Transductive conformal inference with adaptive scores

arXiv:2310.18108v2 Announce Type: replace-cross
Abstract: Conformal inference is a fundamental and versatile tool that provides distribution-free guarantees for many machine learning tasks. We consider the transductive setting, where decisions are made on a test sample of $m$ new points, giving rise to $m$ conformal $p$-values. While classical results only concern their marginal distribution, we show that their joint distribution follows a P\'olya urn model, and establish a concentration inequality for their empirical distribution function. The results hold for arbitrary exchangeable …

abstract arxiv cs.lg decisions distribution free giving inference machine machine learning results sample stat.me tasks test tool type values