This research addresses the performance gap of LLMs in interactive medical diagnostics through a novel clinical experience learning approach.

• LLMs show promising results in static medical diagnosis but perform poorly in interactive settings requiring active information gathering
• The study identifies fundamental mechanisms behind diagnostic performance degradation
• Researchers developed a systematic framework to enhance AI diagnostic abilities through simulated clinical experiences
• Results demonstrate significant improvements in diagnostic accuracy and efficiency compared to baseline models

This innovation matters because it enables more reliable AI-assisted medical diagnostics in real-world healthcare settings, potentially enhancing patient care while reducing physician workload.

Improving Interactive Diagnostic Ability of a Large Language Model Agent Through Clinical Experience Learning