Problem
Research questions and friction points this paper is trying to address.
explainable fake news detection
large language models
unverified reports
comprehensible explanation
breaking news
A Graph-Enhanced Defense Framework for Explainable Fake News Detection with LLM
🤖 AI Summary
This work addresses critical limitations in existing explainable fake news detection methods—namely, their inefficiency, poor adaptability to breaking news, reliance on unverified external sources that may introduce errors, and lack of fine-grained explanations for news claims. To overcome these challenges, the authors propose G-Defense, a novel framework that integrates graph-structured representation with argumentation-inspired reasoning. G-Defense constructs a claim-centered sub-claim dependency graph, leverages retrieval-augmented generation to gather evidence and produce competing explanations, and employs an argumentation-style reasoning module to assess overall veracity. This process guides large language models to generate intuitive, human-interpretable explanation graphs. Notably, G-Defense achieves comprehensive, fine-grained interpretability using only unverified reports, significantly outperforming state-of-the-art methods in both veracity assessment accuracy and explanation quality.
📝 Abstract
Explainable fake news detection aims to assess the veracity of news claims while providing human-friendly explanations. Existing methods incorporating investigative journalism are often inefficient and struggle with breaking news. Recent advances in large language models (LLMs) enable leveraging externally retrieved reports as evidence for detection and explanation generation, but unverified reports may introduce inaccuracies. Moreover, effective explainable fake news detection should provide a comprehensible explanation for all aspects of a claim to assist the public in verifying its accuracy. To address these challenges, we propose a graph-enhanced defense framework (G-Defense) that provides fine-grained explanations based solely on unverified reports. Specifically, we construct a claim-centered graph by decomposing the news claim into several sub-claims and modeling their dependency relationships. For each sub-claim, we use the retrieval-augmented generation (RAG) technique to retrieve salient evidence and generate competing explanations. We then introduce a defense-like inference module based on the graph to assess the overall veracity. Finally, we prompt an LLM to generate an intuitive explanation graph. Experimental results demonstrate that G-Defense achieves state-of-the-art performance in both veracity detection and the quality of its explanations.
Innovation
Methods, ideas, or system contributions that make the work stand out.
graph-enhanced defense
explainable fake news detection
retrieval-augmented generation
claim decomposition
explanation graph
🔎 Similar Papers
No similar papers found.
