arXiv cs.CL· Hsin-Ling Hsu, Zizheng Wang, Donghua Zhang, Nai-Chia Chen, Jerry Wang, Jun-En Ding, Chia-Hsuan Hsu, Guoan Wang, Feng Liu, Fang-Ming Hung, Chenwei Wu, Liyue Shen·· 12 小时前AI 评分35
MedAction：面向主动多轮临床诊断的大语言模型
MedAction: Towards Active Multi-turn Clinical Diagnostic LLMs
AI 导读
研究团队提出 MedAction，一个树状结构蒸馏流程，通过 LLM 与环境交互合成多轮诊断轨迹，并构建了含 32,681 条轨迹、来自 2,896 个 PMC 病例的 MedAction-32K 数据集。
正文 · 原文
Authors:Hsin-Ling Hsu, Zizheng Wang, Donghua Zhang, Nai-Chia Chen, Jerry Wang, Jun-En Ding, Chia-Hsuan Hsu, Guoan Wang, Feng Liu, Fang-Ming Hung, Chenwei Wu, Liyue Shen
Abstract:Most existing LLM diagnoses are evaluated on static, single-turn settings where complete patient information is provided upfront, an oversimplification of real clinical practice. We study active diagnosis: the real-life clinical process of starting from initial observation, ordering tests, interpreting results, and updating a differential diagnosis across multiple turns. Through systematic analysis, we identify three recurring failure modes in current LLMs: ungrounded test ordering, unreliable diagnostic update, and degraded multi-turn coherence. Together, these failures reveal a core deficit: existing medical training data teaches models to reason from complete information but not to act under evolving, partial evidence. To address this gap, we introduce MedAction, a tree-structured distillation pipeline that synthesizes diverse and high-quality multi-turn diagnostic trajectories via LLM-environment interaction. We propose two knowledge-graph-grounded metrics to filter trajectory quality: Disease Trajectory Consistency (DTC), which tracks whether the model's hypothesis converges toward the correct diagnosis, and Reasoning-Action Consistency (RAC), which verifies that belief updates are driven by gathered evidence. Using this pipeline, we construct MedAction-32K, a dataset of 32,681 trajectories from 2,896 PMC cases. Fine-tuning an 8B model on MedAction-32K achieves state-of-the-art performance among open-source models on both MedR-Bench and our curated MedAction-300-Hard benchmark, pushing the edge for open-source medical LLMs.
| Comments: | COLM 2026. v2 adds an RL extension (Appendix D) |
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2605.07305 [cs.CL] |
| (or arXiv:2605.07305v2 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2605.07305 arXiv-issued DOI via DataCite |
Submission history
From: Hsin-Ling Hsu [view email]
[v1]
Fri, 8 May 2026 06:15:24 UTC (1,008 KB)
[v2]
Thu, 22 Oct 2026 18:11:13 UTC (1,328 KB)
来源：arXiv cs.CL · arxiv.org