This research introduces a novel graph-based fact verification framework that addresses key limitations in LLM-based fact-checking systems.

• Solves the insufficient decomposition problem by using triplet-based claim graphs
• Reduces mention ambiguity that often leads to incorrect verification
• Improves fact-checking precision through structured representation of claims
• Enhances security by providing more reliable misinformation detection

This framework has significant implications for information security, helping organizations combat the spread of false information and protect data integrity in an increasingly complex information landscape.

A Graph-based Verification Framework for Fact-Checking