拓展了风险控制方法，可处理多维参数下的非单调损失。
Conformal Risk Control for Non-Monotonic Losses
- 针对非单调损失设计通用风险控制算法
- 算法稳定性决定控制效果强弱
- 适用于图像分类、肿瘤分割与公平预测
置信风险控制是置信预测的扩展，用于控制超出误覆盖率的风险函数。原始算法仅能控制一维参数下单调损失的期望值。本文提出对通用算法在可能非单调且多维参数损失下的风险控制保证。这些保证依赖于算法的稳定性——不稳定的算法控制效果较弱。该技术应用于选择性图像分类、肿瘤分割的假阳性发现率（FDR）与交并比（IOU）控制，以及使用经验风险最小化在重叠种族与性别群体中实现累犯预测的多组公平性去偏。
原文摘要 · Abstract (English)
Conformal risk control is an extension of conformal prediction for controlling risk functions beyond miscoverage. The original algorithm controls the expected value of a loss that is monotonic in a one-dimensional parameter. Here, we present risk control guarantees for generic algorithms applied to possibly non-monotonic losses with multidimensional parameters. The guarantees depend on the stability of the algorithm -- unstable algorithms have looser guarantees. We give applications of this technique to selective image classification, FDR and IOU control of tumor segmentations, and multigroup debiasing of recidivism predictions across overlapping race and sex groups using empirical risk minimization.
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。