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

Deep learning is usually data starved, and the unsupervised domain adaptation(UDA) is developed to introduce the knowledge in the labeled source domain tothe unlabeled target domain. Recently, deep self-training presents a powerfulmeans for UDA, involving an iterative process of predicting the target domainand then taking the confident predictions as hard pseudo-labels for retraining.However, the pseudo-labels are usually unreliable, thus easily leading todeviated solutions with propagated errors. In this paper, we resort to theenergy-based model and constrain the training of the unlabeled target samplewith an energy function minimization objective. It can be achieved via a simpleadditional regularization or an energy-based loss. This framework allows us togain the benefits of the energy-based model, while retaining strongdiscriminative performance following a plug-and-play fashion. The convergenceproperty and its connection with classification expectation minimization areinvestigated. We deliver extensive experiments on the most popular andlarge-scale UDA benchmarks of image classification as well as semanticsegmentation to demonstrate its generality and effectiveness.

Constraining Pseudo-label in Self-training Unsupervised Domain Adaptation with Energy-based Model