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

Contrastive self-supervised learning (CSL) has managed to match or surpassthe performance of supervised learning in image and video classification.However, it is still largely unknown if the nature of the representationsinduced by the two learning paradigms is similar. We investigate this under thelens of adversarial robustness. Our analysis of the problem reveals that CSLhas intrinsically higher sensitivity to perturbations over supervised learning.We identify the uniform distribution of data representation over a unithypersphere in the CSL representation space as the key contributor to thisphenomenon. We establish that this is a result of the presence of falsenegative pairs in the training process, which increases model sensitivity toinput perturbations. Our finding is supported by extensive experiments forimage and video classification using adversarial perturbations and other inputcorruptions. We devise a strategy to detect and remove false negative pairsthat is simple, yet effective in improving model robustness with CSL training.We close up to 68% of the robustness gap between CSL and its supervisedcounterpart. Finally, we contribute to adversarial learning by incorporatingour method in CSL. We demonstrate an average gain of about 5% over twodifferent state-of-the-art methods in this domain.

Contrastive Self-Supervised Learning Leads to Higher Adversarial Susceptibility