ReAgent: Reversible Multi-Agent Reasoning for Knowledge-Enhanced Multi-Hop QA

2025-03-10

AI Summary

In multi-hop question answering, conventional chain-of-thought (CoT) reasoning is inherently irreversible, leading to error accumulation and undermining both robustness and interpretability. To address this, we propose the first explicitly reversible multi-agent collaborative framework: it introduces a backtrackable reasoning control flow that enables dynamic error detection, backward error localization, and on-the-fly path correction; integrates text retrieval, information aggregation, cross-validation, and knowledge enhancement to support real-time verification and iterative refinement of reasoning steps. Evaluated on three major benchmarks, our framework achieves an average improvement of approximately 6% over strong baselines. Our core contribution is the first explicit incorporation of reversibility into multi-hop reasoning—establishing a novel QA paradigm that simultaneously ensures fault tolerance and full process interpretability.

Technology Category

• Knowledge Representation and Reasoning: Computational Complexity of Reasoning
• Multiagent Systems: Multiagent Planning
• Reasoning under Uncertainty: Other Foundations of Reasoning under Uncertainty

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Recent advances in large language models (LLMs) have significantly improved multi-hop question answering (QA) through direct Chain-of-Thought (CoT) reasoning. However, the irreversible nature of CoT leads to error accumulation, making it challenging to correct mistakes in multi-hop reasoning. This paper introduces ReAgent: a Reversible multi-Agent collaborative framework augmented with explicit backtracking mechanisms, enabling reversible multi-hop reasoning. By incorporating text-based retrieval, information aggregation and validation, our system can detect and correct errors mid-reasoning, leading to more robust and interpretable QA outcomes. The framework and experiments serve as a foundation for future work on error-tolerant QA systems. Empirical evaluations across three benchmarks indicate ReAgent's efficacy, yielding average about 6% improvements against baseline models.

Problem

• Addresses error accumulation in multi-hop QA reasoning
• Introduces reversible reasoning with backtracking mechanisms
• Enhances QA robustness and interpretability through error correction

Innovation

• Reversible multi-Agent collaborative reasoning framework
• Explicit backtracking mechanisms for error correction
• Text-based retrieval and information aggregation validation