Problem
Research questions and friction points this paper is trying to address.
Multimodal Fake News Detection
Evidence Retrieval
Harm Awareness
Event-level Noise
RAEGNet: Relation-Aware Evidence Graph Network for Harm-Aware Multimodal Fake News Detection
🤖 AI Summary
This study addresses the limitations of entity-level retrieval in multimodal fake news detection, which often introduces irrelevant noise and overlooks varying degrees of harmfulness. To this end, we propose an event-level evidence retrieval framework coupled with a relation-aware evidence graph network. Methodologically, event-level retrieval replaces conventional entity-level approaches to effectively eliminate extraneous noise. Architecturally, a directed graph neural network is employed for joint modeling, alongside a novel conditional harm branch that enables the simultaneous assessment of veracity and harmfulness severity. Experimental results demonstrate that the proposed approach comprehensively outperforms existing baselines on datasets such as Weibo-21, yielding significant improvements in both detection accuracy and harm-awareness capabilities.
📝 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.
Innovation
Methods, ideas, or system contributions that make the work stand out.
Multimodal Fake News Detection
Event-Level Evidence Retrieval
Relation-Aware Evidence Graph Network
Harm-Aware Modeling
Stance Relations
🔎 Similar Papers
No similar papers found.
💼 Related Jobs
No related jobs found.
