A study of uncertainty quantification in overparametrized high-dimensional models - Archive ouverte HAL Accéder directement au contenu
Pré-Publication, Document De Travail Année : 2022

A study of uncertainty quantification in overparametrized high-dimensional models

Résumé

Uncertainty quantification is a central challenge in reliable and trustworthy machine learning. Naive measures such as last-layer scores are well-known to yield overconfident estimates in the context of overparametrized neural networks. Several methods, ranging from temperature scaling to different Bayesian treatments of neural networks, have been proposed to mitigate overconfidence, most often supported by the numerical observation that they yield better calibrated uncertainty measures. In this work, we provide a sharp comparison between popular uncertainty measures for binary classification in a mathematically tractable model for overparametrized neural networks: the random features model. We discuss a trade-off between classification accuracy and calibration, unveiling a double descent like behavior in the calibration curve of optimally regularized estimators as a function of overparametrization. This is in contrast with the empirical Bayes method, which we show to be well calibrated in our setting despite the higher generalization error and overparametrization.
Fichier principal
Vignette du fichier
2210.12760.pdf (635.15 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04019034 , version 1 (08-03-2023)

Identifiants

Citer

Lucas Clarté, Bruno Loureiro, Florent Krzakala, Lenka Zdeborová. A study of uncertainty quantification in overparametrized high-dimensional models. 2022. ⟨hal-04019034⟩
20 Consultations
35 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More