Medical multi-agent system that learns from experience improves diagnosis

EMR: Self-Evolving Medical Multi-Agent System via Experience Mining and Reuse

Computation and LanguageArtificial Intelligence

Summary

Medical diagnosis is complicated, and computer programs that try to help often rely on fixed methods that don’t improve with time. The authors developed EMR, a system where different AI agents work together like doctors in different specialties consulting each other. EMR remembers past successes and mistakes by organizing knowledge in a layered way and uses these experiences to get better over time. Tests show EMR beats previous systems at medical reasoning and can adapt to new situations and different AI models.

What this means in practice

Authors

Dongsheng Shi, Yue Li, Xin Yi, Linlin Wang

Abstract

Large language model (LLM) driven multi-agent systems have shown promise in complex clinical reasoning, yet existing approaches rely on static strategies and lack persistent clinical memory, preventing self-evolving from prior diagnostic successes and failures. We present EMR, a self-evolving medical multi-agent system via Experience Mining and Reuse. EMR introduces a hierarchical clinical experience library that organizes accumulated knowledge into three levels: clinical principles, diagnostic patterns, and representative cases. During inference, EMR emulates multidisciplinary consultation: a planner agent coordinates domain-specific department agents for specialized reasoning, while a summary agent synthesizes their analyses into a final decision. Critically, EMR automatically extracts correct diagnostic insights and failure-related warnings from multi-agent reasoning trajectories, incrementally updating the experience library to guide future cases. Experiments on medical reasoning benchmarks demonstrate that EMR consistently outperforms state-of-the-art medical multi-agent baselines. Further analysis reveals that the hierarchical experience enables cross-specialty generalization and transfer across diverse LLM backbones, offering a scalable and in