder1", "~Dennis_Frauen1", "~Jonas_Schweisthal1", "~Konstantin_Hess1", "~Valentyn_Melnychuk1", "~Stefan_Feuerriegel1"], "authors_source": "OpenReview API", "abstract": "Uncertainty quantification of causal effects is crucial for safety-critical applications such as personalized medicine. A powerful approach for this is conformal prediction, which has several practical benefits due to model-agnostic finite-sample guarantees. Yet, existing methods for conformal prediction of causal effects are limited to binary/discrete treatments and make highly restrictive assumptions such as known propensity scores. In this work, we provide a novel conformal prediction method for potential outcomes of continuous treatments. We account for the additional uncertainty introduced through propensity estimation so that our conformal prediction intervals are valid even if the propensity score is unknown. Our contributions are three-fold: (1) We derive finite-sample prediction intervals for potential outcomes of continuous treatments. (2) We provide an algorithm for calculating the derived intervals. (3) We demonstrate the effectiveness of the conformal prediction intervals in experiments on synthetic and medical datasets. To the best of our knowledge, we are the first to propose conformal prediction for continuous treatments when the propensity score is unknown and must be estimated from data.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=pVL4bYKOGM", "arxiv_id": "2407.03094", "paper_pdf": "papers/pVL4bYKOGM.pdf", "paper_pdf_sha256": "6e0fe044a34f2471fc6cc544cf6f01e0748186b0e043027539398867ee64b184", "paper_pdf_bytes": 1848233, "paper_pdf_source": "openreview", "code_url": "https://github.com/m-schroder/ContinuousCausalCP", "code_repository": "m-schroder/ContinuousCausalCP", "code_commit": "d557555d6bb38239809b2c4d846c2768e9a279b7", "code_archive": "repos/pVL4bYKOGM.zip", "code_archive_sha256": "8e999ad3baa99b7926d1c2268094e4f43dd6cbd3c351846b9fd04d06103de02a", "code_archive_bytes": 96126, "code_file_count": 9, "code_extensions": {".py": 9}, "github_disk_usage_kb": 374, "github_languages": {"Python": 58734}, "github_archived": false, "github_pushed_at": "2024-07-04T11:28:49Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/conformal-prediction-for-causal-effects-of"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "npf3gREtf7", "year": 2024, "status": "rejected", "title": "Which Examples to Annotate for In-Context Learning? Towards Effective and Efficient Selection", "authors": ["Costas Mavromatis", "Balasubramaniam Srinivasan", "Zhengyuan Shen", "Jiani Zhang", "Huzefa Rangwala", "Christos Faloutsos", "George Karypis"], "authorids": ["~Costas_Mavromatis1", "~Balasubramaniam_Srinivasan1", "~Zhengyuan_Shen1", "~Jiani_Zhang2", "~Huzefa_Rangwala2", "~Christos_Faloutsos1", "~George_Karypis1"], "authors_source": "OpenReview API", "abstract": "Large Language Models (LLMs) can adapt to new tasks via in-context learning (ICL). ICL is efficient as it does not require any parameter updates to the trained LLM, but only few annotated examples as input for the LLM. In this work, we investigate an active learning approach for ICL, where there is a limited budget for annotating examples. We propose a model-adaptive optimization-free algorithm, termed AdaICL, which identifies examples that the model is uncertain about, and performs semantic diversity-based example selection. Diversity-based sampling improves overall effectiveness, while uncertainty sampling improves budget efficiency and helps the LLM learn new information. Moreover, AdaICL poses its sampling strategy as a Maximum Coverage problem, that dynamically adapts based on the model’s feedback and can be approximately solved via greedy algorithms. Extensive experiments on nine datasets and seven LLMs show that AdaICL improves performance by 4.4% accuracy points over SOTA (7.7% relative improvement), is up to 3× more budget-efficient than performing annotations uniformly at random, while it outperforms SOTA with 2× fewer ICL examples.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=npf3gREtf7", "arxiv_id": "2310.20046", "paper_pdf": "papers/npf3gREtf7.pdf", "paper_pdf_sha256": "3a2c953209a509c4d3882001a9bf84bb744ad7e4246ababaa33a0eeda22bdeb6", "paper_pdf_bytes": 1364854, "paper_pdf_source": "openreview", "code_url": "https://github.com/amazon-science/adaptive-in-context-learning", "code_repository": "amazon-science/adaptive-in-context-learning", "code_commit": "d0ea1c7d333517de592b4c1bf3862ab2ddfa027a", "code_archive": "repos/npf3gREtf7.zip", "code_archive_sha256": "2dbbfd33c298268ae85ac003d2ebd7988b899983ea75beb21986e42e30d990ed", "code_archive_bytes": 374045, "code_file_count": 13, "code_extensions": {".py": 12, ".sh": 1}, "github_disk_usage_kb": 354, "github_languages": {"Python": 234643, "Shell": 690}, "github_archived": true, "github_pushed_at": "2023-10-30T20:14:11Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/which-examples-to-annotate-for-in-context"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "MtGmCCPJD-", "year": 2023, "status": "rejected", "title": "Repository-Level Prompt Generation for Large Language Models of Code", "authors": ["Disha Shrivastava", "Hugo Larochelle", "Daniel Tarlow"], "authorids": ["~Disha_Shrivastava1", "~Hugo_Larochelle1", "~Daniel_Tarlow1"], "authors_source": "OpenReview API", "abstract": "With the success of large language models (LLMs) of code and their use as code assistants (e.g.\\ Codex used in GitHub Copilot, techniques for introducing domain-specific knowledge in the prompt design process become important. In this work, we propose a framework called Repo-Level Prompt Generator that learns to generate example-specific prompts using prompt proposals. The prompt proposals take context from the entire repository, thereby incorporating both the structure of the repository and the context from other relevant files (e.g.\\ imports, parent class files). Our technique doesn't require any access to the weights of the LLM, making it applicable in cases where we only have black-box access to the LLM. We conduct experiments on the task of single-line code-autocompletion using code repositories taken from Google Code archives. We demonstrate that an oracle constructed from our prompt proposals gives a remarkably high relative improvement of 36\\% over Codex, showing the quality of these proposals. Further, we show that when we train a model to select the best prompt proposal, we can achieve significant performance gains over Codex and other baselines.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=MtGmCCPJD-", "arxiv_id": "2206.12839", "paper_pdf": "papers/MtGmCCPJD-.pdf", "paper_pdf_sha256": "2d88ae7c3f1764c4ba10c08def0acb0edc46217ab8b3b3dc2fc55ac169f42238", "paper_pdf_bytes": 509561, "paper_pdf_source": "openreview", "code_url": "https://github.com/shrivastavadisha/repo_level_prompt_generation", "code_repository": "shrivastavadisha/repo_level_prompt_generation", "code_commit": "3af5f3424740448d8e325b3726e61944f6eec8b6", "code_archive": "repos/MtGmCCPJD-.zip", "code_archive_sha256": "d220dde993f61f5712b042b3f1077b29abdb317a01728c84e79a48276920b50d", "code_archive_bytes": 459465, "code_file_count": 21, "code_extensions": {".py": 21}, "github_disk_usage_kb": 865, "github_languages": {"Python": 150598}, "github_archived": false, "github_pushed_at": "2023-04-22T15:23:11Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/repository-level-prompt-generation-for-large"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "G-7GlfTneYg", "year": 2022, "status": "rejected", "title": "VoiceFixer: Toward General Speech Restoration with Neural Vocoder", "authors": ["Haohe Liu", "Qiuqiang Kong", "Qiao Tian", "Yan Zhao", "DeLiang Wang", "Chuanzeng Huang", "Yuxuan Wang"], "authorids": ["~Haohe_Liu1", "~Qiuqiang_Kong1", "~Qiao_Tian1", "~Yan_Zhao6", "~DeLiang_Wang1", "~Chuanzeng_Huang1", "~Yuxuan_Wang1"], "authors_source": "OpenReview API", "abstract": "Speech restoration aims to remove distortions in speech signals. Prior methods mainly focus on single-task speech restoration (SSR), such as speech denoising or speech declipping. However, SSR systems only focus on one task and do not address the general speech restoration problem. In addition, previous SSR systems show limited performance in some speech restoration tasks such as speech super-resolution. To overcome those limitations, we propose a general speech restoration (GSR) task that attempts to remove multiple distortions \u0002simultaneously. Furthermore, we propose VoiceFixer, a generative framework to address the GSR task. VoiceFixer consists of an analysis stage and a synthesis stage to mimic the speech analysis and comprehension of the human auditory system. We employ a ResUNet to model the analysis stage and a neural vocoder to model the synthesis stage. We evaluate VoiceFixer with additive noise, room \u0002reverberation, low-resolution, and clipping distortions. Our baseline GSR model achieves a 0.499 higher mean opinion score (MOS) than the speech denoising SSR model. VoiceFixer further surpasses the GSR baseline model on the MOS score by 0.256. Moreover, we observe that VoiceFixer generalizes well to severely degraded real speech recordings, indicating its potential in restoring old movies and historical speeches.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=G-7GlfTneYg", "arxiv_id": "2109.13731", "paper_pdf": "papers/G-7GlfTneYg.pdf", "paper_pdf_sha256": "4b934e4e19c82adad798f4c54fec4282c8584972e63bd3c18134883a867721a1", "paper_pdf_bytes": 18109096, "paper_pdf_source": "openreview", "code_url": "https://github.com/haoheliu/voicefixer", "code_repository": "haoheliu/voicefixer", "code_commit": "aae2253c85f97a87b844b6832384de249c23ab37", "code_archive": "repos/G-7GlfTneYg.zip", "code_archive_sha256": "2d2b90c23a7f767ea796ed71f057066a5309b3c4ed4b120a41d9015a4da048be", "code_archive_bytes": 2290754, "code_file_count": 32, "code_extensions": {".py": 31, ".sh": 1}, "github_disk_usage_kb": 3946, "github_languages": {"Python": 166428, "Dockerfile": 1335, "Shell": 52, "Batchfile": 33}, "github_archived": false, "github_pushed_at": "2025-02-17T14:13:03Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/voicefixer-toward-general-speech-restoration"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "jYkO_0z2TAr", "year": 2021, "status": "rejected", "title": "Zero-Shot Learning with Common Sense Knowledge Graphs", "authors": ["Nihal Nayak", "Stephen Bach"], "authorids": ["~Nihal_Nayak1", "~Stephen_Bach1"], "authors_source": "OpenReview API", "abstract": "Zero-shot learning relies on semantic class representations such as hand-engineered attributes or learned embeddings to predict classes without any labeled examples. We propose to learn class representations from common sense knowledge graphs. Common sense knowledge graphs are an untapped source of explicit high-level knowledge that requires little human effort to apply to a range of tasks. To capture the knowledge in the graph, we introduce ZSL-KG, a general-purpose framework with a novel transformer graph convolutional network (TrGCN) to generate class representations. Our proposed TrGCN architecture computes non-linear combinations of the node neighbourhood and leads to significant improvements on zero-shot learning tasks. We report new state-of-the-art accuracies on six zero-shot benchmark datasets in object classification, intent classification, and fine-grained entity typing tasks. ZSL-KG outperforms the specialized state-of-the-art method for each task by an average 1.7 accuracy points and outperforms the general-purpose method with the best average accuracy by 5.3 points. Our ablation study on ZSL-KG with alternate graph neural networks shows that our transformer-based aggregator adds up to 2.8 accuracy points improvement on these tasks.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=jYkO_0z2TAr", "arxiv_id": "2006.10713", "paper_pdf": "papers/jYkO_0z2TAr.pdf", "paper_pdf_sha256": "4e858af206ee3c02b4724cc720e1600cfdbad41c109938ba1b32dafd77d88ae4", "paper_pdf_bytes": 490474, "paper_pdf_source": "openreview", "code_url": "https://github.com/BatsResearch/nayak-tmlr22-code", "code_repository": "BatsResearch/nayak-tmlr22-code", "code_commit": "fd86c1d2dfb2dd540fd744ce8c49bf58719abadf", "code_archive": "repos/jYkO_0z2TAr.zip", "code_archive_sha256": "3cabbebf859e82a84ee72c4ba043d96edfca5aacfbc7698b95174201b9e6b70a", "code_archive_bytes": 2280638, "code_file_count": 55, "code_extensions": {".py": 55}, "github_disk_usage_kb": 2212, "github_languages": {"Python": 302147}, "github_archived": false, "github_pushed_at": "2022-07-25T11:24:21Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/zero-shot-learning-with-common-sense"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "3DZeEUTwhq", "year": 2026, "status": "rejected", "title": "Are You Getting What You Pay For? Auditing Model Substitution in LLM APIs", "authors": ["Will Cai", "Tianneng Shi", "Xuandong Zhao", "Dawn Song"], "authorids": ["~Will_Cai1", "~Tianneng_Shi1", "~Xuandong_Zhao1", "~Dawn_Song1"], "authors_source": "OpenReview API", "abstract": "Commercial Large Language Model (LLM) APIs create a fundamental trust problem: users pay for specific models but have no guarantee that providers deliver them faithfully. Providers may covertly substitute cheaper alternatives (e.g., quantized versions, smaller models) to reduce costs while maintaining advertised pricing. We formalize this model substitution problem and systematically evaluate detection methods under realistic adversarial conditions. Our empirical analysis reveals that software-only methods are fundamentally unreliable: statistical tests on text outputs are query-intensive and fail against subtle substitutions, while methods using log probabilities are defeated by inherent inference nondeterminism in production environments. We argue that this verification gap can be more effectively closed with hardware-level security. We propose and evaluate the use of Trusted Execution Environments (TEEs) as one practical and robust solution. Our findings demonstrate that TEEs can provide provable cryptographic guarantees of model integrity with only a modest performance overhead, offering a clear and actionable path to ensure users get what they pay for.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=3DZeEUTwhq", "arxiv_id": "2504.04715", "paper_pdf": "papers/3DZeEUTwhq.pdf", "paper_pdf_sha256": "d1d2e34ab04395637bdf86f1e4148b3dcc2003f87913e613caa9b313cc7810ea", "paper_pdf_bytes": 688098, "paper_pdf_source": "openreview", "code_url": "https://github.com/sunblaze-ucb/llm-api-audit", "code_repository": "sunblaze-ucb/llm-api-audit", "code_commit": "21ffeef65c4c92689e4db5a6a8ec2a779d2bc049", "code_archive": "repos/3DZeEUTwhq.zip", "code_archive_sha256": "ec41d15b8c3142121872c13b845e74f8d6377f973d8ca42663456af21a085add", "code_archive_bytes": 11433031, "code_file_count": 409, "code_extensions": {".py": 393, ".sh": 11, ".ipynb": 4, ".cpp": 1}, "github_disk_usage_kb": 7772, "github_languages": {"Python": 1873356, "Jupyter Notebook": 85561, "Shell": 11893, "C++": 6748}, "github_archived": false, "github_pushed_at": "2025-04-10T06:22:13Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/are-you-getting-what-you-pay-for-auditing"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "uLAAVg0ymc", "year": 2025, "status": "rejected", "title": "Do we need rebalancing strategies? A theoretical and empirical study around SMOTE and its variants", "authors": ["Abdoulaye SAKHO", "Emmanuel Malherbe", "Erwan Scornet"], "authorids": ["~Abdoulaye_SAKHO1", "~Emmanuel_Malherbe3", "~Erwan_Scornet1"], "authors_source": "OpenReview API", "abstract": "Synthetic Minority Oversampling Technique (SMOTE) is a common rebalancing strategy for handling imbalanced tabular data sets. However, few works analyze SMOTE theoretically. In this paper, we prove that SMOTE  (with default parameter) tends to copy the original minority samples asymptotically. We also prove that SMOTE exhibits boundary artifacts, thus justifying existing SMOTE variants. Then we introduce two new SMOTE-related strategies, and compare them with state-of-the-art rebalancing procedures. Surprisingly, for most data sets, we observe that applying no rebalancing strategy is competitive in terms of predictive performances, with tuned random forests, logistic regression or LightGBM. For highly imbalanced data sets, our new methods, named CV-SMOTE and Multivariate Gaussian SMOTE, are competitive. Besides, our analysis sheds some lights on the behavior of common rebalancing strategies, when used in conjunction with random forests.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=uLAAVg0ymc", "arxiv_id": "2402.03819", "paper_pdf": "papers/uLAAVg0ymc.pdf", "paper_pdf_sha256": "534af05494452613ed9d4eb4c133140dc887364d31ed35afa35e5b8740f70947", "paper_pdf_bytes": 530463, "paper_pdf_source": "openreview", "code_url": "https://github.com/artefactory/smote_strategies_study", "code_repository": "artefactory/smote_strategies_study", "code_commit": "476f902a7aafbea99432d0ac5de95509e2ff275a", "code_archive": "repos/uLAAVg0ymc.zip", "code_archive_sha256": "9fc8edb25052a82b46e9453ef84830192bd0043d71d931030e4e9e502614727c", "code_archive_bytes": 143969, "code_file_count": 10, "code_extensions": {".ipynb": 6, ".py": 4}, "github_disk_usage_kb": 375, "github_languages": {"Jupyter Notebook": 254510, "Python": 98462}, "github_archived": false, "github_pushed_at": "2025-09-02T11:37:07Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/theoretical-and-experimental-study-of-smote"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "K7KQkiHanD", "year": 2024, "status": "rejected", "title": "One-for-All: Generalized LoRA for Parameter-Efficient Fine-tuning", "authors": ["Arnav Chavan", "Zhuang Liu", "Deepak Gupta", "Eric Xing", "Zhiqiang Shen"], "authorids": ["~Arnav_Chavan1", "~Zhuang_Liu1", "~Deepak_Gupta2", "~Eric_Xing1", "~Zhiqiang_Shen1"], "authors_source": "OpenReview API", "abstract": "We present Generalized LoRA (GLoRA), a flexible approach for universal parameter-efficient fine-tuning tasks. Enhancing Low-Rank Adaptation (LoRA), GLoRA employs a generalized prompt module to optimize pre-trained model weights and adjust intermediate activations, providing more flexibility and capability across diverse tasks and datasets. Moreover, GLoRA facilitates efficient parameter adaptation by employing a scalable, modular, layer-wise structure search that learns individual adapter of each layer. \nOriginating from a unified mathematical formulation, GLoRA exhibits strong transfer learning, few-shot learning and domain generalization abilities, as it adapts to new tasks through not only weights but also additional dimensions like activations. Comprehensive experiments demonstrate that GLoRA outperforms all previous methods in natural, specialized, and structured benchmarks in the vision field, achieving superior accuracy with fewer parameters and computations. Our models on LLaMA-1 and 2 also show considerable enhancements compared to the original LoRA in the language domain. Furthermore, our structural re-parameterization design ensures that GLoRA incurs no extra inference cost, rendering it a practical solution for resource-limited applications.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=K7KQkiHanD", "arxiv_id": "2306.07967", "paper_pdf": "papers/K7KQkiHanD.pdf", "paper_pdf_sha256": "fe6fd2a0e19a44a4b39b4fb2ade6c68b50e2e87f99bf25b6d0ab7f357ddc586e", "paper_pdf_bytes": 786191, "paper_pdf_source": "openreview", "code_url": "https://github.com/Arnav0400/ViT-Slim", "code_repository": "Arnav0400/ViT-Slim", "code_commit": "390467ddf529d747168709ce2f007a85a0399cca", "code_archive": "repos/K7KQkiHanD.zip", "code_archive_sha256": "046de9581cfd6cd723aeec9fd014f3bc55f6f862c9970e3585046ec4e695998c", "code_archive_bytes": 352907, "code_file_count": 22, "code_extensions": {".py": 22}, "github_disk_usage_kb": 367, "github_languages": {"Python": 171073}, "github_archived": false, "github_pushed_at": "2025-08-24T17:10:22Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/one-for-all-generalized-lora-for-parameter"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ESR6hysKDsW", "year": 2023, "status": "rejected", "title": "Class-Incremental Learning with Repetition", "authors": ["Hamed Hemati", "Andrea Cossu", "Antonio Carta", "Julio Hurtado", "Lorenzo Pellegrini", "Davide Bacciu", "Vincenzo Lomonaco", "Damian Borth"], "authorids": ["~Hamed_Hemati1", "~Andrea_Cossu1", "~Antonio_Carta1", "~Julio_Hurtado1", "~Lorenzo_Pellegrini1", "~Davide_Bacciu1", "~Vincenzo_Lomonaco1", "~Damian_Borth1"], "authors_source": "OpenReview API", "abstract": "Real-world data streams naturally include the repetition of previous concepts. From a Continual Learning (CL) perspective, repetition is a property of the environment and, unlike replay, cannot be controlled by the user. Nowadays, Class-Incremental scenarios represent the leading test-bed for assessing and comparing CL strategies. This family of scenarios is very easy to use, but it never allows revisiting previously seen classes, thus completely disregarding the role of repetition. We focus on the family of Class-Incremental with Repetition (CIR) scenarios, where repetition is embedded in the definition of the stream. We propose two stochastic scenario generators that produce a wide range of CIR scenarios starting from a single dataset and a few control parameters. We conduct the first comprehensive evaluation of repetition in CL by studying the behavior of existing CL strategies under different CIR scenarios. We then present a novel replay strategy that exploits repetition and counteracts the natural imbalance present in the stream. On both CIFAR100 and TinyImageNet, our strategy outperforms other replay approaches, which are not designed for environments with repetition.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=ESR6hysKDsW", "arxiv_id": "2301.11396", "paper_pdf": "papers/ESR6hysKDsW.pdf", "paper_pdf_sha256": "1e207e1d3488917781efede8cde237c154479521c6254024c9734223b9d14d53", "paper_pdf_bytes": 1545202, "paper_pdf_source": "openreview", "code_url": "https://github.com/HamedHemati/CIR", "code_repository": "HamedHemati/CIR", "code_commit": "ca3305aba37e08ed1067f0da4847dba247c1569d", "code_archive": "repos/ESR6hysKDsW.zip", "code_archive_sha256": "8f99cff9675234bee1bd98559a4c68b9e043863f6d4decee5920f01f2c496e04", "code_archive_bytes": 829247, "code_file_count": 47, "code_extensions": {".py": 33, ".ipynb": 7, ".sh": 7}, "github_disk_usage_kb": 877, "github_languages": {"Jupyter Notebook": 901350, "Python": 133802, "Shell": 12069}, "github_archived": true, "github_pushed_at": "2024-02-15T13:11:42Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/class-incremental-learning-with-repetition"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "tJCwZBHm-jW", "year": 2022, "status": "rejected", "title": "Image2Point: 3D Point-Cloud Understanding with 2D Image Pretrained Models", "authors": ["Chenfeng Xu", "Shijia Yang", "Bohan Zhai", "Bichen Wu", "Xiangyu Yue", "Wei Zhan", "Peter Vajda", "Kurt Keutzer", "Masayoshi Tomizuka"], "authorids": ["~Chenfeng_Xu1", "~Shijia_Yang1", "~Bohan_Zhai1", "~Bichen_Wu1", "~Xiangyu_Yue1", "~Wei_Zhan2", "~Peter_Vajda1", "~Kurt_Keutzer3", "~Masayoshi_Tomizuka2"], "authors_source": "OpenReview API", "abstract": "3D point-clouds and 2D images are different visual representations of the physical world. While human vision can understand both representations, computer vision models designed for 2D image and 3D point-cloud understanding are quite different.\nOur paper explores the potential for transferring between these two representations by empirically investigating the feasibility of the transfer, the benefits of the transfer, and shedding light on why the transfer works.\nWe discovered that we can indeed use the same architecture and pretrained weights of a neural net model to understand both images and point-clouds. Specifically, we can transfer the pretrained image model to a point-cloud model by \\textit{inflating} 2D convolutional filters to 3D and then \\textbf{f}inetuning the \\textbf{i}mage-\\textbf{p}retrained models (FIP). \nWe discover that, surprisingly, models with minimal finetuning efforts --- only on input, output, and optionally batch normalization layers, can achieve competitive performance on 3D point-cloud classification, beating a wide range of point-cloud models that adopt task-specific architectures and use a variety of tricks. When finetuning the whole model, the performance further improves significantly. Meanwhile, we also find that FIP improves data efficiency, achieving up to 10.0 points top-1 accuracy gain on few-shot classification. It also speeds up the training of point-cloud models by up to 11.1x to reach a target accuracy.", "decision": "Reject", "meta_review": null, "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=tJCwZBHm-jW", "arxiv_id": "2106.04180", "paper_pdf": "papers/tJCwZBHm-jW.pdf", "paper_pdf_sha256": "03bd1b1923a6b2f9b09a42124957ccb97ee999331e20e2d49d59ad4f01879307", "paper_pdf_bytes": 8133728, "paper_pdf_source": "openreview", "code_url": "https://github.com/chenfengxu714/image2point", "code_repository": "chenfengxu714/image2point", "code_commit": "6efba318bcd4316dbe91e3a0daec6d8c9769cc8f", "code_archive": "repos/tJCwZBHm-jW.zip", "code_archive_sha256": "eb71a80727a67584ccc039dfcf8fbe20bfe0e3281b7c4648a3697fb9e3487b27", "code_archive_bytes": 647661, "code_file_count": 27, "code_extensions": {".py": 27}, "github_disk_usage_kb": 4153, "github_languages": {"Python": 136558}, "github_archived": false, "github_pushed_at": "2022-11-16T23:56:47Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/image2point-3d-point-cloud-understanding-with"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Ns8v4jHGyAV", "year": 2021, "status": "rejected", "title": "Matrix Shuffle-Exchange Networks for Hard 2D Tasks", "authors": ["Emīls Ozoliņš", "Karlis Freivalds", "Agris Šostaks"], "authorids": ["~Emīls_Ozoliņš1", "~Karlis_Freivalds1", "agris.sostaks@lumii.lv"], "authors_source": "OpenReview API", "abstract": "Convolutional neural networks have become the main tools for processing two-dimensional data. They work well for images, yet convolutions have a limited receptive field that prevents its applications to more complex 2D tasks. We propose a new neural model, called Matrix Shuffle-Exchange network, that can efficiently exploit long-range dependencies in 2D data and has comparable speed to a convolutional neural network. It is derived from  Neural Shuffle-Exchange network and has $\\mathcal{O}( \\log{n})$ layers and $\\mathcal{O}( n^2 \\log{n})$ total time and space complexity for processing a $n \\times n$ data matrix. We show that the Matrix Shuffle-Exchange network is well-suited for algorithmic and logical reasoning tasks on matrices and dense graphs, exceeding convolutional and graph neural network baselines. Its distinct advantage is the capability of retaining full long-range dependency modelling when generalizing to larger instances -- much larger than could be processed with models equipped with a dense attention mechanism.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=Ns8v4jHGyAV", "arxiv_id": "2006.15892", "paper_pdf": "papers/Ns8v4jHGyAV.pdf", "paper_pdf_sha256": "9136d045c232fdb281cbe7a75fb2c890965df6d7e516c48cc8284355f5ecd69f", "paper_pdf_bytes": 530449, "paper_pdf_source": "openreview", "code_url": "https://github.com/LUMII-Syslab/Matrix-SE", "code_repository": "LUMII-Syslab/Matrix-SE", "code_commit": "f398589d23d973836b78cf7dd5cf0872bff42f9c", "code_archive": "repos/Ns8v4jHGyAV.zip", "code_archive_sha256": "bc0c75dd22d020a8a06a52ae8e85e45b7bb7507701f01342f15218377210b614", "code_archive_bytes": 469630, "code_file_count": 26, "code_extensions": {".py": 26}, "github_disk_usage_kb": 2526, "github_languages": {"Python": 120235}, "github_archived": false, "github_pushed_at": "2021-02-03T09:21:45Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/switchblade-a-neural-network-for-hard-2d"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "DdHrylM8Tr", "year": 2026, "status": "rejected", "title": "AgentAlign: Navigating Safety Alignment in the Shift from Informative to Agentic Large Language Models", "authors": ["Jinchuan Zhang", "Lu Yin", "Yan Zhou", "Songlin Hu"], "authorids": ["~Jinchuan_Zhang1", "~Lu_Yin7", "~Yan_Zhou8", "~Songlin_Hu2"], "authors_source": "OpenReview API", "abstract": "The emergence of agentic capabilities in large language models fundamentally transforms their risk profile from passive information providers to autonomous action executors, introducing unprecedented safety challenges that existing alignment methods fail to address. Current approaches lack systematic frameworks for understanding and modeling the behavioral patterns underlying malicious agentic\nactivities, leading to brittle safety measures that collapse when confronted with multi-step harmful requests. We introduce AgentAlign, a novel behavioral modeling framework for agentic alignment that systematically captures malicious activity patterns through abstract behavior chains – structured representations of action sequences that characterize how harmful objectives are pursued across diverse tool-use scenarios. By instantiating these behavioral abstractions within comprehensive simulated environments, our framework enables scalable generation of authentic, executable training scenarios that preserve complex multi-step dynamics while avoiding real-world risks. Extensive evaluation across three model families demonstrates substantial safety improvements (35.8% to 79.5% enhancement in refusal rates) while maintaining or improving utility on benign tasks, significantly outperforming existing prompting-based defenses.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=DdHrylM8Tr", "arxiv_id": "2505.23020", "paper_pdf": "papers/DdHrylM8Tr.pdf", "paper_pdf_sha256": "25f041c7e9edc5ae49d76dbc7c1318b3b3a4ed4484996d5ffe15b374702b1cb5", "paper_pdf_bytes": 601537, "paper_pdf_source": "openreview", "code_url": "https://github.com/jc-ryan/AgentAlign", "code_repository": "jc-ryan/AgentAlign", "code_commit": "efb728f7211163a39eacdeab77b4d4e0618a5bc7", "code_archive": "repos/DdHrylM8Tr.zip", "code_archive_sha256": "c25f2b40b8789c2f7c666427112b43ee0e6164f1b4bbed7f6c1e325f9698cc0f", "code_archive_bytes": 15115598, "code_file_count": 24, "code_extensions": {".py": 24}, "github_disk_usage_kb": 7934, "github_languages": {"Python": 289689}, "github_archived": false, "github_pushed_at": "2025-06-02T04:42:57Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/agentalign-navigating-safety-alignment-in-the"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "WkpqUVcSTy", "year": 2025, "status": "rejected", "title": "SlowFast-LLaVA: A strong training-free baseline for video large language models", "authors": ["Mingze Xu", "Mingfei Gao", "Zhe Gan", "Hong-You Chen", "Zhengfeng Lai", "Haiming Gang", "Kai Kang", "Afshin Dehghan"], "authorids": ["~Mingze_Xu2", "~Mingfei_Gao1", "~Zhe_Gan1", "~Hong-You_Chen1", "~Zhengfeng_Lai1", "~Haiming_Gang1", "~Kai_Kang2", "~Afshin_Dehghan5"], "authors_source": "OpenReview API", "abstract": "We propose SlowFast-LLaVA (or SF-LLaVA for short), a training-free video large language model (LLM) that can jointly capture the detailed spatial semantics and long-range temporal context without exceeding the token budget of commonly used LLMs. This is realized by using a two-stream SlowFast design of inputs for Video LLMs to aggregate features from sampled video frames in an effective way. Specifically, the Slow pathway extracts features at a low frame rate while keeping as many spatial details as possible (e.g., with 24x24 tokens), and the Fast pathway operates on a high frame rate but uses a larger spatial pooling stride (e.g., downsampling 6x) to focus on the motion cues. As a result, this design allows us to adequately capture both spatial and temporal features that are beneficial for understanding details in the video. Experimental results show that SF-LLaVA outperforms existing training-free methods on a wide range of video tasks. On some benchmarks, it achieves comparable or even better performance compared to state-of-the-art Video LLMs that are fine-tuned on video datasets.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=WkpqUVcSTy", "arxiv_id": "2407.15841", "paper_pdf": "papers/WkpqUVcSTy.pdf", "paper_pdf_sha256": "aab8f27afbd00e073502aa56244af6190f70832ebf3507cd5c363979e1296cc8", "paper_pdf_bytes": 2435222, "paper_pdf_source": "openreview", "code_url": "https://github.com/apple/ml-slowfast-llava", "code_repository": "apple/ml-slowfast-llava", "code_commit": "bea0f73c106b91404ef403353e184b278fcc64a7", "code_archive": "repos/WkpqUVcSTy.zip", "code_archive_sha256": "58950ff192384e32750ad4f3698b8a232189e8020ae4ef3227b9ec21e8c678c4", "code_archive_bytes": 422991, "code_file_count": 52, "code_extensions": {".py": 38, ".sh": 14}, "github_disk_usage_kb": 384, "github_languages": {"Python": 165263, "Shell": 35913}, "github_archived": false, "github_pushed_at": "2024-09-16T21:44:18Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/slowfast-llava-a-strong-training-free"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "XqLcFMMwNb", "year": 2024, "status": "rejected", "title": "MM-LDM: Multi-Modal Latent Diffusion Model for Sounding Video Generation", "authors": ["Mingzhen Sun", "Weining Wang", "Yanyuan Qiao", "Longteng Guo", "Jiahui Sun", "Xinxin Zhu", "Jing Liu"], "authorids": ["~Mingzhen_Sun1", "~Weining_Wang3", "~Yanyuan_Qiao1", "~Longteng_Guo1", "~Jiahui_Sun2", "~Xinxin_Zhu1", "~Jing_Liu1"], "authors_source": "OpenReview API", "abstract": "Sounding video generation (SVG) is a challenging audio-video joint generation task that requires both single-modal realism and cross-modal consistency.\nPrevious diffusion-based methods tackled SVG within the original signal space, resulting in a huge computation burden.\nIn this paper, we introduce a novel multi-modal latent diffusion model (MM-LDM), which establishes a perceptual latent space that is perceptually equivalent to the original audio-video signal space but drastically reduces computational complexity.\nWe unify the representation of audio and video signals and construct a shared high-level semantic feature space to bridge the information gap between audio and video modalities.\nFurthermore, we present a novel cross-modal sampling guidance that extends our generative models to audio-to-video and video-to-audio conditional generation tasks. \nWe obtain the new state-of-the-art results with significant quality and efficiency gains.\nIn particular, our method achieves an overall improvement in all evaluation metrics and a faster training and sampling speed.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=XqLcFMMwNb", "arxiv_id": "2410.01594", "paper_pdf": "papers/XqLcFMMwNb.pdf", "paper_pdf_sha256": "2839738e9086a2bd55bf678ca29282ed4eef5f72ab62d796a7dd725431bba470", "paper_pdf_bytes": 3860841, "paper_pdf_source": "openreview", "code_url": "https://github.com/mzsun01/MM-LDM", "code_repository": "mzsun01/MM-LDM", "code_commit": "5db569aef6d3f637022a2dc6b7947cbc016f01c4", "code_archive": "repos/XqLcFMMwNb.zip", "code_archive_sha256": "7352b51b13776acf37ab580ccde85174c8434b87a219786a402aba4e20ab32e9", "code_archive_bytes": 271400, "code_file_count": 43, "code_extensions": {".py": 42, ".sh": 1}, "github_disk_usage_kb": 371, "github_languages": {"Python": 560854, "Shell": 166}, "github_archived": false, "github_pushed_at": "2024-04-12T07:34:16Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/mm-ldm-multi-modal-latent-diffusion-model-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "HZJje06x6IO", "year": 2023, "status": "rejected", "title": "Global Context Vision Transformers", "authors": ["Ali Hatamizadeh", "Hongxu Yin", "Jan Kautz", "Pavlo Molchanov"], "authorids": ["~Ali_Hatamizadeh1", "~Hongxu_Yin2", "~Jan_Kautz1", "~Pavlo_Molchanov1"], "authors_source": "OpenReview API", "abstract": "We propose global context vision transformer (GC ViT), a novel architecture that enhances parameter and compute utilization for computer vision tasks. The core of the novel model are  global context self-attention modules, joint with standard local self-attention, to effectively yet efficiently model both long and short-range spatial interactions, as an alternative to complex operations such as an attention masks or local windows shifting. While the local self-attention modules are responsible for modeling short-range information, the global query tokens are shared across all global self-attention modules to interact with local key and values. In addition, we address the lack of inductive bias in ViTs and improve the modeling of inter-channel dependencies by proposing a novel downsampler which leverages a parameter-efficient fused inverted residual block. The proposed GC ViT achieves new state-of-the-art performance across image classification, object detection and semantic segmentation tasks. On ImageNet-1K dataset for classification, the tiny, small and base variants of GC ViT with 28M, 51M and 90M parameters achieve 83.4%, 83.9% and 84.4% Top-1 accuracy, respectively, surpassing comparably-sized prior art such as CNN-based ConvNeXt and ViT-based Swin Transformer. Pre-trained GC ViT backbones in downstream tasks of object detection, instance segmentation, and semantic segmentation on MS COCO and ADE20K datasets outperform prior work consistently, sometimes by large margins.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=HZJje06x6IO", "arxiv_id": "2206.09959", "paper_pdf": "papers/HZJje06x6IO.pdf", "paper_pdf_sha256": "0b7b5944d09a2db51ae20680ee0b5b3206a78a56f7241bf25ca1ac7288bd3d03", "paper_pdf_bytes": 892774, "paper_pdf_source": "openreview", "code_url": "https://github.com/NVlabs/GCVit", "code_repository": "NVlabs/GCVit", "code_commit": "8e1941f2099eeb99ee9e8472a5ddbd6b94a8958f", "code_archive": "repos/HZJje06x6IO.zip", "code_archive_sha256": "82adf96e1c063295ebc79d12403e52ac0d85f027928269464b3682a3852bef2d", "code_archive_bytes": 636848, "code_file_count": 29, "code_extensions": {".py": 27, ".sh": 2}, "github_disk_usage_kb": 879, "github_languages": {"Python": 198144, "Shell": 541}, "github_archived": false, "github_pushed_at": "2023-12-22T13:04:04Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/global-context-vision-transformers"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "in1ynkrXyMH", "year": 2022, "status": "rejected", "title": "Introspective Learning : A Two-Stage approach for Inference in Neural Networks", "authors": ["Mohit Prabhushankar", "Ghassan AlRegib"], "authorids": ["~Mohit_Prabhushankar1", "~Ghassan_AlRegib1"], "authors_source": "OpenReview API", "abstract": "In this paper, we advocate for two stages in a neural network's decision making process. The first is the existing feed-forward inference framework where patterns in given data are sensed and associated with previously learned patterns. The second stage is a slower reflection stage where we ask the network to reflect on its feed-forward decision by considering and evaluating all available choices. Together, we term the two stages as introspective learning. We use gradients of trained neural networks as a measurement of this reflection. We perceptually visualize the explanations from both stages to provide a visual grounding to introspection. For the application of recognition, we show that an introspective network is $4\\%$ more robust and $42\\%$ less prone to calibration errors when generalizing to noisy data. We also illustrate the value of introspective networks in downstream tasks that require generalizability and calibration including active learning and out-of-distribution detection. Finally, we ground the proposed machine introspection to human introspection in the application of image quality assessment.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=in1ynkrXyMH", "arxiv_id": "2209.08425", "paper_pdf": "papers/in1ynkrXyMH.pdf", "paper_pdf_sha256": "d8bbd42572fa5fa1d3142db44be8a2ca2e130ad0b512f3aba86bd218fe0d69d2", "paper_pdf_bytes": 6856615, "paper_pdf_source": "openreview", "code_url": "https://github.com/olivesgatech/Introspective-Learning", "code_repository": "olivesgatech/Introspective-Learning", "code_commit": "fd9a358f45d7b417d1687d3bcdcce18facba2b2a", "code_archive": "repos/in1ynkrXyMH.zip", "code_archive_sha256": "e2b0417600deb54e46f3d5fcb6b1f10cd27197ea91f6e2ec521ce7c869e9f238", "code_archive_bytes": 4259454, "code_file_count": 3, "code_extensions": {".py": 3}, "github_disk_usage_kb": 4165, "github_languages": {"Python": 24392}, "github_archived": false, "github_pushed_at": "2022-09-26T13:40:21Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/introspective-learning-a-two-stage-approach-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "GVNGAaY2Dr1", "year": 2021, "status": "rejected", "title": "Multi-Agent Collaboration via Reward Attribution Decomposition", "authors": ["Tianjun Zhang", "Huazhe Xu", "Xiaolong Wang", "Yi Wu", "Kurt Keutzer", "Joseph E. Gonzalez", "Yuandong Tian"], "authorids": ["~Tianjun_Zhang1", "~Huazhe_Xu1", "~Xiaolong_Wang3", "~Yi_Wu1", "~Kurt_Keutzer1", "~Joseph_E._Gonzalez1", "~Yuandong_Tian1"], "authors_source": "OpenReview API", "abstract": "Recent advances in multi-agent reinforcement learning (MARL) have achieved super-human performance in games like Quake 3 and Dota 2. Unfortunately, these techniques require orders-of-magnitude more training rounds than humans and don't generalize to new agent configurations even on the same game. In this work, we propose Collaborative Q-learning (CollaQ) that achieves state-of-the-art performance in the StarCraft multi-agent challenge and supports ad hoc team play. We first formulate multi-agent collaboration as a joint optimization on reward assignment and show that each agent has an approximately optimal policy that decomposes into two parts: one part that only relies on the agent's own state, and the other part that is related to states of nearby agents. Following this novel finding, CollaQ decomposes the Q-function of each agent into a self term and an interactive term, with a Multi-Agent Reward Attribution (MARA) loss that regularizes the training. CollaQ is evaluated on various StarCraft maps and shows that it outperforms existing state-of-the-art techniques  (i.e., QMIX, QTRAN, and VDN) by improving the win rate by 40% with the same number of samples. In the more challenging ad hoc team play setting (i.e., reweight/add/remove units without re-training or finetuning), CollaQ outperforms previous SoTA by over 30%. ", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=GVNGAaY2Dr1", "arxiv_id": "2010.08531", "paper_pdf": "papers/GVNGAaY2Dr1.pdf", "paper_pdf_sha256": "af3be00a2a42879548e1f83bb21506e9282fd1eacc565b8ecdd0de404fbe59a8", "paper_pdf_bytes": 3480548, "paper_pdf_source": "openreview", "code_url": "https://github.com/facebookresearch/CollaQ", "code_repository": "facebookresearch/CollaQ", "code_commit": "ac43314fcf955a21fd21cf644bea864bf4b5013a", "code_archive": "repos/GVNGAaY2Dr1.zip", "code_archive_sha256": "a93407365f39234746ece9400acaa03da19de4ddb6dd41122ac88cc52cfa6158", "code_archive_bytes": 470533, "code_file_count": 7, "code_extensions": {".py": 7}, "github_disk_usage_kb": 2578, "github_languages": {"Python": 124283}, "github_archived": true, "github_pushed_at": "2023-08-14T21:56:29Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/multi-agent-collaboration-via-reward-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "VD7GaNY1tJ", "year": 2026, "status": "rejected", "title": "Robustly Improving LLM Fairness in Realistic Settings via Interpretability", "authors": ["Adam Karvonen", "Samuel Marks"], "authorids": ["~Adam_Karvonen1", "~Samuel_Marks1"], "authors_source": "OpenReview API", "abstract": "Large language models (LLMs) are increasingly deployed in high-stakes hiring applications, making decisions that directly impact people's careers and livelihoods. While prior studies suggest simple anti-bias prompts can eliminate demographic biases in controlled evaluations, we find these mitigations fail when realistic contextual details are introduced. We address these failures through internal bias mitigation: by identifying and neutralizing sensitive attribute directions within model activations, we achieve robust bias reduction across all tested scenarios.\nAcross leading commercial (GPT-4o, Claude 4 Sonnet, Gemini 2.5 Flash) and open-source models (Gemma-2 27B, Gemma-3, Mistral-24B), we find that adding realistic context such as company names, culture descriptions from public careers pages, and selective hiring constraints (e.g.,``only accept candidates in the top 10\\%\") induces significant racial and gender biases (up to 12\\% differences in interview rates). When these biases emerge, they consistently favor Black over White candidates and female over male candidates across all tested models and scenarios. Moreover, models can infer demographics and become biased from subtle cues like college affiliations, with these biases remaining invisible even when inspecting the model's chain-of-thought reasoning.\nTo address these limitations, our internal bias mitigation identifies race and gender-correlated directions and applies affine concept editing at inference time. Despite using directions from a simple synthetic dataset, the intervention generalizes robustly, consistently reducing bias to very low levels (typically under 1\\%, always below 2.5\\%) while largely maintaining model performance.\nOur findings suggest that practitioners deploying LLMs for hiring should adopt more realistic evaluation methodologies and consider internal mitigation strategies for equitable outcomes.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=VD7GaNY1tJ", "arxiv_id": "2506.10922", "paper_pdf": "papers/VD7GaNY1tJ.pdf", "paper_pdf_sha256": "108fc54bdbe9240959f5c11a2bc8c4bcfadc28a3504aefc740780cc8bf177985", "paper_pdf_bytes": 481133, "paper_pdf_source": "openreview", "code_url": "https://github.com/adamkarvonen/llm_bias", "code_repository": "adamkarvonen/llm_bias", "code_commit": "11ce50baccd7fdf4f8eca6936a5e8682a5df4c1b", "code_archive": "repos/VD7GaNY1tJ.zip", "code_archive_sha256": "2a1229891c5e75001cc47e715f36fe7af4a910f797d563810c49376e6cce0add", "code_archive_bytes": 8546158, "code_file_count": 35, "code_extensions": {".py": 27, ".ipynb": 8}, "github_disk_usage_kb": 8359, "github_languages": {"Python": 277378, "Jupyter Notebook": 217687}, "github_archived": false, "github_pushed_at": "2025-10-03T15:14:53Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/robustly-improving-llm-fairness-in-realistic"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "DQfHkEcUqV", "year": 2025, "status": "rejected", "title": "Learning Extrapolative Sequence Transformations from Markov Chains", "authors": ["Sophia Hager", "Aleem Khan", "Andrew Wang", "Nicholas Andrews"], "authorids": ["~Sophia_Hager1", "~Aleem_Khan1", "~Andrew_Wang3", "~Nicholas_Andrews2"], "authors_source": "OpenReview API", "abstract": "Most successful applications of deep learning involve similar training and test conditions. However, for some generative tasks, samples should improve desirable properties beyond previously known values, which requires the ability to generate novel hypotheses that extrapolate beyond training data. While large language models have been successfully extended to a variety of sequence modeling problems, greedy autoregressive sampling can struggle to explore the solution space sufficiently to extrapolate, especially when the properties of interest are global to the sequence. On the other hand, sequence-level sampling methods such as Markov chain Monte Carlo (MCMC) offer theoretical guarantees about capturing the distribution of interest, but suffer from the curse of dimensionality in discrete structured spaces. We propose a new approach that bridges the gap between MCMC and autoregressive sampling, which may be viewed as off-policy reinforcement learning. Our approach uses selected states from Markov chains as a source of training data for an autoregressive inference network, which is then able to generate novel sequences at test time that extrapolate along the sequence-level properties of interest. The proposed approach is validated on three problems: protein sequence design, text sentiment control, and text anonymization. We find that the learned inference network confers many of the same (and sometimes better) generalization benefits compared to the slow sampling process, but with the additional benefit of high sample efficiency.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=DQfHkEcUqV", "arxiv_id": "2505.20251", "paper_pdf": "papers/DQfHkEcUqV.pdf", "paper_pdf_sha256": "35d14ebf9a3beec549c00b78a9ac00f3651b1b269837fb3bc091747148d218ba", "paper_pdf_bytes": 354935, "paper_pdf_source": "openreview", "code_url": "https://github.com/sophia-hager/learning-MCMC-extrapolation", "code_repository": "sophia-hager/learning-MCMC-extrapolation", "code_commit": "1692bbd894c0a9b226888006f2b1e67e95c10229", "code_archive": "repos/DQfHkEcUqV.zip", "code_archive_sha256": "9177f4d2078fda4565068800ead9be9264b524d1fbc3e1d899c13ffe45f20216", "code_archive_bytes": 416653, "code_file_count": 43, "code_extensions": {".py": 35, ".sh": 8}, "github_disk_usage_kb": 388, "github_languages": {"Python": 221029, "Shell": 2777}, "github_archived": false, "github_pushed_at": "2025-08-07T23:26:25Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/learning-extrapolative-sequence"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "RxhOEngX8s", "year": 2024, "status": "rejected", "title": "Expecting The Unexpected: Towards Broad Out-Of-Distribution Detection", "authors": ["Charles Guille-Escuret", "Pierre-Andre Noel", "Ioannis Mitliagkas", "David Vazquez", "Joao Monteiro"], "authorids": ["~Charles_Guille-Escuret1", "~Pierre-Andre_Noel1", "~Ioannis_Mitliagkas1", "~David_Vazquez1", "~Joao_Monteiro1"], "authors_source": "OpenReview API", "abstract": "Deployed machine learning systems can be improved using methods detecting out-of-distribution (OOD) inputs. Existing research mainly focuses on one type of distribution shift: detecting samples from novel classes, absent from the training set. However, real-world systems encounter a broad variety of anomalous inputs, and the OOD literature neglects this diversity. This work categorizes five distinct types of distribution shifts and critically evaluates the performance of recent OOD detection methods on each of them. We publicly release our benchmark under the name BROAD (Benchmarking Resilience Over Anomaly Diversity). We find that while these methods excel in detecting novel classes, their performances are inconsistent across other types of distribution shifts. In other words, they can only reliably detect unexpected inputs that they have been specifically designed to expect. As a first step toward broad OOD detection, we learn a Gaussian mixture generative model for existing detection scores, enabling an ensemble detection approach that is more consistent and comprehensive for broad OOD detection, with improved performances over existing methods. Our code to download BROAD and reproduce our experiments will be released upon publication.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=RxhOEngX8s", "arxiv_id": "2308.11480", "paper_pdf": "papers/RxhOEngX8s.pdf", "paper_pdf_sha256": "0d1b319ec90c1738e8473cef1b3bf4c7f17f021c3203106db75ea2b769c19338", "paper_pdf_bytes": 2040858, "paper_pdf_source": "openreview", "code_url": "https://github.com/ServiceNow/broad-openood", "code_repository": "ServiceNow/broad-openood", "code_commit": "7a9414801544c1a4b759e453873442be9600fbe1", "code_archive": "repos/RxhOEngX8s.zip", "code_archive_sha256": "fabf981968cbe5eac8c6fabe51b49999f2959f63732864d5ac83962bd7c00857", "code_archive_bytes": 561816, "code_file_count": 427, "code_extensions": {".sh": 238, ".py": 186, ".ipynb": 3}, "github_disk_usage_kb": 372, "github_languages": {"Python": 647435, "Jupyter Notebook": 276787, "Shell": 158806}, "github_archived": false, "github_pushed_at": "2023-08-22T11:04:29Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/expecting-the-unexpected-towards-broad-out-of"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "luajgSjRlew", "year": 2023, "status": "rejected", "title": "Social and environmental impact of recent developments in machine learning on biology and chemistry research", "authors": ["Daniel Probst"], "authorids": ["~Daniel_Probst2"], "authors_source": "OpenReview API", "abstract": "Potential societal and environmental effects such as the rapidly increasing resource use and the associated environmental impact, reproducibility issues, and exclusivity, the privatization of ML research leading to a public research brain-drain, a narrowing of the research effort caused by a focus on deep learning, and the introduction of biases through a lack of sociodemographic diversity in data and personnel caused by recent developments in machine learning are a current topic of discussion and scientific publications. However, these discussions and publications focus mainly on computer science-adjacent fields, including computer vision and natural language processing or basic ML research. Using bibliometric analysis of the complete and full-text analysis of the open-access literature, we show that the same observations can be made for applied machine learning in chemistry and biology. These developments can potentially affect basic and applied research, such as drug discovery and development, beyond the known issue of biased data sets.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=luajgSjRlew", "arxiv_id": "2210.00356", "paper_pdf": "papers/luajgSjRlew.pdf", "paper_pdf_sha256": "dcc1ee623895a8498778699700f9e1b72419c3090dd29c22687b51dc1038b4bd", "paper_pdf_bytes": 423657, "paper_pdf_source": "openreview", "code_url": "https://github.com/daenuprobst/anon_aichem", "code_repository": "daenuprobst/anon_aichem", "code_commit": "1fed4455d480029851d7311b30808ddefc50654d", "code_archive": "repos/luajgSjRlew.zip", "code_archive_sha256": "583381e7daaa78a14c6dbf0d3f50a698cfc14ee750e40908f7dc3520a62080d6", "code_archive_bytes": 989642, "code_file_count": 12, "code_extensions": {".py": 8, ".ipynb": 4}, "github_disk_usage_kb": 977, "github_languages": {"Jupyter Notebook": 1948436, "Python": 42528}, "github_archived": false, "github_pushed_at": "2022-09-29T11:43:18Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/social-and-environmental-impact-of-recent"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "UTTrevGchy", "year": 2022, "status": "rejected", "title": "Learning Diverse Options via InfoMax Termination Critic", "authors": ["Yuji Kanagawa", "Tomoyuki Kaneko"], "authorids": ["~Yuji_Kanagawa1", "~Tomoyuki_Kaneko1"], "authors_source": "OpenReview API", "abstract": "We consider the problem of autonomously learning reusable temporally extended actions, or options, in reinforcement learning. While options can speed up transfer learning by serving as reusable building blocks, learning reusable options for unknown task distribution remains challenging. Motivated by the recent success of mutual information (MI) based skill learning, we hypothesize that more diverse options are more reusable. To this end, we propose a method for learning termination conditions of options by maximizing MI between options and corresponding state transitions. We derive a scalable approximation of this MI maximization via gradient ascent, yielding the InfoMax Termination Critic (IMTC) algorithm. Our experiments demonstrate that IMTC significantly improves the diversity of learned options without rewards, combined with an intrinsic option learning method. Moreover, we test the reusability of learned options by transferring options into various tasks, confirming that IMTC helps quick adaptation, especially in complex domains where an agent needs to manipulate objects.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=UTTrevGchy", "arxiv_id": "2010.02756", "paper_pdf": "papers/UTTrevGchy.pdf", "paper_pdf_sha256": "fa886e840c90f0a4921bbced03dbf0cdd122b5d0f8b317d4ee38dcfd82fd0221", "paper_pdf_bytes": 7825823, "paper_pdf_source": "openreview", "code_url": "https://github.com/kngwyu/infomax-option-critic", "code_repository": "kngwyu/infomax-option-critic", "code_commit": "9d907c041c1d0280db9b23eb2fdf9e0033e33bf3", "code_archive": "repos/UTTrevGchy.zip", "code_archive_sha256": "dd5ed1420d3d685e407d63b957414d00e8da56dba9d9e49fd7bbb6a9bd0f099d", "code_archive_bytes": 4359739, "code_file_count": 12, "code_extensions": {".py": 12}, "github_disk_usage_kb": 4256, "github_languages": {"Python": 94250}, "github_archived": false, "github_pushed_at": "2020-10-07T04:19:32Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/diverse-exploration-via-infomax-options-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Jf24xdaAwF9", "year": 2021, "status": "rejected", "title": "Self-Activating Neural Ensembles for Continual Reinforcement Learning", "authors": ["Sam Powers", "Abhinav Gupta"], "authorids": ["~Sam_Powers1", "~Abhinav_Gupta1"], "authors_source": "OpenReview API", "abstract": "The ability for an agent to continuously learn new skills without catastrophically forgetting existing knowledge is of critical importance for the development of generally intelligent agents. Most methods devised to address this problem depend heavily on well-defined task boundaries which simplify the problem considerably. Our task-agnostic method, Self-Activating Neural Ensembles (SANE), uses a hierarchical modular architecture designed to avoid catastrophic forgetting without making any such assumptions. At each timestep a path through the SANE tree is activated; during training only activated nodes are updated, ensuring that unused nodes do not undergo catastrophic forgetting. Additionally, new nodes are created as needed, allowing the system to leverage and retain old skills while growing and learning new ones. We demonstrate our approach on MNIST and a set of grid world environments, demonstrating that SANE does not undergo catastrophic forgetting where existing methods do.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=Jf24xdaAwF9", "arxiv_id": "2301.00141", "paper_pdf": "papers/Jf24xdaAwF9.pdf", "paper_pdf_sha256": "0eee8234ca66b80d4b4cbb8a116b278961d35b81e7d05e13adcd1baee7aac1d3", "paper_pdf_bytes": 2871749, "paper_pdf_source": "openreview", "code_url": "https://github.com/AGI-Labs/continual_rl", "code_repository": "AGI-Labs/continual_rl", "code_commit": "f2754bb282757829765beb4703f24b87efa13ff9", "code_archive": "repos/Jf24xdaAwF9.zip", "code_archive_sha256": "efbc4a642120339684e9f830a7a2108670d7a6ed80ab446c4a3718ba15a19373", "code_archive_bytes": 677532, "code_file_count": 88, "code_extensions": {".py": 88}, "github_disk_usage_kb": 2633, "github_languages": {"Python": 480548}, "github_archived": false, "github_pushed_at": "2023-07-06T14:04:25Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/self-activating-neural-ensembles-for-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Hd1Wciuyka", "year": 2026, "status": "rejected", "title": "Jodi: Unification of Visual Generation and Understanding via Joint Modeling", "authors": ["Yifeng Xu", "Zhenliang He", "Meina Kan", "Shiguang Shan", "Xilin Chen"], "authorids": ["~Yifeng_Xu1", "~Zhenliang_He2", "~Meina_Kan1", "~Shiguang_Shan2", "~Xilin_Chen1"], "authors_source": "OpenReview API", "abstract": "Visual generation and understanding are two deeply interconnected aspects of human intelligence, yet they have been traditionally treated as separate tasks in machine learning. In this paper, we propose Jodi, a diffusion framework that unifies visual generation and understanding by jointly modeling the image domain and multiple label domains. Specifically, Jodi is built upon a linear diffusion transformer along with a Role-Switch mechanism, which enables it to perform three particular types of tasks: (1) joint generation, where the model simultaneously generates images and multiple labels; (2) controllable generation, where images are generated conditioned on any combination of labels; and (3) image perception, where multiple labels can be predicted at once from a given image. Furthermore, we present the Joint-1.6M dataset, which contains 200K high-quality images collected from public sources, automatic labels for 7 visual domains, and LLM-generated captions. Extensive experiments demonstrate that our Jodi excels in generation tasks and performs competitively in understanding tasks. Besides, Jodi exhibits strong extensibility to new visual domains. Codes, data, and model weights will be publicly available.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=Hd1Wciuyka", "arxiv_id": "2505.19084", "paper_pdf": "papers/Hd1Wciuyka.pdf", "paper_pdf_sha256": "7e113f1eb5e658bf8bbc3e34d4e2aaccd272d507edfd06b5760315be4b56ed16", "paper_pdf_bytes": 33196323, "paper_pdf_source": "openreview", "code_url": "https://github.com/VIPL-GENUN/Jodi", "code_repository": "VIPL-GENUN/Jodi", "code_commit": "c21bdcb8817284980f68321fa7375c5376c37d6d", "code_archive": "repos/Hd1Wciuyka.zip", "code_archive_sha256": "0ff0a1794f60f41eaadf7593950e0402bb0586fd822b1df416006c934d568acc", "code_archive_bytes": 8646852, "code_file_count": 72, "code_extensions": {".py": 70, ".sh": 2}, "github_disk_usage_kb": 8435, "github_languages": {"Python": 500890, "Shell": 870}, "github_archived": false, "github_pushed_at": "2026-03-06T02:46:10Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/jodi-unification-of-visual-generation-and"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "EBBeSbmAyh", "year": 2025, "status": "rejected", "title": "Towards Constraint-aware Learning for Resource Allocation in NFV-enabled Networks", "authors": ["Tianfu Wang", "Long Yang", "Chao Wang", "Chuan Qin", "Liwei Deng", "Li Shen", "Hui Xiong"], "authorids": ["~Tianfu_Wang4", "~Long_Yang4", "~Chao_Wang14", "~Chuan_Qin1", "~Liwei_Deng2", "~Li_Shen1", "~Hui_Xiong1"], "authors_source": "OpenReview API", "abstract": "Virtual Network Embedding (VNE) is a challenging combinatorial optimization problem that refers to resource allocation associated with hard and multifaceted constraints in network function virtualization (NFV). Existing works for VNE struggle to handle such complex constraints, leading to compromised system performance and stability. In this paper, we propose a \\textbf{CON}straint-\\textbf{A}ware \\textbf{L}earning framework for VNE, named \\textbf{CONAL}, to achieve efficient constraint management. Concretely, we formulate the VNE problem as a constrained Markov decision process with violation tolerance. This modeling approach aims to improve both resource utilization and solution feasibility by precisely evaluating solution quality and the degree of constraint violation. We also propose a reachability-guided optimization with an adaptive reachability budget method that dynamically assigns budget values. This method achieves persistent zero violation to guarantee the feasibility of VNE solutions and more stable policy optimization by handling instances without any feasible solution. Furthermore, we propose a constraint-aware graph representation method to efficiently learn cross-graph relations and constrained path connectivity in VNE. Finally, extensive experimental results demonstrate the superiority of our proposed method over state-of-the-art baselines. Our code is available at \\href{https://anonymous.4open.science/r/iclr25-conal}{https://anonymous.4open.science/r/iclr25-conal}.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=EBBeSbmAyh", "arxiv_id": "2410.22999", "paper_pdf": "papers/EBBeSbmAyh.pdf", "paper_pdf_sha256": "1fd5a143bda9078bf1b4442865b055aff1e04de08fe1c354dca8382bbb389656", "paper_pdf_bytes": 1182935, "paper_pdf_source": "openreview", "code_url": "https://github.com/GeminiLight/conal-vne", "code_repository": "GeminiLight/conal-vne", "code_commit": "768b322f314d4d33f9ceddeb2a27bf031ba5ee10", "code_archive": "repos/EBBeSbmAyh.zip", "code_archive_sha256": "f3ecb0c7c22d5fa24e82db96834d40bc827cd4fb8dcb4ecd047766945181edaf", "code_archive_bytes": 435026, "code_file_count": 98, "code_extensions": {".py": 91, ".sh": 7}, "github_disk_usage_kb": 390, "github_languages": {"Python": 729337, "Shell": 28174}, "github_archived": false, "github_pushed_at": "2024-10-11T07:01:53Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/towards-constraint-aware-learning-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "xsts7MRLey", "year": 2024, "status": "rejected", "title": "DEEP UNSUPERVISED DOMAIN ADAPTATION FOR TIME SERIES CLASSIFICATION: A BENCHMARK", "authors": ["Hassan Ismail Fawaz", "Ganesh Del Grosso", "Tanguy Kerdoncuff", "Aurelie Boisbunon", "Illyyne Saffar"], "authorids": ["~Hassan_Ismail_Fawaz1", "~Ganesh_Del_Grosso1", "~Tanguy_Kerdoncuff1", "~Aurelie_Boisbunon1", "~Illyyne_Saffar1"], "authors_source": "OpenReview API", "abstract": "Unsupervised Domain Adaptation (UDA) aims to harness labeled source data to train models for unlabeled target data. Despite extensive research in domains like computer vision and natural language processing, UDA remains underexplored for time series data, which has widespread real-world applications ranging from medicine and manufacturing to earth observation and human activity recognition. Our paper addresses this gap by introducing a comprehensive benchmark for evaluating UDA techniques for time series classification, with a focus on deep learning methods. We provide seven new benchmark datasets covering various domain shifts and temporal dynamics, facilitating fair and standardized UDA method assessments with state of the art neural network backbones (e.g. Inception) for time series data. This benchmark offers insights into the strengths and limitations of the evaluated approaches while preserving the unsupervised nature of domain adaptation, making it directly applicable to practical problems. Our paper serves as a vital resource for researchers and practitioners, advancing domain adaptation solutions for time series data and fostering innovation in this critical field.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=xsts7MRLey", "arxiv_id": "2312.09857", "paper_pdf": "papers/xsts7MRLey.pdf", "paper_pdf_sha256": "a57654ccc36c0d8d4e3c007c2876ba573824fefac66092b0bc31a220e2cd76ba", "paper_pdf_bytes": 2251960, "paper_pdf_source": "openreview", "code_url": "https://github.com/EricssonResearch/UDA-4-TSC", "code_repository": "EricssonResearch/UDA-4-TSC", "code_commit": "c402cfc94673b8436e3dde4849342fded29e36a0", "code_archive": "repos/xsts7MRLey.zip", "code_archive_sha256": "509c9b272f11bd3773db2011aaec39b8e34185a86d7cca06819842b732a0ca78", "code_archive_bytes": 349475, "code_file_count": 90, "code_extensions": {".py": 89, ".sh": 1}, "github_disk_usage_kb": 374, "github_languages": {"Python": 349020, "Shell": 158}, "github_archived": false, "github_pushed_at": "2023-12-19T09:08:18Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/deep-unsupervised-domain-adaptation-for-time"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "mPxsHDgsimT", "year": 2023, "status": "rejected", "title": "Subclass-balancing Contrastive Learning for Long-tailed Recognition", "authors": ["Chengkai Hou", "Jieyu Zhang", "Haonan Wang", "Tianyi Zhou"], "authorids": ["~Chengkai_Hou1", "~Jieyu_Zhang1", "~Haonan_Wang1", "~Tianyi_Zhou1"], "authors_source": "OpenReview API", "abstract": "Long-tailed recognition with imbalanced classes naturally emerges in practical machine learning applications. Existing methods such as data reweighing, resampling, and supervised contrastive learning enforce the class balance with a price of introducing imbalance between instances of head class and tail class, which may ignore the underlying rich semantic substructures of the former and exaggerate the biases in the latter. We overcome these drawbacks by a novel \"subclass-balancing contrastive learning (SBCL)'' approach that clusters each head class into multiple subclasses of similar sizes as the tail classes and enforce representations to capture the two-layer class hierarchy between the original classes and their subclasses. Since the clustering is conducted in the representation space and updated during the course of training, the subclass labels preserve the semantic substructures of head classes. Meanwhile, it does not overemphasize tail class samples so each individual instance contribute to the representation learning equally. Hence, our method achieves both the instance- and subclass-balance, while the original class labels are also learned through contrastive learning among subclasses from different classes. We evaluate SBCL over a list of long-tailed benchmark datasets and it achieves the state-of-the-art performance. In addition, we present extensive analyses and ablation studies of SBCL to verify its advantages. ", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=mPxsHDgsimT", "arxiv_id": "2306.15925", "paper_pdf": "papers/mPxsHDgsimT.pdf", "paper_pdf_sha256": "79b49f8a1698c018405b3ea53ba5e9fae7213cbfe0ba041645313f169ca61fc8", "paper_pdf_bytes": 3434710, "paper_pdf_source": "openreview", "code_url": "https://github.com/JackHck/SBCL", "code_repository": "JackHck/SBCL", "code_commit": "3b9047a31d5c54a0ac14cde351ab557d2833611e", "code_archive": "repos/mPxsHDgsimT.zip", "code_archive_sha256": "f3b0b31352b49e0e9982cf2de3aef5283807a44f57bdd756751d35eb37c58edc", "code_archive_bytes": 911392, "code_file_count": 19, "code_extensions": {".py": 19}, "github_disk_usage_kb": 1039, "github_languages": {"Python": 156976}, "github_archived": false, "github_pushed_at": "2023-10-30T09:23:27Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/subclass-balancing-contrastive-learning-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "-uPIaaZdMLF", "year": 2022, "status": "rejected", "title": "Attentional meta-learners for few-shot polythetic classification", "authors": ["Ben Day", "Ramon Viñas Torné", "Nikola Simidjievski", "Pietro Lio"], "authorids": ["~Ben_Day1", "~Ramon_Viñas_Torné1", "~Nikola_Simidjievski1", "~Pietro_Lio1"], "authors_source": "OpenReview API", "abstract": "Polythetic classifications, based on shared patterns of features that need neither be universal nor constant among members of a class, are common in the natural world and greatly outnumber monothetic classifications over a set of features. We show that threshold meta-learners, such as Prototypical Networks, require an embedding dimension that is exponential in the number of features to emulate these functions. In contrast, attentional classifiers, such as Matching Networks, are polythetic by default and able to solve these problems with a linear embedding dimension. However, we find that in the presence of task-irrelevant features, inherent to meta-learning problems, attentional models are susceptible to misclassification. To address this challenge, we propose a self-attention feature-selection mechanism that adaptively dilutes non-discriminative features. We demonstrate the effectiveness of our approach in meta-learning Boolean functions, and synthetic and real-world few-shot learning tasks.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=-uPIaaZdMLF", "arxiv_id": "2106.05317", "paper_pdf": "papers/-uPIaaZdMLF.pdf", "paper_pdf_sha256": "eba9d7222c7ac6cfe16aa36bc37152d58c64f1d5e7c3e555dafc4abc0c6a11dc", "paper_pdf_bytes": 3574651, "paper_pdf_source": "openreview", "code_url": "https://github.com/rvinas/polythetic_metalearning", "code_repository": "rvinas/polythetic_metalearning", "code_commit": "97f510ae0810035e51be09c8542e6eed2df66d69", "code_archive": "repos/-uPIaaZdMLF.zip", "code_archive_sha256": "ade1343564d7e653ab3e919e52d915d570a9801db018d4db21e4097aaa6f97d6", "code_archive_bytes": 4553622, "code_file_count": 26, "code_extensions": {".py": 18, ".ipynb": 8}, "github_disk_usage_kb": 4460, "github_languages": {"Jupyter Notebook": 4813353, "Python": 110683}, "github_archived": false, "github_pushed_at": "2022-10-02T01:07:13Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/attentional-meta-learners-are-polythetic"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "1AyPW2Emp6", "year": 2021, "status": "rejected", "title": "Tight Second-Order Certificates for Randomized Smoothing", "authors": ["Alexander Levine", "Aounon Kumar", "Tom Goldstein", "Soheil Feizi"], "authorids": ["~Alexander_Levine2", "aounon@umd.edu", "~Tom_Goldstein1", "~Soheil_Feizi2"], "authors_source": "OpenReview API", "abstract": "Randomized smoothing is a popular way of providing robustness guarantees against adversarial attacks: randomly-smoothed functions have a universal Lipschitz-like bound, allowing for robustness certificates to be easily computed. In this work, we show that there also exists a universal curvature-like bound for Gaussian random smoothing: given the exact value and gradient of a smoothed function, we compute a lower bound on the distance of a point to its closest adversarial example, called the Second-order  Smoothing (SoS) robustness certificate. In addition to proving the correctness of this novel certificate, we show that SoS certificates are realizable and therefore tight. Interestingly, we show that the maximum achievable benefits, in terms of certified robustness, from using the additional information of the gradient norm are relatively small: because our bounds are tight, this is a fundamental negative result. The gain of SoS certificates further diminishes if we consider the estimation error of the gradient norms, for which we have developed an estimator. We therefore additionally develop a variant of Gaussian smoothing, called Gaussian dipole smoothing, which provides similar bounds to randomized smoothing with gradient information, but with much-improved sample efficiency. This allows us to achieve (marginally) improved robustness certificates on high-dimensional datasets such as CIFAR-10 and ImageNet.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=1AyPW2Emp6", "arxiv_id": "2010.10549", "paper_pdf": "papers/1AyPW2Emp6.pdf", "paper_pdf_sha256": "2641e8b2bb4d104d85069fcd8502bfb0d6d79206103d40ca3e3777f6fc8192f6", "paper_pdf_bytes": 5942360, "paper_pdf_source": "openreview", "code_url": "https://github.com/alevine0/smoothing_second_order", "code_repository": "alevine0/smoothing_second_order", "code_commit": "ae4b93c5666c3c967f28d62e770696fc2a417435", "code_archive": "repos/1AyPW2Emp6.zip", "code_archive_sha256": "ecec5e74447b6fb7ef37cc7605032c64ae9353d88899bcf93dbaa810f31a0d97", "code_archive_bytes": 2818537, "code_file_count": 18, "code_extensions": {".py": 18}, "github_disk_usage_kb": 2703, "github_languages": {"Python": 85059}, "github_archived": false, "github_pushed_at": "2022-08-22T15:50:16Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/tight-second-order-certificates-for-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "8NtEHw8c8C", "year": 2026, "status": "rejected", "title": "Probing Neural Topology of Large Language Models", "authors": ["Yu Zheng", "Yuan Yuan", "Yue Zhuo", "Yong Li", "Paolo Santi"], "authorids": ["~Yu_Zheng7", "~Yuan_Yuan15", "~Yue_Zhuo2", "~Yong_Li7", "~Paolo_Santi2"], "authors_source": "OpenReview API", "abstract": "Probing large language models (LLMs) has yielded valuable insights into their internal mechanisms by linking neural activations to interpretable semantics. However, the complex mechanisms that link neuron’s functional co-activation with the emergent model capabilities remains largely unknown, hindering a deeper understanding and safer development of LLMs. In this work, we introduce graph probing,\na method for uncovering the functional connectivity of LLM neurons and relating it to language generation performance. By probing models across diverse LLM families and scales, we discover a universal predictability of next-token prediction performance using only neural topology, which persists even when retaining just 1% of neuron connections. Strikingly, probing on topology outperforms probing on activation by up to 130.4% and 67.7% on perplexity and space/time semantic regression respectively, suggesting that neural topology contains orders of richer information of LLM performance than neural activation, which can be easily extracted with simple linear or MLP probes. To explain the dependence between neural topology and language performance, we identify default networks and hub neurons in LLMs and provide causal evidence by interventional experiments on multiple benchmarks, showing that LLMs actually exploit these topological information. Further analyses suggest that neural topology can be effectively leveraged to improve the efficiency, reliability, and safety of LLMs through proof-of-concept applications in model pruning, hallucination detection, and LLM fingerprinting. Codes and data for the graph probing toolbox are available at https://anonymous.4open.science/r/llm-graph-probing-71BD/.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=8NtEHw8c8C", "arxiv_id": "2506.01042", "paper_pdf": "papers/8NtEHw8c8C.pdf", "paper_pdf_sha256": "b6ae534da36f3dc7c6198bfd532c43ef7274663bcb4826c9fa0fd1ec303e4d7e", "paper_pdf_bytes": 2985489, "paper_pdf_source": "openreview", "code_url": "https://github.com/DavyMorgan/llm-graph-probing", "code_repository": "DavyMorgan/llm-graph-probing", "code_commit": "370e592d67aee6bb4bd3ccf38999476094788d4f", "code_archive": "repos/8NtEHw8c8C.zip", "code_archive_sha256": "5157a7a4c381783e7f252f109bf20e653a5fc9578bddcf981d412ec33ec2d1b1", "code_archive_bytes": 3776330, "code_file_count": 34, "code_extensions": {".py": 34}, "github_disk_usage_kb": 8626, "github_languages": {"Python": 218668}, "github_archived": false, "github_pushed_at": "2026-02-27T02:34:09Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/probing-neural-topology-of-large-language"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "EXaKfdsw04", "year": 2025, "status": "rejected", "title": "StepProof: Step-by-step verification of natural language mathematical proofs", "authors": ["Xiaolin Hu", "Qinghua Zhou", "Bogdan Grechuk", "Ivan Y Tyukin", "Oliver Sutton"], "authorids": ["~Xiaolin_Hu7", "~Qinghua_Zhou1", "~Bogdan_Grechuk1", "~Ivan_Y_Tyukin1", "~Oliver_Sutton1"], "authors_source": "OpenReview API", "abstract": "Interactive theorem provers (ITPs) are powerful tools for the formal verification of mathematical proofs down to the axiom level. However, their lack of a natural language interface remains a significant limitation. Recent advancements in large language models (LLMs) have enhanced the understanding of natural language inputs, paving the way for autoformalization—the process of translating natural language proofs into formal proofs that can be verified. Despite these advancements, existing autoformalization approaches are limited to verifying complete proofs and lack the capability for finer, sentence-level verification. To address this gap, we propose StepProof, a novel autoformalization method designed for granular, step-by-step verification. StepProof breaks down complete proofs into multiple verifiable subproofs, enabling sentence-level verification. Experimental results demonstrate that StepProof significantly improves proof success rates and efficiency compared to traditional methods. Additionally, we found that minor manual adjustments to the natural language proofs, tailoring them for step-level verification, further enhanced StepProof’s performance in autoformalization.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=EXaKfdsw04", "arxiv_id": "2506.10558", "paper_pdf": "papers/EXaKfdsw04.pdf", "paper_pdf_sha256": "48437bb844c05dc0a6d09bb9861d6c3497293dfa30306d562793cf4f8f9e7604", "paper_pdf_bytes": 1365628, "paper_pdf_source": "openreview", "code_url": "https://github.com/r1nIGa/STEP-PROOF", "code_repository": "r1nIGa/STEP-PROOF", "code_commit": "e5da5c3e115713e0072a6f62acef0340d8ab8f02", "code_archive": "repos/EXaKfdsw04.zip", "code_archive_sha256": "6b730c88c5c7b0c480c8bb8156c4210d44ea4e716fff607993fa85eb05052e10", "code_archive_bytes": 404606, "code_file_count": 3, "code_extensions": {".py": 3}, "github_disk_usage_kb": 392, "github_languages": {"Python": 40394, "Isabelle": 504}, "github_archived": false, "github_pushed_at": "2025-05-06T16:15:48Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/2506-10558"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "tf6nR1B8Nt", "year": 2024, "status": "rejected", "title": "No Wrong Turns: The Simple Geometry Of Neural Networks Optimization Paths", "authors": ["Charles Guille-Escuret", "Hiroki Naganuma", "Kilian FATRAS", "Ioannis Mitliagkas"], "authorids": ["~Charles_Guille-Escuret1", "~Hiroki_Naganuma1", "~Kilian_FATRAS1", "~Ioannis_Mitliagkas1"], "authors_source": "OpenReview API", "abstract": "Understanding the optimization dynamics of neural networks is necessary for closing the gap between theory and practice. Stochastic first-order optimization algorithms are known to efficiently locate favorable minima in deep neural networks. This efficiency, however, contrasts with the non-convex and seemingly complex structure of neural loss landscapes. In this study, we delve into the fundamental geometric properties of sampled gradients along optimization paths. We focus on two key quantities, which appear in the restricted secant inequality and error bound. Both hold high significance for first-order optimization. Our analysis reveals that these quantities exhibit predictable, consistent behavior throughout training, despite the stochasticity induced by sampling minibatches. Our findings suggest that not only do optimization trajectories never encounter significant obstacles, but they also maintain stable dynamics during the majority of training. These observed properties are sufficiently expressive to theoretically guarantee linear convergence and prescribe learning rate schedules mirroring empirical practices. We conduct our experiments on image classification, semantic segmentation and language modeling across different batch sizes, network architectures, datasets, optimizers, and initialization seeds. We discuss the impact of each factor.\nOur work provides novel insights into the properties of neural network loss functions, and opens the door to theoretical frameworks more relevant to prevalent practice.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=tf6nR1B8Nt", "arxiv_id": "2306.11922", "paper_pdf": "papers/tf6nR1B8Nt.pdf", "paper_pdf_sha256": "4cf7f53c580bcd280882a892c38b7dd790e4d313a57458a02fc0f00d19f0f24f", "paper_pdf_bytes": 713796, "paper_pdf_source": "openreview", "code_url": "https://github.com/Hiroki11x/LossLandscapeGeometry", "code_repository": "Hiroki11x/LossLandscapeGeometry", "code_commit": "adb17b0b342305c7b19ec68c39de0c0bef78d10e", "code_archive": "repos/tf6nR1B8Nt.zip", "code_archive_sha256": "f4a7abad49e9c3fff28c2fd645ac6dfb29b019fba50e7d06bb1320dc473a92b4", "code_archive_bytes": 384564, "code_file_count": 36, "code_extensions": {".py": 21, ".sh": 15}, "github_disk_usage_kb": 384, "github_languages": {"Shell": 2236497, "Python": 113312}, "github_archived": false, "github_pushed_at": "2024-06-10T20:08:40Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/no-wrong-turns-the-simple-geometry-of-neural"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "S80I3NwbbpS", "year": 2023, "status": "rejected", "title": "CAB: Comprehensive Attention Benchmarking on Long Sequence Modeling", "authors": ["Jun Zhang", "Shuyang Jiang", "Jiangtao Feng", "Lin Zheng", "Lingpeng Kong"], "authorids": ["~Jun_Zhang27", "~Shuyang_Jiang2", "~Jiangtao_Feng1", "~Lin_Zheng1", "~Lingpeng_Kong1"], "authors_source": "OpenReview API", "abstract": "Transformer has achieved remarkable success in language, image, and speech processing. Recently, various efficient attention architectures have been proposed to improve transformer's efficiency while largely preserving its efficacy, especially in modeling long sequences. A widely-used benchmark to test these efficient methods' capability on long-range modeling is Long Range Arena (LRA). However, LRA only focuses on the standard bidirectional (or noncausal) self attention, and completely ignores cross attentions and unidirectional (or causal) attentions, which are equally important to downstream applications. Although designing cross and causal variants of an attention method is straightforward for vanilla attention, it is often challenging for efficient attentions with subquadratic time and memory complexity. In this paper, we propose Comprehensive Attention Benchmark (CAB) under a fine-grained attention taxonomy with four distinguishable attention patterns, namely, noncausal self, causal self, noncausal cross, and causal cross attentions. CAB collects seven real-world tasks from different research areas to evaluate efficient attentions under the four attention patterns. Among these tasks, CAB validates efficient attentions in eight backbone networks to show their generalization across neural architectures. We conduct exhaustive experiments to benchmark the performances of nine widely-used efficient attention architectures designed with different philosophies on CAB. Extensive experimental results also shed light on the fundamental problems of efficient attentions, such as efficiency length against vanilla attention, performance consistency across attention patterns, the benefit of attention mechanisms, and interpolation/extrapolation on long-context language modeling.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=S80I3NwbbpS", "arxiv_id": "2210.07661", "paper_pdf": "papers/S80I3NwbbpS.pdf", "paper_pdf_sha256": "7e8413358dbe7bc2d7220fcfc45794e973f83d67a155cf34359c8e128fe77ef3", "paper_pdf_bytes": 29193277, "paper_pdf_source": "openreview", "code_url": "https://github.com/Shark-NLP/CAB", "code_repository": "Shark-NLP/CAB", "code_commit": "0e496969545d5cb7c76b87de56d9d8b7600d0ba4", "code_archive": "repos/S80I3NwbbpS.zip", "code_archive_sha256": "da98e7625841fb76bb0538b4e6a886ec505c219e38ef8c4dd4bbeaf7b0d7d13d", "code_archive_bytes": 1025078, "code_file_count": 33, "code_extensions": {".py": 27, ".sh": 6}, "github_disk_usage_kb": 1046, "github_languages": {"Python": 204944, "Shell": 1948}, "github_archived": false, "github_pushed_at": "2023-07-02T11:23:47Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/cab-comprehensive-attention-benchmarking-on"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ljCoTzUsdS", "year": 2022, "status": "rejected", "title": "Distinguishing rule- and exemplar-based generalization in learning systems", "authors": ["Ishita Dasgupta", "Erin Grant", "Thomas L. Griffiths"], "authorids": ["~Ishita_Dasgupta1", "~Erin_Grant1", "~Thomas_L._Griffiths1"], "authors_source": "OpenReview API", "abstract": "Despite the increasing scale of datasets in machine learning, generalization to unseen regions of the data distribution remains crucial. Such extrapolation is by definition underdetermined and is dictated by a learner’s inductive biases. Machine learning systems often do not share the same inductive biases as humans and, as a result, extrapolate in ways that are inconsistent with our expectations. We investigate two distinct such inductive biases: feature-level bias (differences in which features are more readily learned) and exemplar-vs-rule bias (differences in how these learned features are used for generalization). Exemplar- vs. rule-based generalization has been studied extensively in cognitive psychology, and in this work we present a protocol inspired by these experimental approaches for directly probing this trade-off in learning systems. The measures we propose characterize changes in extrapolation behavior when feature coverage is manipulated in a combinatorial setting. We present empirical results across a range of models and across both expository and real-world image and language domains. We demonstrate that measuring the exemplar-rule trade-off while controlling for feature-level bias provides a more complete picture of extrapolation behavior than existing formalisms. We find that most standard neural network models have a propensity towards exemplar-based extrapolation and discuss the implications of these findings for research on data augmentation, fairness, and systematic generalization.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=ljCoTzUsdS", "arxiv_id": "2110.04328", "paper_pdf": "papers/ljCoTzUsdS.pdf", "paper_pdf_sha256": "9d5f5cbe3ec8678839d853e81fb3fe6e9971d6cc1148f0606440c19023af2bdb", "paper_pdf_bytes": 2408901, "paper_pdf_source": "openreview", "code_url": "https://github.com/eringrant/icml-2022-rules-vs-exemplars", "code_repository": "eringrant/icml-2022-rules-vs-exemplars", "code_commit": "c63fc90aa5862eeb929f66f29e9837fba96be8d4", "code_archive": "repos/ljCoTzUsdS.zip", "code_archive_sha256": "1fccbf918ec31e074311383e62fb136206c74fcc5cee25f2c92b80e44f025f5d", "code_archive_bytes": 4767240, "code_file_count": 20, "code_extensions": {".py": 12, ".ipynb": 7, ".r": 1}, "github_disk_usage_kb": 4648, "github_languages": {"Jupyter Notebook": 6727174, "Python": 58534, "R": 9956}, "github_archived": false, "github_pushed_at": "2023-11-15T15:48:12Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/distinguishing-rule-and-exemplar-based-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "IU8QxEiG4hR", "year": 2021, "status": "rejected", "title": "SBEVNet: End-to-End Deep Stereo Layout Estimation", "authors": ["Divam Gupta", "Wei Pu", "Trenton Tabor", "Jeff Schneider"], "authorids": ["~Divam_Gupta1", "wpu@nrec.ri.cmu.edu", "~Trenton_Tabor1", "~Jeff_Schneider1"], "authors_source": "OpenReview API", "abstract": "Accurate layout estimation is crucial for planning and navigation, for robotics applications such as self driving. In this paper, we introduce stereo bird's eye view network SBEVNet, a novel supervised end-to-end framework for estimation of bird's eye view layout from a pair of stereo images. Although our network reuses the building blocks from the state-of-the-art deep learning networks for disparity estimation, we show that accurate depth estimation is neither sufficient nor necessary. Instead, the learning of a good internal bird's eye view feature representation is essential for layout estimation. Specifically, we first generate a disparity feature volume using the features of the stereo images and then project it to the bird's eye view coordinates. This gives us coarse grained scene structural information. We also apply inverse perspective mapping (IPM) to map the input images and their features to the bird's eye view. This gives us fine grained texture information. The concatenated IPM features with the projected feature volume creates a rich bird's eye view representation which is capable of spatial reasoning. We use this representation to estimate the BEV semantic map. Additionally, we show that using the IPM features as a supervisory signal for stereo features can give an improvement in performance. We demonstrate our approach on two datasets: KITTI dataset and synthetically generated dataset using the CARLA simulator. For both of the datasets, we establish state-of-the-art performance beyond other baselines.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=IU8QxEiG4hR", "arxiv_id": "2105.11705", "paper_pdf": "papers/IU8QxEiG4hR.pdf", "paper_pdf_sha256": "4253842fa38ef10ace1666b6c0e3d97a4f42036658048558156879b86191aa44", "paper_pdf_bytes": 2489012, "paper_pdf_source": "openreview", "code_url": "https://github.com/divamgupta/sbevnet-stereo-layout-estimation", "code_repository": "divamgupta/sbevnet-stereo-layout-estimation", "code_commit": "ef9377df95041399bc3a243fefdbb343effca12c", "code_archive": "repos/IU8QxEiG4hR.zip", "code_archive_sha256": "647abfef2e29f3563a69a49f81dda893dd8c1987de7b366f1b9f3072333a33d4", "code_archive_bytes": 2849843, "code_file_count": 10, "code_extensions": {".py": 10}, "github_disk_usage_kb": 2783, "github_languages": {"Python": 45584}, "github_archived": false, "github_pushed_at": "2023-01-15T17:05:49Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/sbevnet-end-to-end-deep-stereo-layout-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Oo5t7b1jQu", "year": 2026, "status": "rejected", "title": "SituatedThinker: Grounding LLM Reasoning with Real-World through Situated Thinking", "authors": ["Junnan Liu", "Linhao Luo", "Thuy-Trang Vu", "Gholamreza Haffari"], "authorids": ["~Junnan_Liu1", "~Linhao_Luo2", "~Thuy-Trang_Vu1", "~Gholamreza_Haffari2"], "authors_source": "OpenReview API", "abstract": "Recent advances in large language models (LLMs) demonstrate their impressive reasoning capabilities. However, the reasoning confined to internal parametric space limits LLMs' access to real-time information and understanding of the physical world. To overcome this constraint, we introduce SituatedThinker, a novel framework that enables LLMs to ground their reasoning in real-world contexts through situated thinking, which adaptively combines both internal knowledge and external information with predefined interfaces. By utilizing reinforcement learning, SituatedThinker incentivizes deliberate reasoning with the real world to acquire information and feedback, allowing LLMs to surpass their knowledge boundaries and enhance reasoning. Experimental results demonstrate significant performance improvements on multi-hop question-answering and mathematical reasoning benchmarks. Furthermore, SituatedThinker demonstrates strong performance on unseen tasks, such as KBQA, TableQA, and text-based games, showcasing the generalizable real-world grounded reasoning capability.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=Oo5t7b1jQu", "arxiv_id": "2505.19300", "paper_pdf": "papers/Oo5t7b1jQu.pdf", "paper_pdf_sha256": "c063c7b1bca364a3725ed5cc1c94c8c8a3f478686346260922b5c4a1ea0dbb09", "paper_pdf_bytes": 765445, "paper_pdf_source": "openreview", "code_url": "https://github.com/jnanliu/SituatedThinker", "code_repository": "jnanliu/SituatedThinker", "code_commit": "6e44597f4a9e8725206106e2aebfb0cefe31c596", "code_archive": "repos/Oo5t7b1jQu.zip", "code_archive_sha256": "1eca409b74dd4f9c216f72ba7715c63ab749921c48ce1fa4279adc19d143f3b6", "code_archive_bytes": 10721296, "code_file_count": 221, "code_extensions": {".py": 215, ".sh": 6}, "github_disk_usage_kb": 9318, "github_languages": {"Python": 1635369, "Shell": 5281}, "github_archived": false, "github_pushed_at": "2025-06-08T18:17:29Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/situatedthinker-grounding-llm-reasoning-with"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "4RRmy9iw3c", "year": 2025, "status": "rejected", "title": "AutoAL: Automated Active Learning with Differentiable Query Strategy Search", "authors": ["Yifeng Wang", "Xueying Zhan", "Siyu Huang"], "authorids": ["~Yifeng_Wang2", "~Xueying_Zhan1", "~Siyu_Huang2"], "authors_source": "OpenReview API", "abstract": "As deep learning continues to evolve, the need for data efficiency becomes increasingly important. Considering labeling large datasets is both time-consuming and expensive, active learning (AL) provides a promising solution to this challenge by iteratively selecting the most informative subsets of examples to train deep neural networks, thereby reducing the labeling cost. However, the effectiveness of different AL algorithms can vary significantly across data scenarios, and determining which AL algorithm best fits a given task remains a challenging problem. This work presents the first differentiable AL strategy search method, named AutoAL, which is designed on top of existing AL sampling strategies. AutoAL consists of two neural nets, named SearchNet and FitNet, which are optimized concurrently under a differentiable bi-level optimization framework. For any given task, SearchNet and FitNet are iteratively co-optimized using the labeled data, learning how well a set of candidate AL algorithms perform on that task. With the optimal AL strategies identified, SearchNet selects a small subset from the unlabeled pool for querying their annotations, enabling efficient training of the task model. Experimental results demonstrate that AutoAL consistently achieves superior accuracy compared to all candidate AL algorithms and other selective AL approaches, showcasing its potential for adapting and integrating multiple existing AL methods across diverse tasks and domains.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=4RRmy9iw3c", "arxiv_id": "2410.13853", "paper_pdf": "papers/4RRmy9iw3c.pdf", "paper_pdf_sha256": "5e3b9a1dff3f9910694a4ea5f8407ab6a5102e18105bd949830937cdc32684b6", "paper_pdf_bytes": 540568, "paper_pdf_source": "openreview", "code_url": "https://github.com/haizailache999/AutoAL", "code_repository": "haizailache999/AutoAL", "code_commit": "5200ce7b2780467e292001c3583b59e50fcfe6bc", "code_archive": "repos/4RRmy9iw3c.zip", "code_archive_sha256": "80f54138388674310a5525fd94bba6e782d044bb727c78eb1ce0f29ec3ac6dd9", "code_archive_bytes": 386662, "code_file_count": 50, "code_extensions": {".py": 50}, "github_disk_usage_kb": 396, "github_languages": {"Python": 189805}, "github_archived": false, "github_pushed_at": "2025-05-18T23:50:02Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/autoal-automated-active-learning-with"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ICSvW69W5K", "year": 2024, "status": "rejected", "title": "Semantic Parsing with Candidate Expressions for Knowledge Base Question Answering", "authors": ["Daehwan Nam", "Gary Lee"], "authorids": ["~Daehwan_Nam1", "~Gary_Lee1"], "authors_source": "OpenReview API", "abstract": "Semantic parsers convert natural language to logical forms, which can then be evaluated on knowledge bases (KBs) to produce denotations.\nEarly neural semantic parsers used grammars that define actions, such as production rules, then the semantic parsers could sequentially take actions to construct well-typed logical forms.\nIn contrast, recent neural semantic parsers have been developed with pre-trained sequence-to-sequence (seq2seq) models, such as BART and T5, which treat logical forms as sequences of tokens.\nHowever, the seq2seq models have difficulty in learning to generate logical forms that contain components drawn from large KBs.\nIn this work, we propose a grammar augmented with candidate expressions for seq2seq semantic parsing on large KBs.\nThe grammar defines actions as production rules, and our semantic parser predicts actions during inference under the constraints by types and candidate expressions.\nWe apply the grammar to knowledge base question answering, where the constraints by candidate expressions assist a semantic parser to generate valid KB components.\nExperiments on the KQA Pro benchmark showed that the constraints by candidate expressions increased the accuracy of our semantic parser, and our semantic parser achieved state-of-the-art performance on KQA Pro.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=ICSvW69W5K", "arxiv_id": "2410.00414", "paper_pdf": "papers/ICSvW69W5K.pdf", "paper_pdf_sha256": "e9f74dd2245bdb9384e5caa87a0c8712cb258033bf112f5a99f9c055584ec706", "paper_pdf_bytes": 413947, "paper_pdf_source": "openreview", "code_url": "https://github.com/daehwannam/candexpr-sp", "code_repository": "daehwannam/candexpr-sp", "code_commit": "0e6a168e3565c40672970ad28ee7caa88d0a7936", "code_archive": "repos/ICSvW69W5K.zip", "code_archive_sha256": "6ca9ff2b9aaadd46ec9e094a13da7f10c1e28ce437eee5d3c7d99503c6616c6f", "code_archive_bytes": 166866, "code_file_count": 142, "code_extensions": {".py": 139, ".sh": 3}, "github_disk_usage_kb": 412, "github_languages": {"Python": 402632, "Emacs Lisp": 26132, "Shell": 1263, "TeX": 557}, "github_archived": false, "github_pushed_at": "2025-12-31T04:42:15Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/semantic-parsing-with-candidate-expressions"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "dF0g-5k05h_", "year": 2023, "status": "rejected", "title": "The Vendi Score: A Diversity Evaluation Metric for Machine Learning", "authors": ["Dan Friedman", "Adji Bousso Dieng"], "authorids": ["~Dan_Friedman2", "~Adji_Bousso_Dieng1"], "authors_source": "OpenReview API", "abstract": "Diversity is an important criterion for many areas of machine learning (ML), including generative modeling and dataset curation. Yet little work has gone into understanding, formalizing, and measuring diversity in ML. In this paper we address the diversity evaluation problem by proposing the Vendi Score, which connects and extends ideas from ecology and quantum statistical mechanics to ML. The Vendi Score is defined as the exponential of the Shannon entropy of the eigenvalues of a similarity matrix. This matrix is induced by a user-defined similarity function applied to the sample to be evaluated for diversity. In taking a similarity function as input, the Vendi Score enables its user to specify any desired form of diversity. Importantly, unlike many existing metrics in ML, the Vendi Score doesn’t require a reference dataset or distribution over samples or labels, it is therefore general and applicable to any generative model, decoding algorithm, and dataset from any domain where similarity can be defined. We showcase the Vendi Score on molecular generative modeling where we found it addresses shortcomings of the current diversity metric of choice in that domain. We also applied the Vendi Score to generative models of images and decoding algorithms of text where we found it confirms known results about diversity in those domains. Furthermore, we used the Vendi Score to measure mode collapse, a known shortcoming of generative adversarial networks (GANs). In particular, the Vendi Score revealed that even GANs that capture all the modes of a labelled dataset can be less diverse than the original dataset. Finally, the interpretability of the Vendi Score allowed us to diagnose several benchmark ML datasets for diversity, opening the door for diversity-informed data augmentation", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=dF0g-5k05h_", "arxiv_id": "2210.02410", "paper_pdf": "papers/dF0g-5k05h_.pdf", "paper_pdf_sha256": "fadffb85538c503eea30faad855d8fd857e7dc8a77a1eef2c5b08f8029b4a9b5", "paper_pdf_bytes": 8534544, "paper_pdf_source": "openreview", "code_url": "https://github.com/vertaix/Vendi-Score", "code_repository": "vertaix/Vendi-Score", "code_commit": "ff1dfdbe6356b98a6087540f215b9a9db6db7c11", "code_archive": "repos/dF0g-5k05h_.zip", "code_archive_sha256": "65289de42582014b90d3e9d265cc4f2ed4bddee1c9e75f2893d2e406e6ee5650", "code_archive_bytes": 1132711, "code_file_count": 9, "code_extensions": {".py": 6, ".ipynb": 3}, "github_disk_usage_kb": 1139, "github_languages": {"Python": 17064}, "github_archived": false, "github_pushed_at": "2025-07-27T11:18:24Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/the-vendi-score-a-diversity-evaluation-metric"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "m22XrToDacC", "year": 2022, "status": "rejected", "title": "Distributionally Robust Recourse Action", "authors": ["Duy Nguyen", "Ngoc Bui", "Viet Anh Nguyen"], "authorids": ["~Duy_Nguyen2", "~Ngoc_Bui1", "~Viet_Anh_Nguyen2"], "authors_source": "OpenReview API", "abstract": "Recourse actions, also known as counterfactual explanations, aim to explain a particular algorithmic decision by showing one or multiple ways in which the instance could be modified to receive an alternate outcome. Existing recourse recommendations often assume that the machine learning models do not change over time. However, this assumption does not always hold in practice because of data distribution shifts, and in this case, the recourse actions may become invalid. To redress this shortcoming, we propose the Distributionally Robust Recourse Action framework, which generates a recourse action that has high probability of being valid under a mixture of model shifts. We show that the robust recourse can be found efficiently using a projected gradient descent algorithm and we discuss several extensions of our framework. Numerical experiments with both synthetic and real-world datasets demonstrate the benefits of our proposed framework.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=m22XrToDacC", "arxiv_id": "2302.11211", "paper_pdf": "papers/m22XrToDacC.pdf", "paper_pdf_sha256": "3b727f15e569b77d6befdee9902200ead6f72e5c4b2785f2cd8c28f506197afe", "paper_pdf_bytes": 619372, "paper_pdf_source": "openreview", "code_url": "https://github.com/duykhuongnguyen/DiRRAc", "code_repository": "duykhuongnguyen/DiRRAc", "code_commit": "928e021fed815b63b351cb79964530c6bd3d35b8", "code_archive": "repos/m22XrToDacC.zip", "code_archive_sha256": "ae30c2fc7c948680f0a20460bb8555ad2b713e2721a9589973ff3c48638d86da", "code_archive_bytes": 5183492, "code_file_count": 56, "code_extensions": {".py": 48, ".ipynb": 8}, "github_disk_usage_kb": 4779, "github_languages": {"Jupyter Notebook": 741114, "Python": 420207, "Shell": 2772}, "github_archived": false, "github_pushed_at": "2023-02-13T03:03:40Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/distributionally-robust-recourse-action-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "pwwVuSICBgt", "year": 2021, "status": "rejected", "title": "Enabling Binary Neural Network Training on the Edge", "authors": ["Erwei Wang", "James J. Davis", "Daniele Moro", "Piotr Zielinski", "Claudionor Coelho", "Satrajit Chatterjee", "Peter Y. K. Cheung", "George Anthony Constantinides"], "authorids": ["~Erwei_Wang1", "james.davis@imperial.ac.uk", "danielemoro@google.com", "~Piotr_Zielinski1", "claudionor.coelho@alumni.stanford.edu", "~Satrajit_Chatterjee1", "p.cheung@imperial.ac.uk", "~George_Anthony_Constantinides1"], "authors_source": "OpenReview API", "abstract": "The ever-growing computational demands of increasingly complex machine learning models frequently necessitate the use of powerful cloud-based infrastructure for their training. Binary neural networks are known to be promising candidates for on-device inference due to their extreme compute and memory savings over higher-precision alternatives. In this paper, we demonstrate that they are also strongly robust to gradient quantization, thereby making the training of modern models on the edge a practical reality. We introduce a low-cost binary neural network training strategy exhibiting sizable memory footprint reductions and energy savings vs Courbariaux & Bengio's standard approach. Against the latter, we see coincident memory requirement and energy consumption drops of 2--6$\\times$, while reaching similar test accuracy, across a range of small-scale models trained to classify popular datasets. We also showcase ImageNet training of ResNetE-18, achieving a 3.12$\\times$ memory reduction over the aforementioned standard. Such savings will allow for unnecessary cloud offloading to be avoided, reducing latency and increasing energy efficiency while also safeguarding user privacy.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=pwwVuSICBgt", "arxiv_id": "2102.04270", "paper_pdf": "papers/pwwVuSICBgt.pdf", "paper_pdf_sha256": "f4e1e877ae21f7ad9963c980bd59656a790b0979441226ac6a7f7c536784be73", "paper_pdf_bytes": 457318, "paper_pdf_source": "openreview", "code_url": "https://github.com/awai54st/Enabling-Binary-Neural-Network-Training-on-the-Edge", "code_repository": "awai54st/Enabling-Binary-Neural-Network-Training-on-the-Edge", "code_commit": "b2c026b09e81ea01c778a4d500af99df612a7a63", "code_archive": "repos/pwwVuSICBgt.zip", "code_archive_sha256": "1e9caef5bc332ac902f9b23fa3ff6240afff13b531d0b7090b86363f1805b295", "code_archive_bytes": 3917263, "code_file_count": 190, "code_extensions": {".cpp": 93, ".py": 48, ".h": 45, ".ipynb": 4}, "github_disk_usage_kb": 2901, "github_languages": {"C++": 548945, "Python": 291996, "Jupyter Notebook": 283149, "C": 1099}, "github_archived": false, "github_pushed_at": "2022-03-06T22:03:47Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/enabling-binary-neural-network-training-on-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "rVrWNb2XLi", "year": 2026, "status": "rejected", "title": "SpecOffload: Unlocking Latent GPU Capacity for LLM Inference on Resource-Constrained Devices", "authors": ["Xiangwen Zhuge", "Xu Shen", "Zeyu Wang", "Fan Dang", "Xuan Ding", "Danyang Li", "Yahui Han", "Tianxiang Hao", "Zheng Yang"], "authorids": ["~Xiangwen_Zhuge1", "~Xu_Shen5", "~Zeyu_Wang28", "~Fan_Dang1", "~Xuan_Ding3", "~Danyang_Li3", "~Yahui_Han1", "~Tianxiang_Hao1", "~Zheng_Yang1"], "authors_source": "OpenReview API", "abstract": "Efficient LLM inference on resource-constrained devices (i.e., PCs with a single commodity GPU) presents significant challenges in compute and memory utilization. Due to limited GPU memory, existing systems offload model weights to CPU memory, incurring substantial I/O overhead between the CPU and GPU. This leads to two major inefficiencies: (1) GPU cores are underutilized, often remaining idle while waiting for data to be loaded; and (2) GPU memory has a low impact on performance, as reducing its capacity has minimal effect on overall throughput. In this paper, we propose SpecOffload, a high-throughput inference engine that embeds speculative decoding into offloading. Our key idea is to unlock latent GPU resources for storing and executing a draft model used for speculative decoding, thus accelerating inference at near-zero additional cost. To support this, we carefully orchestrate the interleaved execution of target and draft models in speculative decoding within the offloading pipeline, and propose a planner to manage tensor placement and select optimal parameters. Compared with the best baseline, SpecOffload improves GPU core utilization by 4.49× and boosts inference throughput by 2.54×. Anonymous repo is at https://anonymous.4open.science/r/SpecOffload-F3F2/.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=rVrWNb2XLi", "arxiv_id": "2505.10259", "paper_pdf": "papers/rVrWNb2XLi.pdf", "paper_pdf_sha256": "6f7983de61f4ab2a0458a44f9bfd9e608ef149560697e72ae2e5746c6facd019", "paper_pdf_bytes": 775442, "paper_pdf_source": "openreview", "code_url": "https://github.com/MobiSense/SpecOffload-public", "code_repository": "MobiSense/SpecOffload-public", "code_commit": "b8b68c9f4278fd9ef9b585b3e48ede301747a212", "code_archive": "repos/rVrWNb2XLi.zip", "code_archive_sha256": "7392869efc35f918254a472b51466eed964b47c215476934254c16c8aa0e8d4a", "code_archive_bytes": 11599378, "code_file_count": 1872, "code_extensions": {".py": 1834, ".h": 13, ".cu": 8, ".cpp": 7, ".sh": 6, ".cuh": 4}, "github_disk_usage_kb": 9693, "github_languages": {"Python": 44652281, "Cuda": 327808, "C++": 25815, "Shell": 18501, "C": 7703, "Cython": 3635}, "github_archived": false, "github_pushed_at": "2026-02-03T02:04:10Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/specoffload-unlocking-latent-gpu-capacity-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ln2k0PqVQA", "year": 2025, "status": "rejected", "title": "Online Intrinsic Rewards for Decision Making Agents from Large Language Model Feedback", "authors": ["Qinqing Zheng", "Mikael Henaff", "Amy Zhang", "Aditya Grover", "Brandon Amos"], "authorids": ["~Qinqing_Zheng1", "~Mikael_Henaff1", "~Amy_Zhang1", "~Aditya_Grover1", "~Brandon_Amos1"], "authors_source": "OpenReview API", "abstract": "Automatically synthesizing dense rewards from natural language descriptions is a promising paradigm in reinforcement learning (RL), with applications to sparse reward problems, open-ended exploration, and hierarchical skill design. Recent works have made promising steps by exploiting the prior knowledge of large language models (LLMs). However, these approaches suffer from important limitations: they are either not scalable to problems requiring billions of environment samples, due to requiring LLM annotations for each observation, or they require a diverse offline dataset, which may not exist or be impossible to collect. In this work, we address these limitations through a combination of algorithmic and systems-level contributions. We propose ONI, a distributed architecture that simultaneously learns an RL policy and an intrinsic reward function using LLM feedback. Our approach annotates the agent's collected experience via an asynchronous LLM server,  which is then distilled into an intrinsic reward model. We explore a range of algorithmic choices for reward modeling with varying complexity, including hashing, classification, and ranking models. By studying their relative tradeoffs, we shed light on questions regarding intrinsic reward design for sparse reward problems. Our approach achieves state-of-the-art performance across a range of challenging, sparse reward tasks from the NetHack Learning Environment in a simple unified process, solely using the agent's gathered experience, without requiring external datasets.\nWe make our code available at \\url{URL} (coming soon).", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=ln2k0PqVQA", "arxiv_id": "2410.23022", "paper_pdf": "papers/ln2k0PqVQA.pdf", "paper_pdf_sha256": "71573254cebaf2c16434171e4fa902a3a50e4cb51a85bd32f1aa3d6969b08229", "paper_pdf_bytes": 19114039, "paper_pdf_source": "openreview", "code_url": "https://github.com/facebookresearch/oni", "code_repository": "facebookresearch/oni", "code_commit": "f0f86e8abea877e045e74f4bf786d246df0638a4", "code_archive": "repos/ln2k0PqVQA.zip", "code_archive_sha256": "526b5ec20098373b53045c5d381c356896654e9b0c09f708a00b8192ee5adf74", "code_archive_bytes": 463653, "code_file_count": 135, "code_extensions": {".py": 132, ".sh": 3}, "github_disk_usage_kb": 400, "github_languages": {"Python": 783791, "Shell": 2112}, "github_archived": true, "github_pushed_at": "2024-12-17T22:07:27Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/online-intrinsic-rewards-for-decision-making"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "xbXASfz8MD", "year": 2024, "status": "rejected", "title": "Latent Space Symmetry Discovery", "authors": ["Jianke Yang", "Nima Dehmamy", "Robin Walters", "Rose Yu"], "authorids": ["~Jianke_Yang2", "~Nima_Dehmamy1", "~Robin_Walters1", "~Rose_Yu1"], "authors_source": "OpenReview API", "abstract": "Equivariant neural networks require explicit knowledge of the symmetry group. Automatic symmetry discovery methods aim to relax this constraint and learn invariance and equivariance from data. However, existing symmetry discovery methods are limited to linear symmetries in their search space and cannot handle the complexity of symmetries in real-world, often high-dimensional data. We propose a novel generative model, Latent LieGAN (LaLiGAN), which can discover nonlinear symmetries from data. It learns a mapping from data to a latent space where the symmetries become linear and simultaneously discovers symmetries in the latent space. Theoretically, we show that our method can express any nonlinear symmetry under certain conditions. Experimentally, our method can capture the intrinsic symmetry in high-dimensional observations, which results in a well-structured latent space that is useful for other downstream tasks. We demonstrate the use cases for LaLiGAN in improving equation discovery and long-term forecasting for various dynamical systems.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=xbXASfz8MD", "arxiv_id": "2310.00105", "paper_pdf": "papers/xbXASfz8MD.pdf", "paper_pdf_sha256": "bc47a63f3f0c138b41b433b22193b53cb15113b36d3d64755775fa1bccc7f09a", "paper_pdf_bytes": 4034124, "paper_pdf_source": "openreview", "code_url": "https://github.com/jiankeyang/LaLiGAN", "code_repository": "jiankeyang/LaLiGAN", "code_commit": "030457ae5b23555f175bf8422f8d3fb99942973e", "code_archive": "repos/xbXASfz8MD.zip", "code_archive_sha256": "035ee4299e271ef64edae1bc517d9180379fbd565c4983c6ab33371892534599", "code_archive_bytes": 385964, "code_file_count": 14, "code_extensions": {".py": 14}, "github_disk_usage_kb": 423, "github_languages": {"Python": 83625}, "github_archived": false, "github_pushed_at": "2024-07-12T23:25:40Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/latent-space-symmetry-discovery"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "YP4QEmqh6Ia", "year": 2023, "status": "rejected", "title": "Which Invariance Should We Transfer? A Causal Minimax Learning Approach", "authors": ["Mingzhou Liu", "Xiangyu Zheng", "Xinwei Sun", "Fang Fang", "Yizhou Wang"], "authorids": ["~Mingzhou_Liu1", "~Xiangyu_Zheng1", "~Xinwei_Sun1", "~Fang_Fang1", "~Yizhou_Wang1"], "authors_source": "OpenReview API", "abstract": "A major barrier to deploy current machine learning models lies in their sensitivity to dataset shifts. To resolve this problem, most existing studies attempted to transfer stable information to unseen environments. Among these, graph-based methods causally decomposed the data generating process into stable and mutable mechanisms. By removing the effect of mutable generation, they identified a set of stable predictors. However, a key question regarding robustness remains: which subset of the whole stable information should the model transfer, in order to achieve optimal generalization ability? To answer this question, we provide a comprehensive minimax analysis that fully characterizes conditions for a subset to be optimal. Particularly in general cases, we propose to maximize over mutable mechanisms (i.e., the source of dataset shifts), which is provable to identify the worst-case risk over all environments. This ensures us to select the optimal subset with the minimal worst-case risk. To reduce computational costs, we propose to search over only equivalent classes in terms of worst-case risk, instead of over all subsets. In cases when the searching space is still large, we turn this subset selection problem into a sparse min-max optimization scheme, which enjoys the simplicity and efficiency of implementation. The utility of our methods is demonstrated on the diagnosis of Alzheimer's Disease and gene function prediction. ", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=YP4QEmqh6Ia", "arxiv_id": "2107.01876", "paper_pdf": "papers/YP4QEmqh6Ia.pdf", "paper_pdf_sha256": "d60bf2b3bc7b8ae1901261511537a029607437d18640717e635cb74956258376", "paper_pdf_bytes": 1611158, "paper_pdf_source": "openreview", "code_url": "https://github.com/lmz123321/which_invariance", "code_repository": "lmz123321/which_invariance", "code_commit": "707f2772ad121011f49ce8bcc7b04da65b13f911", "code_archive": "repos/YP4QEmqh6Ia.zip", "code_archive_sha256": "41f602cc266e04fb58e39800bee3948905cd3d1e1c637a4aaf5e295fcdc77281", "code_archive_bytes": 2811752, "code_file_count": 31, "code_extensions": {".py": 25, ".ipynb": 5, ".r": 1}, "github_disk_usage_kb": 1179, "github_languages": {"Jupyter Notebook": 163253, "Python": 158335, "R": 15179}, "github_archived": false, "github_pushed_at": "2024-05-27T03:18:07Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/causally-invariant-predictor-with-shift"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "FYUzzBPh_j", "year": 2022, "status": "rejected", "title": "Communicating via Markov Decision Processes", "authors": ["Samuel Sokota", "Christian Schroeder de Witt", "Maximilian Igl", "Luisa M Zintgraf", "Philip Torr", "J Zico Kolter", "Shimon Whiteson", "Jakob Nicolaus Foerster"], "authorids": ["~Samuel_Sokota1", "~Christian_Schroeder_de_Witt1", "~Maximilian_Igl1", "~Luisa_M_Zintgraf1", "~Philip_Torr1", "~J_Zico_Kolter1", "~Shimon_Whiteson1", "~Jakob_Nicolaus_Foerster1"], "authors_source": "OpenReview API", "abstract": "We consider the problem of communicating exogenous information by means of Markov decision process trajectories. This setting, which we call a Markov coding game (MCG), generalizes both source coding and a large class of referential games. MCGs also isolate a problem that is important in decentralized control settings in which cheap-talk is not available---namely, they require balancing communication with the associated cost of communicating. We contribute a theoretically grounded approach to MCGs based on maximum entropy reinforcement learning and minimum entropy coupling that we call greedy minimum entropy coupling (GME). We show both that GME is able to outperform a relevant baseline on small MCGs and that GME is able to scale efficiently to extremely large MCGs. To the latter point, we demonstrate that GME is able to losslessly communicate binary images via trajectories of Cartpole and Pong, while simultaneously achieving the maximal or near maximal expected returns, and that it is even capable of performing well in the presence of actuator noise.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=FYUzzBPh_j", "arxiv_id": "2107.08295", "paper_pdf": "papers/FYUzzBPh_j.pdf", "paper_pdf_sha256": "7749fda12891515acd2506b592a93dd19cf0b3639f78540d6fadd42b0273bec5", "paper_pdf_bytes": 531562, "paper_pdf_source": "openreview", "code_url": "https://github.com/schroederdewitt/meme", "code_repository": "schroederdewitt/meme", "code_commit": "d94690ea7f4026a93d5e116d9ba643504f8f71fe", "code_archive": "repos/FYUzzBPh_j.zip", "code_archive_sha256": "ca98203639850ec49a54b357ccb59029b70caa168241b9b0068a866735d3b684", "code_archive_bytes": 5799169, "code_file_count": 132, "code_extensions": {".py": 121, ".sh": 10, ".cpp": 1}, "github_disk_usage_kb": 4936, "github_languages": {"Python": 1226433, "C++": 13482, "Shell": 3704, "Makefile": 1976, "Dockerfile": 1566}, "github_archived": false, "github_pushed_at": "2023-05-08T13:56:32Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/implicit-communication-as-minimum-entropy"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "hecuSLbL_vC", "year": 2021, "status": "rejected", "title": "Generalisation Guarantees For Continual Learning With Orthogonal Gradient Descent", "authors": ["Mehdi Abbana Bennani", "Thang Doan", "Masashi Sugiyama"], "authorids": ["~Mehdi_Abbana_Bennani1", "~Thang_Doan1", "~Masashi_Sugiyama1"], "authors_source": "OpenReview API", "abstract": "In Continual Learning settings, deep neural networks are prone to Catastrophic Forgetting. Orthogonal Gradient Descent (Farajtabar et al., 2019) was proposed to tackle the challenge. However, no theoretical guarantees have been proven yet. We present a theoretical framework to study Continual Learning algorithms in the NTK regime. This framework comprises closed form expression of the model through tasks and proxies for transfer learning, generalisation and tasks similarity. In this framework, we prove that OGD is robust to Catastrophic Forgetting then derive the first generalisation bound for SGD and OGD for Continual Learning. Finally, we study the limits of this framework in practice for OGD and highlight the importance of the NTK variation for Continual Learning.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=hecuSLbL_vC", "arxiv_id": "2006.11942", "paper_pdf": "papers/hecuSLbL_vC.pdf", "paper_pdf_sha256": "371ddf000178cf42582268efe4af3f3e48d762a91925b523fedd2331d2173a0e", "paper_pdf_bytes": 1224837, "paper_pdf_source": "openreview", "code_url": "https://github.com/MehdiAbbanaBennani/continual-learning-ogdplus", "code_repository": "MehdiAbbanaBennani/continual-learning-ogdplus", "code_commit": "b633f2c5949f6ea165e2e76aab3d043065d770a8", "code_archive": "repos/hecuSLbL_vC.zip", "code_archive_sha256": "ea0cead224a8b2e96836d96bd726cdbdeef1c9b45eb1f3abaa7e5fd29a3351f6", "code_archive_bytes": 2995954, "code_file_count": 77, "code_extensions": {".py": 56, ".sh": 21}, "github_disk_usage_kb": 2924, "github_languages": {"Shell": 753395, "Python": 169648}, "github_archived": false, "github_pushed_at": "2022-10-10T12:10:51Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/generalisation-guarantees-for-continual"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ycxzArIvgF", "year": 2026, "status": "rejected", "title": "Effective Data Pruning through Score Extrapolation", "authors": ["Sebastian Schmidt", "Prasanga Dhungel", "Christoffer Löffler", "Björn Nieth", "Stephan Günnemann", "Leo Schwinn"], "authorids": ["~Sebastian_Schmidt2", "~Prasanga_Dhungel1", "~Christoffer_Löffler1", "~Björn_Nieth1", "~Stephan_Günnemann1", "~Leo_Schwinn1"], "authors_source": "OpenReview API", "abstract": "Training advanced machine learning models demands massive datasets, resulting in prohibitive computational costs. To address this challenge, data pruning techniques identify and remove redundant training samples while preserving model performance. Yet, existing pruning techniques predominantly require a full initial training pass to identify removable samples, negating any efficiency benefits for single training runs. To overcome this limitation, we introduce a novel importance score extrapolation framework that requires training on only a small subset of data. We present two initial approaches in this framework—k-nearest neighbors and graph neural networks—to accurately predict sample importance for the entire dataset using patterns learned from this minimal subset. We demonstrate the effectiveness of our approach for 2 state-of-the-art pruning methods (Dynamic Uncertainty and TDDS), 4 different datasets (CIFAR10, CIFAR100, Places-365, and ImageNet), and\ntraining paradigms (supervised, unsupervised, adversarial). Our results indicate that score extrapolation is a promising direction to scale expensive score calculation methods, such as pruning, data attribution, or other tasks.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=ycxzArIvgF", "arxiv_id": "2506.09010", "paper_pdf": "papers/ycxzArIvgF.pdf", "paper_pdf_sha256": "31d3ed279e251af304d61c1ef614fac09379f1030a640da367a915bbd751b19d", "paper_pdf_bytes": 4263230, "paper_pdf_source": "openreview", "code_url": "https://github.com/prasangadhungel/Data-Pruning-with-Extrapolated-Scores", "code_repository": "prasangadhungel/Data-Pruning-with-Extrapolated-Scores", "code_commit": "655e48fbb48e5856da9052808222388cfc6d99f2", "code_archive": "repos/ycxzArIvgF.zip", "code_archive_sha256": "228a0a556f1c6d16a13b65163b1b9b76ed08d5fa42ec4bd2a2a58ee2e18ebb78", "code_archive_bytes": 112358, "code_file_count": 34, "code_extensions": {".py": 32, ".ipynb": 2}, "github_disk_usage_kb": 9833, "github_languages": {"Python": 267787, "Jupyter Notebook": 98163, "Makefile": 674}, "github_archived": false, "github_pushed_at": "2026-07-29T10:58:03Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/effective-data-pruning-through-score"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "LC2KxRwC3n", "year": 2025, "status": "rejected", "title": "A is for Absorption: Studying Feature Splitting and Absorption in Sparse Autoencoders", "authors": ["David Chanin", "James Wilken-Smith", "Tomáš Dulka", "Hardik Bhatnagar", "Joseph Isaac Bloom"], "authorids": ["~David_Chanin1", "~James_Wilken-Smith1", "~Tomáš_Dulka1", "~Hardik_Bhatnagar1", "~Joseph_Isaac_Bloom1"], "authors_source": "OpenReview API", "abstract": "Sparse Autoencoders (SAEs) have emerged as a promising approach to decompose the activations of Large Language Models (LLMs) into human-interpretable latents. In this paper, we pose two questions. First, to what extent do SAEs extract monosemantic and interpretable latents? Second, to what extent does varying the sparsity or the size of the SAE affect monosemanticity / interpretability? By investigating these questions in the context of a simple first-letter identification task where we have complete access to ground truth labels for all tokens in the vocabulary, we are able to provide more detail than prior investigations. Critically, we identify a problematic form of feature-splitting we call \"feature absorption\" where seemingly monosemantic latents fail to fire in cases where they clearly should. Our investigation suggests that varying SAE size or sparsity is insufficient to solve this issue, and that there are deeper conceptual issues in need of resolution.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=LC2KxRwC3n", "arxiv_id": "2409.14507", "paper_pdf": "papers/LC2KxRwC3n.pdf", "paper_pdf_sha256": "97ffad0178188e8cc48a0bb2614ca50956afcfd9623bb44026f37ff13f16a238", "paper_pdf_bytes": 3838697, "paper_pdf_source": "openreview", "code_url": "https://github.com/lasr-spelling/sae-spelling", "code_repository": "lasr-spelling/sae-spelling", "code_commit": "515648a5b96dc389ba317c4053fd8aa6eaca9929", "code_archive": "repos/LC2KxRwC3n.zip", "code_archive_sha256": "85ed4b95247f9f8ef69fee70664f70d3b44ef86c802c4d3f4565bac05a7adb2c", "code_archive_bytes": 234672, "code_file_count": 32, "code_extensions": {".py": 32}, "github_disk_usage_kb": 400, "github_languages": {"Python": 252157, "Makefile": 223}, "github_archived": false, "github_pushed_at": "2025-12-28T16:51:55Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/a-is-for-absorption-studying-feature"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "I5lcjmFmlc", "year": 2024, "status": "rejected", "title": "Robust Classification via a Single Diffusion Model", "authors": ["Huanran Chen", "Yinpeng Dong", "Zhengyi Wang", "Xiao Yang", "Chengqi Duan", "Hang Su", "Jun Zhu"], "authorids": ["~Huanran_Chen1", "~Yinpeng_Dong2", "~Zhengyi_Wang1", "~Xiao_Yang4", "~Chengqi_Duan1", "~Hang_Su3", "~Jun_Zhu2"], "authors_source": "OpenReview API", "abstract": "Recently, diffusion models have been successfully applied to improving adversarial robustness of image classifiers by purifying the adversarial noises or generating realistic data for adversarial training. However, the diffusion-based purification can be evaded by stronger adaptive attacks while adversarial training does not perform well under unseen threats, exhibiting inevitable limitations of these methods. To better harness the expressive power of diffusion models, in this paper we propose Robust Diffusion Classifier (RDC), a generative classifier that is constructed from a pre-trained diffusion model to be adversarially robust. Our method first maximizes the data likelihood of a given input and then predicts the class probabilities of the optimized input using the conditional likelihood estimated by the diffusion model through Bayes' theorem. To further reduce the computational complexity, we propose a new diffusion backbone called multi-head diffusion and develop efficient sampling strategies. As our method does not require training on particular adversarial attacks, we demonstrate that it is more generalizable to defend against multiple unseen threats. In particular, RDC achieves 75.67% robust accuracy against $\\ell_\\infty$ norm-bounded perturbations with $\\epsilon_\\infty=8/255$ on CIFAR-10, surpassing the previous state-of-the-art adversarial training models by +4.77%. The findings highlight the potential of generative classifiers by employing diffusion models for adversarial robustness compared with the commonly studied discriminative classifiers.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=I5lcjmFmlc", "arxiv_id": "2305.15241", "paper_pdf": "papers/I5lcjmFmlc.pdf", "paper_pdf_sha256": "8609580e25373aaec4239effe0d3e7ca16318c1f5f77c311aa9ba6d21bcdab2c", "paper_pdf_bytes": 2716088, "paper_pdf_source": "openreview", "code_url": "https://github.com/huanranchen/DiffusionClassifier", "code_repository": "huanranchen/DiffusionClassifier", "code_commit": "8a825045d91ab50c98503933c874a4296c5b2bec", "code_archive": "repos/I5lcjmFmlc.zip", "code_archive_sha256": "ff237233e91228f5d66396b1d92ff3fece9b3b6a963447c16057a6b505a29054", "code_archive_bytes": 549418, "code_file_count": 282, "code_extensions": {".py": 278, ".cpp": 2, ".cu": 2}, "github_disk_usage_kb": 437, "github_languages": {"Python": 4123690, "Cuda": 14488, "C++": 2555}, "github_archived": false, "github_pushed_at": "2025-03-07T09:08:25Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/robust-classification-via-a-single-diffusion"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "C9sU3Tnnki8", "year": 2023, "status": "rejected", "title": "Exploring Transformer Backbones for Heterogeneous Treatment Effect Estimation", "authors": ["YiFan Zhang", "Hanlin Zhang", "Zachary Chase Lipton", "Li Erran Li", "Eric Xing"], "authorids": ["~YiFan_Zhang8", "~Hanlin_Zhang1", "~Zachary_Chase_Lipton1", "~Li_Erran_Li1", "~Eric_Xing1"], "authors_source": "OpenReview API", "abstract": "Previous works on Treatment Effect Estimation (TEE) are not in widespread use because they are predominantly theoretical, where strong parametric assumptions are made but untractable for practical application. Recent works use Multilayer Perceptron (MLP) for modeling casual relationships, however, MLPs lag far behind recent advances in ML methodology, which limits their applicability and generalizability. To extend beyond the single domain formulation and towards more realistic learning scenarios, we explore model design spaces beyond MLPs, i.e., transformer backbones, which provide flexibility where attention layers govern interactions among treatments and covariates to exploit structural similarities of potential outcomes for confounding control. Through careful model design, Transformers as Treatment Effect Estimators (TransTEE) is proposed. We show empirically that TransTEE can: (1) serve as a general-purpose treatment effect estimator which significantly outperforms competitive baselines on a variety of challenging TEE problems (e.g., discrete, continuous, structured, or dosage-associated treatments.) and is applicable to both when covariates are tabular and when they consist of structural data (e.g., texts, graphs); (2) yield multiple advantages: compatibility with propensity score modeling, parameter efficiency, robustness to continuous treatment value distribution shifts, explainable in covariate adjustment, and real-world utility in auditing pre-trained language models. ", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=C9sU3Tnnki8", "arxiv_id": "2202.01336", "paper_pdf": "papers/C9sU3Tnnki8.pdf", "paper_pdf_sha256": "d79793737fd52fbf61c0c613fd81558970a912d4679d85fc6055ecc156fcb819", "paper_pdf_bytes": 1659403, "paper_pdf_source": "openreview", "code_url": "https://github.com/hlzhang109/TransTEE", "code_repository": "hlzhang109/TransTEE", "code_commit": "3e6883a7edd7d780185d524553b13ae633b6a229", "code_archive": "repos/C9sU3Tnnki8.zip", "code_archive_sha256": "99650922ea0a0c2cdfb009bc72b9bb866c0a7aef2052a42962b158fc1ff77c50", "code_archive_bytes": 1274747, "code_file_count": 123, "code_extensions": {".py": 123}, "github_disk_usage_kb": 1192, "github_languages": {"Python": 872932, "HTML": 2479}, "github_archived": false, "github_pushed_at": "2025-09-13T04:32:03Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/can-transformers-be-strong-treatment-effect"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "qEGBB9YB31", "year": 2022, "status": "rejected", "title": "Where do Models go Wrong? Parameter-Space Saliency Maps for Explainability", "authors": ["Roman Levin", "Manli Shu", "Eitan Borgnia", "Furong Huang", "Micah Goldblum", "Tom Goldstein"], "authorids": ["~Roman_Levin1", "~Manli_Shu1", "~Eitan_Borgnia1", "~Furong_Huang1", "~Micah_Goldblum1", "~Tom_Goldstein1"], "authors_source": "OpenReview API", "abstract": "Conventional saliency maps highlight input features to which neural network predictions are highly sensitive. We take a different approach to saliency, in which we identify and analyze the network parameters, rather than inputs, which are responsible for erroneous decisions.  We first verify that identified salient parameters are indeed responsible for misclassification by showing that turning these parameters off improves predictions on the associated samples, more than pruning the same number of random or least salient parameters.  We further validate the link between salient parameters and network misclassification errors by observing that fine-tuning a small number of the most salient parameters on a single sample results in error correction on other samples which were misclassified for similar reasons -- nearest neighbors in the saliency space. After validating our parameter-space saliency maps, we demonstrate that samples which cause similar parameters to malfunction are semantically similar.  Further, we introduce an input-space saliency counterpart which reveals how image features cause specific network components to malfunction.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=qEGBB9YB31", "arxiv_id": "2108.01335", "paper_pdf": "papers/qEGBB9YB31.pdf", "paper_pdf_sha256": "b68ead11c290531cc1a8fe901907a7727fab46ac5d5941c90224129bcbf3efa9", "paper_pdf_bytes": 13377231, "paper_pdf_source": "openreview", "code_url": "https://github.com/LevinRoman/parameter-space-saliency", "code_repository": "LevinRoman/parameter-space-saliency", "code_commit": "0e3b3d69c6e222aee6af0264d7ce3ddc6d19744e", "code_archive": "repos/qEGBB9YB31.zip", "code_archive_sha256": "1c97e98c4908db622223bf0705626adcfe9a2d6d060d70e3e47d5814a7ac8938", "code_archive_bytes": 5409635, "code_file_count": 3, "code_extensions": {".py": 3}, "github_disk_usage_kb": 5219, "github_languages": {"Python": 33081}, "github_archived": false, "github_pushed_at": "2023-03-21T23:02:02Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/where-do-models-go-wrong-parameter-space"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "tf8a4jDRFCv", "year": 2021, "status": "rejected", "title": "Learning Aggregation Functions", "authors": ["Giovanni Pellegrini", "Alessandro Tibo", "Paolo Frasconi", "Andrea Passerini", "Manfred Jaeger"], "authorids": ["~Giovanni_Pellegrini1", "~Alessandro_Tibo1", "~Paolo_Frasconi1", "~Andrea_Passerini2", "jaeger@cs.aau.dk"], "authors_source": "OpenReview API", "abstract": " Learning on sets is increasingly gaining attention in the machine learning community, due to its widespread applicability. Typically, representations over sets are computed by using fixed aggregation functions such as sum or maximum. However, recent results showed that universal function representation by sum- (or max-) decomposition requires either highly discontinuous (and thus poorly learnable) mappings, or a latent dimension equal to the maximum number of elements in the set. To mitigate this problem, we introduce LAF (Learning Aggregation Functions), a learnable aggregator for sets of arbitrary cardinality. LAF can approximate several extensively used aggregators (such as average, sum, maximum)  as well as more complex functions (e.g. variance and skewness). We report experiments on semi-synthetic and real data showing that LAF  outperforms state-of-the-art sum- (max-) decomposition architectures such as DeepSets and library-based architectures like Principal Neighborhood Aggregation.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=tf8a4jDRFCv", "arxiv_id": "2012.08482", "paper_pdf": "papers/tf8a4jDRFCv.pdf", "paper_pdf_sha256": "21ae87a965314846c24a5d6a76ea2e46fb7044864a64c16bece07908fc5dfd7e", "paper_pdf_bytes": 2388924, "paper_pdf_source": "openreview", "code_url": "https://github.com/alessandro-t/laf", "code_repository": "alessandro-t/laf", "code_commit": "670aafbad87292343de04fa7a16211fded484ce0", "code_archive": "repos/tf8a4jDRFCv.zip", "code_archive_sha256": "38e63ceb85e6265530147eed80f5d1f9ab58d94ca2c84c764177ac12314c466d", "code_archive_bytes": 2981004, "code_file_count": 10, "code_extensions": {".py": 9, ".sh": 1}, "github_disk_usage_kb": 2952, "github_languages": {"Python": 42444, "Shell": 432}, "github_archived": false, "github_pushed_at": "2021-05-17T15:03:00Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/learning-aggregation-functions-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "RZIfy4Qzxa", "year": 2026, "status": "rejected", "title": "MedSentry: Understanding and Mitigating Safety Risks in Medical LLM Multi-Agent Systems", "authors": ["Kai Chen", "Taihang Zhen", "Hewei Wang", "Kailai Liu", "Xinfeng Li", "Jing Huo", "Tianpei Yang", "Jinfeng Xu", "Wei Dong", "Yang Gao"], "authorids": ["~Kai_Chen37", "~Taihang_Zhen1", "~Hewei_Wang1", "~Kailai_Liu1", "~Xinfeng_Li1", "~Jing_Huo2", "~Tianpei_Yang1", "~Jinfeng_Xu2", "~Wei_Dong5", "~Yang_Gao3"], "authors_source": "OpenReview API", "abstract": "As large language models are increasingly adopted in healthcare, ensuring their safety is critical, particularly in collaborative multi-agent settings. This paper develops an end-to-end attack–defense evaluation workflow to systematically analyze how four representative multi-agent topologies (Layers, SharedPool, Centralized, and Decentralized) behave under attacks from ``dark-personality'' agents. To support the evaluation, we curate MedSentry, a data resource containing 5,000 adversarial medical prompts that span 25 threat topics and 100 subtopics. Our study reveals critical differences in how these architectures handle information contamination and maintain robust decision-making, exposing their underlying vulnerability mechanisms. For example, SharedPool is highly susceptible due to open information sharing, whereas Decentralized exhibits stronger resilience owing to inherent redundancy and isolation. To mitigate these risks, we propose a personality-scale detection and correction mechanism that identifies and rehabilitates malicious agents, restoring safety to near-baseline levels. Taken together, MedSentry provides a rigorous evaluation framework alongside actionable defense strategies, offering guidance for the design of safer LLM-based multi-agent systems in medical contexts.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=RZIfy4Qzxa", "arxiv_id": "2505.20824", "paper_pdf": "papers/RZIfy4Qzxa.pdf", "paper_pdf_sha256": "193101017c3a2ba53a647aca86c9d14ec99a7c308215be9119c26af27b32f27a", "paper_pdf_bytes": 3930456, "paper_pdf_source": "openreview", "code_url": "https://github.com/KaiChenNJ/MedSentry", "code_repository": "KaiChenNJ/MedSentry", "code_commit": "044ba8b36e1bb30a78098b9e9a075072094081c4", "code_archive": "repos/RZIfy4Qzxa.zip", "code_archive_sha256": "7d5d6e75013ac52a428326c48d013c25acc814823bc7de50b92aecaf617d6a9e", "code_archive_bytes": 8939884, "code_file_count": 16, "code_extensions": {".py": 16}, "github_disk_usage_kb": 9979, "github_languages": {}, "github_archived": false, "github_pushed_at": "2025-05-27T05:58:53Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/medsentry-understanding-and-mitigating-safety"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "EWNH3QTSxd", "year": 2025, "status": "rejected", "title": "Which Experiences Are Influential for RL Agents? Efficiently Estimating The Influence of Experiences", "authors": ["Takuya Hiraoka", "Guanquan Wang", "Takashi Onishi", "Yoshimasa Tsuruoka"], "authorids": ["~Takuya_Hiraoka1", "~Guanquan_Wang1", "~Takashi_Onishi1", "~Yoshimasa_Tsuruoka1"], "authors_source": "OpenReview API", "abstract": "In reinforcement learning (RL) with experience replay, experiences stored in a replay buffer influence the RL agent's performance. \nInformation about how these experiences influence the agent's performance is valuable for various purposes, such as identifying experiences that negatively influence underperforming agents. \nOne method for estimating the influence of experiences is the leave-one-out (LOO) method. \nHowever, this method is usually computationally prohibitive. \nIn this paper, we present Policy Iteration with Turn-over Dropout (PIToD), which efficiently estimates the influence of experiences. \nWe evaluate how accurately PIToD estimates the influence of experiences and its efficiency compared to LOO. \nWe then apply PIToD to amend underperforming RL agents, i.e., we use PIToD to estimate negatively influential experiences for the RL agents and to delete the influence of these experiences. \nWe show that RL agents' performance is significantly improved via amendments with PIToD.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=EWNH3QTSxd", "arxiv_id": "2405.14629", "paper_pdf": "papers/EWNH3QTSxd.pdf", "paper_pdf_sha256": "0793ef11fd5ef27c26c488eed081c286bba735152ac3a53263d215c431a8be7e", "paper_pdf_bytes": 6434349, "paper_pdf_source": "openreview", "code_url": "https://github.com/TakuyaHiraoka/Which-Experiences-Are-Influential-for-RL-Agents", "code_repository": "TakuyaHiraoka/Which-Experiences-Are-Influential-for-RL-Agents", "code_commit": "085d8fd2791a50e1aafc2309c35ac05c7863b7be", "code_archive": "repos/EWNH3QTSxd.zip", "code_archive_sha256": "91afbba8af4b0ba18a5f3a6af3e72f0775b7137cb8b683c8add7c70863f91d7a", "code_archive_bytes": 126149, "code_file_count": 18, "code_extensions": {".py": 15, ".sh": 3}, "github_disk_usage_kb": 400, "github_languages": {"Python": 142706, "Shell": 3381}, "github_archived": false, "github_pushed_at": "2025-08-22T19:41:30Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/which-experiences-are-influential-for-rl"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "EPfGHb9Y68", "year": 2024, "status": "rejected", "title": "Continual Offline Reinforcement Learning via Diffusion-based Dual Generative Replay", "authors": ["Jinmei Liu", "Wenbin Li", "Xiangyu Yue", "Chunlin Chen", "Zhi Wang"], "authorids": ["~Jinmei_Liu2", "~Wenbin_Li5", "~Xiangyu_Yue1", "~Chunlin_Chen1", "~Zhi_Wang7"], "authors_source": "OpenReview API", "abstract": "We study continual offline reinforcement learning, a practical paradigm that facilitates forward transfer and mitigates catastrophic forgetting to tackle sequential offline tasks. We propose a dual generative replay framework that retains previous knowledge by concurrent replay of generated pseudo-data. First, we decouple the continual learning policy into a diffusion-based generative behavior model and a multi-head action evaluation model, allowing the policy to inherit distributional expressivity for encompassing a progressive range of diverse behaviors. Second, we train a task-conditioned diffusion model to mimic state distributions of past tasks. Generated states are paired with corresponding responses from the behavior generator to represent old tasks with high-fidelity replayed samples. Finally, by interleaving pseudo samples with real ones of the new task, we continually update the state and behavior generators to model progressively diverse behaviors, and regularize the multi-head critic in a behavior cloning manner to mitigate forgetting. Experiments on various benchmarks demonstrate that our method achieves better forward transfer with less forgetting, and closely approximates results of using previous ground-truth data due to its high-fidelity replay of the sample space.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=EPfGHb9Y68", "arxiv_id": "2404.10662", "paper_pdf": "papers/EPfGHb9Y68.pdf", "paper_pdf_sha256": "e223ceead279b0da07cba3ab31719f26777c65ce83a1fcbb4fe9fcac5665b330", "paper_pdf_bytes": 1330490, "paper_pdf_source": "openreview", "code_url": "https://github.com/NJU-RL/CuGRO", "code_repository": "NJU-RL/CuGRO", "code_commit": "f3f09a4ac96e4e52a92cea946c98813480740af1", "code_archive": "repos/EPfGHb9Y68.zip", "code_archive_sha256": "92addfb40285eabf2d908b9c68798989216a044c430c2a8098964577fab19294", "code_archive_bytes": 516457, "code_file_count": 30, "code_extensions": {".py": 29, ".sh": 1}, "github_disk_usage_kb": 440, "github_languages": {"Python": 224631, "Shell": 1195}, "github_archived": false, "github_pushed_at": "2024-04-14T13:55:38Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/continual-offline-reinforcement-learning-via"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "UhEJz3wgLnG", "year": 2023, "status": "rejected", "title": "Revealing Single Frame Bias for Video-and-Language Learning", "authors": ["Jie Lei", "Tamara L Berg", "Mohit Bansal"], "authorids": ["~Jie_Lei3", "~Tamara_L_Berg1", "~Mohit_Bansal2"], "authors_source": "OpenReview API", "abstract": "Training an effective video-and-language model intuitively requires multiple frames as model inputs. \nHowever, it is unclear whether using multiple frames is beneficial to downstream tasks, and if yes, whether the performance gain is worth the drastically-increased computation and memory costs resulting from using more frames.\nIn this work, we explore single-frame models for video-and-language learning.\nOn a diverse set of video-and-language tasks (including text-to-video retrieval and video question answering), we show the surprising result that, with large-scale pre-training and a proper frame ensemble strategy at inference time, a single-frame trained model that does not consider temporal information can achieve better performance than existing methods that use multiple frames for training.\nThis result reveals the existence of a strong ``static appearance bias'' in popular video-and-language datasets.\nTherefore, to allow for a more comprehensive evaluation of video-and-language models, we propose two new retrieval tasks based on existing fine-grained action recognition datasets that encourage temporal modeling.\nFull code and models will be made publicly available upon acceptance.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=UhEJz3wgLnG", "arxiv_id": "2206.03428", "paper_pdf": "papers/UhEJz3wgLnG.pdf", "paper_pdf_sha256": "05c4ee57035651a5f85ee84636374344f8eab67989d550e50bd340d8e35667c3", "paper_pdf_bytes": 1282037, "paper_pdf_source": "openreview", "code_url": "https://github.com/jayleicn/singularity", "code_repository": "jayleicn/singularity", "code_commit": "10a7f7062ab8af187b3a47803851b4e462191492", "code_archive": "repos/UhEJz3wgLnG.zip", "code_archive_sha256": "b11098042acc3621001a2cfc26f34a781e51be82ddb73a605a27aaa1e6213e4e", "code_archive_bytes": 1216638, "code_file_count": 34, "code_extensions": {".py": 28, ".sh": 6}, "github_disk_usage_kb": 1202, "github_languages": {"Python": 290881, "Shell": 18415}, "github_archived": false, "github_pushed_at": "2023-05-05T17:22:22Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/revealing-single-frame-bias-for-video-and"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "B0JH7vR2iGh", "year": 2022, "status": "rejected", "title": "PMIC: Improving Multi-Agent Reinforcement Learning with Progressive Mutual Information Collaboration", "authors": ["Pengyi Li", "Hongyao Tang", "Tianpei Yang", "Xiaotian Hao", "Sang Tong", "YAN ZHENG", "Jianye HAO", "Matthew E. Taylor", "Jinyi Liu"], "authorids": ["~Pengyi_Li1", "~Hongyao_Tang1", "~Tianpei_Yang1", "~Xiaotian_Hao1", "~Sang_Tong1", "~YAN_ZHENG1", "~Jianye_HAO1", "~Matthew_E._Taylor2", "~Jinyi_Liu1"], "authors_source": "OpenReview API", "abstract": "Learning to collaborate is critical in multi-agent reinforcement learning (MARL). A branch of previous works proposes to promote collaboration by maximizing the correlation of agents’ behaviors, which is typically characterised by mutual information (MI) in different forms. However, simply maximizing the MI of agents’ behaviors cannot guarantee achieving better collaboration because suboptimal collaboration can also lead to high MI. In this paper, we first propose a new collaboration criterion to evaluate collaboration from three perspectives, which arrives at a form of the mutual information between global state and joint policy. This bypasses the introduction of explicit additional input of policies and mitigates the scalability issue meanwhile. Moreover, to better leverage MI-based collaboration signals, we propose a novel MARL framework, called Progressive Mutual Information Collaboration (PMIC) which contains two main components. The first component is Dual Progressive Collaboration Buffer (DPCB) which separately stores superior and inferior trajectories in a progressive manner. The second component is Dual Mutual Information Estimator (DMIE), including two neural estimators of our new designed MI based on separate samples in DPCB. We then make use of the neural MI estimates to improve agents' policies: to maximize the MI lower bound associated with superior collaboration to facilitate better collaboration and to minimize the MI upper bound associated with inferior collaboration to avoid falling into local optimal. PMIC is general and can be combined with existing MARL algorithms. Experiments on a wide range of MARL benchmarks show the superior performance of PMIC compared with other MARL algorithms.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=B0JH7vR2iGh", "arxiv_id": "2203.08553", "paper_pdf": "papers/B0JH7vR2iGh.pdf", "paper_pdf_sha256": "0ad7a1596fb2cb8ad546fe7166678a956a7760fa6adb1dc5f14e649b8430c7e5", "paper_pdf_bytes": 17889723, "paper_pdf_source": "openreview", "code_url": "https://github.com/yeshenpy/PMIC", "code_repository": "yeshenpy/PMIC", "code_commit": "ab3fee60b1c75d65c3c3f209397130c059eedb41", "code_archive": "repos/B0JH7vR2iGh.zip", "code_archive_sha256": "23663713a6b2dd533928d8dbd6b131569352fded2dc3d5a7025cff66d2fc2888", "code_archive_bytes": 5562356, "code_file_count": 234, "code_extensions": {".py": 232, ".ipynb": 1, ".sh": 1}, "github_disk_usage_kb": 5374, "github_languages": {"Python": 1131343, "HTML": 592640, "Jupyter Notebook": 580298, "Shell": 1069, "Dockerfile": 465}, "github_archived": false, "github_pushed_at": "2024-03-26T06:41:25Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/pmic-improving-multi-agent-reinforcement-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "bVzUDC_4ls", "year": 2021, "status": "rejected", "title": "Exploiting Verified Neural Networks via Floating Point Numerical Error", "authors": ["Kai Jia", "Martin Rinard"], "authorids": ["~Kai_Jia2", "~Martin_Rinard1"], "authors_source": "OpenReview API", "abstract": "Motivated by the need to reliably characterize the robustness of deep neural networks, researchers have developed verification algorithms for deep neural networks. Given a neural network, the verifiers aim to answer whether certain properties are guaranteed with respect to all inputs in a space. However, little attention has been paid to floating point numerical error in neural network verification.\n\nWe exploit floating point errors in the inference and verification implementations to construct adversarial examples for neural networks that a verifier claims to be robust with respect to certain inputs. We argue that, to produce sound verification results, any verification system must accurately (or conservatively) model the effects of any float point computations in the network inference or verification system.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=bVzUDC_4ls", "arxiv_id": "2003.03021", "paper_pdf": "papers/bVzUDC_4ls.pdf", "paper_pdf_sha256": "847acf7563e5ed238a4031f77096ca39ad7b86d698ee3517d0572851eac092d3", "paper_pdf_bytes": 411648, "paper_pdf_source": "openreview", "code_url": "https://github.com/jia-kai/realadv", "code_repository": "jia-kai/realadv", "code_commit": "f72d119b023deee83fe06d11b22adecdc104253c", "code_archive": "repos/bVzUDC_4ls.zip", "code_archive_sha256": "dfb60e8a7aac38e4f8277b6a8d4ac54d520fd2bc003f7bd0bd149c7b755445e3", "code_archive_bytes": 3381036, "code_file_count": 38, "code_extensions": {".py": 28, ".sh": 5, ".h": 2, ".jl": 2, ".cpp": 1}, "github_disk_usage_kb": 3146, "github_languages": {"Python": 156906, "C++": 8184, "Julia": 5982, "Shell": 2641, "Cython": 1584, "Makefile": 126}, "github_archived": false, "github_pushed_at": "2021-10-16T15:33:03Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/exploiting-verified-neural-networks-via"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "mDEYl0Ucgr", "year": 2025, "status": "rejected", "title": "Influencing Humans to Conform to Preference Models for RLHF", "authors": ["Stephane Hatgis-Kessell", "W. Bradley Knox", "Serena Booth", "Scott Niekum", "Peter Stone"], "authorids": ["~Stephane_Hatgis-Kessell1", "~W._Bradley_Knox2", "~Serena_Booth1", "~Scott_Niekum1", "~Peter_Stone1"], "authors_source": "OpenReview API", "abstract": "Designing a reinforcement learning from human feedback (RLHF) algorithm for learning from preferences requires assuming a preference model, sometimes implicitly.  A preference model that poorly describes how humans generate preferences risks learning a poor approximation of the human’s unobservable reward function. In this paper, we conduct three human studies to assess whether one can influence the expression of real human preferences to more closely conform to a desired preference model. Importantly, our approach does not seek to alter the human's unobserved reward function. Rather, we change how humans use this reward function to generate preferences, such that they better match whatever preference model is assumed by a particular RLHF algorithm. We introduce three interventions: showing humans the quantities that underlie a preference model, which is normally unobservable information derived from the reward function; training people to follow a specific preference model; and modifying the preference elicitation question. All intervention types show significant effects, providing practical tools to improve preference data quality and the resultant alignment of learned reward functions.Overall we establish a novel research direction in model alignment: training humans and designing interfaces to increase human conformance with the assumptions of the algorithm that will learn from their input.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=mDEYl0Ucgr", "arxiv_id": "2501.06416", "paper_pdf": "papers/mDEYl0Ucgr.pdf", "paper_pdf_sha256": "604b567c1eb9a349480d52bd84748c6223e961da6757b89cb9066528ff9d4c68", "paper_pdf_bytes": 4269937, "paper_pdf_source": "openreview", "code_url": "https://github.com/Stephanehk/InfluencingHumanPrefs", "code_repository": "Stephanehk/InfluencingHumanPrefs", "code_commit": "54ad2720c1d5638574fc5d788a53008aef2c1c12", "code_archive": "repos/mDEYl0Ucgr.zip", "code_archive_sha256": "d8288df4964091f57cabcce041c6af1cde1e0219d53c01654584bcef7248a4ce", "code_archive_bytes": 480186, "code_file_count": 47, "code_extensions": {".py": 44, ".sh": 3}, "github_disk_usage_kb": 405, "github_languages": {"Python": 319533, "Shell": 1124}, "github_archived": false, "github_pushed_at": "2024-12-30T06:25:37Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/influencing-humans-to-conform-to-preference"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "idpV2AqusC", "year": 2024, "status": "rejected", "title": "Improving SAM Requires Rethinking its Optimization Formulation", "authors": ["Wanyun Xie", "Fabian Latorre", "Kimon Antonakopoulos", "Thomas Pethick", "Volkan Cevher"], "authorids": ["~Wanyun_Xie1", "~Fabian_Latorre1", "~Kimon_Antonakopoulos1", "~Thomas_Pethick1", "~Volkan_Cevher1"], "authors_source": "OpenReview API", "abstract": "This paper rethinks Sharpness-Aware Minimization (SAM), which is originally formulated as a zero-sum game where the weights of a network and a bounded perturbation try to minimize/maximize, respectively, the same differentiable loss. We argue that SAM should instead be reformulated using the 0-1 loss, as this provides a tighter bound on its generalization gap. As a continuous relaxation, we follow the simple conventional approach where the minimizing (maximizing) player uses an upper bound (lower bound) surrogate to the 0-1 loss. This leads to a novel formulation of SAM as a bilevel optimization problem, dubbed as BiSAM. Through numerical evidence, we show that BiSAM consistently results in improved performance when compared to the original SAM and variants, while enjoying similar computational complexity.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=idpV2AqusC", "arxiv_id": "2407.12993", "paper_pdf": "papers/idpV2AqusC.pdf", "paper_pdf_sha256": "043920848e62cae574048dcad12a84129687839c6117ce8fb755c2ff31f6f07a", "paper_pdf_bytes": 558257, "paper_pdf_source": "openreview", "code_url": "https://github.com/LIONS-EPFL/BiSAM", "code_repository": "LIONS-EPFL/BiSAM", "code_commit": "c6595703430d062f36d9f111c10d7be44b0fb22e", "code_archive": "repos/idpV2AqusC.zip", "code_archive_sha256": "ab1beb6d0b704f0ba877699e997800cab6e64a79a38143fe519cf2c0e3223753", "code_archive_bytes": 449352, "code_file_count": 17, "code_extensions": {".py": 17}, "github_disk_usage_kb": 442, "github_languages": {"Python": 41821}, "github_archived": false, "github_pushed_at": "2026-08-28T13:56:51Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/improving-sam-requires-rethinking-its"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "wysXxmukfCA", "year": 2023, "status": "rejected", "title": "Towards Robust Model Watermark via Reducing Parametric Vulnerability", "authors": ["Guanhao Gan", "Yiming Li", "Dongxian Wu", "Shu-Tao Xia"], "authorids": ["~Guanhao_Gan1", "~Yiming_Li1", "~Dongxian_Wu1", "~Shu-Tao_Xia1"], "authors_source": "OpenReview API", "abstract": "Deep neural networks are valuable assets considering their commercial benefits and huge demands for costly annotation and computation resources. To protect the copyright of these deep models, backdoor-based ownership verification becomes popular recently, in which the model owner can watermark the model by embedding a specific behavior before releasing it. The defender (usually the model owner) can identify whether a suspicious third-party model is ``stolen'' from it based on the presence of the behavior. Unfortunately, these watermarks are proven to be vulnerable to removal attacks even like fine-tuning. To further explore this vulnerability, we investigate the parametric space and find there exist many watermark-removed models in the vicinity of the watermarked one, which may be easily used by removal attacks. Inspired by this finding, we propose a minimax formulation to find these watermark-removed models and recover their watermark behavior. Extensive experiments demonstrate that our method improves the robustness of the model watermarking against parametric changes and numerous watermark-removal attacks.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=wysXxmukfCA", "arxiv_id": "2309.04777", "paper_pdf": "papers/wysXxmukfCA.pdf", "paper_pdf_sha256": "c2eb899938f77250097769f1e6beab07bdb6ca566f0b2978b82ad4b968636f45", "paper_pdf_bytes": 1905528, "paper_pdf_source": "openreview", "code_url": "https://github.com/GuanhaoGan/robust-model-watermarking", "code_repository": "GuanhaoGan/robust-model-watermarking", "code_commit": "e20c87526a87c86ca9613122cf264166a20b8e6f", "code_archive": "repos/wysXxmukfCA.zip", "code_archive_sha256": "c6b7174641238a2005104d70ea27bdb171f797a2e9d829e19a89178042edfc21", "code_archive_bytes": 1137889, "code_file_count": 15, "code_extensions": {".py": 15}, "github_disk_usage_kb": 1222, "github_languages": {"Python": 141175}, "github_archived": false, "github_pushed_at": "2024-06-03T14:53:27Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/towards-robust-model-watermark-via-reducing"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "JmU7lyDxTpc", "year": 2022, "status": "rejected", "title": "Multi-scale Feature Learning Dynamics: Insights for Double Descent", "authors": ["Mohammad Pezeshki", "Amartya Mitra", "Yoshua Bengio", "Guillaume Lajoie"], "authorids": ["~Mohammad_Pezeshki1", "~Amartya_Mitra1", "~Yoshua_Bengio1", "~Guillaume_Lajoie1"], "authors_source": "OpenReview API", "abstract": "A key challenge in building theoretical foundations for deep learning is the complex optimization dynamics of neural networks, resulting from the high-dimensional interactions between the large number of network parameters. Such non-trivial interactions lead to intriguing model behaviors such as the phenomenon  of \"double descent\" of the generalization error. The more commonly studied aspect of this phenomenon corresponds to model-wise double descent where the test error exhibits a second descent with increasing model complexity, beyond the classical U-shaped error curve. In this work, we investigate the origins of the less studied epoch-wise double descent in which the test error undergoes two non-monotonous transitions, or descents as the training time increases. We study a linear teacher-student setup exhibiting epoch-wise double descent similar to that in deep neural networks. In this setting, we derive closed-form analytical expressions for the evolution of generalization error over training. We find that double descent can be attributed to distinct features being learned at different scales: as fast-learning features overfit, slower-learning features start to fit, resulting in a second descent in test error. We validate our findings through numerical experiments where our theory accurately predicts empirical findings and remains consistent with observations in deep neural networks.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=JmU7lyDxTpc", "arxiv_id": "2112.03215", "paper_pdf": "papers/JmU7lyDxTpc.pdf", "paper_pdf_sha256": "d0d734b323f9a08c5b8c9d5ff07245f68e13ac91ea1a3590e45fb3d6a446ed3f", "paper_pdf_bytes": 2529304, "paper_pdf_source": "openreview", "code_url": "https://github.com/NNdoubledescent/doubledescent", "code_repository": "NNdoubledescent/doubledescent", "code_commit": "3e2ef3e1247462824df8057db2605ef5519a0525", "code_archive": "repos/JmU7lyDxTpc.zip", "code_archive_sha256": "c32570ed33c779ee82bb6d7498f58195d4c8813a651b16f4119ede88dfb73702", "code_archive_bytes": 5544380, "code_file_count": 13, "code_extensions": {".py": 12, ".sh": 1}, "github_disk_usage_kb": 5414, "github_languages": {"Python": 35946, "Shell": 110}, "github_archived": false, "github_pushed_at": "2021-11-29T20:24:16Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/multi-scale-feature-learning-dynamics-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "N9oPAFcuYWX", "year": 2021, "status": "rejected", "title": "Understanding and Mitigating Accuracy Disparity in Regression", "authors": ["Jianfeng Chi", "Han Zhao", "Geoff Gordon", "Yuan Tian"], "authorids": ["~Jianfeng_Chi1", "~Han_Zhao1", "~Geoff_Gordon2", "~Yuan_Tian2"], "authors_source": "OpenReview API", "abstract": "With the widespread deployment of large-scale prediction systems in high-stakes domains, e.g., face recognition, criminal justice, etc., disparity on prediction accuracy between different demographic subgroups has called for fundamental understanding on the source of such disparity and algorithmic intervention to mitigate it. In this paper, we study the accuracy disparity problem in regression. To begin with, we first propose an error decomposition theorem, which decomposes the accuracy disparity into the distance between label populations and the distance between conditional representations, to help explain why such accuracy disparity appears in practice. Motivated by this error decomposition and the general idea of distribution alignment with statistical distances, we then propose an algorithm to reduce this disparity, and analyze its game-theoretic optima of the proposed objective function. We conduct experiments on four real-world datasets. The experimental results suggest that our proposed algorithms can effectively mitigate accuracy disparity while maintaining the predictive power of the regression models.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=N9oPAFcuYWX", "arxiv_id": "2102.12013", "paper_pdf": "papers/N9oPAFcuYWX.pdf", "paper_pdf_sha256": "5abb64976ef9f17c2070ed1a7233127d981c7d16325765f47e46466e62e395c0", "paper_pdf_bytes": 531827, "paper_pdf_source": "openreview", "code_url": "https://github.com/JFChi/Understanding-and-Mitigating-Accuracy-Disparity-in-Regression", "code_repository": "JFChi/Understanding-and-Mitigating-Accuracy-Disparity-in-Regression", "code_commit": "30bd1c82a5ccbcb84a23eb12da9071596796283b", "code_archive": "repos/N9oPAFcuYWX.zip", "code_archive_sha256": "912385ce8d3e776d18c76e46e45ec6fa50ec8c35b3a68525dff02b25892419c4", "code_archive_bytes": 3045097, "code_file_count": 12, "code_extensions": {".py": 12}, "github_disk_usage_kb": 3317, "github_languages": {"Python": 109101}, "github_archived": false, "github_pushed_at": "2021-12-29T21:48:21Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/understanding-and-mitigating-accuracy-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "IgrLJslvxa", "year": 2025, "status": "rejected", "title": "PoisonBench: Assessing Large Language Model Vulnerability to Data Poisoning", "authors": ["Tingchen Fu", "Mrinank Sharma", "Philip Torr", "Yonadav G Shavit", "Shay B Cohen", "David Krueger", "Fazl Barez"], "authorids": ["~Tingchen_Fu1", "~Mrinank_Sharma1", "~Philip_Torr1", "~Yonadav_G_Shavit1", "~Shay_B_Cohen1", "~David_Krueger1", "~Fazl_Barez1"], "authors_source": "OpenReview API", "abstract": "Preference learning is a central component for aligning current LLMs, but this process can be vulnerable to data poisoning attacks. To address this concern, we introduce PoisonBench, a benchmark for evaluating large language models' susceptibility to data poisoning during preference learning. Data poisoning attacks can manipulate large language model responses to include hidden malicious content or biases, potentially causing the model to generate harmful or unintended outputs while appearing to function normally. We deploy two distinct attack types across eight realistic scenarios, assessing 22 widely-used models. Our findings reveal concerning trends: (1) Scaling up parameter size does not always enhance resilience against poisoning attacks and the influence on model resilience varies among different model suites. (2) There exists a log-linear relationship between the effects of the attack and the data poison ratio; (3) The effect of data poisoning can generalize to extrapolated triggers that are not included in the poisoned data. \nThese results expose weaknesses in current preference learning techniques, highlighting the urgent need for more robust defenses against malicious models and data manipulation.", "decision": "Reject", "meta_review": "", "num_reviews": 6, "openreview_url": "https://openreview.net/forum?id=IgrLJslvxa", "arxiv_id": "2410.08811", "paper_pdf": "papers/IgrLJslvxa.pdf", "paper_pdf_sha256": "6a1b8cfb2f98a2fe76c0be3ce895402e97f740527d02ff266aeaabfcc1f7ecd4", "paper_pdf_bytes": 847280, "paper_pdf_source": "openreview", "code_url": "https://github.com/TingchenFu/PoisonBench", "code_repository": "TingchenFu/PoisonBench", "code_commit": "61096ab8b740e782a666f365bd0003c98660dc41", "code_archive": "repos/IgrLJslvxa.zip", "code_archive_sha256": "60f8341ead2f2e0eb646be8fbec23d684cb1e0de9de11a1e9e66192fc623f925", "code_archive_bytes": 409770, "code_file_count": 10, "code_extensions": {".py": 7, ".sh": 3}, "github_disk_usage_kb": 406, "github_languages": {"Python": 26147, "Shell": 7374}, "github_archived": false, "github_pushed_at": "2024-10-19T19:04:15Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/poisonbench-assessing-large-language-model"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "WO4BCqEyWc", "year": 2024, "status": "rejected", "title": "Augmentation-aware Self-Supervised Learning with Conditioned Projector", "authors": ["Marcin Przewięźlikowski", "Mateusz Pyla", "Bartosz Michał Zieliński", "Bartłomiej Twardowski", "Jacek Tabor", "Marek Śmieja"], "authorids": ["~Marcin_Przewięźlikowski1", "~Mateusz_Pyla1", "~Bartosz_Michał_Zieliński1", "~Bartłomiej_Twardowski1", "~Jacek_Tabor1", "~Marek_Śmieja1"], "authors_source": "OpenReview API", "abstract": "Self-supervised learning (SSL) is a powerful technique for learning robust representations from unlabeled data. By learning to remain invariant to applied data augmentations, methods such as SimCLR and MoCo are able to reach quality on par with supervised approaches. However, this invariance may be harmful to solving some downstream tasks which depend on traits affected by augmentations used during pretraining, such as color. In this paper, we propose to foster sensitivity to such characteristics in the representation space by modifying the projector network, a common component of self-supervised architectures. Specifically, we supplement the projector with information about augmentations applied to images. In order for the projector to take advantage of this auxiliary conditioning when solving the SSL task, the feature extractor learns to preserve the augmentation information in its representations. Our approach, coined Conditional Augmentation-aware Self-supervised Learning (CASSLE), is directly applicable to typical joint-embedding SSL methods regardless of their objective functions. Moreover, it does not require major changes in the network architecture or prior knowledge of downstream tasks. In addition to an analysis of sensitivity towards different data augmentations, we conduct a series of experiments, which show that CASSLE improves over various SSL methods, reaching state-of-the-art performance in multiple downstream tasks.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=WO4BCqEyWc", "arxiv_id": "2306.06082", "paper_pdf": "papers/WO4BCqEyWc.pdf", "paper_pdf_sha256": "8e00967f8258c9acb8baa2a1e2df372d1fe052f741e6e0519e618828ed555c68", "paper_pdf_bytes": 6528454, "paper_pdf_source": "openreview", "code_url": "https://github.com/gmum/CASSLE", "code_repository": "gmum/CASSLE", "code_commit": "6e345133c7525644f2dd98cf9324b55d0014fbde", "code_archive": "repos/WO4BCqEyWc.zip", "code_archive_sha256": "5ea97108385012ef2db88d8dd18e5ab33dd2166d8d29575861784873ae50441f", "code_archive_bytes": 356973, "code_file_count": 20, "code_extensions": {".py": 20}, "github_disk_usage_kb": 470, "github_languages": {"Python": 230379}, "github_archived": false, "github_pushed_at": "2026-07-16T08:52:45Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/augmentation-aware-self-supervised-learning"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "vNrmEgfGIg3", "year": 2023, "status": "rejected", "title": "Filtered Semi-Markov CRF", "authors": ["Urchade Zaratiana", "Nadi Tomeh", "Niama Elkhbir", "Pierre Holat", "Thierry Charnois"], "authorids": ["~Urchade_Zaratiana1", "~Nadi_Tomeh1", "~Niama_Elkhbir1", "~Pierre_Holat1", "~Thierry_Charnois2"], "authors_source": "OpenReview API", "abstract": "Semi-Markov CRF \\citep{semicrf} has been proposed as an alternative to the traditional Linear Chain CRF\\citep{crf} for text segmentation tasks such as Named Entity Recognition. In contrast to CRF, which treats text segmentation as token-level prediction, Semi-CRF considers spans as the task's basic unit, which makes it more expressive. However, Semi-CRF has two major drawbacks: (1) it has quadratic complexity over sequence length as it operates on every span of the input sequence, and (2) empirically, it performs worse than classical CRF for sequence labeling tasks such as NER. In our work, we propose Filtered Semi-Markov CRF, a Semi-CRF variant that addresses the aforementioned issues. Our model extends Semi-CRF by incorporating a filtering step for eliminating irrelevant segments, which helps in reducing the complexity and allows to dramatically reduce the search space. On a variety of NER benchmarks, we find that our approach outperforms both CRF and Semi-CRF models while being significantly faster. We will make our code available to the public. ", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=vNrmEgfGIg3", "arxiv_id": "2311.18028", "paper_pdf": "papers/vNrmEgfGIg3.pdf", "paper_pdf_sha256": "25e7e8eb6bdb4345bd9c55f1f48d59346e88a55768a9a3859d15455ac2ca58fa", "paper_pdf_bytes": 524729, "paper_pdf_source": "openreview", "code_url": "https://github.com/urchade/Filtered-Semi-Markov-CRF", "code_repository": "urchade/Filtered-Semi-Markov-CRF", "code_commit": "77da8cdc5f3ab8fbe805dce115e67688b11fd130", "code_archive": "repos/vNrmEgfGIg3.zip", "code_archive_sha256": "f90b3b684ccd1cf84fc3408f47ffab751559ac0b5e9afec9ab230ca83c9756fb", "code_archive_bytes": 1024684, "code_file_count": 15, "code_extensions": {".py": 15}, "github_disk_usage_kb": 1223, "github_languages": {"Python": 48627}, "github_archived": false, "github_pushed_at": "2024-01-05T19:48:13Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/filtered-semi-markov-crf"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "X3WxnuzAYyE", "year": 2022, "status": "rejected", "title": "PKCAM: Previous Knowledge Channel Attention Module", "authors": ["Eslam Mohamed BAKR", "Ahmad A. Al Sallab", "Mohsen Rashwan"], "authorids": ["~Eslam_Mohamed_BAKR1", "~Ahmad_A._Al_Sallab1", "~Mohsen_Rashwan1"], "authors_source": "OpenReview API", "abstract": "Attention mechanisms have been explored with CNNs, both across the spatial and channel dimensions. \nHowever, all the existing methods devote the attention modules to capture local interactions from the current feature map only, disregarded the valuable previous knowledge that is acquired by the earlier layers. \nThis paper tackles the following question: Can one incorporate previous knowledge aggregation while learning channel attention more efficiently? To this end, we propose a Previous Knowledge Channel Attention Module( PKCAM), that captures channel-wise relations across different layers to model the global context. \nOur proposed module PKCAM is easily integrated into any feed-forward CNN architectures and trained in an end-to-end fashion with a negligible footprint due to its lightweight property. We validate our novel architecture through extensive experiments on image classification and object detection tasks with different backbones. \nOur experiments show consistent improvements in performances against their counterparts. We also conduct experiments that probe the robustness of the learned representations.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=X3WxnuzAYyE", "arxiv_id": "2211.07521", "paper_pdf": "papers/X3WxnuzAYyE.pdf", "paper_pdf_sha256": "99fb3ad6524e4930e8663b5ecf321a061f4fd50035253a0fc3a41163075674d4", "paper_pdf_bytes": 1128222, "paper_pdf_source": "openreview", "code_url": "https://github.com/eslambakr/EMCA", "code_repository": "eslambakr/EMCA", "code_commit": "76bae427ea2f66979aed6a9dcef5e84ed922d14e", "code_archive": "repos/X3WxnuzAYyE.zip", "code_archive_sha256": "d62d9ceb8af92d4b877636f76f7f7a63d1cfef19891d0dde257f65ccc6e294f6", "code_archive_bytes": 2366697, "code_file_count": 28, "code_extensions": {".py": 28}, "github_disk_usage_kb": 5864, "github_languages": {"Python": 294765}, "github_archived": false, "github_pushed_at": "2021-11-24T18:02:44Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/pkcam-previous-knowledge-channel-attention-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "E4PK0rg2eP", "year": 2021, "status": "rejected", "title": "Parameter-Efficient Transfer Learning with Diff Pruning", "authors": ["Demi Guo", "Alexander M Rush", "Yoon Kim"], "authorids": ["~Demi_Guo1", "~Alexander_M_Rush1", "~Yoon_Kim1"], "authors_source": "OpenReview API", "abstract": "While task-specific finetuning of deep networks pretrained with self-supervision has led to significant empirical advances in NLP, their large size makes the standard finetuning approach difficult to apply to multi-task, memory-constrained settings, as storing the full model parameters for each task become prohibitively expensive. We propose $\\textit{diff pruning}$ as a simple approach to enable parameter-efficient transfer learning within the pretrain-finetune framework. This approach views finetuning as learning a task-specific diff vector that is applied on top of the pretrained parameter vector, which remains fixed and is shared across different tasks. The diff vector is adaptively pruned during training with a differentiable approximation to the $L_0$-norm penalty to encourage sparsity. Diff pruning becomes parameter-efficient as the number of tasks increases, as it requires storing only the nonzero positions and weights of the diff vector for each task, while the cost of storing the shared pretrained model remains constant. We find that models finetuned with diff pruning can match the performance of fully finetuned baselines on the GLUE benchmark while only modifying 0.5$\\%$ of the pretrained model's parameters per task.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=E4PK0rg2eP", "arxiv_id": "2012.07463", "paper_pdf": "papers/E4PK0rg2eP.pdf", "paper_pdf_sha256": "a3966e58b21caeca135ff1ac346dd6e1587ca0c648c5fb61992cc59cfecb155d", "paper_pdf_bytes": 424130, "paper_pdf_source": "openreview", "code_url": "https://github.com/dguo98/DiffPruning", "code_repository": "dguo98/DiffPruning", "code_commit": "085fa7d9e2bcbcd19f705a08b1c0c6c74c54bada", "code_archive": "repos/E4PK0rg2eP.zip", "code_archive_sha256": "f11d18a2726c04ebd656b7d948cd42ba78b6ac11635e7c6383e40ac4dbf32af5", "code_archive_bytes": 6843736, "code_file_count": 471, "code_extensions": {".py": 452, ".sh": 10, ".ipynb": 7, ".js": 2}, "github_disk_usage_kb": 3373, "github_languages": {"Python": 6386215, "Jupyter Notebook": 596080, "Shell": 9196, "CSS": 5477, "Dockerfile": 3967, "Makefile": 1937}, "github_archived": false, "github_pushed_at": "2021-02-03T08:34:18Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/parameter-efficient-transfer-learning-with-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ZPZ4eCQU9k", "year": 2025, "status": "rejected", "title": "xLSTM-Mixer: Multivariate Time Series Forecasting by Mixing via Scalar Memories", "authors": ["Maurice Kraus", "Felix Divo", "Devendra Singh Dhami", "Kristian Kersting"], "authorids": ["~Maurice_Kraus1", "~Felix_Divo1", "~Devendra_Singh_Dhami1", "~Kristian_Kersting1"], "authors_source": "OpenReview API", "abstract": "Time series data is prevalent across numerous fields, necessitating the development of robust and accurate forecasting models. Capturing patterns both within and between temporal and multivariate components is crucial for reliable predictions.\nWe introduce xLSTM-Mixer, a model designed to effectively integrate temporal sequences, joint time-variate information, and multiple perspectives for robust forecasting. Our approach begins with a linear forecast shared across variates, which is then refined by xLSTM blocks. They serve as key elements for modeling the complex dynamics of challenging time series data. xLSTM-Mixer ultimately\nreconciles two distinct views to produce the final forecast. Our extensive evaluations demonstrate its superior long-term forecasting performance compared to recent state-of-the-art methods. A thorough model analysis provides further insights into its key components and confirms its robustness and effectiveness. This work contributes to the resurgence of recurrent models in time series forecasting.", "decision": "Reject", "meta_review": "", "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=ZPZ4eCQU9k", "arxiv_id": "2410.16928", "paper_pdf": "papers/ZPZ4eCQU9k.pdf", "paper_pdf_sha256": "5f4b67ff8b6aec7b74f60e40a0cef8756d22cf79cfee7e2a3d46ef2ef045df1c", "paper_pdf_bytes": 587538, "paper_pdf_source": "openreview", "code_url": "https://github.com/mauricekraus/xLSTM-Mixer", "code_repository": "mauricekraus/xLSTM-Mixer", "code_commit": "730b0531aa9456e498765028f3c22ca3677de42e", "code_archive": "repos/ZPZ4eCQU9k.zip", "code_archive_sha256": "02df029162721aa1433c38ddee9fed02f661ca724ef35e42e00623115a93abf8", "code_archive_bytes": 434103, "code_file_count": 88, "code_extensions": {".py": 80, ".sh": 7, ".ipynb": 1}, "github_disk_usage_kb": 407, "github_languages": {"Python": 582249, "Jupyter Notebook": 53199, "Shell": 47738, "Dockerfile": 1137}, "github_archived": false, "github_pushed_at": "2025-06-12T15:21:40Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/xlstm-mixer-multivariate-time-series"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "PoBB8n52oi", "year": 2024, "status": "rejected", "title": "SummaryMixing: A Linear-Complexity Alternative to Self-Attention for Speech Recognition and Understanding", "authors": ["Titouan Parcollet", "Rogier van Dalen", "Shucong Zhang", "Sourav Bhattacharya"], "authorids": ["~Titouan_Parcollet1", "~Rogier_van_Dalen2", "~Shucong_Zhang2", "~Sourav_Bhattacharya1"], "authors_source": "OpenReview API", "abstract": "Modern speech processing systems rely on self-attention. Unfortunately, token\nmixing with self-attention takes quadratic time in the length of the speech utterance,\nslowing down inference as well as training and increasing memory consumption.\nCheaper alternatives to self-attention for ASR have been developed, but they fail to\nconsistently reach the same level of accuracy. However, attention layers in trained\nspeech recognizers tend to not capture fine-grained pair-wise information. This\npaper, therefore, proposes a novel linear-time alternative to self-attention. It sum-\nmarises a whole utterance with the mean over vectors for all time steps. This single\nsummary is then combined with time-specific information. We call this method\n“SummaryMixing”. Introducing SummaryMixing in state-of-the-art ASR models\nmakes it feasible to preserve or exceed previous speech recognition performance\nwhile lowering the training and inference times by up to 28% and reducing the\nmemory budget by a factor of two. The benefits of SummaryMixing can also be\ngeneralized to other speech-processing tasks, such as speech understanding.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=PoBB8n52oi", "arxiv_id": "2307.07421", "paper_pdf": "papers/PoBB8n52oi.pdf", "paper_pdf_sha256": "83e672dec3700172524e1d3c3deb24b78e26f07578abc6a132c6ec2bf6c4d83e", "paper_pdf_bytes": 293159, "paper_pdf_source": "openreview", "code_url": "https://github.com/SamsungLabs/SummaryMixing", "code_repository": "SamsungLabs/SummaryMixing", "code_commit": "d1b1f425149cea28f0e0318de82e525af2523b70", "code_archive": "repos/PoBB8n52oi.zip", "code_archive_sha256": "d534a7932c656f9138cd44286daf34eca03cb6ad206a64e31688a5439f84c857", "code_archive_bytes": 460239, "code_file_count": 7, "code_extensions": {".py": 7}, "github_disk_usage_kb": 491, "github_languages": {"Python": 148286}, "github_archived": false, "github_pushed_at": "2025-06-24T09:23:38Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/sumformer-a-linear-complexity-alternative-to"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "me09xlTmm8", "year": 2023, "status": "rejected", "title": "Transport with Support: Data-Conditional Diffusion Bridges", "authors": ["Ella Tamir", "Martin Trapp", "Arno Solin"], "authorids": ["~Ella_Tamir1", "~Martin_Trapp2", "~Arno_Solin1"], "authors_source": "OpenReview API", "abstract": "The dynamic Schrödinger bridge problem provides an appealing setting for posing optimal transport problems as learning non-linear diffusion processes and enables efficient iterative solvers. Recent works have demonstrated state-of-the-art results (eg, in modelling single-cell embryo RNA sequences or sampling from complex posteriors) but are typically limited to learning bridges with only initial and terminal constraints. Our work extends this paradigm by proposing the Iterative Smoothing Bridge (ISB). We combine learning diffusion models with Bayesian filtering and optimal control, allowing for constrained stochastic processes governed by sparse observations at intermediate stages and terminal constraints. We assess the effectiveness of our method on synthetic and real-world data and show that the ISB generalises well to high-dimensional data, is computationally efficient, and provides accurate estimates of the marginals at intermediate and terminal times. ", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=me09xlTmm8", "arxiv_id": "2301.13636", "paper_pdf": "papers/me09xlTmm8.pdf", "paper_pdf_sha256": "1ace3aef22405df3a03de89bcbab28e1ab9a95ffbba75c4075fafc332c8b0d10", "paper_pdf_bytes": 3063866, "paper_pdf_source": "openreview", "code_url": "https://github.com/AaltoML/iterative-smoothing-bridge", "code_repository": "AaltoML/iterative-smoothing-bridge", "code_commit": "12263a6f0be10786e4ff05c567fdc5b750f1b297", "code_archive": "repos/me09xlTmm8.zip", "code_archive_sha256": "170804791aea8d023588f021026982e977918238411f1401853615ee7139ccbf", "code_archive_bytes": 1299216, "code_file_count": 31, "code_extensions": {".py": 31}, "github_disk_usage_kb": 1225, "github_languages": {"Python": 150201}, "github_archived": false, "github_pushed_at": "2023-11-24T08:58:52Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/transport-with-support-data-conditional"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "dHJtoaE3yRP", "year": 2022, "status": "rejected", "title": "NAFS: A Simple yet Tough-to-Beat Baseline for Graph Representation Learning", "authors": ["Wentao Zhang", "Zeang Sheng", "Mingyu Yang", "Yang Li", "Yu Shen", "Zhi Yang", "Zichao Yang", "Bin CUI"], "authorids": ["~Wentao_Zhang1", "~Zeang_Sheng1", "~Mingyu_Yang2", "~Yang_Li36", "~Yu_Shen3", "~Zhi_Yang4", "~Zichao_Yang1", "~Bin_CUI2"], "authors_source": "OpenReview API", "abstract": "Recently, graph neural networks (GNNs) have shown prominent performance in graph representation learning by leveraging knowledge from both graph structure and node features. However, most of them have two major limitations. First, GNNs can learn higher-order structural information by stacking more layers but can not deal with large depth due to the over-smoothing issue.  Second, it is not easy to apply these methods on large graphs due to the expensive computation cost and high memory usage. In this paper, we present node-adaptive feature smoothing (NAFS), a simple non-parametric method that constructs node representations without parameter learning. NAFS first extracts the features of each node with its neighbors of different hops by feature smoothing, and then adaptively combines the smoothed features. Besides, the constructed node representation can further be enhanced by the ensemble of smoothed features extracted via different smoothing strategies. We conduct experiments on four benchmark datasets on two different application scenarios: node clustering and link prediction. Remarkably, NAFS with feature ensemble outperforms the state-of-the-art GNNs on these tasks and mitigates the aforementioned two limitations of most learning-based GNN counterparts. ", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=dHJtoaE3yRP", "arxiv_id": "2206.08583", "paper_pdf": "papers/dHJtoaE3yRP.pdf", "paper_pdf_sha256": "5be907ba1067a1c11b4ec9d2370f2356011d2955424006bd5f54c294e0de35ab", "paper_pdf_bytes": 1524273, "paper_pdf_source": "openreview", "code_url": "https://github.com/PKU-DAIR/NAFS", "code_repository": "PKU-DAIR/NAFS", "code_commit": "966cc58a1df85f2179f69f9785c00cf572f0c121", "code_archive": "repos/dHJtoaE3yRP.zip", "code_archive_sha256": "d34bf65108da728217f64bd47eaa34ae2786848d9b8fa033db0e36a639bb067e", "code_archive_bytes": 5710826, "code_file_count": 11, "code_extensions": {".py": 10, ".sh": 1}, "github_disk_usage_kb": 5879, "github_languages": {"Python": 34978, "Shell": 1867}, "github_archived": false, "github_pushed_at": "2022-06-20T02:49:28Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/nafs-a-simple-yet-tough-to-beat-baseline-for-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Jq8JGA89sDa", "year": 2021, "status": "rejected", "title": "Detecting Hallucinated Content in Conditional Neural Sequence Generation", "authors": ["Chunting Zhou", "Jiatao Gu", "Mona T. Diab", "Paco Guzmán", "Luke Zettlemoyer", "Marjan Ghazvininejad"], "authorids": ["~Chunting_Zhou1", "~Jiatao_Gu1", "~Mona_T._Diab1", "fguzman@fb.com", "~Luke_Zettlemoyer1", "~Marjan_Ghazvininejad1"], "authors_source": "OpenReview API", "abstract": "Neural sequence models can generate highly fluent sentences but recent studies have also shown that they are also prone to hallucinate additional content not supported by the input, which can cause a lack of trust in the model.\nTo better assess the faithfulness of the machine outputs, we propose a new task to predict whether each token in the output sequence is hallucinated conditioned on the source input, and collect new manually annotated evaluation sets for this task.\nWe also introduce a novel method for learning to model hallucination detection, based on pretrained language models fine tuned on synthetic data that includes automatically inserted hallucinations.  \nExperiments on machine translation and abstract text summarization demonstrate the effectiveness of our proposed approach -- we obtain an average F1 of around 0.6 across all the benchmark datasets.\nFurthermore, we demonstrate how to use the token-level hallucination labels to define a fine-grained loss over the target sequence in the low-resource machine translation and achieve significant improvements over strong baseline methods.\nWe will also release our annotated data and code for future research.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=Jq8JGA89sDa", "arxiv_id": "2011.02593", "paper_pdf": "papers/Jq8JGA89sDa.pdf", "paper_pdf_sha256": "5800aeaa63139b80c697661e6c98c643365dc9e079863fcf0f4f287b01b1a587", "paper_pdf_bytes": 582867, "paper_pdf_source": "openreview", "code_url": "https://github.com/violet-zct/fairseq-detect-hallucination", "code_repository": "violet-zct/fairseq-detect-hallucination", "code_commit": "be14c9f378f8b9c0447c3b3d080263966d157b8c", "code_archive": "repos/Jq8JGA89sDa.zip", "code_archive_sha256": "6b5b079ca2db61b113e6dd514c10e6ccfd7850ab728161159e6d32a4b66bf6f1", "code_archive_bytes": 6020127, "code_file_count": 467, "code_extensions": {".py": 418, ".sh": 33, ".cpp": 7, ".cu": 4, ".cuh": 2, ".lua": 2, ".h": 1}, "github_disk_usage_kb": 4052, "github_languages": {"Python": 2018748, "Cuda": 36414, "Shell": 32370, "C++": 15854, "Cython": 8858, "Lua": 4210}, "github_archived": false, "github_pushed_at": "2022-04-15T00:52:31Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/detecting-hallucinated-content-in-conditional-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "RfrdbJVvVf", "year": 2025, "status": "rejected", "title": "MatMamba: A Matryoshka State Space Model", "authors": ["Abhinav Shukla", "Sai Vemprala", "Aditya Kusupati", "Ashish Kapoor"], "authorids": ["~Abhinav_Shukla1", "~Sai_Vemprala1", "~Aditya_Kusupati1", "~Ashish_Kapoor1"], "authors_source": "OpenReview API", "abstract": "State Space Models (SSMs) like Mamba2 are a promising alternative to Transformers, with faster theoretical training and inference times -- especially for long context lengths. Recent work on Matryoshka Representation Learning -- and its application to Transformer backbones in works like MatFormer --  showed how to introduce nested granularities of smaller submodels in one universal elastic model. In this work, we present MatMamba: a state space model which combines Matryoshka-style learning with Mamba2, by modifying the block to contain nested dimensions to enable joint training and adaptive inference. MatMamba allows for efficient and adaptive deployment across various model sizes. We train a single large MatMamba model and are able to get a number of smaller nested models for free -- while maintaining or improving upon the performance of a baseline smaller model trained from scratch. We train language and image models at a variety of parameter sizes from 35M to 1.4B. Our results on ImageNet and FineWeb show that MatMamba models scale comparably to Transformers, while having more efficient inference characteristics. This makes MatMamba a practically viable option for deploying large-scale models in an elastic way based on the available inference compute.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=RfrdbJVvVf", "arxiv_id": "2410.06718", "paper_pdf": "papers/RfrdbJVvVf.pdf", "paper_pdf_sha256": "97bf99cb9e07b4ffc98779861f48251b0230e99179dd8dab3948c505c32c5043", "paper_pdf_bytes": 653017, "paper_pdf_source": "openreview", "code_url": "https://github.com/GenRobo/MatMamba", "code_repository": "GenRobo/MatMamba", "code_commit": "3233d4c7e7652ec13f203eb7b03f339c181660f6", "code_archive": "repos/RfrdbJVvVf.zip", "code_archive_sha256": "f711545a72879fdb8851900a919098262fa920a9e88beceb643330d81858c844", "code_archive_bytes": 381431, "code_file_count": 21, "code_extensions": {".sh": 11, ".py": 10}, "github_disk_usage_kb": 408, "github_languages": {"Python": 159571, "Shell": 7875}, "github_archived": false, "github_pushed_at": "2024-11-21T16:45:50Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/matmamba-a-matryoshka-state-space-model"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "fH9eqpCcR3", "year": 2024, "status": "rejected", "title": "Multiple Physics Pretraining for Physical Surrogate Models", "authors": ["Michael McCabe", "Bruno Régaldo-Saint Blancard", "Liam Holden Parker", "Ruben Ohana", "Miles Cranmer", "Alberto Bietti", "Michael Eickenberg", "Siavash Golkar", "Geraud Krawezik", "Francois Lanusse", "Mariel Pettee", "Tiberiu Tesileanu", "Kyunghyun Cho", "Shirley Ho"], "authorids": ["~Michael_McCabe2", "~Bruno_Régaldo-Saint_Blancard1", "~Liam_Holden_Parker1", "~Ruben_Ohana1", "~Miles_Cranmer2", "~Alberto_Bietti1", "~Michael_Eickenberg5", "~Siavash_Golkar1", "gkrawezik@flatironinstitute.org", "~Francois_Lanusse2", "~Mariel_Pettee1", "~Tiberiu_Tesileanu1", "~Kyunghyun_Cho1", "~Shirley_Ho2"], "authors_source": "OpenReview API", "abstract": "We introduce multiple physics pretraining (MPP), an autoregressive task-agnostic pretraining approach for physical surrogate modeling. MPP involves training large surrogate models to predict the dynamics of multiple heterogeneous physical systems simultaneously by learning features that are broadly useful across diverse physical tasks. In order to learn effectively in this setting, we introduce a shared embedding and normalization strategy that projects the fields of multiple systems into a single shared embedding space. We validate the efficacy of our approach on both pretraining and downstream tasks over a broad fluid mechanics-oriented benchmark. We show that a single MPP-pretrained transformer is able to match or outperform task-specific baselines on all pretraining sub-tasks without the need for finetuning. For downstream tasks, we demonstrate that finetuning MPP-trained models results in more accurate predictions across multiple time-steps on new physics compared to training from scratch or finetuning pretrained video foundation models. We open-source our code and model weights trained at multiple scales for reproducibility and community experimentation. Video examples are included in the supplementary materials.", "decision": "Reject", "meta_review": "", "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=fH9eqpCcR3", "arxiv_id": "2310.02994", "paper_pdf": "papers/fH9eqpCcR3.pdf", "paper_pdf_sha256": "01f4ed804cd79222738933baacfeabf298df378d1de69f75c675cbcdc2dbae83", "paper_pdf_bytes": 3414080, "paper_pdf_source": "openreview", "code_url": "https://github.com/PolymathicAI/multiple_physics_pretraining", "code_repository": "PolymathicAI/multiple_physics_pretraining", "code_commit": "e751fc25ae5274c3ddc869bf73afd6fcd4163cb5", "code_archive": "repos/fH9eqpCcR3.zip", "code_archive_sha256": "8e05c07948bb5c1375c139da4fe39e90043a16db99ea08345f870d315f87d7e6", "code_archive_bytes": 458749, "code_file_count": 13, "code_extensions": {".py": 11, ".sh": 2}, "github_disk_usage_kb": 495, "github_languages": {"Python": 84340, "Shell": 1229}, "github_archived": false, "github_pushed_at": "2024-12-06T15:43:29Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/multiple-physics-pretraining-for-physical"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "3leZITnUE9r", "year": 2023, "status": "rejected", "title": "An Empirical Study of Metrics to Measure Representational Harms in Pre-Trained Language Models", "authors": ["Saghar Hosseini", "Ahmed Hassan Awadallah", "Hamid Palangi"], "authorids": ["~Saghar_Hosseini1", "~Ahmed_Hassan_Awadallah1", "~Hamid_Palangi1"], "authors_source": "OpenReview API", "abstract": "Large-scale Pre-Trained Language Models (PTLMs) capture knowledge from massive human-written data which contains latent societal biases and toxic contents. In this paper, we leverage the primary task of PTLMs, i.e. language modeling, and propose a new metric to quantify manifested implicit representational harms in PTLMs towards 13 marginalized demographics. Using this metric, we conducted an empirical analysis of 24 widely used PTLMs. Our analysis provides insights into the correlation between the proposed metric in this work and other related fairness metrics. We observe that our metric correlates with the majority of gender-specific fairness metrics in the literature. Through extensive experiments, we explore the connections between PTLMs architectures and representational harms across two dimensions: depth and width of the networks. We found that prioritizing depth over width, mitigates representational harms in some PTLMs.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=3leZITnUE9r", "arxiv_id": "2301.09211", "paper_pdf": "papers/3leZITnUE9r.pdf", "paper_pdf_sha256": "111fd283ddaa801cffd2f85b56deae23dff3c7e54620549ca4b4184b1c9f4663", "paper_pdf_bytes": 453389, "paper_pdf_source": "openreview", "code_url": "https://github.com/microsoft/SafeNLP", "code_repository": "microsoft/SafeNLP", "code_commit": "e351b6cb892284cf843b88e363b1a6dbaded584e", "code_archive": "repos/3leZITnUE9r.zip", "code_archive_sha256": "caa7a49a41212bc290f462894c1d69d61a9743bbae58bea2f9ac055a6a0bd397", "code_archive_bytes": 1698691, "code_file_count": 3, "code_extensions": {".py": 2, ".sh": 1}, "github_disk_usage_kb": 1254, "github_languages": {"Python": 7551, "Shell": 517}, "github_archived": true, "github_pushed_at": "2023-10-18T20:57:47Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/an-empirical-study-of-metrics-to-measure"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "FD8xldQIgdq", "year": 2022, "status": "rejected", "title": "Robust Models Are More Interpretable Because Attributions Look Normal", "authors": ["Zifan Wang", "Matt Fredrikson", "Anupam Datta"], "authorids": ["~Zifan_Wang1", "~Matt_Fredrikson1", "~Anupam_Datta1"], "authors_source": "OpenReview API", "abstract": "Recent work has found that adversarially-robust deep networks used for image classification are more interpretable: their feature attributions tend to be sharper, and are more concentrated on the objects associated with the image's ground-truth class. We show that smooth decision boundaries play an important role in this enhanced interpretability, as the model's input gradients around data points will more closely align with boundaries' normal vectors when they are smooth. Thus, because robust models have smoother boundaries, the results of gradient-based attribution methods will capture more accurate information about nearby decision boundaries. This understanding of robust interpretability leads to our second contribution: \\emph{boundary attributions}, which aggregate information about the normal vectors of local decision boundaries to explain a classification outcome. We show that by leveraging the key factors underpinning robust interpretability, boundary attributions produce sharper, more concentrated visual explanations---even on non-robust models.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=FD8xldQIgdq", "arxiv_id": "2103.11257", "paper_pdf": "papers/FD8xldQIgdq.pdf", "paper_pdf_sha256": "b45d21710e26232a8d2442332c71d84584d7baae403893002c0134d92079f124", "paper_pdf_bytes": 25874427, "paper_pdf_source": "openreview", "code_url": "https://github.com/zifanw/boundary", "code_repository": "zifanw/boundary", "code_commit": "a901b7e75ee8cdf683c490ba076a9308301db2b9", "code_archive": "repos/FD8xldQIgdq.zip", "code_archive_sha256": "0ebf7a00af4ba389f96dca33844c234c254a89f1a46a19ee0ce0f0393a98de6c", "code_archive_bytes": 6457814, "code_file_count": 8, "code_extensions": {".py": 5, ".ipynb": 2, ".sh": 1}, "github_disk_usage_kb": 6321, "github_languages": {"Jupyter Notebook": 1957190, "Python": 47304, "Shell": 61}, "github_archived": false, "github_pushed_at": "2021-08-13T20:28:44Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/boundary-attributions-provide-normal-vector"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "SnhmiKUPWL", "year": 2021, "status": "rejected", "title": "Leveraging Class Hierarchies with Metric-Guided Prototype Learning", "authors": ["Vivien Sainte Fare Garnot", "Loic Landrieu"], "authorids": ["~Vivien_Sainte_Fare_Garnot1", "~Loic_Landrieu1"], "authors_source": "OpenReview API", "abstract": "In many classification tasks, the set of classes can be organized according to a meaningful hierarchy. This structure can be used to assess the severity of confusing each pair of classes, and summarized  under the form of a cost matrix which also defines a finite metric. We propose to integrate this metric in the supervision of a prototypical network in order to model the hierarchical class structure. Our method relies on jointly learning a feature-extracting network and a set of class representations, or prototypes, which incorporate the error metric into their relative arrangement in the embedding space. We show that this simultaneous training allows for consistent improvement of the severity of the network's errors with regard to the class hierarchy when compared to traditional methods and other prototype-based strategies. Furthermore, when the induced metric contains insight on the data structure, our approach improves the overall precision as well. Experiments on four different public datasets—from agricultural time series classification to depth image semantic segmentation—validate our approach.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=SnhmiKUPWL", "arxiv_id": "2007.03047", "paper_pdf": "papers/SnhmiKUPWL.pdf", "paper_pdf_sha256": "2ef02da2fb1d48418933787b103bd90527527938a3d2b97b645fb320f40463df", "paper_pdf_bytes": 1110853, "paper_pdf_source": "openreview", "code_url": "https://github.com/VSainteuf/metric-guided-prototypes-pytorch", "code_repository": "VSainteuf/metric-guided-prototypes-pytorch", "code_commit": "fae90235b57a71f3ab682917eaf5517799fbffed", "code_archive": "repos/SnhmiKUPWL.zip", "code_archive_sha256": "4fa50ce76fd27a5e61f291f897bfb9724e900ea39be93371509fa912754abf63", "code_archive_bytes": 1257982, "code_file_count": 10, "code_extensions": {".py": 9, ".ipynb": 1}, "github_disk_usage_kb": 4169, "github_languages": {"Jupyter Notebook": 320324, "Python": 27733}, "github_archived": false, "github_pushed_at": "2021-10-20T11:30:19Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/metric-guided-prototype-learning"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "74QmBTV0Zf", "year": 2025, "status": "rejected", "title": "Late Chunking: Contextual Chunk Embeddings Using Long-Context Embedding Models", "authors": ["Michael Günther", "Isabelle Mohr", "Daniel James Williams", "Bo Wang", "Han Xiao"], "authorids": ["~Michael_Günther1", "~Isabelle_Mohr1", "~Daniel_James_Williams1", "~Bo_Wang31", "~Han_Xiao8"], "authors_source": "OpenReview API", "abstract": "Many use cases require retrieving smaller portions of text, and dense vector-based retrieval systems often perform better with shorter text segments, as the semantics are less likely to be \"over-compressed\" in the embeddings. Consequently, practitioners often split text documents into smaller chunks and encode them separately. However, chunk embeddings created in this way can lose contextual information from surrounding chunks, resulting in sub-optimal representations. In this paper, we introduce a novel method called \"late chunking, which leverages long context embedding models to first embed all tokens of the long text, with chunking applied after the transformer model and just before mean pooling - hence the term \"late\" in its naming. The resulting chunk embeddings capture the full contextual information, leading to superior results across various retrieval tasks.  The method is generic enough to be applied to a wide range of long-context embedding models and works without additional training. To further increase the effectiveness of late chunking, we propose a dedicated fine-tuning approach for embedding models.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=74QmBTV0Zf", "arxiv_id": "2409.04701", "paper_pdf": "papers/74QmBTV0Zf.pdf", "paper_pdf_sha256": "7f40499529248f5d920fb3d021e728b68c17c8c5cb8219426f908bb01d742c1b", "paper_pdf_bytes": 435456, "paper_pdf_source": "openreview", "code_url": "https://github.com/jina-ai/late-chunking", "code_repository": "jina-ai/late-chunking", "code_commit": "1d3bb02bf091becd0771455e4e7959463935e26c", "code_archive": "repos/74QmBTV0Zf.zip", "code_archive_sha256": "0003f6c5611ca3e991f1d3d284e179f5c87f66db7110df1868248973ff5159b0", "code_archive_bytes": 381516, "code_file_count": 12, "code_extensions": {".py": 11, ".ipynb": 1}, "github_disk_usage_kb": 421, "github_languages": {"Python": 80840, "Jupyter Notebook": 7185}, "github_archived": false, "github_pushed_at": "2024-12-23T13:38:26Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/late-chunking-contextual-chunk-embeddings"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "gusHSc09zj", "year": 2024, "status": "rejected", "title": "Discovering Mixtures of Structural Causal Models from Time Series Data", "authors": ["Sumanth Varambally", "Yian Ma", "Rose Yu"], "authorids": ["~Sumanth_Varambally1", "~Yian_Ma1", "~Rose_Yu1"], "authors_source": "OpenReview API", "abstract": "In fields such as finance, climate science, and neuroscience, inferring causal relationships from time series data poses a formidable challenge. While contemporary techniques can handle non-linear relationships between variables and flexible noise distributions, they rely on the simplifying assumption that data originates from the same underlying causal model. In this work, we relax this assumption and perform causal discovery from time series data originating from mixtures of different causal models. We infer both the underlying structural causal models and the posterior probability for each sample belonging to a specific mixture component. Our approach employs an end-to-end training process that maximizes an evidence-lower bound for data likelihood. Through extensive experimentation on both synthetic and real-world datasets, we demonstrate that our method surpasses state-of-the-art benchmarks in causal discovery tasks, particularly when the data emanates from diverse underlying causal graphs. Theoretically, we prove the identifiability of such a model under some mild assumptions.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=gusHSc09zj", "arxiv_id": "2310.06312", "paper_pdf": "papers/gusHSc09zj.pdf", "paper_pdf_sha256": "a1653ffb9495cfc06f9376bc1a39359d2396e43200e55f806d03b5ebdcdef87b", "paper_pdf_bytes": 918134, "paper_pdf_source": "openreview", "code_url": "https://github.com/Rose-STL-Lab/MCD", "code_repository": "Rose-STL-Lab/MCD", "code_commit": "a9de419a6293b34d43b7519621a100dda51fc5a3", "code_archive": "repos/gusHSc09zj.zip", "code_archive_sha256": "1c3d37399f2867c53e5d954df863f49646d4a8786c50928580161958835f1737", "code_archive_bytes": 511566, "code_file_count": 49, "code_extensions": {".py": 43, ".sh": 6}, "github_disk_usage_kb": 500, "github_languages": {"Python": 218323, "Shell": 1152}, "github_archived": false, "github_pushed_at": "2025-06-26T01:59:50Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/discovering-mixtures-of-structural-causal"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "_AkC4QYxF5", "year": 2023, "status": "rejected", "title": "Closing the Gap Between SVRG and TD-SVRG with Gradient Splitting", "authors": ["Arsenii Mustafin", "Ioannis Paschalidis", "Alex Olshevsky"], "authorids": ["~Arsenii_Mustafin1", "~Ioannis_Paschalidis1", "~Alex_Olshevsky1"], "authors_source": "OpenReview API", "abstract": "Temporal difference (TD) learning is a simple algorithm for policy evaluation\nin reinforcement learning. The performance of TD learning is affected by high\nvariance and it can be naturally enhanced with variance reduction techniques, such\nas the Stochastic Variance Reduced Gradient (SVRG) method. Recently, multiple\nworks have sought to fuse TD learning with SVRG to obtain a policy evaluation\nmethod with a linear rate of convergence. However, the resulting convergence rate\nis significantly weaker than what is achieved by SVRG in the setting of convex\noptimization. In this work we utilize a recent interpretation of TD-learning as the\nsplitting of the gradient of an appropriately chosen function, thus simplifying the\nalgorithm and fusing TD with SVRG. We prove a linear convergence bound that\nis identical to the convergence bound available for SVRG in the convex setting.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=_AkC4QYxF5", "arxiv_id": "2211.16237", "paper_pdf": "papers/_AkC4QYxF5.pdf", "paper_pdf_sha256": "f73f4731b2cfd3962e1983dec269f68dfae7e7a21dd16cae2479c1ac8ec1355f", "paper_pdf_bytes": 552054, "paper_pdf_source": "openreview", "code_url": "https://github.com/gaarsmu/SVRG_for_TD_learning", "code_repository": "gaarsmu/SVRG_for_TD_learning", "code_commit": "0f8ddee6a59e541254bb1a80463825b10ce11762", "code_archive": "repos/_AkC4QYxF5.zip", "code_archive_sha256": "b145a42e75033fd203b05618fd6ad99f7373785f1dcafc72894402a85bb9e957", "code_archive_bytes": 555452, "code_file_count": 17, "code_extensions": {".py": 17}, "github_disk_usage_kb": 1279, "github_languages": {"Python": 55640}, "github_archived": false, "github_pushed_at": "2023-03-19T15:20:21Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/closing-the-gap-between-svrg-and-td-svrg-with"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "jxTRL-VOoQo", "year": 2022, "status": "rejected", "title": "Evaluating Deep Graph Neural Networks", "authors": ["Wentao Zhang", "Zeang Sheng", "Jiang Yuezihan", "Yikuan Xia", "Jun Gao", "Zhi Yang", "Bin CUI"], "authorids": ["~Wentao_Zhang1", "~Zeang_Sheng1", "~Jiang_Yuezihan1", "~Yikuan_Xia1", "~Jun_Gao6", "~Zhi_Yang4", "~Bin_CUI2"], "authors_source": "OpenReview API", "abstract": "Graph Neural Networks (GNNs) have already been widely applied in various graph mining tasks. However, most GNNs only have shallow architectures, which limits performance improvement. In this paper, we conduct a systematic experimental evaluation on the fundamental limitations of current architecture designs. Based on the experimental results, we answer the following two essential questions: (1) what actually leads to the compromised performance of deep GNNs; (2) how to build deep GNNs. The answers to the above questions provide empirical insights and guidelines for researchers to design deep GNNs. Further, we present Deep Graph Multi-Layer Perceptron (DGMLP), a powerful approach implementing our proposed guidelines. Experimental results demonstrate three advantages of DGMLP: 1) high accuracy -- it achieves state-of-the-art node classification performance on various datasets; 2) high flexibility -- it can flexibly choose different propagation and transformation depths according to certain graph properties; 3) high scalability and efficiency -- it supports fast training on large-scale graphs.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=jxTRL-VOoQo", "arxiv_id": "2108.00955", "paper_pdf": "papers/jxTRL-VOoQo.pdf", "paper_pdf_sha256": "5ec0879f24472d76bd525ec41441a0c1ed47d95c40cc32bb48ee784c5db14370", "paper_pdf_bytes": 788758, "paper_pdf_source": "openreview", "code_url": "https://github.com/zwt233/AIR", "code_repository": "zwt233/AIR", "code_commit": "d5cc2325b071ad062c7a7855f6a42336d21a7c3d", "code_archive": "repos/jxTRL-VOoQo.zip", "code_archive_sha256": "42fcf9644b138ac0ad9f6121d2aaa59e85dc982771548cf4b212c6a1e718f3f8", "code_archive_bytes": 5414890, "code_file_count": 14, "code_extensions": {".py": 14}, "github_disk_usage_kb": 6351, "github_languages": {"Python": 58516}, "github_archived": false, "github_pushed_at": "2022-09-26T03:30:59Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/evaluating-deep-graph-neural-networks"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "iEcqwosBEgx", "year": 2021, "status": "rejected", "title": "Novel Policy Seeking with Constrained Optimization", "authors": ["Hao Sun", "Zhenghao Peng", "Bo Dai", "Jian Guo", "Dahua Lin", "Bolei Zhou"], "authorids": ["~Hao_Sun3", "~Zhenghao_Peng1", "~Bo_Dai2", "guoj@pcl.ac.cn", "~Dahua_Lin1", "~Bolei_Zhou5"], "authors_source": "OpenReview API", "abstract": "We address the problem of seeking novel policies in reinforcement learning tasks. Instead of following the multi-objective framework commonly used in existing methods, we propose to rethink the problem under a novel perspective of constrained optimization. We at first introduce a new metric to evaluate the difference between policies, and then design two practical novel policy seeking methods following the new perspective, namely the Constrained Task Novel Bisector (CTNB), and the Interior Policy Differentiation (IPD), corresponding to the feasible direction method and the interior point method commonly known in the constrained optimization literature. Experimental comparisons on the MuJuCo control suite show our methods can achieve substantial improvements over previous novelty-seeking methods in terms of both the novelty of policies and their performances in the primal task.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=iEcqwosBEgx", "arxiv_id": "2005.10696", "paper_pdf": "papers/iEcqwosBEgx.pdf", "paper_pdf_sha256": "6ea4d66dde148372a8c31d0edd502732266d6e6adc5568bc4a7476c321853d8b", "paper_pdf_bytes": 5812241, "paper_pdf_source": "openreview", "code_url": "https://github.com/holarissun/NPSCO", "code_repository": "holarissun/NPSCO", "code_commit": "28106089d65e392746ece741a735510c96eb65f9", "code_archive": "repos/iEcqwosBEgx.zip", "code_archive_sha256": "f61b43633dd0461c9accf1268ddbc5a0f8024a5d4062c5d4be330fc3423ba2c9", "code_archive_bytes": 4463029, "code_file_count": 5, "code_extensions": {".py": 5}, "github_disk_usage_kb": 4357, "github_languages": {"Python": 119589}, "github_archived": false, "github_pushed_at": "2020-08-11T03:09:48Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/novel-policy-seeking-with-constrained"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "I1VCj1l1Zn", "year": 2025, "status": "rejected", "title": "DLP-LoRA: Efficient Task-Specific LoRA Fusion with a Dynamic, Lightweight Plugin for Large Language Models", "authors": ["Yuxuan Zhang", "Ruizhe Li"], "authorids": ["~Yuxuan_Zhang15", "~Ruizhe_Li2"], "authors_source": "OpenReview API", "abstract": "Recent advancements in Large Language Models (LLMs) have achieved robust performance across diverse tasks, but fine-tuning these models for specific domains remains resource-intensive. Parameter-Efficient Fine-Tuning (PEFT) methods like Low-Rank Adaptation (LoRA) address this challenge by fine-tuning a small subset of parameters. However, existing methods for fusing multiple LoRAs lack dynamic fusion based on contextual inputs and often increase inference time due to token-level operations. We propose DLP-LoRA, a Dynamic Lightweight Plugin that employs a mini-MLP module with only 5M parameters to dynamically fuse multiple LoRAs at the sentence level using top-$p$ sampling strategies. This approach reduces inference time to less than twice that of single LoRA inference by leveraging parallel computation. Evaluations across 26 tasks—including multiple-choice questions and question answering—demonstrate that DLP-LoRA achieves an average accuracy of 92.34\\% on multiple-choice datasets and significant improvements in BLEU and ROUGE scores on QA datasets, outperforming different LLMs backbones under composite task settings. DLP-LoRA effectively balances performance and efficiency, making it a practical solution for dynamic multi-task adaptation in LLMs.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=I1VCj1l1Zn", "arxiv_id": "2410.01497", "paper_pdf": "papers/I1VCj1l1Zn.pdf", "paper_pdf_sha256": "8323c066f332b87ab0dc5e6680565091367367241c598687e772f2fe98e167a6", "paper_pdf_bytes": 2722444, "paper_pdf_source": "openreview", "code_url": "https://github.com/MeCuping/DLP-LoRA", "code_repository": "MeCuping/DLP-LoRA", "code_commit": "d68fd38f36d9762f3e60041472193d379257cfb9", "code_archive": "repos/I1VCj1l1Zn.zip", "code_archive_sha256": "d7af434a61dce55ca98f67239b40ba01a2c39a8f2e4f1b72d8ac136ac51f5a84", "code_archive_bytes": 462891, "code_file_count": 18, "code_extensions": {".py": 18}, "github_disk_usage_kb": 426, "github_languages": {"Python": 62697, "Batchfile": 28}, "github_archived": false, "github_pushed_at": "2024-10-03T01:48:29Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/dlp-lora-efficient-task-specific-lora-fusion"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "96nX9xIIx2", "year": 2024, "status": "rejected", "title": "Visual Prompting Upgrades Neural Network Sparsification: A Data-Model Perspective", "authors": ["Can Jin", "Tianjin Huang", "Yihua Zhang", "Mykola Pechenizkiy", "Sijia Liu", "Shiwei Liu", "Tianlong Chen"], "authorids": ["~Can_Jin1", "~Tianjin_Huang1", "~Yihua_Zhang1", "~Mykola_Pechenizkiy1", "~Sijia_Liu1", "~Shiwei_Liu2", "~Tianlong_Chen1"], "authors_source": "OpenReview API", "abstract": "The rapid development of large-scale deep learning models questions the affordability of hardware platforms, which necessitates the pruning to reduce their computational and memory footprints. Sparse neural networks as the product, have demonstrated numerous favorable benefits like low complexity, undamaged generalization, $\\textit{etc}$. Most of the prominent pruning strategies are invented from a $\\textit{model-centric}$ perspective, focusing on searching and preserving crucial weights by analyzing network topologies. However, the role of data and its interplay with model-centric pruning has remained relatively unexplored. In this research, we introduce a novel $\\textit{data-model co-design}$ perspective: to promote superior weight sparsity by learning important model topology and adequate input data in a synergetic manner. Specifically, customized $\\textbf{V}$isual $\\textbf{P}$rompts are mounted to upgrade neural $\\textbf{N}$etwork $\\textbf{s}$parsification in our proposed $\\textbf{\\texttt{VPNs}}$ framework. As a pioneering effort, this paper conducts systematic investigations about the impact of different visual prompts on model pruning and suggests an effective joint optimization approach. Extensive experiments with $3$ network architectures and $8$ datasets evidence the substantial performance improvements from $\\textbf{\\texttt{VPNs}}$ over existing start-of-the-art pruning algorithms. Furthermore, we find that subnetworks discovered by $\\textbf{\\texttt{VPNs}}$ from pre-trained models enjoy better transferability across diverse downstream scenarios. These insights shed light on new promising possibilities of data-model co-designs for vision model sparsification. Codes are in the supplement.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=96nX9xIIx2", "arxiv_id": "2312.01397", "paper_pdf": "papers/96nX9xIIx2.pdf", "paper_pdf_sha256": "56d73c5360d9ba99f924049be362260485caf6bcf5ffdd2e9b8c63a2353e7ceb", "paper_pdf_bytes": 4569192, "paper_pdf_source": "openreview", "code_url": "https://github.com/UNITES-Lab/VPNs", "code_repository": "UNITES-Lab/VPNs", "code_commit": "912072c4ec2156a7d0414f00c463f3bd48279ab5", "code_archive": "repos/96nX9xIIx2.zip", "code_archive_sha256": "0a3ffaa4fa693c4cf0b2fb82907aeb748c395973f623e242209ea07fe400df99", "code_archive_bytes": 501483, "code_file_count": 9, "code_extensions": {".py": 6, ".sh": 3}, "github_disk_usage_kb": 500, "github_languages": {"Python": 61870, "Shell": 4708}, "github_archived": false, "github_pushed_at": "2023-12-05T02:36:39Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/visual-prompting-upgrades-neural-network"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "0DwzMsUNIr", "year": 2023, "status": "rejected", "title": "From Points to Functions: Infinite-dimensional Representations in Diffusion Models", "authors": ["Sarthak Mittal", "Guillaume Lajoie", "Stefan Bauer", "Arash Mehrjou"], "authorids": ["~Sarthak_Mittal1", "~Guillaume_Lajoie1", "~Stefan_Bauer1", "~Arash_Mehrjou1"], "authors_source": "OpenReview API", "abstract": "Diffusion-based generative models learn to iteratively transfer unstructured noise to a complex target distribution as opposed to Generative Adversarial Networks (GANs) or the decoder of Variational Autoencoders (VAEs) which produce samples from the target distribution in a single step. Thus, in diffusion models every sample is naturally connected to a random trajectory which is a solution to a learned stochastic differential equation (SDE). Generative models are only concerned with the final state of this trajectory that delivers samples from the desired distribution. \\cite{abstreiter2021diffusion} showed that these stochastic trajectories can be seen as continuous filters that wash out information along the way. Consequently, it is reasonable to ask if there is an intermediate time step at which the preserved information is optimal for a given downstream task. In this work, we show that a combination of information content from different time steps gives a strictly better representation for the downstream task. We introduce an attention and recurrence based modules that ``learn to mix'' information content of various time-steps such that the resultant representation leads to superior performance in downstream tasks.\n", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=0DwzMsUNIr", "arxiv_id": "2210.13774", "paper_pdf": "papers/0DwzMsUNIr.pdf", "paper_pdf_sha256": "89b6b10c9c3512e38d1e040c81c307cecc37f331200fe531c134311659914ef5", "paper_pdf_bytes": 3225151, "paper_pdf_source": "openreview", "code_url": "https://github.com/sarthmit/traj_drl", "code_repository": "sarthmit/traj_drl", "code_commit": "61aef79dfa4d88dcb6f91c5e80ebc6e91e8df481", "code_archive": "repos/0DwzMsUNIr.zip", "code_archive_sha256": "dd80fa7b82867a4ae01b6e8788e8d6b5f841241ccaaac054ef41c03d98fbeaa1", "code_archive_bytes": 1422515, "code_file_count": 113, "code_extensions": {".py": 92, ".sh": 17, ".cpp": 2, ".cu": 2}, "github_disk_usage_kb": 1302, "github_languages": {"Python": 494902, "Cuda": 14488, "Shell": 7432, "C++": 1792}, "github_archived": false, "github_pushed_at": "2022-10-25T14:48:48Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/from-points-to-functions-infinite-dimensional"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "hq7vLjZTJPk", "year": 2022, "status": "rejected", "title": "A Communication-Efficient Distributed Gradient Clipping Algorithm for Training Deep Neural Networks", "authors": ["Chunyang Liao", "Zhenxun Zhuang", "Mingrui Liu"], "authorids": ["~Chunyang_Liao1", "~Zhenxun_Zhuang1", "~Mingrui_Liu2"], "authors_source": "OpenReview API", "abstract": "In distributed training of deep neural networks or Federated Learning (FL), people usually run Stochastic Gradient Descent (SGD) or its variants on each machine and communicate with other machines periodically. However, SGD might converge slowly in training some deep neural networks (e.g., RNN, LSTM) because of the exploding gradient issue. Gradient clipping is usually employed to address this issue in the single machine setting, but exploring this technique in the FL setting is still in its infancy: it remains mysterious whether the gradient clipping scheme can take advantage of multiple machines to enjoy parallel speedup in the FL setting. The main technical difficulty lies at dealing with nonconvex loss function, non-Lipschitz continuous gradient, and skipping communication rounds simultaneously. In this paper, we explore a relaxed-smoothness assumption of the loss landscape which LSTM was shown to satisfy in previous works, and design a communication-efficient gradient clipping algorithm. This algorithm can be run on multiple machines, where each machine employs a gradient clipping scheme and communicate with other machines after multiple steps of gradient-based updates. Our algorithm is proved to have $O\\left(\\frac{1}{N\\epsilon^4}\\right)$ iteration complexity for finding an $\\epsilon$-stationary point, where $N$ is the number of machines. This indicates that our algorithm enjoys linear speedup. Our experiments on several benchmark datasets demonstrate that our algorithm indeed exhibits fast convergence speed in practice and validate our theory.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=hq7vLjZTJPk", "arxiv_id": "2205.05040", "paper_pdf": "papers/hq7vLjZTJPk.pdf", "paper_pdf_sha256": "5b9ba9370cbffa344aada27734a1f8d5f515c13c1ecf440b8c17bcf2119fd204", "paper_pdf_bytes": 3780791, "paper_pdf_source": "openreview", "code_url": "https://github.com/MingruiLiu-ML-Lab/Communication-Efficient-Local-Gradient-Clipping", "code_repository": "MingruiLiu-ML-Lab/Communication-Efficient-Local-Gradient-Clipping", "code_commit": "5f572e368492acc30732dbf47121761de70aacf1", "code_archive": "repos/hq7vLjZTJPk.zip", "code_archive_sha256": "b0b0f08554e5f4dfdaa8bd01b91538cbe0feafdb7c17467ec1d0471be2382a24", "code_archive_bytes": 6518207, "code_file_count": 26, "code_extensions": {".py": 20, ".sh": 6}, "github_disk_usage_kb": 6368, "github_languages": {"Python": 94320, "Shell": 10441}, "github_archived": false, "github_pushed_at": "2022-09-21T16:12:17Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/a-communication-efficient-distributed-3"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "PGmqOzKEPZN", "year": 2021, "status": "rejected", "title": "Non-Negative Bregman Divergence Minimization for Deep Direct Density Ratio Estimation", "authors": ["Masahiro Kato", "Takeshi Teshima"], "authorids": ["~Masahiro_Kato1", "~Takeshi_Teshima1"], "authors_source": "OpenReview API", "abstract": "The estimation of the ratio of two probability densities has garnered attention as the density ratio is useful in various machine learning tasks, such as anomaly detection and domain adaptation. To estimate the density ratio, methods collectively known as direct density ratio estimation (DRE) have been explored. These methods are based on the minimization of the Bregman (BR) divergence between a density ratio model and the true density ratio. However, existing direct DRE suffers from serious overfitting when using flexible models such as neural networks. In this paper, we introduce a non-negative correction for empirical risk using only the prior knowledge of the upper bound of the density ratio. This correction makes a DRE method more robust against overfitting and enables the use of flexible models. In the theoretical analysis, we discuss the consistency of the empirical risk. In our experiments, the proposed estimators show favorable performance in inlier-based outlier detection and covariate shift adaptation.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=PGmqOzKEPZN", "arxiv_id": "2006.06979", "paper_pdf": "papers/PGmqOzKEPZN.pdf", "paper_pdf_sha256": "45cab1b2e9f942bbf8b96da30a2eca81b84ca95333cd3fa4f07961595057cdc9", "paper_pdf_bytes": 4267784, "paper_pdf_source": "openreview", "code_url": "https://github.com/MasaKat0/D3RE", "code_repository": "MasaKat0/D3RE", "code_commit": "fa008291ec2b30654081f8c80c0b682fe5df1f76", "code_archive": "repos/PGmqOzKEPZN.zip", "code_archive_sha256": "c0eb5c7feb70334b25bae3d11ddf9219317c9274cba90b5fde7e5e4346d1c823", "code_archive_bytes": 7703983, "code_file_count": 90, "code_extensions": {".py": 79, ".sh": 11}, "github_disk_usage_kb": 4639, "github_languages": {"Python": 382981, "Shell": 13989}, "github_archived": false, "github_pushed_at": "2022-11-29T19:56:39Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/non-negative-bregman-divergence-minimization"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "996aKQIom0", "year": 2025, "status": "rejected", "title": "PingPong: A Benchmark for Role-Playing Language Models with User Emulation and Multi-Model Evaluation", "authors": ["Ilya Gusev"], "authorids": ["~Ilya_Gusev1"], "authors_source": "OpenReview API", "abstract": "We introduce a benchmark for evaluating the role-playing capabilities of language models. Our approach leverages language models themselves to emulate users in dynamic, multi-turn conversations and to assess the resulting dialogues. The framework consists of three main components: a player model assuming a specific character role, an interrogator model simulating user behavior, and several judge models evaluating conversation quality. We conducted experiments comparing automated evaluations with human annotations to validate our approach, demonstrating strong correlations across multiple criteria. This work provides a foundation for a robust and dynamic evaluation of model capabilities in interactive scenarios.", "decision": "Reject", "meta_review": "", "num_reviews": 6, "openreview_url": "https://openreview.net/forum?id=996aKQIom0", "arxiv_id": "2409.06820", "paper_pdf": "papers/996aKQIom0.pdf", "paper_pdf_sha256": "2ae8962d19fe1d7fc1fa37655a36874da21a01792ac9f62f5c7d95a688166ae5", "paper_pdf_bytes": 1389322, "paper_pdf_source": "openreview", "code_url": "https://github.com/IlyaGusev/ping_pong_bench", "code_repository": "IlyaGusev/ping_pong_bench", "code_commit": "4e3b0e0b6b093670f67b80d7d62b5a5e8e7b13fd", "code_archive": "repos/996aKQIom0.zip", "code_archive_sha256": "820ead7e8ec5e0157c0ef08d3c83cbf1af0ecef7b8099a373d69aefc1577b0ee", "code_archive_bytes": 290873, "code_file_count": 22, "code_extensions": {".py": 22}, "github_disk_usage_kb": 429, "github_languages": {"Python": 100806, "Jinja": 13368, "HTML": 9651}, "github_archived": false, "github_pushed_at": "2025-05-25T20:17:29Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/pingpong-a-benchmark-for-role-playing"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "E296x0YpML", "year": 2024, "status": "rejected", "title": "Fooling the Textual Fooler via Randomizing Latent Representations", "authors": ["Duy Cao Hoang", "Nguyen Hung-Quang", "Saurav Manchanda", "Minlong Peng", "Kok-Seng Wong", "Khoa D Doan"], "authorids": ["~Duy_Cao_Hoang1", "~Nguyen_Hung-Quang1", "~Saurav_Manchanda1", "~Minlong_Peng1", "~Kok-Seng_Wong1", "~Khoa_D_Doan1"], "authors_source": "OpenReview API", "abstract": "Despite outstanding performance in a variety of NLP tasks, recent studies have revealed that NLP models are vulnerable to adversarial attacks that slightly perturb the input to cause the models to misbehave. Among these attacks, adversarial word-level perturbations are well-studied and effective attack strategies. These attacks involve querying the victim model many times to determine the most important words in an input text and to replace these words with their corresponding synonyms. Query-based attacks as such work in black-box settings, which can be detrimental to NLP applications that can be accessed publicly. In this work, we propose a lightweight and attack-agnostic defense whose main goal is to perplex the process of generating an adversarial example in these query-based black-box attacks; that is to fool the textual fooler. This defense, named AdvFooler, works by randomizing the latent representation of the input at inference time. Different from existing defenses, AdvFooler does not necessitate additional computational overhead during training nor relies on assumptions about the potential adversarial perturbation set while having a negligible impact on the model's accuracy. Our theoretical and empirical analyses highlight the significance of robustness resulting from confusing the adversary via randomizing the latent space, as well as the impact of randomization on clean accuracy. Finally, we empirically demonstrate the near state-of-the-art robustness of AdvFooler against representative adversarial word-level attacks on two benchmark datasets.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=E296x0YpML", "arxiv_id": "2310.01452", "paper_pdf": "papers/E296x0YpML.pdf", "paper_pdf_sha256": "81d777153b5ddf419736c45453d6df8d4c7e59ad199d5fe460677826182ac410", "paper_pdf_bytes": 338325, "paper_pdf_source": "openreview", "code_url": "https://github.com/mail-research/AdvFooler-text-defender", "code_repository": "mail-research/AdvFooler-text-defender", "code_commit": "29b69d7f38f73ab65eea78467c993d0a4d0e3035", "code_archive": "repos/E296x0YpML.zip", "code_archive_sha256": "af74518c78f6dd7e0aa9a6ef6724c328e01d8f44db0656063061444858b0f8ce", "code_archive_bytes": 613774, "code_file_count": 282, "code_extensions": {".py": 282}, "github_disk_usage_kb": 508, "github_languages": {"Python": 1849165}, "github_archived": false, "github_pushed_at": "2024-07-22T06:03:45Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/fooling-the-textual-fooler-via-randomizing"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "QP02DQ-FG-8", "year": 2023, "status": "rejected", "title": "Incomplete to complete multiphysics forecasting - a hybrid approach for learning unknown phenomena", "authors": ["Nilam Nandkishor Tathawadekar", "Nguyen Anh Khoa Doan", "Camilo Fernando Silva", "Nils Thuerey"], "authorids": ["~Nilam_Nandkishor_Tathawadekar2", "~Nguyen_Anh_Khoa_Doan1", "~Camilo_Fernando_Silva1", "~Nils_Thuerey1"], "authors_source": "OpenReview API", "abstract": "Modeling complex dynamical systems where only partial knowledge of their physical mechanisms is available is a crucial problem across all scientific and engineering disciplines. Purely data-driven approaches, which only make use of an artificial neural network and data, often fail to accurately simulate the evolution of the system dynamics over a sufficiently long time and in a physically consistent manner. Therefore, we propose a hybrid approach that uses a neural network model in combination with an incomplete PDE solver that provides known but incomplete physical information. In this study, we demonstrate that the results obtained from the incomplete PDEs can be efficiently corrected at every time step by the proposed hybrid neural network – PDE solver model, so that the effect of the unknown physics present in the system is correctly accounted for. For validation purposes, the obtained simulations of the hybrid model are successfully compared against results coming from the complete set of PDEs describing the full physics of the considered system. We demonstrate the validity of the proposed approach on a reactive flow, an archetypal multi-physics system that combines fluid mechanics and chemistry, the latter being the physics considered unknown. Experiments are made on planar and Bunsen-type flames at various operating conditions. The hybrid neural network - PDE approach correctly models the flame evolution of the cases under study for significantly long time windows, yields improved generalization, and allows for larger simulation time steps. ", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=QP02DQ-FG-8", "arxiv_id": "2111.11185", "paper_pdf": "papers/QP02DQ-FG-8.pdf", "paper_pdf_sha256": "ca651ffdcb1b721f03052392620772eb02244738ad2c859419391ace2c1e6524", "paper_pdf_bytes": 2072512, "paper_pdf_source": "openreview", "code_url": "https://github.com/tum-pbs/Hybrid-Solver-for-Reactive-Flows", "code_repository": "tum-pbs/Hybrid-Solver-for-Reactive-Flows", "code_commit": "75d836ff82c30da698b8abd12fe0e2c5b6d79da2", "code_archive": "repos/QP02DQ-FG-8.zip", "code_archive_sha256": "911e669bcbe2cd9192a255a8da0ff270896a0e4398746ca051b007d73ce96d2b", "code_archive_bytes": 1410465, "code_file_count": 31, "code_extensions": {".py": 31}, "github_disk_usage_kb": 1319, "github_languages": {"Python": 443412, "Makefile": 10048}, "github_archived": false, "github_pushed_at": "2024-01-10T13:44:28Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/hybrid-neural-network-pde-solvers-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "e_D6AmszH4P", "year": 2022, "status": "rejected", "title": "ViViT: Curvature access through the generalized Gauss-Newton's low-rank structure", "authors": ["Felix Dangel", "Lukas Tatzel", "Philipp Hennig"], "authorids": ["~Felix_Dangel1", "~Lukas_Tatzel1", "~Philipp_Hennig1"], "authors_source": "OpenReview API", "abstract": "Curvature in form of the Hessian or its generalized Gauss-Newton (GGN) approximation is valuable for algorithms that rely on a local model for the loss to train, compress, or explain deep networks. Existing methods based on implicit multiplication via automatic differentiation or Kronecker-factored block diagonal approximations do not consider noise in the mini-batch. We present ViViT, a curvature model that leverages the GGN's low-rank structure without further approximations. It allows for efficient computation of eigenvalues, eigenvectors, as well as per-sample first- and second-order directional derivatives. The representation is computed in parallel with gradients in one backward pass and offers a fine-grained cost-accuracy trade-off, which allows it to scale. As examples for ViViT's usefulness, we investigate the directional first- and second-order derivatives during training, and how noise information can be used to improve the stability of second-order methods.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=e_D6AmszH4P", "arxiv_id": "2106.02624", "paper_pdf": "papers/e_D6AmszH4P.pdf", "paper_pdf_sha256": "a01c3b769fce368d47389399bcb8219d51cbb9033b0da8c7b04bf16f1c69107c", "paper_pdf_bytes": 1165873, "paper_pdf_source": "openreview", "code_url": "https://github.com/PwLo3K46/vivit", "code_repository": "PwLo3K46/vivit", "code_commit": "937642975be2ade122632d4eaef273461992d7ab", "code_archive": "repos/e_D6AmszH4P.zip", "code_archive_sha256": "044f27de9661d83d3c97bd25375261f3bc26558c802be2b34baf2f6d9c2dee7c", "code_archive_bytes": 6680925, "code_file_count": 106, "code_extensions": {".py": 101, ".sh": 5}, "github_disk_usage_kb": 6492, "github_languages": {"Python": 393933, "Makefile": 2117, "Shell": 567}, "github_archived": false, "github_pushed_at": "2021-10-03T20:35:05Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/vivit-curvature-access-through-the"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "oWy06SBgt4", "year": 2025, "status": "rejected", "title": "1-Bit FQT: Pushing the Limit of Fully Quantized Training to 1-bit", "authors": ["Chang Gao", "JingRen Hou", "Kang Zhao", "Jiaqi Wang", "Jianfei Chen", "Liping Jing"], "authorids": ["~Chang_Gao4", "~JingRen_Hou1", "~Kang_Zhao5", "~Jiaqi_Wang8", "~Jianfei_Chen1", "~Liping_Jing3"], "authors_source": "OpenReview API", "abstract": "Fully quantized training (FQT) accelerates the training of deep neural networks by quantizing the activations, weights, and gradients into lower precision. To explore the ultimate limit of FQT (the lowest achievable precision), we make a first attempt to 1-bit FQT. We provide a theoretical analysis of FQT based on Adam and SGD, revealing that the gradient variance influences the convergence of FQT. Building on these theoretical results, we introduce an Average 1-bit Quantization (AQ) strategy. The strategy leverages the heterogeneity of gradients to mitigate gradient variance by pruning less informative gradients and enhancing the numerical precision of remaining gradients. Additionally, we propose Sample Channel joint Quantization (SCQ), which utilizes different quantization strategies in the computation of weight gradients and activation gradients to ensure that the method is friendly to low-bitwidth hardware. Finally, we present a framework to deploy our algorithm. For fine-tuning VGGNet-16 and ResNet-18 on multiple datasets, our algorithm achieves an average accuracy improvement of approximately 6\\%, compared to per-sample quantization. Moreover, our training speedup can reach a maximum of 5.13× compared to full precision training.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=oWy06SBgt4", "arxiv_id": "2408.14267", "paper_pdf": "papers/oWy06SBgt4.pdf", "paper_pdf_sha256": "8870b4a096c6d1b07e84f5c6f0c286ad92b11ada3914a34f25f7174abf611cc0", "paper_pdf_bytes": 1276291, "paper_pdf_source": "openreview", "code_url": "https://github.com/Gaochang-bjtu/1-bit-FQT", "code_repository": "Gaochang-bjtu/1-bit-FQT", "code_commit": "008b6188e4f1da2e6870f6c896afd89113ab2be4", "code_archive": "repos/oWy06SBgt4.zip", "code_archive_sha256": "fab8dfa4ea68f187dd85a6eafda5599a7769cf65604d8d8b7b78779a0f8296ef", "code_archive_bytes": 490626, "code_file_count": 63, "code_extensions": {".py": 51, ".h": 6, ".cu": 2, ".c": 2, ".cpp": 1, ".sh": 1}, "github_disk_usage_kb": 433, "github_languages": {"Python": 237989, "C": 73394, "Cuda": 11419, "Makefile": 956, "C++": 944, "Shell": 174}, "github_archived": false, "github_pushed_at": "2024-08-26T12:22:33Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/1-bit-fqt-pushing-the-limit-of-fully"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "qud5pDnpzo", "year": 2024, "status": "rejected", "title": "ViP: A Differentially Private Foundation Model for Computer Vision", "authors": ["Yaodong Yu", "Maziar Sanjabi", "Yi Ma", "Kamalika Chaudhuri", "Chuan Guo"], "authorids": ["~Yaodong_Yu4", "~Maziar_Sanjabi1", "~Yi_Ma4", "~Kamalika_Chaudhuri1", "~Chuan_Guo1"], "authors_source": "OpenReview API", "abstract": "Artificial intelligence (AI) has seen a tremendous surge in capabilities thanks to the use of foundation models trained on internet-scale data. On the flip side, the uncurated nature of internet-scale data also poses significant privacy and legal risks, as they often contain personal information or copyrighted material that should not be trained on without permission. In this work, we propose as a mitigation measure a recipe to train foundation vision models with differential privacy (DP) guarantee. We identify masked autoencoders as a suitable learning algorithm that aligns well with DP-SGD, and train ViP---a Vision transformer with differential Privacy---under a strict privacy budget of $\\epsilon=8$ on the LAION400M dataset. We evaluate the quality of representation learned by ViP using standard downstream vision tasks; in particular, ViP achieves a (non-private) linear probing accuracy of 55.7% on ImageNet, comparable to that of end-to-end trained AlexNet (trained and evaluated on ImageNet). Our result suggests that scaling to internet-scale data can be practical for private learning.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=qud5pDnpzo", "arxiv_id": "2306.08842", "paper_pdf": "papers/qud5pDnpzo.pdf", "paper_pdf_sha256": "c80967407f7be8f70c7184b94dfe867c5017182668b760d8ffe295bef486e6ee", "paper_pdf_bytes": 1619764, "paper_pdf_source": "openreview", "code_url": "https://github.com/facebookresearch/ViP-MAE", "code_repository": "facebookresearch/ViP-MAE", "code_commit": "d87fd2de456a07f2c21d9017850448d1caffd3ce", "code_archive": "repos/qud5pDnpzo.zip", "code_archive_sha256": "2a0fc2d815abab75df09a87afb520c51859a97120797bb2e3f4acdbe1f6be77b", "code_archive_bytes": 538430, "code_file_count": 19, "code_extensions": {".py": 19}, "github_disk_usage_kb": 509, "github_languages": {"Python": 121841}, "github_archived": true, "github_pushed_at": "2023-06-27T23:36:29Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/vip-a-differentially-private-foundation-model"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "3jBXX9Xb1iz", "year": 2023, "status": "rejected", "title": "Multi-Label Knowledge Distillation", "authors": ["Peng-Hui Yang", "Ming-Kun Xie", "Chen-Chen Zong", "Lei Feng", "Gang Niu", "Masashi Sugiyama", "Sheng-Jun Huang"], "authorids": ["~Peng-Hui_Yang1", "~Ming-Kun_Xie1", "~Chen-Chen_Zong1", "~Lei_Feng1", "~Gang_Niu1", "~Masashi_Sugiyama1", "~Sheng-Jun_Huang1"], "authors_source": "OpenReview API", "abstract": "Existing knowledge distillation methods typically work by enforcing the consistency of output logits or intermediate feature maps between the teacher network and student network. Unfortunately, these methods can hardly be extended to the multi-label learning scenario. Because each instance is associated with multiple semantic labels, neither the prediction logits nor the feature maps obtained from the whole example can accurately transfer knowledge for each label. In this paper, we propose a novel multi-label knowledge distillation method. On one hand, it exploits the informative semantic knowledge from the logits by label decoupling with the one-versus-all reduction strategy; on the other hand, it enhances the distinctiveness of the learned feature representations by leveraging the structural information of label-wise embeddings. Experimental results on multiple benchmark datasets validate that the proposed method can avoid knowledge counteraction among labels, and achieve superior performance against diverse comparing methods.", "decision": "Reject", "meta_review": null, "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=3jBXX9Xb1iz", "arxiv_id": "2308.06453", "paper_pdf": "papers/3jBXX9Xb1iz.pdf", "paper_pdf_sha256": "d4f01bc2e752d95bd47b969bb88d5f5e5384b19c5600a454b7b999267efe7e28", "paper_pdf_bytes": 1642188, "paper_pdf_source": "openreview", "code_url": "https://github.com/penghui-yang/L2D", "code_repository": "penghui-yang/L2D", "code_commit": "ea7fa5581de3cd0dba2f4d5ba3e3aaed78837a58", "code_archive": "repos/3jBXX9Xb1iz.zip", "code_archive_sha256": "9169733a62bb098bd068e2bdc00b03d52247427651d2bb544dfd25631748b2e5", "code_archive_bytes": 1271502, "code_file_count": 38, "code_extensions": {".py": 38}, "github_disk_usage_kb": 1370, "github_languages": {"Python": 110295}, "github_archived": false, "github_pushed_at": "2024-04-24T05:30:44Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/multi-label-knowledge-distillation"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "qLqeb9AjD2o", "year": 2022, "status": "rejected", "title": "Confidence-aware Training of Smoothed Classifiers for Certified Robustness", "authors": ["Jongheon Jeong", "Seojin Kim", "Jinwoo Shin"], "authorids": ["~Jongheon_Jeong1", "~Seojin_Kim2", "~Jinwoo_Shin1"], "authors_source": "OpenReview API", "abstract": "Any classifier can be \"smoothed out\" under Gaussian noise to build a new classifier that is provably robust to $\\ell_2$-adversarial perturbations, viz., by averaging its predictions over the noise, namely via randomized smoothing. Under the smoothed classifiers, the fundamental trade-off between accuracy and (adversarial) robustness has been well evidenced in the literature: i.e., increasing the robustness of a classifier for an input can be at the expense of decreased accuracy for some other inputs. In this paper, we propose a simple training method leveraging this trade-off for obtaining more robust smoothed classifiers, in particular, through a sample-wise control of robustness over the training samples. We enable this control feasible by investigating the correspondence between robustness and prediction confidence of smoothed classifiers: specifically, we propose to use the \"accuracy under Gaussian noise\" as an easy-to-compute proxy of adversarial robustness for each input. We differentiate the training objective depending on this proxy to filter out samples that are unlikely to benefit from the worst-case (adversarial) objective. Our experiments following the standard benchmarks consistently show that the proposed method, despite its simplicity, exhibits improved certified robustness upon existing state-of-the-art training methods.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=qLqeb9AjD2o", "arxiv_id": "2212.09000", "paper_pdf": "papers/qLqeb9AjD2o.pdf", "paper_pdf_sha256": "c6511ef06638f442450886893c65398c8a10c07008cf210a63dfb44cb538c2eb", "paper_pdf_bytes": 837194, "paper_pdf_source": "openreview", "code_url": "https://github.com/alinlab/smoothing-catrs", "code_repository": "alinlab/smoothing-catrs", "code_commit": "d4bc576e7d373d158f087ba5744af8bb48466bb7", "code_archive": "repos/qLqeb9AjD2o.zip", "code_archive_sha256": "a7812fb83b825258884964045e66eed124829cfaf6b99a2f1cc6c1d62d5631bc", "code_archive_bytes": 7562839, "code_file_count": 18, "code_extensions": {".py": 18}, "github_disk_usage_kb": 6953, "github_languages": {"Python": 108214}, "github_archived": false, "github_pushed_at": "2023-01-19T06:05:09Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/confidence-aware-training-of-smoothed-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "33TBJachvOX", "year": 2021, "status": "rejected", "title": "How to compare adversarial robustness of classifiers from a global perspective", "authors": ["Niklas Risse", "Jan Philip Göpfert", "Christina Göpfert"], "authorids": ["~Niklas_Risse1", "jgoepfert@techfak.uni-bielefeld.de", "~Christina_Göpfert1"], "authors_source": "OpenReview API", "abstract": "Adversarial robustness of machine learning models has attracted considerable attention over recent years. Adversarial attacks undermine the reliability of and trust in machine learning models, but the construction of more robust models hinges on a rigorous understanding of adversarial robustness as a property of a given model. Point-wise measures for specific threat models are currently the most popular tool for comparing the robustness of classifiers and are used in most recent publications on adversarial robustness. In this work, we use robustness curves to show that point-wise measures fail to capture important global properties that are essential to reliably compare the robustness of different classifiers. We introduce new ways in which robustness curves can be used to systematically uncover these properties and provide concrete recommendations for researchers and practitioners when assessing and comparing the robustness of trained models. Furthermore, we characterize scale as a way to distinguish small and large perturbations, and relate it to inherent properties of data sets, demonstrating that robustness thresholds must be chosen accordingly. We hope that our work contributes to a shift of focus away from point-wise measures of robustness and towards a discussion of the question what kind of robustness could and should reasonably be expected. We release code to reproduce all experiments presented in this paper, which includes a Python module to calculate robustness curves for arbitrary data sets and classifiers, supporting a number of frameworks, including TensorFlow, PyTorch and JAX.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=33TBJachvOX", "arxiv_id": "2004.10882", "paper_pdf": "papers/33TBJachvOX.pdf", "paper_pdf_sha256": "1e8ddcae1a9f2cc9a0d020723415270d632d2874d0ab59100222dc68977ad9b1", "paper_pdf_bytes": 631544, "paper_pdf_source": "openreview", "code_url": "https://github.com/niklasrisse/how-to-compare-adversarial-robustness-of-classifiers-from-a-global-perspective", "code_repository": "niklasrisse/how-to-compare-adversarial-robustness-of-classifiers-from-a-global-perspective", "code_commit": "2e03e661fbe9639913750e6044967d3ddbae4a1d", "code_archive": "repos/33TBJachvOX.zip", "code_archive_sha256": "405d6c804f570ceaf1e0134bf7be236003a5990892b60326e3b44c1212ee4803", "code_archive_bytes": 5190222, "code_file_count": 10, "code_extensions": {".ipynb": 8, ".py": 2}, "github_disk_usage_kb": 5052, "github_languages": {"Jupyter Notebook": 78815, "Python": 17587}, "github_archived": false, "github_pushed_at": "2020-10-02T15:35:44Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/adversarial-examples-and-where-to-find-them"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "YMCtQlm8Bc", "year": 2025, "status": "rejected", "title": "Critical Influence of Overparameterization on Sharpness-aware Minimization", "authors": ["Sungbin Shin", "Dongyeop Lee", "Maksym Andriushchenko", "Namhoon Lee"], "authorids": ["~Sungbin_Shin1", "~Dongyeop_Lee1", "~Maksym_Andriushchenko1", "~Namhoon_Lee1"], "authors_source": "OpenReview API", "abstract": "Training overparameterized neural networks often yields solutions with varying generalization capabilities, even when achieving similar training losses. Recent evidence indicates a strong correlation between the sharpness of a minimum and its generalization error, leading to increased interest in optimization methods that explicitly seek flatter minima for improved generalization. Despite its contemporary relevance to overparameterization, however, this sharpness-aware minimization (SAM) strategy has not been studied much yet as to exactly how it is affected by overparameterization. In this work, we analyze SAM under varying degrees of overparameterization, presenting both empirical and theoretical findings that reveal its critical influence on SAM's effectiveness. First, we conduct extensive numerical experiments across diverse domains and show that SAM consistently improves with overparameterization. Next, we attribute this phenomenon to the interplay between the enlarged solution space and increased implicit bias resulting from overparameterization. Furthermore, we show that this effect is particularly pronounced in practical settings involving label noise and sparsity, and yet, sufficient regularization is necessary. Last but not least, we provide other theoretical insights into how overparameterization helps SAM achieve minima with more uniform Hessian moments compared to SGD, and much faster convergence at a linear rate.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=YMCtQlm8Bc", "arxiv_id": "2311.17539", "paper_pdf": "papers/YMCtQlm8Bc.pdf", "paper_pdf_sha256": "e482abfbc585f1a91c990f9065e43a7f385a070f68a417f6326b84ab0f239203", "paper_pdf_bytes": 1913746, "paper_pdf_source": "openreview", "code_url": "https://github.com/LOG-postech/SAM-overparam", "code_repository": "LOG-postech/SAM-overparam", "code_commit": "a96543a160555481d5db3513daf67370e9062b5e", "code_archive": "repos/YMCtQlm8Bc.zip", "code_archive_sha256": "af46474301cffa44ed514fbb09de6fe0eeb25bfa03baedb7ad7c2557014c3a49", "code_archive_bytes": 292216, "code_file_count": 9, "code_extensions": {".py": 9}, "github_disk_usage_kb": 435, "github_languages": {"Python": 47149}, "github_archived": false, "github_pushed_at": "2025-05-14T07:51:49Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/the-effects-of-overparameterization-on"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "9Kgnvknvwd", "year": 2024, "status": "rejected", "title": "A First-Order Multi-Gradient Algorithm for Multi-Objective Bi-Level Optimization", "authors": ["Feiyang Ye", "Baijiong Lin", "Xiaofeng Cao", "Yu Zhang", "Ivor Tsang"], "authorids": ["~Feiyang_Ye4", "~Baijiong_Lin1", "~Xiaofeng_Cao2", "~Yu_Zhang3", "~Ivor_Tsang1"], "authors_source": "OpenReview API", "abstract": "In this paper, we study the Multi-Objective Bi-Level Optimization (MOBLO) problem, where the upper-level subproblem is a multi-objective optimization problem and the lower-level subproblem is for scalar optimization. Existing gradient-based MOBLO algorithms need to compute the Hessian matrix, causing the computational inefficient problem. To address this, we propose an efficient first-order multi-gradient method for MOBLO, called FORUM. Specifically, we reformulate MOBLO problems as a constrained multi-objective optimization (MOO) problem via the value-function approach. Then we propose a novel multi-gradient aggregation method to solve the challenging constrained MOO problem. Theoretically, we provide the complexity analysis to show the efficiency of the proposed method and a non-asymptotic convergence result. Empirically, extensive experiments demonstrate the effectiveness and efficiency of the proposed FORUM method in different learning problems. In particular, it achieves state-of-the-art performance on three multi-task learning benchmark datasets.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=9Kgnvknvwd", "arxiv_id": "2401.09257", "paper_pdf": "papers/9Kgnvknvwd.pdf", "paper_pdf_sha256": "293c575bdb29f63088782569c988ae4dfda05b2a30a2cf3a7f3863516e93ad44", "paper_pdf_bytes": 506917, "paper_pdf_source": "openreview", "code_url": "https://github.com/Baijiong-Lin/FORUM", "code_repository": "Baijiong-Lin/FORUM", "code_commit": "318607c503ec4ec5ce4f0d297401fc43717b026f", "code_archive": "repos/9Kgnvknvwd.zip", "code_archive_sha256": "f0243dac00e0edb628c5374d155a54ab6b163b3c2b7996967ee4d863f8f6f408", "code_archive_bytes": 522142, "code_file_count": 46, "code_extensions": {".py": 46}, "github_disk_usage_kb": 511, "github_languages": {"Python": 145838}, "github_archived": false, "github_pushed_at": "2024-07-04T08:05:20Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/a-first-order-multi-gradient-algorithm-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "mPzpPv0geS2", "year": 2023, "status": "rejected", "title": "Adan: Adaptive Nesterov Momentum Algorithm for Faster Optimizing Deep Models", "authors": ["Xingyu Xie", "Pan Zhou", "Huan Li", "Zhouchen Lin", "Shuicheng YAN"], "authorids": ["~Xingyu_Xie1", "~Pan_Zhou3", "~Huan_Li1", "~Zhouchen_Lin1", "~Shuicheng_YAN3"], "authors_source": "OpenReview API", "abstract": "Adaptive gradient algorithms combine the moving average idea with heavy ball acceleration to estimate accurate first- and second-order moments of the gradient for accelerating convergence.  However, Nesterov acceleration which converges faster than heavy ball acceleration in theory and also in many empirical cases, is much less investigated under the adaptive gradient setting.  In this work, we propose the ADAptive Nesterov momentum algorithm, Adan for short, to speed up the training of deep neural networks effectively.  Adan first reformulates the vanilla Nesterov acceleration to develop a new Nesterov momentum estimation (NME) method, which avoids the extra computation and memory overhead of computing gradient at the extrapolation point.  Then Adan adopts NME to estimate the first- and second-order moments of the gradient in adaptive gradient algorithms for convergence acceleration. Besides, we prove that Adan finds an $\\epsilon$-approximate first-order stationary point within $O(\\epsilon^{-3.5})$ stochastic gradient complexity on the non-convex stochastic problems (e.g., deep learning problems), matching the best-known lower bound. Extensive experimental results show that  Adan surpasses the corresponding SoTA optimizers on vision, language, and RL tasks and sets new SoTAs for many popular networks and frameworks, e.g., ResNet,  ConvNext, ViT, Swin, MAE, LSTM, Transformer-XL, and BERT.  More surprisingly, Adan can use half of the training cost (epochs) of SoTA optimizers to achieve higher or comparable performance on ViT, ResNet, MAE, etc., and also shows great tolerance to a large range of minibatch size, e.g., from 1k to 32k.  We hope Adan can contribute to developing deep learning by reducing training costs and relieving the engineering burden of trying different optimizers on various architectures.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=mPzpPv0geS2", "arxiv_id": "2208.06677", "paper_pdf": "papers/mPzpPv0geS2.pdf", "paper_pdf_sha256": "20fdc1ac034d4ac8f55c72e5c35542aee29bd3ddbb2a236785be3423996021c6", "paper_pdf_bytes": 687321, "paper_pdf_source": "openreview", "code_url": "https://github.com/sail-sg/Adan", "code_repository": "sail-sg/Adan", "code_commit": "2c65beaf400775753b155da4082c3819fea689b9", "code_archive": "repos/mPzpPv0geS2.zip", "code_archive_sha256": "566e520cdcf502f66f299901533420521442849f4928e604efeb33d877c2b9cb", "code_archive_bytes": 1374021, "code_file_count": 40, "code_extensions": {".py": 32, ".sh": 2, ".cu": 2, ".cuh": 2, ".h": 1, ".cpp": 1}, "github_disk_usage_kb": 1373, "github_languages": {"Python": 269155, "Cuda": 21780, "C++": 12838, "Shell": 2610}, "github_archived": false, "github_pushed_at": "2025-06-08T14:35:41Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/adan-adaptive-nesterov-momentum-algorithm-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "xWRX16GCugt", "year": 2022, "status": "rejected", "title": "Sequoia: A Software Framework to Unify Continual Learning Research", "authors": ["Fabrice Normandin", "Oleksiy Ostapenko", "Pau Rodriguez", "Florian Golemo", "Ryan Lindeborg", "Matthew Riemer", "Lucas Cecchi", "Timothee LESORT", "Khimya Khetarpal", "David Vazquez", "Laurent Charlin", "Irina Rish", "Massimo Caccia"], "authorids": ["~Fabrice_Normandin1", "~Oleksiy_Ostapenko1", "~Pau_Rodriguez2", "~Florian_Golemo1", "~Ryan_Lindeborg1", "~Matthew_Riemer1", "~Lucas_Cecchi1", "~Timothee_LESORT1", "~Khimya_Khetarpal1", "~David_Vazquez1", "~Laurent_Charlin1", "~Irina_Rish1", "~Massimo_Caccia1"], "authors_source": "OpenReview API", "abstract": "The field of Continual Learning (CL) seeks to develop algorithms that accumulate knowledge and skills over time through interaction with non-stationary environments. In practice, a plethora of evaluation procedures (settings) and algorithmic solutions (methods) exist, each with their own potentially disjoint set of assumptions. This variety makes measuring progress in CL difficult. We propose a taxonomy of settings, where each setting is described as a set of assumptions. A tree-shaped hierarchy emerges from this view, where more general settings become the parents of those with more restrictive assumptions. This makes it possible to use inheritance to share and reuse research, as developing a method for a given setting also makes it directly applicable onto any of its children. We instantiate this idea as a publicly available software framework called Sequoia, which features a wide variety of settings from both the Continual Supervised Learning (CSL) and Continual Reinforcement Learning (CRL) domains. Sequoia also includes a growing suite of methods which are easy to extend and customize, in addition to more specialized methods from external libraries. We hope that this new paradigm and its first implementation can help unify and accelerate research in CL. You can help us grow the tree by visiting (this GitHub URL).", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=xWRX16GCugt", "arxiv_id": "2108.01005", "paper_pdf": "papers/xWRX16GCugt.pdf", "paper_pdf_sha256": "21a8c76f54b72c288d72216822848f13692c8c98857abd6e4e9f8fe1d4d71b72", "paper_pdf_bytes": 1437338, "paper_pdf_source": "openreview", "code_url": "https://github.com/lebrice/Sequoia", "code_repository": "lebrice/Sequoia", "code_commit": "7e12ff8ed67fada8cf220c5c396dc26332f558c2", "code_archive": "repos/xWRX16GCugt.zip", "code_archive_sha256": "05456f055e07f93b2e86c5a0e2276450461d8d3697232848c5c22682ecd48cf9", "code_archive_bytes": 926909, "code_file_count": 399, "code_extensions": {".py": 387, ".sh": 11, ".ipynb": 1}, "github_disk_usage_kb": 7144, "github_languages": {"Python": 2192958, "Shell": 11640, "Dockerfile": 4441}, "github_archived": false, "github_pushed_at": "2023-05-30T15:33:28Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/sequoia-a-software-framework-to-unify"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "1-Mh-cWROZ", "year": 2021, "status": "rejected", "title": "Fold2Seq: A Joint Sequence(1D)-Fold(3D) Embedding-based Generative Model for Protein Design", "authors": ["Yue Cao", "Payel Das", "Pin-Yu Chen", "Vijil Chenthamarakshan", "Igor Melnyk", "Yang Shen"], "authorids": ["~Yue_Cao4", "~Payel_Das1", "~Pin-Yu_Chen1", "~Vijil_Chenthamarakshan1", "~Igor_Melnyk1", "~Yang_Shen4"], "authors_source": "OpenReview API", "abstract": "Designing novel protein sequences consistent with a desired 3D structure  or fold, often referred to as the inverse protein folding problem, is a central, but non-trivial, task in protein engineering.  It has a wide range of applications  in energy, biomedicine, and materials science. However, challenges exist due to the complex sequence-fold relationship and difficulties associated with  modeling 3D folds. To overcome these challenges, we propose Fold2Seq, a novel transformer-based generative framework for designing protein sequences conditioned on a specific fold. Our model learns a fold  embedding from the density of the secondary structural elements in 3D voxels, and then models  the complex sequence-structure relationship by learning a joint sequence-fold embedding. Experiments on  high-resolution, complete, and single-structure test set demonstrate improved  performance of Fold2Seq in terms of speed and reliability for sequence design, compared to existing baselines including the state-of-the-art RosettaDesign and other neural net-based approaches. The unique advantages of fold-based Fold2Seq becomes more evident on diverse real-world test sets comprised of low-resolution, incomplete, or ensemble structures,  in comparison to a structure-based model. ", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=1-Mh-cWROZ", "arxiv_id": "2106.13058", "paper_pdf": "papers/1-Mh-cWROZ.pdf", "paper_pdf_sha256": "d83ecfc36418ba1394affebdc567665760fded62a0e85f6fc65b1ebcf908b16f", "paper_pdf_bytes": 6948439, "paper_pdf_source": "openreview", "code_url": "https://github.com/IBM/fold2seq", "code_repository": "IBM/fold2seq", "code_commit": "b9a97d81eac329b5259ad10e2a6f4fe80ade542f", "code_archive": "repos/1-Mh-cWROZ.zip", "code_archive_sha256": "915d3bb6b0d7ae80a24fbac44b57eaf8e952f7d5cd410bd141857458fdbc8fec", "code_archive_bytes": 3877907, "code_file_count": 12, "code_extensions": {".py": 12}, "github_disk_usage_kb": 5093, "github_languages": {"Python": 62341}, "github_archived": true, "github_pushed_at": "2022-05-25T06:49:33Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/fold2seq-a-joint-sequence-1d-fold-3d-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "qVyjN01x4P", "year": 2025, "status": "rejected", "title": "Learning Pattern-Specific Experts for Time Series Forecasting Under Patch-level Distribution Shift", "authors": ["Yanru Sun", "Zongxia Xie", "Emadeldeen Eldele", "Dongyue Chen", "Qinghua Hu", "Min Wu"], "authorids": ["~Yanru_Sun1", "~Zongxia_Xie1", "~Emadeldeen_Eldele1", "~Dongyue_Chen3", "~Qinghua_Hu1", "~Min_Wu2"], "authors_source": "OpenReview API", "abstract": "Time series forecasting, which aims to predict future values based on historical data, has garnered significant attention due to its broad range of applications.  However, real-world time series often exhibit complex non-uniform distribution with varying patterns across segments, such as season, operating condition, or semantic meaning, making accurate forecasting challenging. Existing approaches, which typically train a single model to capture all these diverse patterns, often struggle with the pattern drifts between patches and may lead to poor generalization. To address these challenges, we propose TFPS, a novel architecture that leverages pattern-specific experts for more accurate and adaptable time series forecasting. TFPS employs a dual-domain encoder to capture both time-domain and frequency-domain features, enabling a more comprehensive understanding of temporal dynamics. It then uses subspace clustering to dynamically identify distinct patterns across data patches. Finally, pattern-specific experts model these unique patterns, delivering tailored predictions for each patch. By explicitly learning and adapting to evolving patterns, TFPS achieves significantly improved forecasting accuracy. Extensive experiments on real-world datasets demonstrate that TFPS outperforms state-of-the-art methods, particularly in long-term forecasting, through its dynamic and pattern-aware learning approach. The data and codes are available: https://anonymous.4open.science/r/TFPS-D001.", "decision": "Reject", "meta_review": "", "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=qVyjN01x4P", "arxiv_id": "2410.09836", "paper_pdf": "papers/qVyjN01x4P.pdf", "paper_pdf_sha256": "92a2f1308f3cdb35e6581c53679d0b2b31f6940ae6780b14873147c3e5ee8ee4", "paper_pdf_bytes": 2627191, "paper_pdf_source": "openreview", "code_url": "https://github.com/syrGitHub/TFPS", "code_repository": "syrGitHub/TFPS", "code_commit": "83a11827e27e6617e8c8a8771f0a1dd7e10976a5", "code_archive": "repos/qVyjN01x4P.zip", "code_archive_sha256": "50228f3260690aae63dbbcdc8dc12a2ba011f5b924e2e529078267495c42086b", "code_archive_bytes": 428277, "code_file_count": 37, "code_extensions": {".py": 33, ".sh": 4}, "github_disk_usage_kb": 437, "github_languages": {"Python": 207122, "Shell": 21175}, "github_archived": false, "github_pushed_at": "2024-11-08T02:58:30Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/learning-pattern-specific-experts-for-time"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "sOHVDPqoUJ", "year": 2024, "status": "rejected", "title": "Less is More: Selective Layer Finetuning with SubTuning", "authors": ["Gal Kaplun", "Andrey Gurevich", "Tal Swisa", "Mazor David", "Shai Shalev-Shwartz", "eran malach"], "authorids": ["~Gal_Kaplun1", "~Andrey_Gurevich1", "~Tal_Swisa1", "~Mazor_David1", "~Shai_Shalev-Shwartz1", "~eran_malach1"], "authors_source": "OpenReview API", "abstract": "Finetuning a pretrained model has become the standard approach for training neural networks on novel tasks, leading to rapid convergence and enhanced performance. In this work, we present a parameter-efficient finetuning method, wherein we selectively train a carefully chosen subset of layers while keeping the remaining weights frozen at their initial (pre-trained) values. We observe that not all layers are created equal: different layers across the network contribute variably to the overall performance, and the optimal choice of layers is contingent upon the downstream task and the underlying data distribution. We demonstrate that our proposed method, termed *subset finetuning* (or SubTuning), offers several advantages over conventional finetuning. We show that SubTuning outperforms both finetuning and linear probing in scenarios with scarce or corrupted data, achieving state-of-the-art results compared to competing methods for finetuning on small datasets. When data is abundant, SubTuning often attains performance comparable to finetuning while simultaneously enabling efficient inference in a multi-task setting when deployed alongside other models. We showcase the efficacy of SubTuning across various tasks, diverse network architectures and pre-training methods.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=sOHVDPqoUJ", "arxiv_id": "2302.06354", "paper_pdf": "papers/sOHVDPqoUJ.pdf", "paper_pdf_sha256": "2959d112a8557b8fb0f273b5a40267ddc317203704e6b6013c5decc2ae0cd045", "paper_pdf_bytes": 2800232, "paper_pdf_source": "openreview", "code_url": "https://github.com/talswisa/SubTuning", "code_repository": "talswisa/SubTuning", "code_commit": "68b64c24ad0b5053d4ad2b79d0c6c459c327c5a4", "code_archive": "repos/sOHVDPqoUJ.zip", "code_archive_sha256": "822a43b27e88ac40d5b78c43d7cc274e717be48903654b6bfec7556217e104bd", "code_archive_bytes": 522020, "code_file_count": 16, "code_extensions": {".py": 12, ".sh": 4}, "github_disk_usage_kb": 519, "github_languages": {"Python": 55704, "Shell": 2884, "Dockerfile": 630}, "github_archived": false, "github_pushed_at": "2023-06-11T12:01:56Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/subtuning-efficient-finetuning-for-multi-task"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "3vOtC1t1kF", "year": 2023, "status": "rejected", "title": "Efficient Personalized Federated Learning via Sparse Model-Adaptation", "authors": ["Daoyuan Chen", "Liuyi Yao", "Dawei Gao", "Bolin Ding", "Yaliang Li"], "authorids": ["~Daoyuan_Chen1", "~Liuyi_Yao1", "~Dawei_Gao1", "~Bolin_Ding3", "~Yaliang_Li1"], "authors_source": "OpenReview API", "abstract": "Federated Learning (FL) aims to train machine learning models for multiple clients without sharing their own private data. Due to the heterogeneity of clients' local data distribution, recent studies explore the personalized FL that learns and deploys distinct local models with the help of auxiliary global models. However, the clients can be heterogeneous in terms of not only local data distribution, but also their computation and communication resources. The capacity and efficiency of personalized models are restricted by the lowest-resource clients, leading to sub-optimal performance and limited practicality of personalized FL. To overcome these challenges, we propose a novel approach named pFedGate for efficient personalized FL by adaptively and efficiently learning sparse local models. With a lightweight trainable gating layer, pFedGate enables clients to reach their full potential in model capacity by generating different sparse models accounting for both the heterogeneous data distributions and resource constraints. Meanwhile, the computation and communication efficiency are both improved thanks to the adaptability between the model sparsity and clients' resources. Further, we theoretically show that the proposed pFedGate has superior complexity with guaranteed convergence and generalization error. Extensive experiments show that pFedGate achieves superior global accuracy, individual accuracy and efficiency simultaneously over state-of-the-art methods, by up to 4.53\\% accuracy improvement and 12x smaller model size. We also demonstrate that pFedGate performs better than competitors in the novel clients participation and partial clients participation scenarios, and can learn meaningful sparse local models adapted to different data distributions.", "decision": "Reject", "meta_review": null, "num_reviews": 5, "openreview_url": "https://openreview.net/forum?id=3vOtC1t1kF", "arxiv_id": "2305.02776", "paper_pdf": "papers/3vOtC1t1kF.pdf", "paper_pdf_sha256": "2db5442db797c005ec4a900d973d002d4f400b5716063c6ecbea64994fb61bda", "paper_pdf_bytes": 3128174, "paper_pdf_source": "openreview", "code_url": "https://github.com/yxdyc/pFedGate", "code_repository": "yxdyc/pFedGate", "code_commit": "b8cfa156acdb08207d33641a355959503ca1db78", "code_archive": "repos/3vOtC1t1kF.zip", "code_archive_sha256": "a9bd83e2b07861800568ce1cd6e93a00b7bacee23f7d7d32f4653b52376c5e2a", "code_archive_bytes": 1405277, "code_file_count": 40, "code_extensions": {".py": 39, ".sh": 1}, "github_disk_usage_kb": 1374, "github_languages": {"Python": 332909, "Shell": 1523}, "github_archived": false, "github_pushed_at": "2023-05-26T07:25:36Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/efficient-personalized-federated-learning-via"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "qvUJV2-t_c", "year": 2022, "status": "rejected", "title": "Using a one dimensional parabolic model of the full-batch loss to estimate learning rates during training", "authors": ["Maximus Mutschler", "Kevin Alexander Laube", "Andreas Zell"], "authorids": ["~Maximus_Mutschler1", "~Kevin_Alexander_Laube1", "~Andreas_Zell1"], "authors_source": "OpenReview API", "abstract": "A fundamental challenge in Deep Learning is to find optimal step sizes for stochastic gradient descent automatically. In traditional optimization, line searches are a commonly used method to determine step sizes. One problem in Deep Learning is that finding appropriate step sizes on the full-batch loss is unfeasibly expensive. Therefore, classical line search approaches, designed for losses without inherent noise, are usually not applicable. Recent empirical findings suggest that the full-batch loss behaves locally parabolically in the direction of noisy update step directions. Furthermore, the trend of the optimal update step size changes slowly. By exploiting these findings, this work introduces a line-search method that approximates the full-batch loss with a parabola estimated over several mini-batches. Learning rates are derived from such parabolas during training. In the experiments conducted, our approach mostly outperforms SGD tuned with a piece-wise constant learning rate schedule and other line search approaches for Deep Learning across models, datasets, and batch sizes on validation and test accuracy.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=qvUJV2-t_c", "arxiv_id": "2108.13880", "paper_pdf": "papers/qvUJV2-t_c.pdf", "paper_pdf_sha256": "9f597cea7a98e0aa66702ce4e8dbe59eb0dd032694dd240317aa790df57e14c1", "paper_pdf_bytes": 1981608, "paper_pdf_source": "openreview", "code_url": "https://github.com/cogsys-tuebingen/LABPAL", "code_repository": "cogsys-tuebingen/LABPAL", "code_commit": "f577cf976d5b88ea2dc901c2f71d30a643424a98", "code_archive": "repos/qvUJV2-t_c.zip", "code_archive_sha256": "db9bba8e4e81078624f30af104f2309f8b2a59a7707809d0711fadc47dd73394", "code_archive_bytes": 1145573, "code_file_count": 43, "code_extensions": {".py": 43}, "github_disk_usage_kb": 7363, "github_languages": {"Python": 321710}, "github_archived": false, "github_pushed_at": "2022-09-19T14:10:35Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/using-a-one-dimensional-parabolic-model-of"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "XG1Drw7VbLJ", "year": 2021, "status": "rejected", "title": "Defining Benchmarks for Continual Few-Shot Learning", "authors": ["Antreas Antoniou", "Massimiliano Patacchiola", "Mateusz Ochal", "Amos Storkey"], "authorids": ["~Antreas_Antoniou2", "~Massimiliano_Patacchiola1", "~Mateusz_Ochal1", "~Amos_Storkey1"], "authors_source": "OpenReview API", "abstract": "In recent years there has been substantial progress in few-shot learning, where a model is trained on a small labeled dataset related to a specific task, and in continual learning, where a model has to retain knowledge acquired on a sequence of datasets. Both of these fields are different abstractions of the same real world scenario, where a learner has to adapt to limited information from different changing sources and be able to generalize in and from each of them. Combining these two paradigms, where a model is trained on several sequential few-shot tasks, and then tested on a validation set stemming from all those tasks, helps by explicitly defining the competing requirements for both efficient integration and continuity. In this paper we propose such a setting, naming it Continual Few-Shot Learning (CFSL). We first define a theoretical framework for CFSL, then we propose a range of flexible benchmarks to unify the evaluation criteria. As part of the benchmark, we introduce a compact variant of ImageNet, called SlimageNet64, which retains all original 1000 classes but only contains 200 instances of each one (a total of 200K data-points) downscaled to 64 by 64 pixels. We provide baselines for the proposed benchmarks using a number of popular few-shot and continual learning methods, exposing previously unknown strengths and weaknesses of those algorithms. The dataloader and dataset will be released with an open-source license.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=XG1Drw7VbLJ", "arxiv_id": "2004.11967", "paper_pdf": "papers/XG1Drw7VbLJ.pdf", "paper_pdf_sha256": "485efc10dcca6d91e8c84d41e64f977ac061289cbdaa2abd54b2666a74138cbc", "paper_pdf_bytes": 1703999, "paper_pdf_source": "openreview", "code_url": "https://github.com/AntreasAntoniou/FewShotContinualLearning", "code_repository": "AntreasAntoniou/FewShotContinualLearning", "code_commit": "819b9cc26ef9d2360a040c51f17958e1b8dba8fd", "code_archive": "repos/XG1Drw7VbLJ.zip", "code_archive_sha256": "2ea8a735c9dfabd031f26c81d0868989240fe2d83b3d3589450fe957cbfe12c3", "code_archive_bytes": 937410, "code_file_count": 448, "code_extensions": {".sh": 435, ".py": 13}, "github_disk_usage_kb": 5401, "github_languages": {"Python": 336297, "Shell": 160045}, "github_archived": false, "github_pushed_at": "2020-08-18T12:00:15Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/defining-benchmarks-for-continual-few-shot"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "pjfrGVekwK", "year": 2025, "status": "rejected", "title": "Variational Bayes Gaussian Splatting", "authors": ["Toon Van de Maele", "Ozan Catal", "Alexander Tschantz", "Christopher Buckley", "Tim Verbelen"], "authorids": ["~Toon_Van_de_Maele1", "~Ozan_Catal2", "~Alexander_Tschantz2", "~Christopher_Buckley1", "~Tim_Verbelen1"], "authors_source": "OpenReview API", "abstract": "Recently, 3D Gaussian Splatting has emerged as a promising approach for modeling 3D scenes using mixtures of Gaussians. The predominant optimization method for these models relies on backpropagating gradients through a differentiable rendering pipeline, which struggles with catastrophic forgetting when dealing with continuous streams of data. To address this limitation, we propose Variational Bayes Gaussian Splatting (VBGS), a novel approach that frames training a Gaussian splat as variational inference over model parameters. By leveraging the conjugacy properties of multivariate Gaussians, we derive a closed-form variational update rule, allowing efficient updates from partial, sequential observations without the need for replay buffers. Our experiments show that VBGS not only matches state-of-the-art performance on static datasets, but also enables continual learning from sequentially streamed 2D and 3D data, drastically improving performance in this setting.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=pjfrGVekwK", "arxiv_id": "2410.03592", "paper_pdf": "papers/pjfrGVekwK.pdf", "paper_pdf_sha256": "839a8697315880151fea95573a9d113d2cd1419ca17078de736876bc8abe9119", "paper_pdf_bytes": 49753355, "paper_pdf_source": "openreview", "code_url": "https://github.com/VersesTech/vbgs", "code_repository": "VersesTech/vbgs", "code_commit": "2ae3f4bea6ed3a5d69271c0e2a67322c06c09b9e", "code_archive": "repos/pjfrGVekwK.zip", "code_archive_sha256": "c0149126d32b7301091bec2b96e4035c1594e70ab93474a0de82254a4620e2c9", "code_archive_bytes": 156422, "code_file_count": 41, "code_extensions": {".py": 40, ".sh": 1}, "github_disk_usage_kb": 438, "github_languages": {"Python": 219948, "Shell": 534}, "github_archived": false, "github_pushed_at": "2025-04-11T15:56:18Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/variational-bayes-gaussian-splatting"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "BMw4Cm0gGO", "year": 2024, "status": "rejected", "title": "C-MCTS: Safe Planning with Monte Carlo Tree Search", "authors": ["Dinesh Parthasarathy", "Georgios Kontes", "Axel Plinge", "Christopher Mutschler"], "authorids": ["~Dinesh_Parthasarathy1", "~Georgios_Kontes1", "~Axel_Plinge1", "~Christopher_Mutschler1"], "authors_source": "OpenReview API", "abstract": "The Constrained Markov Decision Process (CMDP) allows to solve safety-critical decision making tasks that are subject to constraints. \nWhile CMDPs have been extensively studied in the Reinforcement Learning literature, little attention has been given to sampling-based planning algorithms such as MCTS for solving them. Previous approaches perform conservatively with respect to costs as they avoid constraint violations by using Monte Carlo cost estimates that suffer from high variance. We propose Constrained MCTS (C-MCTS), which estimates cost using a safety critic that is trained with Temporal Difference learning in an offline phase prior to agent deployment. The critic limits exploration by pruning unsafe trajectories within MCTS during deployment. C-MCTS satisfies cost constraints but operates closer to the constraint boundary, achieving higher rewards than previous work. As a nice byproduct, the planner is more efficient w.r.t. planning steps. Most importantly, under model mismatch between the planner and the real world, C-MCTS is less susceptible to cost violations than previous work.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=BMw4Cm0gGO", "arxiv_id": "2305.16209", "paper_pdf": "papers/BMw4Cm0gGO.pdf", "paper_pdf_sha256": "ae98759604c00dbad971d43d8dbdab354781980c377e42e492bc472337f41886", "paper_pdf_bytes": 483528, "paper_pdf_source": "openreview", "code_url": "https://github.com/mutschcr/C-MCTS", "code_repository": "mutschcr/C-MCTS", "code_commit": "f690100502cf0db548edc318b0f055ecbfcac790", "code_archive": "repos/BMw4Cm0gGO.zip", "code_archive_sha256": "a92a43b669da9eaf6fac00eaa2fce613a01f8fdd1aebb66f17008ded68428b8d", "code_archive_bytes": 629146, "code_file_count": 36, "code_extensions": {".h": 14, ".cpp": 12, ".py": 10}, "github_disk_usage_kb": 537, "github_languages": {"C++": 85212, "Python": 69327, "Makefile": 1200}, "github_archived": false, "github_pushed_at": "2023-05-25T04:37:26Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/c-mcts-safe-planning-with-monte-carlo-tree"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "MpGP-z07TmM", "year": 2023, "status": "rejected", "title": "Learning Specialized Activation Functions for Physics-informed Neural Networks", "authors": ["Honghui Wang", "Lu Lu", "Shiji Song", "Gao Huang"], "authorids": ["~Honghui_Wang1", "~Lu_Lu1", "~Shiji_Song1", "~Gao_Huang1"], "authors_source": "OpenReview API", "abstract": "At the heart of network architectures lie the non-linear activation functions, the choice of which affects the model optimization and task performance. In computer vision and natural language processing, the Rectified Linear Unit is widely adopted across different tasks. However, there is no such default choice of activation functions in the context of physics-informed neural networks (PINNs). It is observed that PINNs exhibit high sensitivity to activation functions due to the various characteristics of each physics system, which makes the choice of the suitable activation function for PINNs a critical issue. Existing works usually choose activation functions in an inefficient trial-and-error manner. To address this problem, we propose to search automatically for the optimal activation function when solving different PDEs. This is achieved by learning an adaptive activation function as linear combinations of a set of candidate functions, whose coefficients can be directly optimized by gradient descent. In addition to its efficient optimization, the proposed method enables the discovery of novel activation function and the incorporation with prior knowledge about the PDE system. We can further enhance its search space with adaptive slope. The effectiveness of the proposed adaptive activation function is demonstrated on a series of benchmarks, including the Poisson's equation, Burgers' equation, Allen-Cahn equation, convection equation, Korteweg–de Vries equation and Cahn-Hilliard equation. The performance gain of the proposed method is further interpreted from the neural tangent kernel perspective. Code will be released. ", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=MpGP-z07TmM", "arxiv_id": "2308.04073", "paper_pdf": "papers/MpGP-z07TmM.pdf", "paper_pdf_sha256": "9c98e682fe8f416b08b23c73ccfd585da392fe9d1a163d93364485b87dc92019", "paper_pdf_bytes": 1036724, "paper_pdf_source": "openreview", "code_url": "https://github.com/LeapLabTHU/AdaAFforPINNs", "code_repository": "LeapLabTHU/AdaAFforPINNs", "code_commit": "2e87a3dd134c6bb359c8330a4d8e84f0a97201ed", "code_archive": "repos/MpGP-z07TmM.zip", "code_archive_sha256": "a659359d57d1659e599d17aa8e8555140585b0a83aee3968e2bf0a4028d5b61f", "code_archive_bytes": 1674830, "code_file_count": 8, "code_extensions": {".py": 7, ".sh": 1}, "github_disk_usage_kb": 1642, "github_languages": {"Python": 125038, "Shell": 7708}, "github_archived": false, "github_pushed_at": "2023-08-09T03:15:18Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/learning-specialized-activation-functions-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "OBwsUF4nFye", "year": 2022, "status": "rejected", "title": "Private Multi-Task Learning: Formulation and Applications to Federated Learning", "authors": ["Shengyuan Hu", "Steven Wu", "Virginia Smith"], "authorids": ["~Shengyuan_Hu2", "~Steven_Wu1", "~Virginia_Smith1"], "authors_source": "OpenReview API", "abstract": "Many problems in machine learning rely on multi-task learning (MTL), in which the goal is to solve multiple related machine learning tasks simultaneously. MTL is particularly relevant for privacy-sensitive applications in areas such as healthcare, finance, and IoT computing, where sensitive data from multiple, varied sources are shared for the purpose of learning. In this work, we formalize notions of task-level privacy for MTL via joint differential privacy(JDP), a relaxation of differential privacy for mechanism design and distributed optimization. We then propose an algorithm for mean-regularized MTL, an objective commonly used for applications in personalized federated learning, subject to JDP. We analyze our objective and solver, providing certifiable guarantees on both privacy and utility.  Empirically, we find that our method allows for improved privacy/utility trade-offs relative to global baselines across common federated learning benchmarks.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=OBwsUF4nFye", "arxiv_id": "2108.12978", "paper_pdf": "papers/OBwsUF4nFye.pdf", "paper_pdf_sha256": "e13fcdb2e1488db3919dd73797bbecbfff096a7ceae6f886befedf0154ffbe38", "paper_pdf_bytes": 1208779, "paper_pdf_source": "openreview", "code_url": "https://github.com/s-huu/PMTL", "code_repository": "s-huu/PMTL", "code_commit": "9e849ada6af2d8df92c40724d81b4b4cf2ac24b0", "code_archive": "repos/OBwsUF4nFye.zip", "code_archive_sha256": "a5faf61ca8aed85f60630498ea84c5f7f0a1fe2dbce016a22a2c7e76a91eb5b7", "code_archive_bytes": 7688047, "code_file_count": 18, "code_extensions": {".py": 15, ".sh": 3}, "github_disk_usage_kb": 7495, "github_languages": {"Python": 136765, "Shell": 3547}, "github_archived": false, "github_pushed_at": "2023-04-08T21:05:08Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/private-multi-task-learning-formulation-and"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "qn_gk5j3PJ", "year": 2021, "status": "rejected", "title": "PIVEN: A Deep Neural Network for Prediction Intervals with Specific Value Prediction", "authors": ["Eli Simhayev", "Gilad Katz", "Lior Rokach"], "authorids": ["~Eli_Simhayev1", "giladkz@post.bgu.ac.il", "liorrk@post.bgu.ac.il"], "authors_source": "OpenReview API", "abstract": "Improving the robustness of neural nets in regression tasks is key to their application in multiple domains. Deep learning-based approaches aim to achieve this goal either by improving their prediction of specific values (i.e., point prediction), or by producing prediction intervals (PIs) that quantify uncertainty. We present PIVEN, a deep neural network  for producing both a PI and a prediction of specific values. Unlike previous studies, PIVEN makes no assumptions regarding data distribution inside the PI, making its point prediction more effective for various real-world problems. Benchmark experiments show that our approach produces tighter uncertainty bounds than the current state-of-the-art approach for producing PIs, while maintaining comparable performance to the state-of-the-art approach for specific value-prediction. Additional evaluation on large image datasets further support our conclusions.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=qn_gk5j3PJ", "arxiv_id": "2006.05139", "paper_pdf": "papers/qn_gk5j3PJ.pdf", "paper_pdf_sha256": "09e5f450f0190b32544022760c49a533f6836077452b5b688d2998e72a161171", "paper_pdf_bytes": 4424000, "paper_pdf_source": "openreview", "code_url": "https://github.com/elisim/piven", "code_repository": "elisim/piven", "code_commit": "ecdfc024f3e2f63b10a930039b3e6ada3c3c74d4", "code_archive": "repos/qn_gk5j3PJ.zip", "code_archive_sha256": "4174c9f5b706247f718cedf24dcbc32858f14a9aab174fbf9e1f69b7772f4cec", "code_archive_bytes": 5189515, "code_file_count": 19, "code_extensions": {".py": 15, ".sh": 3, ".ipynb": 1}, "github_disk_usage_kb": 5518, "github_languages": {"Jupyter Notebook": 309667, "Python": 110570, "Shell": 585}, "github_archived": false, "github_pushed_at": "2023-03-12T05:51:21Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/piven-a-deep-neural-network-for-prediction"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "7iCT2vmYAR", "year": 2025, "status": "rejected", "title": "Contrastive learning of cell state dynamics in response to perturbations", "authors": ["Soorya Pradeep", "Alishba Imran", "Ziwen Liu", "Eduardo Hirata-Miyasaki", "Taylla Milena Theodoro", "Ivan E. Ivanov", "Madhura Bhave", "Sudip Khadka", "Hunter Woosley", "Carolina Arias", "Shalin B. Mehta"], "authorids": ["~Soorya_Pradeep1", "~Alishba_Imran1", "~Ziwen_Liu5", "~Eduardo_Hirata-Miyasaki1", "~Taylla_Milena_Theodoro1", "~Ivan_E._Ivanov1", "~Madhura_Bhave1", "~Sudip_Khadka1", "~Hunter_Woosley1", "~Carolina_Arias1", "~Shalin_B._Mehta1"], "authors_source": "OpenReview API", "abstract": "We introduce dynaCLR, a self-supervised framework for modeling cell and organelle dynamics via contrastive learning of representations of time-lapse datasets. Live cell imaging of cells and organelles is widely used to analyze cellular responses to perturbations. Supervised modeling of dynamic cell states encoded in 3D time-lapse data is laborious and prone to bias. dy- naCLR leverages single-cell tracking and time-aware contrastive sampling to map images of cells at neighboring time points to neighboring embed- dings. We illustrate the features and applications of dynaCLR with the following experiments: analyzing the kinetics of viral infection in human cells, detecting transient changes in cell morphology due to cell division, and mapping the dynamics of organelles due to viral infection. Temporally regularized embeddings computed with dynaCLR models enable efficient and quantitative annotation, classification, clustering, or interpretation of the cell states. The models reliably embed, i.e., generalize to, data from un- seen experiments with different microscopes and imaging contrasts. Models trained with dynaCLR consistently achieve > 95% accuracy in mitosis and infection state classification, enable the detection of transient cell states and reliably embed unseen experiments. dynaCLR provides a flexible framework for comparative analysis of cell state dynamics due to perturbations, such as infection, gene knockouts, and drugs. We provide PyTorch-based implementations of the model training and inference pipeline and a napari plugin user interface for the visualization and annotation of trajectories of cells in the real space and the embedding space.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=7iCT2vmYAR", "arxiv_id": "2410.11281", "paper_pdf": "papers/7iCT2vmYAR.pdf", "paper_pdf_sha256": "a54aa992dc3b37a0890ef584d0b47c63d54628cc7d9851ca799b2dc34f21024c", "paper_pdf_bytes": 33399544, "paper_pdf_source": "openreview", "code_url": "https://github.com/czbiohub-sf/napari-iohub", "code_repository": "czbiohub-sf/napari-iohub", "code_commit": "6a5d13f16af0b6a7865a117e9309a837743ee28c", "code_archive": "repos/7iCT2vmYAR.zip", "code_archive_sha256": "b9ba16113e16ed3347b8b05184ddf21dd3e603702c8b18800bf080cee9ea7930", "code_archive_bytes": 36012, "code_file_count": 12, "code_extensions": {".py": 12}, "github_disk_usage_kb": 442, "github_languages": {"Python": 76355}, "github_archived": false, "github_pushed_at": "2026-05-21T17:36:06Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/contrastive-learning-of-cell-state-dynamics"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Gpp1dfvZYYH", "year": 2022, "status": "rejected", "title": "ProgFed: Effective, Communication, and Computation Efficient Federated Learning by Progressive Training", "authors": ["Hui-Po Wang", "Sebastian U Stich", "Yang He", "Mario Fritz"], "authorids": ["~Hui-Po_Wang1", "~Sebastian_U_Stich1", "yang.he@cispa.saarland", "~Mario_Fritz1"], "authors_source": "OpenReview API", "abstract": "Federated learning is a powerful distributed learning scheme that allows numerous edge devices to collaboratively train a model without sharing their data. However, training is resource-intensive for edge devices, and limited network bandwidth is often the main bottleneck. Prior work often overcomes the constraints by condensing the models or messages into compact formats, e.g., by gradient compression or distillation. In contrast, we propose ProgFed, the first progressive training framework for efficient and effective federated learning. It inherently reduces computation and two-way communication costs while maintaining the strong performance of the final models. We theoretically prove that ProgFed converges at the same asymptotic rate as standard training on full models. Extensive results on a broad range of architectures, including CNNs (VGG, ResNet, ConvNets) and U-nets, and diverse tasks from simple classification to medical image segmentation show that our highly effective training approach saves up to $20\\%$ computation and up to $63\\%$ communication costs for converged models. As our approach is also complimentary to prior work on compression, we can achieve a wide range of trade-offs, showing reduced communication of up to $50\\times$ at only $0.1\\%$ loss in utility. ", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=Gpp1dfvZYYH", "arxiv_id": "2110.05323", "paper_pdf": "papers/Gpp1dfvZYYH.pdf", "paper_pdf_sha256": "7e84b451cfafcd539ff5178c3ca052c1375b64b2460a244981715a8ea3659ab3", "paper_pdf_bytes": 3503773, "paper_pdf_source": "openreview", "code_url": "https://github.com/hui-po-wang/ProgFed", "code_repository": "hui-po-wang/ProgFed", "code_commit": "1d1eb9d83110bc31ad38447288325560fa1d9878", "code_archive": "repos/Gpp1dfvZYYH.zip", "code_archive_sha256": "eeb75970647d74dc71642004cf5ffc37edccc6670f1157cdc5073530d3665b2d", "code_archive_bytes": 145342, "code_file_count": 17, "code_extensions": {".py": 14, ".sh": 3}, "github_disk_usage_kb": 7744, "github_languages": {"Python": 96470, "Shell": 3430}, "github_archived": false, "github_pushed_at": "2022-10-17T15:05:45Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/progfed-effective-communication-and-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "uELnyih9gqb", "year": 2021, "status": "rejected", "title": "WAVEQ: GRADIENT-BASED DEEP QUANTIZATION OF NEURAL NETWORKS THROUGH SINUSOIDAL REGULARIZATION", "authors": ["Ahmed T. Elthakeb", "Prannoy Pilligundla", "Tarek Elgindi", "Fatemehsadat Mireshghallah", "Charles-Alban Deledalle", "Hadi Esmaeilzadeh"], "authorids": ["~Ahmed_T._Elthakeb1", "~Prannoy_Pilligundla1", "telgindi@ucsd.edu", "~Fatemehsadat_Mireshghallah1", "~Charles-Alban_Deledalle2", "~Hadi_Esmaeilzadeh1"], "authors_source": "OpenReview API", "abstract": "Deep quantization of neural networks below eight bits can lead to superlinear benefits in storage and compute efficiency. However, homogeneously quantizing all the layers to the same level does not account for the distinction of the layers and their individual properties. Heterogenous assignment of bitwidths to individual layers is attractive but opens an exponentially large non-contiguous hyperparameter space (${Available Bitwidths}^{\\# Layers}$). As such finding the bitwidth while also quantizing the network to those levels becomes a major challenge. This paper addresses this challenge through a sinusoidal regularization mechanism, dubbed WaveQ. Adding our parametrized sinusoidal regularizer enables us to not only find the quantized weights but also learn the bitwidth of the layers by making the period of the sinusoidal regularizer a trainable parameter. In addition, the sinusoidal regularizer itself is designed to align its minima on the quantization levels. With these two innovations, during training, stochastic gradient descent uses the form of the sinusoidal regularizer and its minima to push the weights to the quantization levels while it is also learning the period which will determine the bitwidth of each layer separately. As such WaveQ is a gradient-based mechanism that jointly learns the quantized weights as well as the heterogeneous bitwidths. We show how WaveQ balance compute efficiency and accuracy, and provide a heterogeneous bitwidth assignment for quantization of a large variety of deep networks (AlexNet, CIFAR-10, MobileNet, ResNet-18, ResNet-20, SVHN, and VGG-11) that virtually preserves the accuracy. WaveQ is versatile and can also be used with predetermined bitwidths by fixing the period of the sinusoidal regularizer. In this case. WaveQ enhances quantized training algorithms (DoReFa and WRPN) with about 4.8% accuracy improvements on average, and outperforms multiple state-of-the-art techniques. Finally, WaveQ applied to quantizing transformers\n", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=uELnyih9gqb", "arxiv_id": null, "paper_pdf": "papers/uELnyih9gqb.pdf", "paper_pdf_sha256": "5a682c3fb2d6b864bfc43325e3e695ec542c92bc507c527d834559d6e44354a7", "paper_pdf_bytes": 6135804, "paper_pdf_source": "openreview", "code_url": "https://github.com/waveq-reg/waveq", "code_repository": "waveq-reg/waveq", "code_commit": "39bbf0dd95f5bf139197152e651acba08f5ccc68", "code_archive": "repos/uELnyih9gqb.zip", "code_archive_sha256": "e62168ffadd6946d47432e0d32637d2b2c34742fd8564bc4035a23797dc70b73", "code_archive_bytes": 10720076, "code_file_count": 344, "code_extensions": {".py": 318, ".sh": 10, ".cpp": 7, ".cu": 4, ".cuh": 2, ".lua": 2, ".h": 1}, "github_disk_usage_kb": 5593, "github_languages": {"Python": 9078893, "Cuda": 36414, "C++": 15854, "Shell": 4215, "Lua": 4210}, "github_archived": false, "github_pushed_at": "2020-06-11T19:08:04Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/waveq-gradient-based-deep-quantization-of"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "Bo5eKnJPML", "year": 2025, "status": "rejected", "title": "A Reasoning-Based Approach to Cryptic Crossword Clue Solving", "authors": ["Martin Andrews", "Sam Witteveen"], "authorids": ["~Martin_Andrews1", "~Sam_Witteveen1"], "authors_source": "OpenReview API", "abstract": "Cryptic crossword clues are challenging language tasks for which new test sets are released daily by major newspapers on a global basis. Each cryptic clue contains both the definition of the answer to be placed in the crossword grid (in common with regular crosswords), and ‘wordplay’ that *proves* that the answer is correct (i.e. a human solver can be confident that an answer is correct without needing crossing words as confirmation). This work describes an LLM-based reasoning system built from open-licensed components that solves cryptic clues by (i) hypothesising answers; (ii) proposing wordplay explanations; and (iii) using a verifier system that operates on codified reasoning steps. Overall, this system establishes a new state-of-the-art performance on the challenging Cryptonite dataset of clues from The Times and The Telegraph newspapers in the UK. Because each proved solution is expressed in Python, interpretable wordplay reasoning for proven answers is available for inspection", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=Bo5eKnJPML", "arxiv_id": "2506.04824", "paper_pdf": "papers/Bo5eKnJPML.pdf", "paper_pdf_sha256": "7a54ae3cbcd5c4d48606e426ede74932347f8bbb56f15cc3e28d452c5ec9f02d", "paper_pdf_bytes": 556499, "paper_pdf_source": "openreview", "code_url": "https://github.com/mdda/cryptic-crossword-reasoning-verifier", "code_repository": "mdda/cryptic-crossword-reasoning-verifier", "code_commit": "0122293fbe640f0aa4b837d37841eb0df34fec38", "code_archive": "repos/Bo5eKnJPML.zip", "code_archive_sha256": "3a2204f9bda6de0b50722aab90bc9b9711ac8344630481e9402ac1cade509954", "code_archive_bytes": 418350, "code_file_count": 23, "code_extensions": {".py": 14, ".ipynb": 9}, "github_disk_usage_kb": 442, "github_languages": {"Jupyter Notebook": 1208319, "Python": 333246}, "github_archived": false, "github_pushed_at": "2026-01-11T18:50:18Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/a-reasoning-based-approach-to-cryptic"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "D6aGz0Zyvn", "year": 2024, "status": "rejected", "title": "Enhancing Kernel Flexibility via Learning Asymmetric Locally-Adaptive Kernels", "authors": ["FAN He", "Mingzhen He", "Lei Shi", "Xiaolin Huang", "Johan Suykens"], "authorids": ["~FAN_He1", "~Mingzhen_He1", "~Lei_Shi7", "~Xiaolin_Huang1", "~Johan_Suykens1"], "authors_source": "OpenReview API", "abstract": "The lack of sufficient flexibility is the key bottleneck of kernel-based learning that relies on manually designed, pre-given, and non-trainable kernels. To enhance kernel flexibility, this paper introduces the concept of Locally-Adaptive-Bandwidths (LAB) as trainable parameters to enhance the Radial Basis Function (RBF) kernel, giving rise to the LAB RBF kernel. The parameters in LAB RBF kernels are data-dependent, and its number can increase with the dataset, allowing for better adaptation to diverse data patterns and enhancing the flexibility of the learned function. This newfound flexibility also brings challenges, particularly with regards to asymmetry and the need for an efficient learning algorithm. To address these challenges, this paper for the first time establishes an asymmetric kernel ridge regression framework and introduces an iterative kernel learning algorithm. This novel approach not only reduces the demand for extensive support data but also significantly improves generalization by training bandwidths on the available training data. Experimental results on real datasets underscore the remarkable performance of the proposed algorithm, showcasing its superior capability in handling large-scale datasets compared to Nyström approximation-based algorithms. Moreover, it demonstrates a significant improvement in regression accuracy over existing kernel-based learning methods and even surpasses residual neural networks.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=D6aGz0Zyvn", "arxiv_id": "2310.05236", "paper_pdf": "papers/D6aGz0Zyvn.pdf", "paper_pdf_sha256": "1ecddbb3608915a6ebb9c6c87cc1163154f3fde7264783fb76afa2aa93518141", "paper_pdf_bytes": 424349, "paper_pdf_source": "openreview", "code_url": "https://github.com/hefansjtu/LABRBF_kernel", "code_repository": "hefansjtu/LABRBF_kernel", "code_commit": "0ed445b3ac5bae7e0cc6bb25a5064ec028bf8436", "code_archive": "repos/D6aGz0Zyvn.zip", "code_archive_sha256": "e784a7e42e8727b25f19d99595aca7b11797ca639b4a757a4d96b69c3a920416", "code_archive_bytes": 553627, "code_file_count": 10, "code_extensions": {".py": 10}, "github_disk_usage_kb": 556, "github_languages": {"Python": 100417}, "github_archived": false, "github_pushed_at": "2023-11-24T12:14:16Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/enhancing-kernel-flexibility-via-learning"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "DvMDIEFtyjV", "year": 2023, "status": "rejected", "title": "CLUTR: Curriculum Learning via Unsupervised Task Representation Learning", "authors": ["Abdus Salam Azad", "Izzeddin Gur", "Aleksandra Faust", "Pieter Abbeel", "Ion Stoica"], "authorids": ["~Abdus_Salam_Azad1", "~Izzeddin_Gur1", "~Aleksandra_Faust1", "~Pieter_Abbeel2", "~Ion_Stoica1"], "authors_source": "OpenReview API", "abstract": "Reinforcement Learning (RL) algorithms are often known for sample inefficiency and difficult generalization. Recently, Unsupervised Environment Design (UED) emerged as a new paradigm for zero-shot generalization by simultaneously learning a task distribution and agent policies on the sampled tasks. This is a non-stationary process where the task distribution evolves along with agent policies; creating an instability over time. While past works demonstrated the potential of such approaches, sampling effectively from the task space remains an open challenge, bottlenecking these approaches. To this end, we introduce CLUTR: a novel curriculum learning algorithm that decouples task representation and curriculum learning into a two-stage optimization. It first trains a recurrent variational autoencoder on randomly generated tasks to learn a latent task manifold. Next, a teacher agent creates a curriculum by optimizing a minimax REGRET-based objective on a set of latent tasks sampled from this manifold. By keeping the task manifold fixed, we show that CLUTR successfully overcomes the non-stationarity problem and improves stability. Our experimental results show CLUTR outperforms PAIRED, a principled and popular UED method, in terms of generalization and sample efficiency in the challenging CarRacing and navigation environments: showing an 18x improvement on the F1 CarRacing benchmark. CLUTR also performs comparably to the non-UED state-of-the-art for CarRacing, outperforming it in nine of the 20 tracks. CLUTR also achieves a 33% higher solved rate than PAIRED on a set of 18 out-of-distribution navigation tasks.", "decision": "Reject", "meta_review": null, "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=DvMDIEFtyjV", "arxiv_id": "2210.10243", "paper_pdf": "papers/DvMDIEFtyjV.pdf", "paper_pdf_sha256": "0aeeb5936b11bae3bd0baf887738a66d9513aed4b572c5673ea99a2bb15c8a32", "paper_pdf_bytes": 2658059, "paper_pdf_source": "openreview", "code_url": "https://github.com/clutr/clutr", "code_repository": "clutr/clutr", "code_commit": "b461ffdfc937a231b7a5086bb520c1d3f9b88c39", "code_archive": "repos/DvMDIEFtyjV.zip", "code_archive_sha256": "b439267c18b401b50d17c52ce9398cd64fef1d7d459dde1f3f2154579254a28b", "code_archive_bytes": 1849371, "code_file_count": 83, "code_extensions": {".py": 80, ".sh": 3}, "github_disk_usage_kb": 1789, "github_languages": {"Python": 4654068, "Shell": 4210}, "github_archived": false, "github_pushed_at": "2022-12-10T00:23:30Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/clutr-curriculum-learning-via-unsupervised"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "luO6l9cP6b6", "year": 2022, "status": "rejected", "title": "Identifying the Limits of Cross-Domain Knowledge Transfer for Pretrained Models", "authors": ["Zhengxuan Wu", "Nelson F. Liu", "Christopher Potts"], "authorids": ["~Zhengxuan_Wu1", "~Nelson_F._Liu1", "~Christopher_Potts1"], "authors_source": "OpenReview API", "abstract": "There is growing evidence that pretrained language models improve task-specific fine-tuning even where the task examples are radically different from those seen in training. What is the nature of this surprising cross-domain transfer? We offer a partial answer via a systematic exploration of how much transfer occurs when models are denied any information about word identity via random scrambling. In four classification tasks and two sequence labeling tasks, we evaluate LSTMs using GloVe embeddings, BERT, and baseline models. Among these models, we find that only BERT shows high rates of transfer into our scrambled domains, and for classification but not sequence labeling tasks. Our analyses seek to explain why transfer succeeds for some tasks but not others, to isolate the separate contributions of pretraining versus fine-tuning, to show that the fine-tuning process is not merely learning to unscramble the scrambled inputs, and to quantify the role of word frequency. These findings help explain where and why cross-domain transfer occurs, which can guide future studies and practical fine-tuning efforts.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=luO6l9cP6b6", "arxiv_id": "2104.08410", "paper_pdf": "papers/luO6l9cP6b6.pdf", "paper_pdf_sha256": "220871086cc7ab4dc0b183640c4e3d76dd90b87f380076f3a8a96029058fea71", "paper_pdf_bytes": 7142459, "paper_pdf_source": "openreview", "code_url": "https://github.com/frankaging/limits-cross-domain-transfer", "code_repository": "frankaging/limits-cross-domain-transfer", "code_commit": "318d5154f2391006709a16ab47fae3fa896c42fc", "code_archive": "repos/luO6l9cP6b6.zip", "code_archive_sha256": "c961790df54567a452fbdbaf6d9d6cb4d7eda53a34a63c20b1c787878a83c153", "code_archive_bytes": 1297450, "code_file_count": 26, "code_extensions": {".ipynb": 14, ".py": 11, ".sh": 1}, "github_disk_usage_kb": 7788, "github_languages": {"Jupyter Notebook": 2070243, "Python": 227190, "Shell": 701}, "github_archived": false, "github_pushed_at": "2021-11-12T06:31:52Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/identifying-the-limits-of-cross-domain"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "ZvvxYyjfvZc", "year": 2021, "status": "rejected", "title": "Correcting Momentum in Temporal Difference Learning", "authors": ["Emmanuel Bengio", "Joelle Pineau", "Doina Precup"], "authorids": ["~Emmanuel_Bengio1", "~Joelle_Pineau1", "~Doina_Precup1"], "authors_source": "OpenReview API", "abstract": "A common optimization tool used in deep reinforcement learning is momentum, which consists in accumulating and discounting past gradients, reapplying them at each iteration. We argue that, unlike in supervised learning, momentum in Temporal Difference (TD) learning accumulates gradients that become doubly stale: not only does the gradient of the loss change due to parameter updates, the loss itself changes due to bootstrapping. We first show that this phenomenon exists, and then propose a first-order correction term to momentum. We show that this correction term improves sample efficiency in policy evaluation by correcting target value drift. An important insight of this work is that deep RL methods are not always best served by directly importing techniques from the supervised setting.", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=ZvvxYyjfvZc", "arxiv_id": "2106.03955", "paper_pdf": "papers/ZvvxYyjfvZc.pdf", "paper_pdf_sha256": "4cfc3a5841809cfedf0d3e1104b1dc0fd87e27f91b738850fbe648a5039acb3a", "paper_pdf_bytes": 709918, "paper_pdf_source": "openreview", "code_url": "https://github.com/bengioe/staleness-corrected-momentum", "code_repository": "bengioe/staleness-corrected-momentum", "code_commit": "ecdc5a3e49c009471c6cac8af8ebd58ed908115c", "code_archive": "repos/ZvvxYyjfvZc.zip", "code_archive_sha256": "83d6a5d81ee47251e9ae49ec17beab073078da29c710bc5bf98abe77bd151831", "code_archive_bytes": 5815991, "code_file_count": 12, "code_extensions": {".py": 12}, "github_disk_usage_kb": 5680, "github_languages": {"Python": 112910}, "github_archived": false, "github_pushed_at": "2021-06-28T17:58:24Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/correcting-momentum-in-temporal-difference-1"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "aYx7JR20sI", "year": 2025, "status": "rejected", "title": "Tropical Expressivity of Neural Networks", "authors": ["Paul Lezeau", "Thomas Walker", "Yueqi Cao", "Shiv Bhatia", "Anthea Monod"], "authorids": ["~Paul_Lezeau1", "~Thomas_Walker2", "~Yueqi_Cao1", "~Shiv_Bhatia1", "~Anthea_Monod1"], "authors_source": "OpenReview API", "abstract": "We propose an algebraic geometric framework to study the expressivity of piecewise linear activation neural networks.  A particular quantity of neural networks that has been actively studied is the number of linear regions, which gives a quantification of the information capacity of the architecture.  To study and evaluate information capacity and expressivity, we work in the setting of tropical geometry---a combinatorial and polyhedral variant of algebraic geometry---where there are known connections between tropical rational maps and feedforward neural networks. Our work builds on and expands this connection to capitalize on the rich theory of tropical geometry to characterize and study various architectural aspects of neural networks. Our contributions are threefold: we provide a novel tropical geometric approach to selecting sampling domains among linear regions; an algebraic result allowing for a guided restriction of the sampling domain for network architectures with symmetries; and a new open source OSCAR library to analyze neural networks symbolically using their tropical representations, where we present a new algorithm that computes the exact number of their linear regions. We provide a comprehensive set of proof-of-concept numerical experiments demonstrating the breadth of neural network architectures to which tropical geometric theory can be applied to reveal insights on expressivity characteristics of a network.  Our work provides the foundations for the adaptation of both theory and existing software from computational tropical geometry and symbolic computation to neural networks and deep learning.", "decision": "Reject", "meta_review": "", "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=aYx7JR20sI", "arxiv_id": "2405.20174", "paper_pdf": "papers/aYx7JR20sI.pdf", "paper_pdf_sha256": "77dd594f146bb00e667ffdf799c9032ceeb5d1490139b3e5f29cfe9d3017d6af", "paper_pdf_bytes": 525659, "paper_pdf_source": "openreview", "code_url": "https://github.com/Paul-Lez/tropicalnn", "code_repository": "Paul-Lez/tropicalnn", "code_commit": "479ec5e96030f46fe5d0783d74ca193196512729", "code_archive": "repos/aYx7JR20sI.zip", "code_archive_sha256": "ceb1903c51542bbc926e575c1322b4f3b31238713f2f846c3dff7746f43dfd85", "code_archive_bytes": 61274, "code_file_count": 15, "code_extensions": {".jl": 14, ".sh": 1}, "github_disk_usage_kb": 451, "github_languages": {"Julia": 66379, "Shell": 1603}, "github_archived": false, "github_pushed_at": "2026-08-28T16:56:29Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/tropical-expressivity-of-neural-networks"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "YGTSLDAPqb", "year": 2024, "status": "rejected", "title": "Connect Later: Improving Fine-Tuning for Robustness with Targeted Augmentations", "authors": ["Helen Qu", "Sang Michael Xie"], "authorids": ["~Helen_Qu1", "~Sang_Michael_Xie1"], "authors_source": "OpenReview API", "abstract": "Models trained on a labeled source domain (e.g., bright, nearby astronomical objects) often generalize poorly when deployed on an out-of-distribution (OOD) target domain (e.g., faint, distant objects). In the domain adaptation setting where unlabeled target data is available, self-supervised pretraining (e.g., masked autoencoding or contrastive learning) is a promising method to mitigate this performance drop. Pretraining improves OOD error when the generic data augmentations used (e.g., masking or cropping) connect the source and target domains, which may be far apart in the input space. In this paper, we show on real-world tasks that standard fine-tuning after pretraining does not consistently improve OOD error over just supervised learning on labeled source data. To better leverage pretraining for distribution shifts, we propose Connect Later: after pretraining with generic augmentations to learn good representations within the source and target domains, fine-tune with targeted augmentations designed with knowledge of the distribution shift to better connect the domains. Connect Later improves average OOD error over standard fine-tuning and supervised learning with targeted augmentations on 3 real-world datasets: astronomical time-series classification (AstroClassification) by 12%, redshift prediction for astronomical time-series (Redshifts) by 0.03 RMSE (11% relative), and wildlife species identification (iWildCam-WILDS) by 0.9%, achieving the state-of-the-art on AstroClassification and on iWildCam-WILDS with ResNet-50.", "decision": "Reject", "meta_review": "", "num_reviews": 3, "openreview_url": "https://openreview.net/forum?id=YGTSLDAPqb", "arxiv_id": "2402.03325", "paper_pdf": "papers/YGTSLDAPqb.pdf", "paper_pdf_sha256": "d62f4a92d56ceb897111d1b57253777107275f060253eeb3d030f22fa0ea4439", "paper_pdf_bytes": 1827859, "paper_pdf_source": "openreview", "code_url": "https://github.com/helenqu/connect-later", "code_repository": "helenqu/connect-later", "code_commit": "6a12eab8918ac0152fa79aa168c028e6f378a478", "code_archive": "repos/YGTSLDAPqb.zip", "code_archive_sha256": "cee7e178d6e53111ce3368e94c9d1eb60da77dfdc95690c4dc139f3817f77456", "code_archive_bytes": 40610, "code_file_count": 23, "code_extensions": {".py": 13, ".sh": 10}, "github_disk_usage_kb": 556, "github_languages": {"Python": 111188, "Shell": 5135}, "github_archived": false, "github_pushed_at": "2024-01-03T18:09:49Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/connect-later-improving-fine-tuning-for"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}
{"forum": "IJwhRE510b", "year": 2023, "status": "rejected", "title": "ELODI: Ensemble Logit Difference Inhibition for Positive-Congruent Training", "authors": ["Yue Zhao", "Yantao Shen", "Yuanjun Xiong", "Shuo Yang", "Wei Xia", "Zhuowen Tu", "Bernt Schiele", "Stefano Soatto"], "authorids": ["~Yue_Zhao4", "~Yantao_Shen2", "~Yuanjun_Xiong3", "~Shuo_Yang2", "~Wei_Xia6", "~Zhuowen_Tu1", "~Bernt_Schiele1", "~Stefano_Soatto1"], "authors_source": "OpenReview API", "abstract": "Negative flips are errors introduced in a classification system when a legacy model is updated. Existing methods to reduce the negative flip rate (NFR) either do so at the expense of overall accuracy by forcing a new model to imitate the old models, or use ensembles, which multiply inference cost prohibitively. We analyze the role of ensembles in reducing NFR and observe that they remove negative flips that are typically not close to the decision boundary, but often exhibit large deviations in the distance among their logits. Based on the observation, we present a method, called Ensemble Logit Difference Inhibition ELODI, to train a classification system that achieves paragon performance in both error rate and NFR, at the inference cost of a single model. The method distills a homogeneous ensemble to a single student model which is used to update the classification system. ELODI also introduces a generalized distillation objective, Logit Difference Inhibition (LDI), which penalizes changes in the logits between the reference ensemble and the student single model.\nOn multiple image classification benchmarks, model updates with ELODI demonstrate superior accuracy retention and NFR reduction. ", "decision": "Reject", "meta_review": null, "num_reviews": 4, "openreview_url": "https://openreview.net/forum?id=IJwhRE510b", "arxiv_id": "2205.06265", "paper_pdf": "papers/IJwhRE510b.pdf", "paper_pdf_sha256": "962ab09db994ac743851b1dfde9a9550e79dc50cd610caef821d32d8383aafa6", "paper_pdf_bytes": 1817942, "paper_pdf_source": "openreview", "code_url": "https://github.com/amazon-science/regression-constraint-model-upgrade", "code_repository": "amazon-science/regression-constraint-model-upgrade", "code_commit": "fb2268c67706834632ec1427945fcf30a7218227", "code_archive": "repos/IJwhRE510b.zip", "code_archive_sha256": "f96d039c962b8974340148d30689577d69def49e44d86c514613f84caa48fcbb", "code_archive_bytes": 1964441, "code_file_count": 45, "code_extensions": {".py": 45}, "github_disk_usage_kb": 1809, "github_languages": {"Python": 349619}, "github_archived": false, "github_pushed_at": "2023-12-08T20:26:18Z", "code_match": {"method": "exact_normalized_title", "pwc_official": true, "pwc_mentioned_in_paper": true, "pwc_url": "https://paperswithcode.com/paper/elodi-ensemble-logit-difference-inhibition"}, "provenance": {"paper_reviews_decision": "OpenReview API", "paper_code_mapping": "Papers With Code archive", "repository_snapshot": "GitHub API commit-pinned ZIP"}}