Test-Time Adaptation with Principal Component Analysis

TL;DR

This paper introduces TTAwPCA, a test-time adaptation method that uses PCA to improve model robustness against distribution shifts with minimal additional parameters.

Contribution

It proposes a novel PCA-based spectral filtering approach for test-time adaptation that requires fewer parameters than existing methods.

Findings

Effective on CIFAR-10-C and CIFAR-100-C datasets

Uses only 2000 parameters for adaptation

Demonstrates robustness to corruptions

Abstract

Machine Learning models are prone to fail when test data are different from training data, a situation often encountered in real applications known as distribution shift. While still valid, the training-time knowledge becomes less effective, requiring a test-time adaptation to maintain high performance. Following approaches that assume batch-norm layer and use their statistics for adaptation, we propose a Test-Time Adaptation with Principal Component Analysis (TTAwPCA), which presumes a fitted PCA and adapts at test time a spectral filter based on the singular values of the PCA for robustness to corruptions. TTAwPCA combines three components: the output of a given layer is decomposed using a Principal Component Analysis (PCA), filtered by a penalization of its singular values, and reconstructed with the PCA inverse transform. This generic enhancement adds fewer parameters than current methods. Experiments on CIFAR-10-C and CIFAR- 100-C demonstrate the effectiveness and limits of our method using a unique filter of 2000 parameters.