This research introduces an innovative retrieval-augmented generation (RAG) framework specifically designed for Traditional Chinese Medicine question answering systems.

• Implements a tree-organized self-reflective retrieval structure to improve knowledge access
• Addresses the critical gap in efficient retrieval frameworks for TCM applications
• Supports both auxiliary diagnosis and medical education use cases
• Evaluated using medical licensing examination datasets with attention to medical safety

This advancement matters because it creates more reliable AI-powered medical support systems that respect the unique knowledge structure of Traditional Chinese Medicine, potentially improving healthcare delivery and practitioner training.

Improving TCM Question Answering through Tree-Organized Self-Reflective Retrieval with LLMs