EvoTrustRAG: Evolution-Aware Conflict Attribution and Evidence Handling for Reliable Retrieval-Augmented Generation

2026-08-08

AI Summary

This work addresses the challenge of reliably handling conflicting evidence in retrieval-augmented generation (RAG) caused by knowledge evolution, adversarial manipulation, or inherent uncertainty. The authors propose a training-free, pre-inference conflict resolution framework that formulates conflict attribution as an interpretable provenance identification problem. By integrating an evolution-aware mechanism to dynamically discern the nature of conflicts, the approach combines passage-based factual graph construction, temporal relation analysis, support structure evaluation, and consistency verification to project local judgments into globally coherent explanations. Experimental results demonstrate that the framework achieves an average accuracy of 81.4% on benchmark datasets, improves attribution macro F1 to 79.1% (+6.9%), and significantly reduces error rates from 31.2% to 16.0% under the strongest coordinated attacks.

Abstract

Retrieval-Augmented Generation (RAG) improves the factuality of large language models with external knowledge, yet conflicting evidence remains a fundamental challenge in dynamic and adversarial environments. Existing approaches often treat conflicts as static inconsistencies and select more reliable knowledge, overlooking that the same conflict may arise from legitimate knowledge evolution, malicious manipulation, or unresolved uncertainty. We formulate conflict origin attribution as a new problem in RAG: identifying which explanation of conflicting evidence is supported by observable context rather than simply which fact should be trusted. We propose EvoTrustRAG, a training-free framework for evolution-aware conflict attribution and evidence handling before answer generation. EvoTrustRAG represents span-grounded retrieved facts as a conflict evidence graph, evaluates grounded evolution and directional intervention hypotheses using temporal relations, support structure, and auxiliary consistency, and projects local decisions onto a globally consistent explanation of each conflict group. The attribution determines whether earlier and later states are preserved as temporal knowledge, an intervention candidate is separated from the primary context, or an unresolved conflict remains visible to the generator. Unlike provenance-based approaches focused on post-hoc analysis, EvoTrustRAG determines during inference whether conflicting evidence follows plausible knowledge evolution, exhibits intervention-like support, or cannot be reliably attributed. Experiments show that EvoTrustRAG achieves 81.4% average accuracy on benchmark-native conflict settings, improves attribution macro-F1 from 72.2% to 79.1% over the strongest baseline, and reduces the error rate under the strongest coordinated attack from 31.2% to 16.0%.

Problem

• Retrieval-Augmented Generation
• conflict attribution
• knowledge evolution
• evidence handling
• factuality

Innovation

• conflict attribution
• knowledge evolution
• retrieval-augmented generation
• evidence graph
• temporal reasoning