ArbGraph: Conflict-Aware Evidence Arbitration for Reliable Long-Form Retrieval-Augmented Generation

2026-04-20

AI Summary

This work addresses the challenge of factual inconsistency in long-context retrieval-augmented generation (RAG), often caused by noisy or contradictory retrieved evidence. The authors propose ArbGraph, a novel framework that decouples evidence verification from generation for the first time. ArbGraph constructs a conflict-aware evidence graph by extracting atomic claims and introduces a strength-driven iterative arbitration mechanism to explicitly model supportive and contradictory relationships among evidence pieces at the evidence level. By propagating credibility signals through this graph, the framework effectively resolves factual conflicts. Evaluated on the LongFact and RAGChecker benchmarks, ArbGraph significantly improves factual recall and information density while reducing hallucination and sensitivity to retrieval noise, demonstrating particularly strong performance in scenarios involving conflicting or ambiguous evidence.

Technology Category

• Knowledge Representation and Reasoning: Argumentation
• Reasoning under Uncertainty: Other Foundations of Reasoning under Uncertainty
• Natural Language Processing: Fact-Checking / Misinformation Detection (NLP Focus)

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Retrieval-augmented generation (RAG) remains unreliable in long-form settings, where retrieved evidence is noisy or contradictory, making it difficult for RAG pipelines to maintain factual consistency. Existing approaches focus on retrieval expansion or verification during generation, leaving conflict resolution entangled with generation. To address this limitation, we propose ArbGraph, a framework for pre-generation evidence arbitration in long-form RAG that explicitly resolves factual conflicts. ArbGraph decomposes retrieved documents into atomic claims and organizes them into a conflict-aware evidence graph with explicit support and contradiction relations. On top of this graph, we introduce an intensity-driven iterative arbitration mechanism that propagates credibility signals through evidence interactions, enabling the system to suppress unreliable and inconsistent claims before final generation. In this way, ArbGraph separates evidence validation from text generation and provides a coherent evidence foundation for downstream long-form generation. We evaluate ArbGraph on two widely used long-form RAG benchmarks, LongFact and RAGChecker, using multiple large language model backbones. Experimental results show that ArbGraph consistently improves factual recall and information density while reducing hallucinations and sensitivity to retrieval noise. Additional analyses show that these gains are evident under conflicting or ambiguous evidence, highlighting the effectiveness of evidence-level conflict resolution for improving the reliability of long-form RAG. The implementation is publicly available at https://github.com/1212Judy/ArbGraph.

Problem

• Retrieval-Augmented Generation
• factual consistency
• evidence conflict
• long-form generation
• hallucination

Innovation

• evidence arbitration
• conflict-aware graph
• retrieval-augmented generation
• factual consistency
• pre-generation validation