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

Many research efforts have been committed to unsupervised domain adaptation(DA) problems that transfer knowledge learned from a labeled source domain toan unlabeled target domain. Various DA methods have achieved remarkable resultsrecently in terms of predicting ability, which implies the effectiveness of theaforementioned knowledge transferring. However, state-of-the-art methods rarelyprobe deeper into the transferred mechanism, leaving the true essence of suchknowledge obscure. Recognizing its importance in the adaptation process, wepropose an interpretive model of unsupervised domain adaptation, as the firstattempt to visually unveil the mystery of transferred knowledge. Adapting theexisting concept of the prototype from visual image interpretation to the DAtask, our model similarly extracts shared information from the domain-invariantrepresentations as prototype vectors. Furthermore, we extend the currentprototype method with our novel prediction calibration and knowledge fidelitypreservation modules, to orientate the learned prototypes to the actualtransferred knowledge. By visualizing these prototypes, our method not onlyprovides an intuitive explanation for the base model's predictions but alsounveils transfer knowledge by matching the image patches with the samesemantics across both source and target domains. Comprehensive experiments andin-depth explorations demonstrate the efficacy of our method in understandingthe transferred mechanism and its potential in downstream tasks including modeldiagnosis.

Visualizing Transferred Knowledge: An Interpretive Model of Unsupervised Domain Adaptation