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

Real-world classification tasks are frequently required to work in anopen-set setting. This is especially challenging for few-shot learning problemsdue to the small sample size for each known category, which prevents existingopen-set methods from working effectively; however, most multiclass few-shotmethods are limited to closed-set scenarios. In this work, we address theproblem of few-shot open-set classification by first proposing methods forfew-shot one-class classification and then extending them to few-shotmulticlass open-set classification. We introduce two independent few-shotone-class classification methods: Meta Binary Cross-Entropy (Meta-BCE), whichlearns a separate feature representation for one-class classification, andOne-Class Meta-Learning (OCML), which learns to generate one-class classifiersgiven standard multiclass feature representation. Both methods can augment anyexisting few-shot learning method without requiring retraining to work in afew-shot multiclass open-set setting without degrading its closed-setperformance. We demonstrate the benefits and drawbacks of both methods indifferent problem settings and evaluate them on three standard benchmarkdatasets, miniImageNet, tieredImageNet, and Caltech-UCSD-Birds-200-2011, wherethey surpass the state-of-the-art methods in the few-shot multiclass open-setand few-shot one-class tasks.

One-Class Meta-Learning: Towards Generalizable Few-Shot Open-Set Classification