EA2N: Evidence-based AMR Attention Network for Fake News Detection
Primary Area: representation learning for computer vision, audio, language, and other modalities
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.
Keywords: Fake News Detection, AMR Network, Natural Language Processing
Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2024/AuthorGuide.
TL;DR: This study proposes a new evidence based representation learning method for fake news detection.
Abstract: Proliferation of fake news has become a critical issue in today's information-driven society. Our study includes external knowledge from Wikidata and deviates from the reliance on social information to detect fake news, that many state-of-the-art (SOTA) fact-checking models adopt. This paper introduces EA$^2$N, an Evidence-based AMR Attention Network for Fake News Detection. EA$^2$N leverages Abstract Meaning Representation (AMR) and incorporates knowledge from Wikidata using proposed evidence linking algorithm, pushing the boundaries of fake news detection. The proposed framework encompasses a combination of novel language encoder and graph encoder to detect the fake news. While the language encoder effectively combines transformer encoded textual features with affective lexical features, the graph encoder encodes AMR with evidence through external knowledge, referred as WikiAMR graph. A path-aware graph learning module is designed to capture crucial semantic relationships among entities over evidences. Extensive experiments supports our model's superior performance, surpassing SOTA methodologies. This research not only advances the field of Fake News Detection but also showcases the potential of AMR and external knowledge for robust NLP applications, promising a more trustworthy information landscape.
Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors' identity.
No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.
Submission Number: 2835
Loading