arXiv cs.AI· Yue Zhou, Haiyang Zhou, Jin Zhang, Kong Wang, Zihao Rao, Maolin Wang, Hanwen Du, Youhua Li·· 12 小时前AI 评分38
CaRDI：以校准风险替代不确定性降低的交互式诊断规划框架
Beyond Uncertainty Reduction: Calibrated Planning for Interactive Diagnosis
AI 导读
研究者提出 CaRDI（Calibrated Risk for Diagnostic Interaction）规划框架，用结合主诊断错误与严重程度加权鉴别遗漏的临床损失替代不确定性降低作为提问选择目标，并通过保形残差校准修正部分证据下的低估。在多个医学基准上，CaRDI 提升准确率的同时所需提问更少；以临床风险替代熵作为规划目标使临床损失降低 21.5%，并减少严重遗漏。代码已开源。
正文 · 原文
Abstract:Interactive medical diagnosis requires an agent to gather patient evidence through sequential questions before committing to a diagnosis. Large language models (LLMs) now make such consultations feasible, and existing approaches typically select questions by how much they reduce uncertainty among candidate diseases. However, this objective is insensitive to the unequal clinical costs of diagnostic errors and omitted severe alternatives, so minimizing uncertainty need not minimize final diagnostic loss. To bridge this gap, we propose CaRDI (Calibrated Risk for Diagnostic Interaction), a planning framework built around a clinical loss combining primary diagnosis error with severity-weighted differential omission. Diagnostic and safety beliefs estimate the two loss components, while conformal residual calibration corrects for underestimation from partial evidence. The calibrated risk then guides question selection and stopping by balancing expected risk reduction against interaction cost. Experiments on multiple medical benchmarks show that CaRDI improves accuracy while requiring fewer questions. Further analysis shows that replacing entropy with clinical risk as the planning objective reduces clinical loss by 21.5% and decreases severe omissions. Our code is available at this https URL.
| Subjects: | Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2608.27847 [cs.AI] |
| (or arXiv:2608.27847v2 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2608.27847 arXiv-issued DOI via DataCite |
Submission history
From: Yue Zhou [view email]
[v1]
Fri, 28 Aug 2026 02:29:38 UTC (2,700 KB)
[v2]
Fri, 16 Oct 2026 09:04:48 UTC (813 KB)
来源：arXiv cs.AI · arxiv.org