iNews: A Multimodal Dataset for Modeling Personalized Affective Responses to News
Paper • 2503.03335 • Published
Request access to the iNews screenshots
These screenshots reproduce public Facebook news posts and may contain third-party copyrighted material. Access is provided only for approved, non-commercial academic research. Requests are reviewed manually.
Log in or Sign Up to review the conditions and access this dataset content.
iNews Facebook Post Screenshots
This package contains the screenshots corresponding to every unique Post_ID
in the five splits currently released by pitehu/inews and
pitehu/inews_public on Hugging Face.
Contents
inews_screenshots_hf_release.zip: the complete release archive containing 2,736 PNG screenshots named<Post_ID>.png, plus this README and manifest.manifest.csv: the filename, phase, Hugging Face split membership, byte size, and SHA-256 checksum for every screenshot.
Joining the data
Extract the archive and join a dataset row to its image using Post_ID. For
example, the row with Post_ID Phase2_15 maps to
inews_screenshots_hf_release/images/Phase2_15.png inside the archive.
Repeated annotations of the same post share one screenshot. The 11,320 rows in the currently released Hugging Face splits therefore map to 2,736 unique image files.
The unreleased withheld evaluation annotations are not included. In particular,
Phase2_652, the only screenshot unique to those withheld rows, is excluded.
Use
The screenshots reproduce public Facebook posts and may contain material owned by Facebook, news publishers, photographers, or other third parties. Access to this package does not transfer ownership of that material or grant permission for redistribution. Use is intended for authorized non-commercial research in accordance with the iNews dataset access conditions.
Access is granted to individual researchers, not laboratories or institutions. Every person who downloads or handles the files must be named in the request or obtain access separately. Approval may be revoked if these conditions are not followed.
Citation
@article{hu2025inews,
title={iNews: A Multimodal Dataset for Modeling Personalized Affective Responses to News},
author={Hu, Tiancheng and Collier, Nigel},
journal={arXiv preprint arXiv:2503.03335},
year={2025}
}
- Downloads last month
- 32