Layerwise Early Stopping for Test Time Adaptation

arXiv:2404.03784v1 Announce Type: new
Abstract: Test Time Adaptation (TTA) addresses the problem of distribution shift by enabling pretrained models to learn new features on an unseen domain at test time. However, it poses a significant challenge to maintain a balance between learning new features and retaining useful pretrained features. In this paper, we propose Layerwise EArly STopping (LEAST) for TTA to address this problem. The key idea is to stop adapting individual layers during TTA if the features being learned …

abstract arxiv balance challenge cs.ai cs.cv cs.lg distribution domain enabling features however learn new features paper pretrained models shift test type