给医疗模型加个不确定度管理层，让高危错误预测自动转给人看。
MedBayes-Lite: A Clinical Uncertainty Governance Layer for Risk-Aware Medical Decision Support
- 不重新训练，用蒙特卡洛丢弃+置信度校准+主动放弃低信度预测
- 在高危错误上将置信度过高率降至接近零，校准误差降0.23至0.33
- 适合需要安全决策的临床问答场景，尤其关注高危误判风险
临床语言模型常对错误预测给出过高置信度，尤其在严重病例和分布外数据上。我们提出MedBayes-Lite，一种无需重训练的不确定性治理层，用于基于Transformer的临床预测器。该层结合蒙特卡洛丢弃、预测校准与置信度引导的主动放弃机制，可自动将低置信度预测转交人工审核，且不引入可训练参数。在MedMCQA和MedQA-USMLE上的评估显示，其将期望校准误差降低0.23至0.33，并使有害过自信错误（高置信、错误、高严重性）趋近于零。在从MedMCQA到MedQA-USMLE的领域偏移下，高置信高严重性错误从约21%降至接近零，同时校准漂移大致减半。我们还提出临床不确定性评分（CUS），其与有害过自信程度高度相关（r≈0.88）。尽管该框架未提升风险覆盖排序性能，且温度缩放或深度集成在校准成本或风险排序上可能更优，但MedBayes-Lite提供了一种实用的校准与主动放弃层，在临床问答基准中显著减少高危错误。
原文摘要 · Abstract (English)
Clinical language models often assign high confidence to incorrect predictions, particularly in high-severity and out-of-distribution cases. We present MedBayes-Lite, a retraining-free uncertainty governance layer for transformer-based clinical predictors. It combines Monte Carlo dropout, predictive calibration, and confidence-guided abstention to defer low-confidence predictions for human review, adding no trainable parameters. Evaluated on MedMCQA and MedQA-USMLE, MedBayes-Lite reduces expected calibration error by 0.23 to 0.33 and drives harmful overconfident errors (confident, incorrect, high-severity predictions) toward zero. Under domain shift from MedMCQA to MedQA-USMLE, it reduces confident high-severity errors from about 21% to near zero while roughly halving calibration drift. We also introduce the Clinical Uncertainty Score (CUS), which strongly correlates with harmful overconfidence (r approximately 0.88). Although the framework does not improve risk-coverage ranking, and temperature scaling or deep ensembles may provide advantages in calibration cost or risk ranking, MedBayes-Lite offers a practical calibration-and-abstention layer that reduces confident high-severity errors in clinical question-answering benchmarks.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。