MedTutor-R1: Socratic Personalized Medical Teaching with Multi-Agent Simulation

2025-12-05

AI Summary

Clinical medical education faces a critical challenge: a severe shortage of expert faculty amid rapidly growing teaching demands. Existing research predominantly focuses on individual knowledge transfer, neglecting the cultivation of collaborative reasoning—the cornerstone of team-based clinical practice. To address this, we propose ClinEdu, the first multi-agent simulation system designed for one-to-many clinical instruction. Its core is MedTutor-R1, a multimodal Socratic tutor model that innovatively simulates both anthropomorphized patients and diverse student cohorts, enabling structured collaborative reasoning training. MedTutor-R1 integrates large language models, multi-agent simulation, instruction tuning, and reinforcement learning optimized against a tri-dimensional benchmark—structural fidelity, analytical quality, and clinical safety. Experiments show MedTutor-R1 achieves over 20% higher teaching evaluation scores than baselines, matches the performance of o3 models, and demonstrates robust generalization across varying cohort sizes—validating the feasibility and efficacy of population-scale Socratic pedagogy in clinical education.

Technology Category

• Multiagent Systems: Multiagent Learning
• Cognitive Modeling & Cognitive Systems: Agent Architectures
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

The significant gap between rising demands for clinical training and the scarcity of expert instruction poses a major challenge to medical education. With powerful capabilities in personalized guidance, Large Language Models (LLMs) offer a promising solution to bridge this gap. However, current research focuses mainly on one-on-one knowledge instruction, overlooking collaborative reasoning, a key skill for students developed in teamwork like ward rounds. To this end, we develop ClinEdu, a multi-agent pedagogical simulator with personality-driven patients and diverse student cohorts, enabling controlled testing of complex pedagogical processes and scalable generation of teaching data. Based on ClinEdu, we construct ClinTeach, a large Socratic teaching dialogue dataset that captures the complexities of group instruction. We then train MedTutor-R1, the first multimodal Socratic tutor designed for one-to-many instruction in clinical medical education. MedTutor-R1 is first instruction-tuned on our ClinTeach dataset and then optimized with reinforcement learning, using rewards derived from a three-axis rubric, covering structural fidelity, analytical quality, and clinical safety, to refine its adaptive Socratic strategies. For authentic in-situ assessment, we use simulation-based interactive evaluation that redeploys the tutor back into ClinEdu. Experimental results demonstrate that our MedTutor-R1 outperforms the base model by over 20% in average pedagogical score and is comparable to o3, while also exhibiting high adaptability in handling a varying number of students. This promising performance underscores the effectiveness of our pedagogical simulator, ClinEdu.

Problem

• Addresses the shortage of expert instructors for clinical training demands.
• Enhances collaborative reasoning skills beyond one-on-one knowledge instruction.
• Develops adaptive Socratic teaching strategies for group medical education.

Innovation

• Multi-agent simulator for personalized medical teaching
• Socratic dialogue dataset for group instruction training
• Reinforcement learning with three-axis rubric optimization