RAG Combined with Instruction Tuning for Traditional Chinese Medicine Syndrome Differentiation Thinking

Chunliang Chen; Ming Guan; Wenjing Yue; Xinyu Wang; Yuanbin Wu; Xiaoling Wang

doi:10.1007/978-981-96-3752-2_6

RAG Combined with Instruction Tuning for Traditional Chinese Medicine Syndrome Differentiation Thinking

Chunliang Chen
, Ming Guan
, Wenjing Yue
, Xinyu Wang
, Yuanbin Wu
, Xiaoling Wang^*

^*Corresponding author for this work

School of Computer Science and Technology

East China Normal University

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

The rapid advancements of large language models (LLMs) have opened new avenues for data processing and knowledge extraction, particularly in the medical domain. This paper investigates the application of LLMs in Traditional Chinese Medicine (TCM), with a focus on enhancing the models’ capabilities in syndrome differentiation thinking tasks. We propose a method that delineates the syndrome differentiation process in TCM into four critical steps: clinical information extraction, pathogenesis inference, syndrome inference, and explanatory summarization, with tailored prompting strategies designed for each step. By integrating Retrieval-Augmented Generation (RAG) with instruction tuning, we generated 800 instruction data entries rich in localized knowledge and instruction tuning of a pre-trained model. Experimental results indicate that our approach significantly improves the models’ performance in TCM syndrome differentiation thinking, achieving top rankings in both the A and B leaderboards, with scores of 45.12 and 44.37, respectively.

Original language	English
Title of host publication	Health Information Processing - 10th China Health Information Processing Conference, CHIP 2024, Proceedings
Editors	Yanchun Zhang, Qingcai Chen, Hongfei Lin, Lei Liu, Xiangwen Liao, Buzhou Tang, Tianyong Hao, Zhengxing Huang
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	78-89
Number of pages	12
ISBN (Print)	9789819637515
DOIs	https://doi.org/10.1007/978-981-96-3752-2_6
State	Published - 2025
Event	10th China Health Information Processing Conference, CHIP 2024 - Fuzhou, China Duration: 15 Nov 2024 → 17 Nov 2024

Publication series

Name	Communications in Computer and Information Science
Volume	2433 CCIS
ISSN (Print)	1865-0929
ISSN (Electronic)	1865-0937

Conference

Conference	10th China Health Information Processing Conference, CHIP 2024
Country/Territory	China
City	Fuzhou
Period	15/11/24 → 17/11/24

Keywords

Instruction Tuning
RAG
TCM LLM

Access to Document

10.1007/978-981-96-3752-2_6

Cite this

Chen, C., Guan, M., Yue, W., Wang, X., Wu, Y., & Wang, X. (2025). RAG Combined with Instruction Tuning for Traditional Chinese Medicine Syndrome Differentiation Thinking. In Y. Zhang, Q. Chen, H. Lin, L. Liu, X. Liao, B. Tang, T. Hao, & Z. Huang (Eds.), Health Information Processing - 10th China Health Information Processing Conference, CHIP 2024, Proceedings (pp. 78-89). (Communications in Computer and Information Science; Vol. 2433 CCIS). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-981-96-3752-2_6

Chen, Chunliang ; Guan, Ming ; Yue, Wenjing et al. / RAG Combined with Instruction Tuning for Traditional Chinese Medicine Syndrome Differentiation Thinking. Health Information Processing - 10th China Health Information Processing Conference, CHIP 2024, Proceedings. editor / Yanchun Zhang ; Qingcai Chen ; Hongfei Lin ; Lei Liu ; Xiangwen Liao ; Buzhou Tang ; Tianyong Hao ; Zhengxing Huang. Springer Science and Business Media Deutschland GmbH, 2025. pp. 78-89 (Communications in Computer and Information Science).

@inproceedings{785f52317123485d899aba9d4d79ba64,

title = "RAG Combined with Instruction Tuning for Traditional Chinese Medicine Syndrome Differentiation Thinking",

abstract = "The rapid advancements of large language models (LLMs) have opened new avenues for data processing and knowledge extraction, particularly in the medical domain. This paper investigates the application of LLMs in Traditional Chinese Medicine (TCM), with a focus on enhancing the models{\textquoteright} capabilities in syndrome differentiation thinking tasks. We propose a method that delineates the syndrome differentiation process in TCM into four critical steps: clinical information extraction, pathogenesis inference, syndrome inference, and explanatory summarization, with tailored prompting strategies designed for each step. By integrating Retrieval-Augmented Generation (RAG) with instruction tuning, we generated 800 instruction data entries rich in localized knowledge and instruction tuning of a pre-trained model. Experimental results indicate that our approach significantly improves the models{\textquoteright} performance in TCM syndrome differentiation thinking, achieving top rankings in both the A and B leaderboards, with scores of 45.12 and 44.37, respectively.",

keywords = "Instruction Tuning, RAG, TCM LLM",

author = "Chunliang Chen and Ming Guan and Wenjing Yue and Xinyu Wang and Yuanbin Wu and Xiaoling Wang",

note = "Publisher Copyright: {\textcopyright} The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.; 10th China Health Information Processing Conference, CHIP 2024 ; Conference date: 15-11-2024 Through 17-11-2024",

year = "2025",

doi = "10.1007/978-981-96-3752-2\_6",

language = "英语",

isbn = "9789819637515",

series = "Communications in Computer and Information Science",

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "78--89",

editor = "Yanchun Zhang and Qingcai Chen and Hongfei Lin and Lei Liu and Xiangwen Liao and Buzhou Tang and Tianyong Hao and Zhengxing Huang",

booktitle = "Health Information Processing - 10th China Health Information Processing Conference, CHIP 2024, Proceedings",

address = "德国",

}

Chen, C, Guan, M, Yue, W, Wang, X, Wu, Y & Wang, X 2025, RAG Combined with Instruction Tuning for Traditional Chinese Medicine Syndrome Differentiation Thinking. in Y Zhang, Q Chen, H Lin, L Liu, X Liao, B Tang, T Hao & Z Huang (eds), Health Information Processing - 10th China Health Information Processing Conference, CHIP 2024, Proceedings. Communications in Computer and Information Science, vol. 2433 CCIS, Springer Science and Business Media Deutschland GmbH, pp. 78-89, 10th China Health Information Processing Conference, CHIP 2024, Fuzhou, China, 15/11/24. https://doi.org/10.1007/978-981-96-3752-2_6

RAG Combined with Instruction Tuning for Traditional Chinese Medicine Syndrome Differentiation Thinking. / Chen, Chunliang; Guan, Ming; Yue, Wenjing et al.
Health Information Processing - 10th China Health Information Processing Conference, CHIP 2024, Proceedings. ed. / Yanchun Zhang; Qingcai Chen; Hongfei Lin; Lei Liu; Xiangwen Liao; Buzhou Tang; Tianyong Hao; Zhengxing Huang. Springer Science and Business Media Deutschland GmbH, 2025. p. 78-89 (Communications in Computer and Information Science; Vol. 2433 CCIS).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - RAG Combined with Instruction Tuning for Traditional Chinese Medicine Syndrome Differentiation Thinking

AU - Chen, Chunliang

AU - Guan, Ming

AU - Yue, Wenjing

AU - Wang, Xinyu

AU - Wu, Yuanbin

AU - Wang, Xiaoling

N1 - Publisher Copyright: © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

PY - 2025

Y1 - 2025

N2 - The rapid advancements of large language models (LLMs) have opened new avenues for data processing and knowledge extraction, particularly in the medical domain. This paper investigates the application of LLMs in Traditional Chinese Medicine (TCM), with a focus on enhancing the models’ capabilities in syndrome differentiation thinking tasks. We propose a method that delineates the syndrome differentiation process in TCM into four critical steps: clinical information extraction, pathogenesis inference, syndrome inference, and explanatory summarization, with tailored prompting strategies designed for each step. By integrating Retrieval-Augmented Generation (RAG) with instruction tuning, we generated 800 instruction data entries rich in localized knowledge and instruction tuning of a pre-trained model. Experimental results indicate that our approach significantly improves the models’ performance in TCM syndrome differentiation thinking, achieving top rankings in both the A and B leaderboards, with scores of 45.12 and 44.37, respectively.

AB - The rapid advancements of large language models (LLMs) have opened new avenues for data processing and knowledge extraction, particularly in the medical domain. This paper investigates the application of LLMs in Traditional Chinese Medicine (TCM), with a focus on enhancing the models’ capabilities in syndrome differentiation thinking tasks. We propose a method that delineates the syndrome differentiation process in TCM into four critical steps: clinical information extraction, pathogenesis inference, syndrome inference, and explanatory summarization, with tailored prompting strategies designed for each step. By integrating Retrieval-Augmented Generation (RAG) with instruction tuning, we generated 800 instruction data entries rich in localized knowledge and instruction tuning of a pre-trained model. Experimental results indicate that our approach significantly improves the models’ performance in TCM syndrome differentiation thinking, achieving top rankings in both the A and B leaderboards, with scores of 45.12 and 44.37, respectively.

KW - Instruction Tuning

KW - RAG

KW - TCM LLM

UR - https://www.scopus.com/pages/publications/105003861870

U2 - 10.1007/978-981-96-3752-2_6

DO - 10.1007/978-981-96-3752-2_6

M3 - 会议稿件

AN - SCOPUS:105003861870

SN - 9789819637515

T3 - Communications in Computer and Information Science

SP - 78

EP - 89

BT - Health Information Processing - 10th China Health Information Processing Conference, CHIP 2024, Proceedings

A2 - Zhang, Yanchun

A2 - Chen, Qingcai

A2 - Lin, Hongfei

A2 - Liu, Lei

A2 - Liao, Xiangwen

A2 - Tang, Buzhou

A2 - Hao, Tianyong

A2 - Huang, Zhengxing

PB - Springer Science and Business Media Deutschland GmbH

T2 - 10th China Health Information Processing Conference, CHIP 2024

Y2 - 15 November 2024 through 17 November 2024

ER -

Chen C, Guan M, Yue W, Wang X, Wu Y, Wang X. RAG Combined with Instruction Tuning for Traditional Chinese Medicine Syndrome Differentiation Thinking. In Zhang Y, Chen Q, Lin H, Liu L, Liao X, Tang B, Hao T, Huang Z, editors, Health Information Processing - 10th China Health Information Processing Conference, CHIP 2024, Proceedings. Springer Science and Business Media Deutschland GmbH. 2025. p. 78-89. (Communications in Computer and Information Science). doi: 10.1007/978-981-96-3752-2_6