让大模型推理与知识图谱对齐，解决问答中的幻觉问题
Reason-Align-Respond: Aligning LLM Reasoning with Knowledge Graphs for KGQA
- 三阶段框架：生成推理链、映射知识图谱路径、合成答案
- 在WebQSP和CWQ上达到93.3%和91.0%的Hit@1精度
- 推理过程可解释，且零样本泛化能力强，适合需要可信推理场景
大语言模型在复杂推理任务中表现优异，但常出现幻觉且缺乏可靠的事实依据。知识图谱虽提供结构化事实知识，却缺乏灵活的推理能力。本文提出一种名为Reason-Align-Respond（RAR）的新框架，系统融合大模型推理与知识图谱以实现知识图谱问答（KGQA）。该方法包含三个核心组件：生成类人推理链的Reasoner、将推理链映射到有效知识图谱路径的Aligner，以及合成最终答案的Responser。我们将其建模为概率模型，并使用期望最大化算法进行优化，迭代精炼推理链与知识路径。大量实验表明，RAR在多个基准测试上取得领先性能，分别在WebQSP和CWQ上达到93.3%和91.0%的Hit@1得分。人工评估证实，RAR生成的推理链质量高、可解释性强，且与知识图谱路径高度对齐。此外，RAR展现出强大的零样本泛化能力，推理过程保持高效。
原文摘要 · Abstract (English)
LLMs have demonstrated remarkable capabilities in complex reasoning tasks, yet they often suffer from hallucinations and lack reliable factual grounding. Meanwhile, knowledge graphs (KGs) provide structured factual knowledge but lack the flexible reasoning abilities of LLMs. In this paper, we present Reason-Align-Respond (RAR), a novel framework that systematically integrates LLM reasoning with knowledge graphs for KGQA. Our approach consists of three key components: a Reasoner that generates human-like reasoning chains, an Aligner that maps these chains to valid KG paths, and a Responser that synthesizes the final answer. We formulate this process as a probabilistic model and optimize it using the Expectation-Maximization algorithm, which iteratively refines the reasoning chains and knowledge paths. Extensive experiments on multiple benchmarks demonstrate the effectiveness of RAR, achieving state-of-the-art performance with Hit@1 scores of 93.3% and 91.0% on WebQSP and CWQ respectively. Human evaluation confirms that RAR generates high-quality, interpretable reasoning chains well-aligned with KG paths. Furthermore, RAR exhibits strong zero-shot generalization capabilities and maintains computational efficiency during inference.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。