学习推迟(L2D)可以通过将困难的/某些病例转交给人类,来使青光眼筛查更安全,但标准公式忽略了专家的可用性,异构读者行为,工作负载不平衡,不对称诊断危害,形态和部署转变带来的病例困难。我们引入 MPD$^2$-Router, 一个面罩感知多专家延迟框架，它将眼科分类重新定义为受约束的人类 - AI 路由: 是否延迟以及哪个可用专家。它将双头延迟/allocation策略与掩模感知Gumbel（S形门控）结合起来，严格执行每个样本的可用性,并融合不确定性,形态,图像质量,和OOD信号。训练使用非对称成本敏感目标，具有增强拉格朗日延迟预算,、特定于组的分布先验, 和排名主要化 JS 正则化器，共同防止专家崩溃，而无需强制统一分配。在三个跨国青光眼队列中，(REFUGE, CHAKSU, ORIGA) 具有冷冻 REFUGE 训练的骨干, MPD$^2$-路由器大大降低了临床成本，并以适度的延迟率改善了仅使用 AI 的 MCC。它在 F1--MCC--cost, 中是帕累托最优，在跨域转移, 下具有鲁棒性，并产生平衡的专家利用率。
Learning-to-defer (L2D) can make glaucoma screening safer by routing difficult/uncertain cases to humans, yet standard formulations overlook expert availability, heterogeneous readers behavior, workload imbalance, asymmetric diagnostic harm, case difficulty from morphology and deployment shift. We introduce MPD$^2$-Router, a mask-aware multi-expert deferral framework that recasts ophthalmic triage as constrained human--AI routing: whether to defer and to which available expert. It couples a dual-head deferral/allocation policy with mask-aware Gumbel--sigmoid gating that strictly enforces per-sample availability, and fuses uncertainty, morphology, image-quality, and OOD signals. Training uses an asymmetric cost-sensitive objective with an augmented-Lagrangian deferral budget, a group-specific distribution prior, and a rank-majorization JS regularizer that jointly prevent expert collapse without forcing uniform allocation. Across three cross-national glaucoma cohorts (REFUGE, CHAKSU, ORIGA) with a frozen REFUGE-trained backbone, MPD$^2$-Router substantially lowers clinical cost and improves MCC over AI-only at a moderate deferral rate. It is Pareto-optimal in F1--MCC--cost, robust under cross-domain shift, and yields balanced expert utilization.
科目:人工智能(cs.AI)
Subjects: Artificial Intelligence (cs.AI)