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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Domain generalization aims to address the domain shift between training andtesting data. To learn the domain invariant representations, the model isusually trained on multiple domains. It has been found that the gradients ofnetwork weight relative to a specific task loss can characterize the taskitself. In this work, with the assumption that the gradients of a specificdomain samples under the classification task could also reflect the property ofthe domain, we propose a Shape Guided Gradient Voting (SGGV) method for domaingeneralization. Firstly, we introduce shape prior via extra inputs of thenetwork to guide gradient descending towards a shape-biased direction forbetter generalization. Secondly, we propose a new gradient voting strategy toremove the outliers for robust optimization in the presence of shape guidance.To provide shape guidance, we add edge/sketch extracted from the training dataas an explicit way, and also use texture augmented images as an implicit way.We conduct experiments on several popular domain generalization datasets inimage classification task, and show that our shape guided gradient updatingstrategy brings significant improvement of the generalization.

Shape Guided Gradient Voting for Domain Generalization