Distill and Align Decomposition for Enhanced Claim Verification

Research Summary

Contribution Summary

• Complex claim verification requires decomposing sentences into verifiable subclaims, yet existing methods struggle to align decomposition quality with verification performance.
• We propose a reinforcement learning (RL) approach that jointly optimizes decomposition quality and verifier alignment using Group Relative Policy Optimization (GRPO).
• Our method integrates: (i) structured sequential reasoning; (ii) supervised finetuning on teacher-distilled exemplars; and (iii) a multi-objective reward balancing format compliance, verifier alignment, and decomposition quality.

Why It Matters For Eval

• Across six evaluation settings, our trained 8B decomposer improves downstream verification performance to (71.75%) macro-F1, outperforming prompt-based approaches ((+1.99), (+6.24)) and existing RL methods ((+5.84)).
• Human evaluation confirms the high quality of the generated subclaims.

Related Papers

• TAROT: Test-driven and Capability-adaptive Curriculum Reinforcement Fine-tuning for Code Generation with Large Language Models Embedding Neighbor
• Explicit Grammar Semantic Feature Fusion for Robust Text Classification Embedding Neighbor
• PVminer: A Domain-Specific Tool to Detect the Patient Voice in Patient Generated Data Embedding Neighbor
• ViTextVQA: A Large-Scale Visual Question Answering Dataset and a Novel Multimodal Feature Fusion Method for Vietnamese Text Comprehension in Images Embedding Neighbor
• PersonalAI: A Systematic Comparison of Knowledge Graph Storage and Retrieval Approaches for Personalized LLM agents Embedding Neighbor
• Vichara: Appellate Judgment Prediction and Explanation for the Indian Judicial System Embedding Neighbor
• STAR: Similarity-guided Teacher-Assisted Refinement for Super-Tiny Function Calling Models Embedding Neighbor
• Cold-Start Personalization via Training-Free Priors from Structured World Models Embedding Neighbor
• The ASIR Courage Model: A Phase-Dynamic Framework for Truth Transitions in Human and AI Systems Embedding Neighbor
• Claim Automation using Large Language Model Tag Overlap
• Peeking inside the Black-Box: Reinforcement Learning for Explainable and Accurate Relation Extraction Tag Overlap
• CARE: An Explainable Computational Framework for Assessing Client-Perceived Therapeutic Alliance Using Large Language Models Tag Overlap