Communication Dans Un Congrès Année : 2024

MaNo: Exploiting Matrix Norm for Unsupervised Accuracy Estimation Under Distribution Shifts

Résumé

Leveraging the models' outputs, specifically the logits, is a common approach to estimating the test accuracy of a pre-trained neural network on out-of-distribution (OOD) samples without requiring access to the corresponding ground truth labels. Despite their ease of implementation and computational efficiency, current logit-based methods are vulnerable to overconfidence issues, leading to prediction bias, especially under the natural shift. In this work, we first study the relationship between logits and generalization performance from the view of low-density separation assumption. Our findings motivate our proposed method MaNo which (1) applies a data-dependent normalization on the logits to reduce prediction bias, and (2) takes the $L_p$ norm of the matrix of normalized logits as the estimation score. Our theoretical analysis highlights the connection between the provided score and the model's uncertainty. We conduct an extensive empirical study on common unsupervised accuracy estimation benchmarks and demonstrate that MaNo achieves state-of-the-art performance across various architectures in the presence of synthetic, natural, or subpopulation shifts. The code is available at \url{https://github.com/Renchunzi-Xie/MaNo}.

Fichier principal
Vignette du fichier
NeurIPS-2024-mano-exploiting-matrix-norm-for-unsupervised-accuracy-estimation-under-distribution-shifts-Paper-Conference.pdf (1.58 Mo) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)
Licence

Dates et versions

hal-05068034 , version 1 (14-05-2025)

Licence

Identifiants

Citer

Renchunzi Xie, Ambroise Odonnat, Vasilii Feofanov, Weijian Deng, Jianfeng Zhang, et al.. MaNo: Exploiting Matrix Norm for Unsupervised Accuracy Estimation Under Distribution Shifts. NeurIPS 2024 - 38th Conference on Neural Information Processing Systems, Dec 2024, Vancouver, Canada. ⟨hal-05068034⟩
40 Consultations
209 Téléchargements

Altmetric

Partager

  • More