Saved in:
Bibliographic Details
Main Authors: Chuah, WeiQin, Tennakoon, Ruwan, Bab-Hadiashar, Alireza
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2405.07171
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929340811313152
author Chuah, WeiQin
Tennakoon, Ruwan
Bab-Hadiashar, Alireza
author_facet Chuah, WeiQin
Tennakoon, Ruwan
Bab-Hadiashar, Alireza
contents Online Test-Time Adaptation (OTTA) has emerged as an effective strategy to handle distributional shifts, allowing on-the-fly adaptation of pre-trained models to new target domains during inference, without the need for source data. We uncovered that the widely studied entropy minimization (EM) method for OTTA, suffers from noisy gradients due to ambiguity near decision boundaries and incorrect low-entropy predictions. To overcome these limitations, this paper introduces a novel cosine alignment optimization approach with a dual-objective loss function that refines the precision of class predictions and adaptability to novel domains. Specifically, our method optimizes the cosine similarity between feature vectors and class weight vectors, enhancing the precision of class predictions and the model's adaptability to novel domains. Our method outperforms state-of-the-art techniques and sets a new benchmark in multiple datasets, including CIFAR-10-C, CIFAR-100-C, ImageNet-C, Office-Home, and DomainNet datasets, demonstrating high accuracy and robustness against diverse corruptions and domain shifts.
format Preprint
id arxiv_https___arxiv_org_abs_2405_07171
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Enhanced Online Test-time Adaptation with Feature-Weight Cosine Alignment
Chuah, WeiQin
Tennakoon, Ruwan
Bab-Hadiashar, Alireza
Computer Vision and Pattern Recognition
Online Test-Time Adaptation (OTTA) has emerged as an effective strategy to handle distributional shifts, allowing on-the-fly adaptation of pre-trained models to new target domains during inference, without the need for source data. We uncovered that the widely studied entropy minimization (EM) method for OTTA, suffers from noisy gradients due to ambiguity near decision boundaries and incorrect low-entropy predictions. To overcome these limitations, this paper introduces a novel cosine alignment optimization approach with a dual-objective loss function that refines the precision of class predictions and adaptability to novel domains. Specifically, our method optimizes the cosine similarity between feature vectors and class weight vectors, enhancing the precision of class predictions and the model's adaptability to novel domains. Our method outperforms state-of-the-art techniques and sets a new benchmark in multiple datasets, including CIFAR-10-C, CIFAR-100-C, ImageNet-C, Office-Home, and DomainNet datasets, demonstrating high accuracy and robustness against diverse corruptions and domain shifts.
title Enhanced Online Test-time Adaptation with Feature-Weight Cosine Alignment
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2405.07171