aiqbench analysis
This research introduces C-MIG, a framework that enhances Retrieval-Augmented Generation (RAG) by using multi-view information gain to better supervise clinical reasoning in large language models. By moving beyond exact-match binary rewards, the method effectively captures semantically relevant but non-verbatim diagnostic steps while improving knowledge recall across diverse medical benchmarks. The approach demonstrates superior performance over existing RAG-reinforcement learning methods and general-purpose LLMs specifically for grounding AI outputs in trustworthy medical evidence.
C-MIG: Multi-view Information Gain-based Retrieval-Augmented Generation for Clinical Diagnosis Reasoning
arXiv:2605.27860v2 Announce Type: replace Abstract: Retrieval-augmented generation combined with reinforcement learning has shown promise for grounding large language models in trustworthy medical evidence. However, existing methods rely on exact-match binary rewards, which in clinical diagnosis cause two issues: (i) semantically relevant but non-verbatim steps receive zero signal, discarding valuable learning signals; and (ii) uni-dimensional rewards cannot effectively supervise heterogeneous reasoning capabilities. To address these issues, we propose C-MIG, a Multi-view Information Gain-based retrieval-augmented generation framework for Clinical diagnosis. C-MIG estimates information gain under a frozen reference model from two complementary views, retrieved-document and document-refinement, to jointly guide what to retrieve and how to refine, alleviating the issues of valuable reward signal loss and credit assignment. We further design a multi-subquery retrieval augmentation strategy that improves knowledge recall coverage in clinical diagnostic scenarios. Comprehensive experiments on four medical benchmarks demonstrate that C-MIG achieves the best performance among all RAG-RL methods on both in-domain and out-of-domain sets, and outperforms state-of-the-art general-purpose LLMs for clinical diagnosis.
Abstract:Retrieval-augmented generation combined with reinforcement learning has shown promise for grounding large language models in trustworthy medical evidence. However, existing methods rely on exact-match binary rewards, which in clinical diagnosis cause two issues: (i) semantically relevant but non-verbatim steps receive zero signal, discarding valuable learning signals; and (ii) uni-dimensional rewards cannot effectively supervise heterogeneous reasoning capabilities. To address these issues, we propose C-MIG, a Multi-view Information Gain-based retrieval-augmented generation framework for Clinical diagnosis. C-MIG estimates information gain under a frozen reference model from two complementary views, retrieved-document and document-refinement, to jointly guide what to retrieve and how to refine, alleviating the issues of valuable reward signal loss and credit assignment. We further design a multi-subquery retrieval augmentation strategy that improves knowledge recall coverage in clinical diagnostic scenarios. Comprehensive experiments on four medical benchmarks demonstrate that C-MIG achieves the best performance among all RAG-RL methods on both in-domain and out-of-domain sets, and outperforms state-of-the-art general-purpose LLMs for clinical diagnosis.
Artificial Intelligence (cs.AI)
arXiv:2605.27860 [cs.AI]
(or arXiv:2605.27860v2 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2605.27860 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 250 words
Related coverage
Introducing Instella-MoE: A State-of-the-Art Fully Open Mixture-of-Experts Language Model
AMD · Jul 24, 2026
Modernizing and scaling support operations with generative AI on AWS
AWS · Sep 2, 2026
Exploring Hierarchical Interest Representation For Meta Ads Deep Funnel Optimization
Meta Engineering · Jul 15, 2026
J++ Lens: Jacobian Filtering Enables More Faithful Workspace Lenses
LessWrong · Oct 8, 2026