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

Understanding generalization is crucial to confidently engineer and deploymachine learning models, especially when deployment implies a shift in the datadomain. For such domain adaptation problems, we seek generalization boundswhich are tractably computable and tight. If these desiderata can be reached,the bounds can serve as guarantees for adequate performance in deployment.However, in applications where deep neural networks are the models of choice,deriving results which fulfill these remains an unresolved challenge; mostexisting bounds are either vacuous or has non-estimable terms, even infavorable conditions. In this work, we evaluate existing bounds from theliterature with potential to satisfy our desiderata on domain adaptation imageclassification tasks, where deep neural networks are preferred. We find thatall bounds are vacuous and that sample generalization terms account for much ofthe observed looseness, especially when these terms interact with measures ofdomain shift. To overcome this and arrive at the tightest possible results, wecombine each bound with recent data-dependent PAC-Bayes analysis, greatlyimproving the guarantees. We find that, when domain overlap can be assumed, asimple importance weighting extension of previous work provides the tightestestimable bound. Finally, we study which terms dominate the bounds and identifypossible directions for further improvement.

Practicality of generalization guarantees for unsupervised domain adaptation with neural networks