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 Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

The findings on open-set recognition (OSR) show that models trained onclassification datasets are capable of detecting unknown classes notencountered during the training process. Specifically, after training, thelearned representations of known classes dissociate from the representations ofthe unknown class, facilitating OSR. In this paper, we investigate thisemergent phenomenon by examining the relationship between the Jacobian norm ofrepresentations and the inter/intra-class learning dynamics. We provide atheoretical analysis, demonstrating that intra-class learning reduces theJacobian norm for known class samples, while inter-class learning increases theJacobian norm for unknown samples, even in the absence of direct exposure toany unknown sample. Overall, the discrepancy in the Jacobian norm between theknown and unknown classes enables OSR. Based on this insight, which highlightsthe pivotal role of inter-class learning, we devise a marginal one-vs-rest(m-OvR) loss function that promotes strong inter-class separation. To furtherimprove OSR performance, we integrate the m-OvR loss with additional strategiesthat maximize the Jacobian norm disparity. We present comprehensiveexperimental results that support our theoretical observations and demonstratethe efficacy of our proposed OSR approach.

Understanding Open-Set Recognition by Jacobian Norm and Inter-Class Separation