Conformal risk control for non-monotonic losses

作者：Anastasios N. Angelopoulos · 发表于：Philosophical Transactions of the Royal Society A Mathematical Physical and Engineering Sciences · 年份：2026 · DOI：10.1098/rsta.2025.0068 · 研究领域：Statistical Methods and Inference、Explainable Artificial Intelligence (XAI)、Imbalanced Data Classification Techniques

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 multi-dimensional 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, false discovery rate and intersection-over-union control of tumour segmentations and multi-group debiasing of recidivism predictions across overlapping race and sex groups using empirical risk minimization. This article is part of the theme issue 'Advancing uncertainty quantification in AI systems'.