Backpropagation-Free Test-Time Adaptation via Probabilistic Gaussian Alignment

TL;DR

This paper introduces ADAPT, a novel test-time adaptation method that models class-conditional distributions probabilistically without backpropagation, enabling scalable, real-time robustness improvements under distribution shifts.

Contribution

ADAPT reframes TTA as a Gaussian inference problem, using class means and covariance, with regularization from CLIP priors, requiring no source data or gradient updates.

Findings

Achieves state-of-the-art performance across diverse benchmarks.

Supports online and transductive settings without source data.

Demonstrates superior scalability and robustness.

Abstract

Test-time adaptation (TTA) enhances the zero-shot robustness under distribution shifts by leveraging unlabeled test data during inference. Despite notable advances, several challenges still limit its broader applicability. First, most methods rely on backpropagation or iterative optimization, which limits scalability and hinders real-time deployment. Second, they lack explicit modeling of class-conditional feature distributions. This modeling is crucial for producing reliable decision boundaries and calibrated predictions, but it remains underexplored due to the lack of both source data and supervision at test time. In this paper, we propose ADAPT, an Advanced Distribution-Aware and backPropagation-free Test-time adaptation method. We reframe TTA as a Gaussian probabilistic inference task by modeling class-conditional likelihoods using gradually updated class means and a shared covariance matrix. This enables closed-form, training-free inference. To correct potential likelihood bias, we introduce lightweight regularization guided by CLIP priors and a historical knowledge bank. ADAPT requires no source data, no gradient updates, and no full access to target data, supporting both online and transductive settings. Extensive experiments across diverse benchmarks demonstrate that our method achieves state-of-the-art performance under a wide range of distribution shifts with superior scalability and robustness.