CoTKR: Chain-of-Thought Enhanced Knowledge Rewriting for Complex Knowledge Graph Question Answering
This work aims to improve the accuracy of Knowledge Graph Question Answering for users by providing more relevant and complete knowledge to Large Language Models, which is an incremental improvement to existing RAG-based KGQA systems.
This paper addresses the issue of irrelevant or incomplete knowledge rewriting for complex Knowledge Graph Question Answering (KGQA) when using Large Language Models (LLMs) with Retrieval Augmented Generation (RAG). They propose CoTKR, a method that generates reasoning traces and knowledge interleaved, and PAQAF, a training strategy that uses feedback from the QA model to optimize the rewriter. Experiments show CoTKR significantly improves LLM performance in KGQA compared to previous rewriting methods.
Recent studies have explored the use of Large Language Models (LLMs) with Retrieval Augmented Generation (RAG) for Knowledge Graph Question Answering (KGQA). They typically require rewriting retrieved subgraphs into natural language formats comprehensible to LLMs. However, when tackling complex questions, the knowledge rewritten by existing methods may include irrelevant information, omit crucial details, or fail to align with the question's semantics. To address them, we propose a novel rewriting method CoTKR, Chain-of-Thought Enhanced Knowledge Rewriting, for generating reasoning traces and corresponding knowledge in an interleaved manner, thereby mitigating the limitations of single-step knowledge rewriting. Additionally, to bridge the preference gap between the knowledge rewriter and the question answering (QA) model, we propose a training strategy PAQAF, Preference Alignment from Question Answering Feedback, for leveraging feedback from the QA model to further optimize the knowledge rewriter. We conduct experiments using various LLMs across several KGQA benchmarks. Experimental results demonstrate that, compared with previous knowledge rewriting methods, CoTKR generates the most beneficial knowledge representation for QA models, which significantly improves the performance of LLMs in KGQA.