Self-Correcting RAG: Enhancing Faithfulness via MMKP Context Selection and NLI-Guided MCTS

2026-04-12

AI Summary

This work addresses the low contextual utilization and frequent hallucinations of retrieval-augmented generation (RAG) in complex reasoning tasks by proposing a novel approach that jointly models retrieval and generation as a constrained optimization and path planning problem. The key innovation lies in formalizing context selection as a multidimensional multiple-choice knapsack problem (MMKP), which maximizes information density under strict token budgets. Additionally, the method introduces natural language inference (NLI)-guided Monte Carlo tree search (MCTS) to dynamically verify the faithfulness of generated content. Experimental results demonstrate that the proposed approach significantly outperforms strong existing baselines across six multi-hop question answering and fact-checking datasets, effectively improving reasoning accuracy while reducing hallucination.

Technology Category

• Knowledge Representation and Reasoning: Computational Complexity of Reasoning
• Search and Optimization: Metareasoning and Metaheuristics
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Retrieval-augmented generation (RAG) substantially extends the knowledge boundary of large language models. However, it still faces two major challenges when handling complex reasoning tasks: low context utilization and frequent hallucinations. To address these issues, we propose Self-Correcting RAG, a unified framework that reformulates retrieval and generation as constrained optimization and path planning. On the input side, we move beyond traditional greedy retrieval and, for the first time, formalize context selection as a multi-dimensional multiple-choice knapsack problem (MMKP), thereby maximizing information density and removing redundancy under a strict token budget. On the output side, we introduce a natural language inference (NLI)-guided Monte Carlo Tree Search (MCTS) mechanism, which leverages test-time compute to dynamically explore reasoning trajectories and validate the faithfulness of generated answers. Experiments on six multi-hop question answering and fact-checking datasets demonstrate that our method significantly improves reasoning accuracy on complex queries while effectively reducing hallucinations, outperforming strong existing baselines.Our code is available at https://github.com/xjiacs/Self-Correcting-RAG .

Problem

• Retrieval-Augmented Generation
• Hallucination
• Context Utilization
• Complex Reasoning
• Faithfulness

Innovation

• MMKP
• NLI-guided MCTS
• Self-Correcting RAG
• retrieval-augmented generation
• faithfulness