Reinforcement Learning for Evidence-Seeking Diagnostic Reasoning with Large Language Models

Researchers formalized medical diagnosis as an iterative evidence-seeking task and used reinforcement learning with verifiable rewards to train large language models to autonomously gather clinical evidence. Their framework, including the RAGES examination simulator, enabled models to match or exceed larger baselines in diagnostic reasoning. The work marks a shift from passive LLM responders to proactive diagnostic assistants.

arXiv:2607.02983v1 Announce Type: new Abstract: Recent reasoning-centric Large Language Models LLMs have made significant strides, yet they predominantly operate on a passive-inference pattern that assumes complete information. In contrast, real-world clinical intelligence is inherently an iterative investigative process requiring strategic evidence acquisition. To bridge this gap, we formalize medical diagnosis as an Iterative Evidence-Seeking Task. We leverage Reinforcement Learning with Verifiable Rewards RLVR to elicit intrinsic reasoning within a closed-loop environment, guided by a novel suite of rewards that enforce diagnostic precision and examination consistency. To facilitate this, we introduce the Retrieval-Augmented Generation-based Examination Simulator RAGES , a high-fidelity clinical oracle that provides realistic, knowledge-grounded follow-up evidence. Empirical results across diverse datasets demonstrate that our framework enables LLMs to transition from passive responders to autonomous assistants. Notably, our model demonstrates comparable performance to larger and reasoning-enhanced baselines, while RAGES proves superior to vanilla LLMs in generating biologically plausible clinical feedback.