Saved in:
Bibliographic Details
Main Authors: Lin, Guoliang, Lai, Hanjiang, Pan, Yan, Yin, Jian
Format: Preprint
Published: 2023
Subjects:
Online Access:https://arxiv.org/abs/2310.20327
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911853540540416
author Lin, Guoliang
Lai, Hanjiang
Pan, Yan
Yin, Jian
author_facet Lin, Guoliang
Lai, Hanjiang
Pan, Yan
Yin, Jian
contents Domain shift is a common problem in the realistic world, where training data and test data follow different data distributions. To deal with this problem, fully test-time adaptation (TTA) leverages the unlabeled data encountered during test time to adapt the model. In particular, entropy-based TTA (EBTTA) methods, which minimize the prediction's entropy on test samples, have shown great success. In this paper, we introduce a new clustering perspective on the EBTTA. It is an iterative algorithm: 1) in the assignment step, the forward process of the EBTTA models is the assignment of labels for these test samples, and 2) in the updating step, the backward process is the update of the model via the assigned samples. This new perspective allows us to explore how entropy minimization influences test-time adaptation. Accordingly, this observation can guide us to put forward the improvement of EBTTA. We propose to improve EBTTA from the assignment step and the updating step, where robust label assignment, similarity-preserving constraint, sample selection, and gradient accumulation are proposed to explicitly utilize more information. Experimental results demonstrate that our method can achieve consistent improvements on various datasets. Code is provided in the supplementary material.
format Preprint
id arxiv_https___arxiv_org_abs_2310_20327
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Improving Entropy-Based Test-Time Adaptation from a Clustering View
Lin, Guoliang
Lai, Hanjiang
Pan, Yan
Yin, Jian
Artificial Intelligence
Domain shift is a common problem in the realistic world, where training data and test data follow different data distributions. To deal with this problem, fully test-time adaptation (TTA) leverages the unlabeled data encountered during test time to adapt the model. In particular, entropy-based TTA (EBTTA) methods, which minimize the prediction's entropy on test samples, have shown great success. In this paper, we introduce a new clustering perspective on the EBTTA. It is an iterative algorithm: 1) in the assignment step, the forward process of the EBTTA models is the assignment of labels for these test samples, and 2) in the updating step, the backward process is the update of the model via the assigned samples. This new perspective allows us to explore how entropy minimization influences test-time adaptation. Accordingly, this observation can guide us to put forward the improvement of EBTTA. We propose to improve EBTTA from the assignment step and the updating step, where robust label assignment, similarity-preserving constraint, sample selection, and gradient accumulation are proposed to explicitly utilize more information. Experimental results demonstrate that our method can achieve consistent improvements on various datasets. Code is provided in the supplementary material.
title Improving Entropy-Based Test-Time Adaptation from a Clustering View
topic Artificial Intelligence
url https://arxiv.org/abs/2310.20327