From Given to Gathered Evidence: Agentic Learning for Longitudinal Medical Reasoning

Hugging Face Daily Papers Papers
medical-ai agentic-learning vision-language-models reinforcement-learning clinical-reasoning benchmark uk-biobank

Summary

该论文提出CASE——一系列用于证据搜集的临床智能体，结合工具使用框架与智能体后训练方法，在基于UK Biobank构建的纵向多模态医学基准上进行评估，其基于Qwen3-VL-8B的智能体在匹配评估条件下答案准确率相对GPT-5.4和Claude Opus 4.8分别提升超16%和10%。

Foundation models can serve as clinical agents through tool-use harnesses. However, conventional medical benchmarks assess reasoning over preselected evidence rather than the ability to seek it across clinical records and longitudinal imaging. We propose CASE: a series of role-specific Clinical Agents for Seeking Evidence, together with a tool-use harness and an agentic post-training framework for compact vision-language policy models. We further introduce a longitudinal multimodal benchmark built on UK Biobank, comprising 50,401 clinical questions derived from real-world ICD-10-coded diagnoses of 4,739 participants. Each question links to a patient-specific environment containing clinical context and multi-sequence MRI from baseline and follow-up visits, where agents autonomously select which visits, organs, modalities, slices, and specialist tools to inspect and compare. Supervised fine-tuning transfers evidence-seeking workflows from 14,734 frontier-model interaction trajectories, followed by agentic reinforcement learning on the learner's own environment interactions. Privileged on-policy self-distillation and rubric-based LLM feedback refine evidence-to-conclusion reasoning without prescribing tool sequences. Experiments show that CASE moves beyond question-answer imitation toward transferable investigation policies, strengthening evidence-grounded longitudinal reasoning. Under matched evaluation conditions, our Qwen3-VL-8B based agent achieves over 16% and 10% relative improvements in answer accuracy over GPT-5.4 and Claude Opus 4.8. Code will be available at https://github.com/VinyehShaw/CASE.
Original Article
View Cached Full Text

Cached at: 10/12/26, 06:54 AM

Paper page - From Given to Gathered Evidence: Agentic Learning for Longitudinal Medical Reasoning

Source: https://huggingface.co/papers/2609.39566

Abstract

Foundationmodelscanserveasclinicalagentsthroughtool-useharnesses.However,conventionalmedicalbenchmarksassessreasoningoverpreselectedevidenceratherthantheabilitytoseekitacrossclinicalrecordsandlongitudinalimaging.WeproposeCASE:aseriesofrole-specificClinicalAgentsforSeekingEvidence,togetherwithatool-useharnessandanagenticpost-trainingframeworkforcompactvision-languagepolicymodels.WefurtherintroducealongitudinalmultimodalbenchmarkbuiltonUKBiobank,comprising50,401clinicalquestionsderivedfromreal-worldICD-10-codeddiagnosesof4,739participants.Eachquestionlinkstoapatient-specificenvironmentcontainingclinicalcontextandmulti-sequenceMRIfrombaselineandfollow-upvisits,whereagentsautonomouslyselectwhichvisits,organs,modalities,slices,andspecialisttoolstoinspectandcompare.Supervisedfine-tuningtransfersevidence-seekingworkflowsfrom14,734frontier-modelinteractiontrajectories,followedbyagenticreinforcementlearningonthelearner’sownenvironmentinteractions.Privilegedon-policyself-distillationandrubric-basedLLMfeedbackrefineevidence-to-conclusionreasoningwithoutprescribingtoolsequences.ExperimentsshowthatCASEmovesbeyondquestion-answerimitationtowardtransferableinvestigationpolicies,strengtheningevidence-groundedlongitudinalreasoning.Undermatchedevaluationconditions,ourQwen3-VL-8Bbasedagentachievesover16%and10%relativeimprovementsinansweraccuracyoverGPT-5.4andClaudeOpus4.8.Codewillbeavailableathttps://github.com/VinyehShaw/CASE.

View arXiv pageView PDFGitHub0Add to collection

Get this paper in your agent:

hf papers read 2609\.39566

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2609.39566 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2609.39566 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2609.39566 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles