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 models learn to generalize to previously unseendomains, but suffer from prediction uncertainty and domain shift. In thispaper, we address both problems. We introduce a probabilistic meta-learningmodel for domain generalization, in which classifier parameters shared acrossdomains are modeled as distributions. This enables better handling ofprediction uncertainty on unseen domains. To deal with domain shift, we learndomain-invariant representations by the proposed principle of meta variationalinformation bottleneck, we call MetaVIB. MetaVIB is derived from novelvariational bounds of mutual information, by leveraging the meta-learningsetting of domain generalization. Through episodic training, MetaVIB learns togradually narrow domain gaps to establish domain-invariant representations,while simultaneously maximizing prediction accuracy. We conduct experiments onthree benchmarks for cross-domain visual recognition. Comprehensive ablationstudies validate the benefits of MetaVIB for domain generalization. Thecomparison results demonstrate our method outperforms previous approachesconsistently.

Learning to Learn with Variational Information Bottleneck for Domain Generalization