Less is More: Pseudo-Label Filtering for Continual Test-Time Adaptation

TL;DR

This paper introduces a novel pseudo-label filtering method called PLF for continual test-time adaptation, improving model performance on unlabeled, evolving target domains by selecting reliable pseudo-labels and encouraging diverse predictions.

Contribution

The paper proposes a new pseudo-label selection technique with adaptive thresholds and class prior alignment, enhancing CTTA performance over existing methods.

Findings

PLF outperforms state-of-the-art CTTA methods in experiments.

Adaptive thresholding improves pseudo-label quality.

Class prior alignment promotes diverse predictions.

Abstract

Continual Test-Time Adaptation (CTTA) aims to adapt a pre-trained model to a sequence of target domains during the test phase without accessing the source data. To adapt to unlabeled data from unknown domains, existing methods rely on constructing pseudo-labels for all samples and updating the model through self-training. However, these pseudo-labels often involve noise, leading to insufficient adaptation. To improve the quality of pseudo-labels, we propose a pseudo-label selection method for CTTA, called Pseudo Labeling Filter (PLF). The key idea of PLF is to keep selecting appropriate thresholds for pseudo-labels and identify reliable ones for self-training. Specifically, we present three principles for setting thresholds during continuous domain learning, including initialization, growth and diversity. Based on these principles, we design Self-Adaptive Thresholding to filter pseudo-labels. Additionally, we introduce a Class Prior Alignment (CPA) method to encourage the model to make diverse predictions for unknown domain samples. Through extensive experiments, PLF outperforms current state-of-the-art methods, proving its effectiveness in CTTA.