English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Domain generalization (DG) aims to improve the generalization ability of themodel trained on several known training domains over unseen test domains.Previous work has shown that self-supervised contrastive pre-training improvesthe robustness of the model on downstream tasks. However, in this paper, wefind that self-supervised models do not exhibit better generalizationperformance than supervised models pre-trained on the same dataset in the DGsetting. We argue that this is owing to the fact that the richer intra-classdiscriminative features extracted by self-supervised contrastive learning,which we term silent features, are suppressed during supervised fine-tuning.These silent features are likely to contain features that are moregeneralizable on the test domain. In this work, we model and analyze thisfeature suppression phenomenon and theoretically prove that preserving silentfeatures can achieve lower expected test domain risk under certain conditions.In light of this, we propose a simple yet effective method termed STEP (SilentFeature Preservation) to improve the generalization performance of theself-supervised contrastive learning pre-trained model by alleviating thesuppression of silent features during the supervised fine-tuning process.Experimental results show that STEP exhibits state-of-the-art performance onstandard DG benchmarks with significant distribution shifts.

Preserving Silent Features for Domain Generalization