STAMP: Outlier-Aware Test-Time Adaptation with Stable Memory Replay

TL;DR

STAMP introduces a memory-based test-time adaptation method that improves recognition and outlier detection in open-world scenarios by dynamically updating a stable memory bank and employing a self-weighted entropy minimization strategy.

Contribution

The paper proposes STAMP, a novel outlier-aware TTA approach using a stable memory replay mechanism and entropy-based sample weighting, addressing open-world inference challenges.

Findings

Outperforms existing TTA methods in recognition accuracy

Enhances outlier detection performance

Demonstrates robustness in open-world inference scenarios

Abstract

Test-time adaptation (TTA) aims to address the distribution shift between the training and test data with only unlabeled data at test time. Existing TTA methods often focus on improving recognition performance specifically for test data associated with classes in the training set. However, during the open-world inference process, there are inevitably test data instances from unknown classes, commonly referred to as outliers. This paper pays attention to the problem that conducts both sample recognition and outlier rejection during inference while outliers exist. To address this problem, we propose a new approach called STAble Memory rePlay (STAMP), which performs optimization over a stable memory bank instead of the risky mini-batch. In particular, the memory bank is dynamically updated by selecting low-entropy and label-consistent samples in a class-balanced manner. In addition, we develop a self-weighted entropy minimization strategy that assigns higher weight to low-entropy samples. Extensive results demonstrate that STAMP outperforms existing TTA methods in terms of both recognition and outlier detection performance. The code is released at https://github.com/yuyongcan/STAMP.