Mask-Conditional Conformal Prediction: Valid Uncertainty For All Missing Data Mechanisms

arXiv:2512.14221 · stat.ML, cs.LG · Submitted 2025-12-16 · Read on arXiv

stat.ML, cs.LG

Submitted: 2025-12-16

Updated: 2026-11-01

Journal ref: Proceedings of the 29th International Conference on Artificial Intelligence and Statistics (AISTATS), PMLR 300:3034-3042, 2026

License: http://creativecommons.org/licenses/by/4.0/

The gist: Conformal prediction (CP) offers a principled framework for uncertainty quantification, but it fails to guarantee mask-conditional coverage when faced with missing covariates.

Terminology

Abstract

Conformal prediction (CP) offers a principled framework for uncertainty quantification, but it fails to guarantee mask-conditional coverage when faced with missing covariates. In addressing the heterogeneity induced by various missing patterns, Mask-Conditional Valid (MCV) Coverage has emerged as a more desirable property than Marginal Coverage. In this work, we adapt split CP to handle missing values by proposing a preimpute-mask-then-correct framework that can offer valid coverage. We show that our method provides guaranteed Marginal Coverage and Mask-Conditional Validity for general missing data mechanisms. A key component of our approach is a reweighted conformal prediction procedure that corrects the prediction sets after distributional imputation (multiple imputation) of the calibration dataset, making our method compatible with standard imputation pipelines. We derive two algorithms and prove that they achieve both marginal validity and MCV. We evaluate them on synthetic and real-world datasets. It reduces significantly the width of prediction intervals w.r.t standard MCV methods, while maintaining the target guarantees.

Sources

Related papers