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/

Zendy Plus

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

Zendy Tools

Zendy Open

Neural networks are vulnerable to adversarial attacks: adding well-crafted,imperceptible perturbations to their input can modify their output. Adversarialtraining is one of the most effective approaches in training robust modelsagainst such attacks. However, it is much slower than vanilla training ofneural networks since it needs to construct adversarial examples for the entiretraining data at every iteration, hampering its effectiveness. Recently, FastAdversarial Training (FAT) was proposed that can obtain robust modelsefficiently. However, the reasons behind its success are not fully understood,and more importantly, it can only train robust models for $\ell_\infty$-boundedattacks as it uses FGSM during training. In this paper, by leveraging thetheory of coreset selection, we show how selecting a small subset of trainingdata provides a general, more principled approach toward reducing the timecomplexity of robust training. Unlike existing methods, our approach can beadapted to a wide variety of training objectives, including TRADES,$\ell_p$-PGD, and Perceptual Adversarial Training (PAT). Our experimentalresults indicate that our approach speeds up adversarial training by 2-3 timeswhile experiencing a slight reduction in the clean and robust accuracy.

$\ell_\infty$-Robustness and Beyond: Unleashing Efficient Adversarial Training