Speaker: 王曦廷

Title: Research on TCM Diagnosis Model with LLM based Multi-Agent Approach

Language: Chinese

Time & Venue: 2024.11.30 19:00-20:00 思源楼415

Abstract: In clinical practice, artificial intelligence (AI) based models exhibit high accuracy in specific tasks, such as tongue image recognition and the differentiation and classification of coronary heart disease. However, these models still face issues with low accuracy and poor robustness in their generalization ability when dealing with unknown data distributions. Moreover, in complex real-world scenarios, it is challenging to standardize the inputs and outputs of small models, leading to the problem of "model islands". Therefore, building intelligent systems for complex scenarios based on small models is a key issue in the current intelligentization of TCM. This report will focus on how to construct an integrated TCM diagnosis and treatment model for real-world medical research, based on large language model multi-agent methods.