搜索医学 AI 论文与资源

ICLR 2026 Poster accepted paper at ICLR 2026. Learning to Defer (L2D) improves AI reliability in decision-critical environments by training AI to either make its own prediction or defer the decision to a human expert. A key challenge is adapting to unseen experts at test time, whose competence can differ from the training population. Current methods for this task, however, can falter when unseen experts are out-of-distribution (OOD) relative to the training population. We identify a core architectural flaw as the cause: they learn identity-conditioned policies by processing class-indexed signals in fixed coordinates, creating shortcuts that violate the problem's inherent permutation symmetry.

系统综述与荟萃分析，评估临床医学中人类与大语言模型协作相对于人类单独工作流的表现，覆盖临床推理、文档和解释等任务；研究指出当前证据仍初步且具有情境依赖性，建议后续开展预注册、务实、多中心并嵌入真实工作流的临床研究。

Nature Communications 论文，提出用于医学文献挖掘的人机协作基础模型。

The ISIC Archive is a large public dermatology image repository for skin lesion analysis. It is widely used for melanoma classification, lesion segmentation, dermoscopic image retrieval, bias and domain shift analysis, and clinical imaging benchmark development.

This SAGE / Design for Augmented Humanity special issue calls for papers on human-centred and augmented healthcare AI, including theories, applications, and methodologies. The CFP lists a paper submission deadline of 2026-05-30 and is relevant to clinical translation, trustworthy and explainable medical AI, human-AI collaboration, and clinical workflow AI.

Stanford Medicine RAISE Health symposium on responsible AI for safe and equitable health care.