Hugging Face's logo

Datasets:
sidneibarbieri
/
topvenues-complete

Tasks:
Text Retrieval
Modalities:
Text
Formats:
parquet
Languages:
English
Size:
100K - 1M
Tags:
cybersecurity
security
bibliography
literature-review
dblp
Libraries:
Datasets
pandas
Polars
License:
Dataset card Data Studio Files
xet
 Community
1
sidneibarbieri commited on
Commit
a572e6a
·
verified ·
1 Parent(s): 0defa14

Complete corpus: 118,817 distinct works across 40 venues

Browse files
Files changed (2) hide show
  1. README.md +123 -0
  2. train-00000-of-00001.parquet +3 -0
README.md ADDED
@@ -0,0 +1,123 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ pretty_name: TopVenues Complete Corpus
3
+ license: other
4
+ language:
5
+ - en
6
+ task_categories:
7
+ - text-retrieval
8
+ tags:
9
+ - cybersecurity
10
+ - security
11
+ - bibliography
12
+ - literature-review
13
+ - dblp
14
+ size_categories:
15
+ - 100K<n<1M
16
+ configs:
17
+ - config_name: default
18
+ data_files:
19
+ - split: train
20
+ path: train-*.parquet
21
+ ---
22
+
23
+ # TopVenues Complete Corpus
24
+
25
+ The widest corpus the TopVenues pipeline produces: **118,817 distinct
26
+ publications** across **40 venues** (2017-2026), every record carrying a
27
+ canonical BibTeX entry and a DBLP key.
28
+
29
+ Built by a monotonic union of the `full-40` and `security-20` profiles. Neither
30
+ was a superset of the other, so the union recovers records and abstracts that
31
+ either profile alone was missing.
32
+
33
+ - **Code / tool:** <https://github.com/sidneibarbieri/topvenues-tool>
34
+ - **Profile:** `complete-40`
35
+ - **Snapshot SHA-256 (gzip):** `271e6d4dd30020cea16801045feddc871b3c69e673f1704307b03fa3d81f5ff5`
36
+
37
+ ## Not a paper denominator
38
+
39
+ This corpus is a tool release, not the denominator of any paper. Two separate
40
+ frozen snapshots back the published work and are **not** replaced by this one:
41
+
42
+ | Release | Records | Cited by |
43
+ |---|---|---|
44
+ | `submitted-11` | 9,925 | the full paper |
45
+ | `security-20` (dataset [`sidneibarbieri/topvenues`](https://huggingface.co/datasets/sidneibarbieri/topvenues)) | 20,305 | the tools-track paper |
46
+
47
+ Any measurement must name the snapshot that produced its denominator. Numbers
48
+ from one corpus do not transfer to another merely because the software is shared.
49
+
50
+ ## One record per work
51
+
52
+ Records are deduplicated by DBLP key: **118,817 rows, 118,817 distinct
53
+ keys**, so a count of rows is a count of distinct publications.
54
+
55
+ ## Abstract coverage is uneven, by construction
56
+
57
+ 14,289 of 118,817 records carry an abstract (12.0%). That
58
+ figure is low because the corpus deliberately includes large AI and systems
59
+ venues whose records were ingested for bibliographic completeness without
60
+ abstract enrichment. Coverage on the security venues is high. 21 of
61
+ 40 venues carry abstracts at all:
62
+
63
+ | Venue | Records | Abstracts | Coverage |
64
+ |---|---:|---:|---:|
65
+ | USENIX Security | 2,103 | 2,056 | 97.8% |
66
+ | ACM CCS | 1,972 | 1,971 | 99.9% |
67
+ | TrustCom | 1,915 | 1,912 | 99.8% |
68
+ | ACM Computing Surveys | 1,813 | 1,812 | 99.9% |
69
+ | IEEE S&P | 1,166 | 1,165 | 99.9% |
70
+ | NDSS | 1,059 | 1,058 | 99.9% |
71
+ | ACM ASIA CCS | 837 | 827 | 98.8% |
72
+ | IEEE Communications Surveys & Tutorials | 833 | 804 | 96.5% |
73
+ | ACSAC | 509 | 509 | 100.0% |
74
+ | IEEE CNS | 381 | 381 | 100.0% |
75
+ | IEEE EURO S&P | 339 | 339 | 100.0% |
76
+ | ACM CODASPY | 316 | 281 | 88.9% |
77
+ | ACM WiSec | 314 | 270 | 86.0% |
78
+ | HotNets | 245 | 245 | 100.0% |
79
+ | RAID | 290 | 222 | 76.6% |
80
+ | ACM SACMAT | 176 | 176 | 100.0% |
81
+ | IEEE SaTML | 127 | 127 | 100.0% |
82
+ | ACM AISec | 77 | 73 | 94.8% |
83
+ | ESORICS | 598 | 39 | 6.5% |
84
+ | Foundations and Trends in Privacy and Security | 20 | 20 | 100.0% |
85
+ | USENIX WOOT | 18 | 2 | 11.1% |
86
+ | AAAI | 20,085 | 0 | 0.0% |
87
+ | NeurIPS | 16,527 | 0 | 0.0% |
88
+ | EMNLP | 12,817 | 0 | 0.0% |
89
+ | ICML | 11,971 | 0 | 0.0% |
90
+ | ACL | 11,294 | 0 | 0.0% |
91
+ | ICLR | 10,680 | 0 | 0.0% |
92
+ | IJCAI | 6,501 | 0 | 0.0% |
93
+ | ACM KDD | 4,110 | 0 | 0.0% |
94
+ | NAACL | 4,107 | 0 | 0.0% |
95
+ | ACM MobiCom | 993 | 0 | 0.0% |
96
+ | ACM SenSys | 808 | 0 | 0.0% |
97
+ | ACM MobiSys | 655 | 0 | 0.0% |
98
+ | USENIX NSDI | 542 | 0 | 0.0% |
99
+ | ACM EuroSys | 518 | 0 | 0.0% |
100
+ | USENIX ATC | 518 | 0 | 0.0% |
101
+ | ACM IMC | 466 | 0 | 0.0% |
102
+ | ACM SIGCOMM | 452 | 0 | 0.0% |
103
+ | ACM SIGMETRICS | 376 | 0 | 0.0% |
104
+ | ACM HotMobile | 289 | 0 | 0.0% |
105
+
106
+ ## Usage
107
+
108
+ ```python
109
+ from datasets import load_dataset
110
+
111
+ corpus = load_dataset("sidneibarbieri/topvenues-complete", split="train")
112
+ security = corpus.filter(lambda r: r["abstract"] and "USENIX" in (r["event"] or ""))
113
+ ```
114
+
115
+ ## Fields
116
+
117
+ `paper_id`, `key` (DBLP), `title`, `authors`, `event`, `venue`, `area`, `year`,
118
+ `paper_type`, `pages`, `url`, `ee` (DOI), `abstract`, `bibtex`.
119
+
120
+ ## License
121
+
122
+ Metadata originates from DBLP (CC0) and open scholarly APIs. Abstracts remain
123
+ under their publishers' terms and are included for research use.
train-00000-of-00001.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2298ba3608fef9472b3d2e7a28e602e9f29d77bac4c7db4c3f2d733fa9b62d98
3
+ size 48548478