AI Paper Atlas

Research map / Large language models / Agents, code and math

Language Agents for Hypothesis-driven Clinical Decision Making with Reinforcement Learning

David Bani-Harouni, Chantal Pellegrini, Ege Özsoy, Nassir Navab, Matthias Keicher

ICLR 2026 · Large language models › Agents, code and math · approach: Agent instruction-following benchmarks · 15 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
train LA-CDM with SFT hypothesis loss and GRPO instead of zero-shotLA-CDM (Qwen2.5-7B-Instruct) · MIMIC-CDMmean accuracy (%): 64.5 -> 81.3Improved
train LA-CDM instead of untrained ReActLA-CDM · MIMIC-CDMaverage test cost ($): 1480.32 -> 1295.61Improved
add test-cost reward (ablation)LA-CDM · MIMIC-CDMaverage test cost ($): 1427.85 -> 1295.61Improved
add test-cost reward (ablation)LA-CDM · MIMIC-CDMmean accuracy (%): 82.3 -> 81.3No clear effect
add hypothesis agent (ablation)LA-CDM decision agent · MIMIC-CDMmean accuracy (%): 78.5 -> 82.3Improved
add hypothesis agent (ablation)LA-CDM decision agent · MIMIC-CDMaverage test cost ($): 1410.01 -> 1427.85Hurt
train hypothesis generation with supervised fine-tuninghypothesis agent · MIMIC-CDMhypothesis accuracy (%): 75.7 -> 81.9Improved
train confidence calibration with GRPO betting rewardhypothesis agent · MIMIC-CDMECE: 0.069 -> 0.037Improved
use all patient data with SFT instead of sequential testingQwen2.5-7B-Instruct · MIMIC-CDMmean accuracy (%): 81.3 -> 92.8 (cost 1295.61 -> 3792.79)Improved
process textual notes and imaging reports instead of tabular-onlyLA-CDM vs SM-DDPO · MIMIC-CDMmean accuracy (%): 37.0 -> 81.3Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy