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

Existing domain adaptation methods generally assume different domains havethe identical label space, which is quite restrict for real-world applications.In this paper, we focus on a more realistic and challenging case of open setdomain adaptation. Particularly, in open set domain adaptation, we allow theclasses from the source and target domains to be partially overlapped. In thiscase, the assumption of conventional distribution alignment does not holdanymore, due to the different label spaces in two domains. To tackle thischallenge, we propose a new approach coined as Known-class Aware Self-Ensemble(KASE), which is built upon the recently developed self-ensemble model. InKASE, we first introduce a Known-class Aware Recognition (KAR) module toidentify the known and unknown classes from the target domain, which isachieved by encouraging a low cross-entropy for known classes and a highentropy based on the source data from the unknown class. Then, we develop aKnown-class Aware Adaptation (KAA) module to better adapt from the sourcedomain to the target by reweighing the adaptation loss based on the likelinessto belong to known classes of unlabeled target samples as predicted by KAR.Extensive experiments on multiple benchmark datasets demonstrate theeffectiveness of our approach.

Known-class Aware Self-ensemble for Open Set Domain Adaptation