DeLIVeR: Decomposed Learning for Information-grounded Veracity Recognition via Reinforced Knowledge Graph Exploration

2026-07-20

AI Summary

This work addresses the inaccuracy in evidence retrieval for automated fact-checking caused by the query fragility of traditional retrieval systems when applied to large language models. To overcome this limitation, the authors propose a reinforcement learning–based evidence retrieval framework that leverages a planner large language model to decompose complex claims into targeted sub-questions, enabling multi-hop exploration over a knowledge graph. The retrieval policy is optimized through a reward mechanism jointly driven by structural diversity and verification accuracy. Using the Group Relative Policy Optimization (GRPO) algorithm, the framework produces auditable and transparent reasoning paths. Evaluated on the LIAR, FEVER, and PolitiFact datasets, the method achieves F1 scores of 83.73, 84.57, and 79.70, respectively—outperforming the current state-of-the-art HippoRAG2 by 10–15%.

Abstract

Automated fact-checking remains a challenge for Large Language Models (LLMs) due to "query brittleness" in traditional retrieval systems. We propose DeLIVeR (Decomposed Learning for Information-grounded Veracity Recognition), a framework that treats evidence retrieval as a reinforced strategic exploration task. DeLIVeR utilizes a Planner LLM to decompose complex claims into targeted question sets, which are used to traverse structured Knowledge Graphs (KGs) for high-precision evidence. We optimize the Planner's policy using Group Relative Policy Optimization (GRPO) with a reward system prioritizing structural diversity and verdict accuracy. Our evaluation on LIAR, FEVER, and PolitiFact shows that DeLIVeR significantly outperforms state-of-the-art baselines. Using Qwen2.5-7B, our framework achieved peak F1-scores of 83.73, 84.57, and 79.70 respectively, representing a 10-15% improvement over HippoRAG2. By shifting to a reinforced question-planning strategy, DeLIVeR effectively bridges multi-hop reasoning gaps and provides an auditable, transparent path for verifiable misinformation detection.

Problem

• fact-checking
• query brittleness
• evidence retrieval
• knowledge graphs
• veracity recognition

Innovation

• reinforced knowledge graph exploration
• decomposed question planning
• information-grounded veracity recognition
• Group Relative Policy Optimization (GRPO)
• multi-hop reasoning