Relation-aware graph network improves harm-aware fake news detection
RAEGNet: Relation-Aware Evidence Graph Network for Harm-Aware Multimodal Fake News Detection
Computation and LanguageComputer Vision and Pattern RecognitionMultimedia
Summary
Detecting fake news often uses extra related information, but existing methods focus too much on individual entities, which can cause irrelevant distractions. The authors designed a way to retrieve evidence at the event level, meaning they look at the whole news story for clues. They then use a special graph network that understands how pieces of evidence relate to each other and to the news, also estimating how harmful the fake news could be. Tests on multiple datasets show their method works better than earlier ones.
What this means in practice
- •For social media platforms: Improve automated detection of harmful fake news by analyzing whole events and their evidence interrelations more accurately.
- •For content moderation teams: Prioritize review and intervention based on the estimated harm level of fake news items detected using event-level evidence analysis.
Authors
Wenbin Shen, Guoxuan Qin, Guangxu Yao, Baodong Wang, Yuanbo Rui, Zhongjie Ba, Zhichao Lian
Abstract
Existing multimodal fake news detection methods often introduce external information to assist detection. However, most of them rely on entity-level retrieval and are therefore prone to introducing event-irrelevant noise. Meanwhile, existing methods mainly focus on improving overall performance and do not account for differences in the degree of harm posed by different instances of fake news. To address these limitations, we design an Event-Level Evidence Retrieval Framework (ELERF) and propose a Relation-Aware Evidence Graph Network (RAEGNet). ELERF retrieves external evidence based on the complete event semantics of a news item. RAEGNet constructs a directed graph that incorporates news-evidence stance relations and evidence-evidence interaction relations, and introduces a conditional-harm branch to jointly model authenticity and potential harm. Experimental results demonstrate that RAEGNet outperforms multiple baseline methods across all evaluated metrics on Weibo-21, Fakeddit, and our self-constructed SSS dataset.