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

Zero-shot learning aims at recognizing unseen classes (no training example)with knowledge transferred from seen classes. This is typically achieved byexploiting a semantic feature space shared by both seen and unseen classes,i.e., attribute or word vector, as the bridge. One common practice in zero-shotlearning is to train a projection between the visual and semantic featurespaces with labeled seen classes examples. When inferring, this learnedprojection is applied to unseen classes and recognizes the class labels by somemetrics. However, the visual and semantic feature spaces are mutuallyindependent and have quite different manifold structures. Under such aparadigm, most existing methods easily suffer from the domain shift problem andweaken the performance of zero-shot recognition. To address this issue, wepropose a novel model called AMS-SFE. It considers the alignment of manifoldstructures by semantic feature expansion. Specifically, we build upon anautoencoder-based model to expand the semantic features from the visual inputs.Additionally, the expansion is jointly guided by an embedded manifold extractedfrom the visual feature space of the data. Our model is the first attempt toalign both feature spaces by expanding semantic features and derives twobenefits: first, we expand some auxiliary features that enhance the semanticfeature space; second and more importantly, we implicitly align the manifoldstructures between the visual and semantic feature spaces; thus, the projectioncan be better trained and mitigate the domain shift problem. Extensiveexperiments show significant performance improvement, which verifies theeffectiveness of our model.

A Novel Perspective to Zero-shot Learning: Towards an Alignment of Manifold Structures via Semantic Feature Expansion