Transductive conformal inference with adaptive scores

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Gazin, Ulysse, Blanchard, Gilles, Roquain, Etienne
Format: Preprint
Veröffentlicht: 2023
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866910372456300544
author Gazin, Ulysse
Blanchard, Gilles
Roquain, Etienne
author_facet Gazin, Ulysse
Blanchard, Gilles
Roquain, Etienne
contents 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 Pólya urn model, and establish a concentration inequality for their empirical distribution function. The results hold for arbitrary exchangeable scores, including adaptive ones that can use the covariates of the test+calibration samples at training stage for increased accuracy. We demonstrate the usefulness of these theoretical results through uniform, in-probability guarantees for two machine learning tasks of current interest: interval prediction for transductive transfer learning and novelty detection based on two-class classification.
format Preprint
id arxiv_https___arxiv_org_abs_2310_18108
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Transductive conformal inference with adaptive scores
Gazin, Ulysse
Blanchard, Gilles
Roquain, Etienne
Methodology
Machine Learning
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 Pólya urn model, and establish a concentration inequality for their empirical distribution function. The results hold for arbitrary exchangeable scores, including adaptive ones that can use the covariates of the test+calibration samples at training stage for increased accuracy. We demonstrate the usefulness of these theoretical results through uniform, in-probability guarantees for two machine learning tasks of current interest: interval prediction for transductive transfer learning and novelty detection based on two-class classification.
title Transductive conformal inference with adaptive scores
topic Methodology
Machine Learning
url https://arxiv.org/abs/2310.18108