Preprint

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

Shengyi Hua Kangzhe Hu Conghui He Xiaofan Zhang Shaoting Zhang

Jul 2026 · 1 citation · 57 references

Computer Science

TL;DR

Empirical results demonstrate that the Reinforcement Learning with Verifiable Rewards framework enables LLMs to transition from passive responders to autonomous assistants, and demonstrates comparable performance to larger and reasoning-enhanced baselines, while RAGES proves superior to vanilla LLMs in generating biologically plausible clinical feedback.

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.

arXiv Semantic Scholar

View source

Similar papers

Dissecting Reinforcement Learning: Mechanisms Behind Compositional Reasoning in LLMs

This thesis proposes a unified two-axis framework that organizes SFT and RL methods along a data axis (off-policy to on-policy) and a loss function axis (positive-only to positive-plus-negative to GRPO) and enables controlled ablations of individual components.

G. Kim, Chair Chenyan Xiong, Aditi Raghunathan · 0 citations

Book Open access Jul 2026

R3Check: Reinforcement Learning for Iterative Retrieval and Structured Reasoning in Complex Fact Checking

R3Check is proposed, a rule-guided reinforcement learning framework that enables LLMs to perform iterative retrieval–reasoning for multi-hop fact-checking under weak supervision, and demonstrates the potential of pure reinforcement learning to induce effective retrieval–reasoning behaviors for fact-checking under weak...

Peng Qi, Yuyang Zhao, W. Hsu et al. · Annual International ACM SIG... · 0 citations

Jul 2026

Learning Latent Reasoning Traces for Scalar Reward Models End-to-End

LatentRM is a reward modeling framework that learns intermediate reasoning traces as discrete latent variables to explicitly maximize the likelihood of downstream scalar rewards through on-policy optimization of the latent reasoning space end-to-end.

Sanwoo Lee, Clive Bai, Hsiu-Yuan Huang et al. · arXiv.org · 0 citations

#artificial intelligence Preprint Aug 2026

Co-RL: Unsupervised Reasoning Emerges from Diverse Cohort in Multi-agent RL

This work introduces Co-RL, a framework in which multiple decoupled models, sharing no parameters, are simultaneously optimized through RL using rewards derived from their peers, and shows that unsupervised reasoning can emerge through cooperative multi-agent training.

Yunhao Yang, Yuexin Bian, Yun-Jie Tian et al. · 1 citation

Preprint Aug 2026

MedCalc-R1: Knowledge-Guided Reward Framework for Medical Mathematical Reasoning

Experimental results demonstrate that the proposed knowledge-guided hybrid reward framework significantly outperforms existing baselines in both reasoning accuracy and generalization capability, validating the effectiveness and applicability in safety-critical domains.

Hao-Tian Wang, Lian Yan, Xingzhi Yao et al. · 2 citations

Preprint Aug 2026

BODHI: Do LLMs Branch Out and Discover Heterogeneous Inferences?

The nature of test-time exploration in RLVR-trained LLMs is investigated by employing controlled maze-solving experiments and extracting a tree structure from mathematical reasoning traces (BODHI-Trees) based on semantic equivalence to delineate between entropy arising from stylistic variations and genuine inferential...

Soumadeep Saha, Krish Sharma, Akshay Chaturvedi et al. · 1 citation