Toward Vibe Medicine: A Self-Evolving Multi-Agent Framework for Clinical Decision Support

VIBEMed, a new multi-agent framework for clinical decision support, introduces a self-evolving mechanism that allows AI systems to learn dynamically from patient interactions and outcomes. It comprises three specialized agents: a Clinical Diagnostic Agent for hypothesis generation, a Therapeutic Execution Agent for treatment planning, and a Clinical Evolution Manager that distills clinical feedback into reusable knowledge. This framework shows superior performance in complex clinical cases, particularly in oncology, by enabling iterative updates to decision-making strategies, thus advancing the capabilities of AI in precision medicine.