SPARNet: Continual Test-Time Adaptation via Sample Partitioning Strategy and Anti-Forgetting Regularization

TL;DR

SPARNet introduces a novel continual test-time adaptation framework that partitions samples and applies anti-forgetting regularization to improve long-term model performance amidst domain shifts and noisy pseudo-labels.

Contribution

It proposes a new sample partitioning strategy combined with anti-forgetting regularization to enhance continual test-time adaptation and mitigate catastrophic forgetting.

Findings

Effective in continual TTA scenarios on CIFAR10-C, CIFAR100-C, ImageNet-C

Reduces error accumulation from noisy pseudo-labels

Maintains long-term performance through regularization

Abstract

Test-time Adaptation (TTA) aims to improve model performance when the model encounters domain changes after deployment. The standard TTA mainly considers the case where the target domain is static, while the continual TTA needs to undergo a sequence of domain changes. This encounters a significant challenge as the model needs to adapt for the long-term and is unaware of when the domain changes occur. The quality of pseudo-labels is hard to guarantee. Noisy pseudo-labels produced by simple self-training methods can cause error accumulation and catastrophic forgetting. In this work, we propose a new framework named SPARNet which consists of two parts, sample partitioning strategy and anti-forgetting regularization. The sample partition strategy divides samples into two groups, namely reliable samples and unreliable samples. According to the characteristics of each group of samples, we choose different strategies to deal with different groups of samples. This ensures that reliable samples contribute more to the model. At the same time, the negative impacts of unreliable samples are eliminated by the mean teacher's consistency learning. Finally, we introduce a regularization term to alleviate the catastrophic forgetting problem, which can limit important parameters from excessive changes. This term enables long-term adaptation of parameters in the network. The effectiveness of our method is demonstrated in continual TTA scenario by conducting a large number of experiments on CIFAR10-C, CIFAR100-C and ImageNet-C.