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

Recent works have shown that self-supervised learning can achieve remarkablerobustness when integrated with adversarial training (AT). However, therobustness gap between supervised AT (sup-AT) and self-supervised AT (self-AT)remains significant. Motivated by this observation, we revisit existing self-ATmethods and discover an inherent dilemma that affects self-AT robustness:either strong or weak data augmentations are harmful to self-AT, and a mediumstrength is insufficient to bridge the gap. To resolve this dilemma, we proposea simple remedy named DYNACL (Dynamic Adversarial Contrastive Learning). Inparticular, we propose an augmentation schedule that gradually anneals from astrong augmentation to a weak one to benefit from both extreme cases. Besides,we adopt a fast post-processing stage for adapting it to downstream tasks.Through extensive experiments, we show that DYNACL can improve state-of-the-artself-AT robustness by 8.84% under Auto-Attack on the CIFAR-10 dataset, and caneven outperform vanilla supervised adversarial training for the first time. Ourcode is available at \url{https://github.com/PKU-ML/DYNACL}.

Rethinking the Effect of Data Augmentation in Adversarial Contrastive Learning