Exploring the Role of Knowledge Graph-Based RAG in Japanese Medical Question Answering with Small-Scale LLMs
This addresses the challenge of applying retrieval-augmented generation in Japanese medical QA for clinical settings, but it is incremental as it builds on existing RAG methods with a focus on a specific domain and language.
The study tackled the problem of limited effectiveness of large language models in Japanese medical question answering due to privacy constraints, by exploring a knowledge graph-based retrieval-augmented generation framework with small-scale open-source LLMs, and found that it had only a limited impact, with effectiveness sensitive to the quality and relevance of retrieved content.
Large language models (LLMs) perform well in medical QA, but their effectiveness in Japanese contexts is limited due to privacy constraints that prevent the use of commercial models like GPT-4 in clinical settings. As a result, recent efforts focus on instruction-tuning open-source LLMs, though the potential of combining them with retrieval-augmented generation (RAG) remains underexplored. To bridge this gap, we are the first to explore a knowledge graph-based (KG) RAG framework for Japanese medical QA small-scale open-source LLMs. Experimental results show that KG-based RAG has only a limited impact on Japanese medical QA using small-scale open-source LLMs. Further case studies reveal that the effectiveness of the RAG is sensitive to the quality and relevance of the external retrieved content. These findings offer valuable insights into the challenges and potential of applying RAG in Japanese medical QA, while also serving as a reference for other low-resource languages.