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

A major drawback of adversarially robust models, in particular for largescale datasets like ImageNet, is the extremely long training time compared tostandard ones. Moreover, models should be robust not only to one $l_p$-threatmodel but ideally to all of them. In this paper we propose Extreme normAdversarial Training (E-AT) for multiple-norm robustness which is based ongeometric properties of $l_p$-balls. E-AT costs up to three times less thanother adversarial training methods for multiple-norm robustness. Using E-AT weshow that for ImageNet a single epoch and for CIFAR-10 three epochs aresufficient to turn any $l_p$-robust model into a multiple-norm robust model. Inthis way we get the first multiple-norm robust model for ImageNet and boost thestate-of-the-art for multiple-norm robustness to more than $51\%$ on CIFAR-10.Finally, we study the general transfer via fine-tuning of adversarialrobustness between different individual $l_p$-threat models and improve theprevious SOTA $l_1$-robustness on both CIFAR-10 and ImageNet. Extensiveexperiments show that our scheme works across datasets and architecturesincluding vision transformers.

Adversarial Robustness against Multiple and Single $l_p$-Threat Models via Quick Fine-Tuning of Robust Classifiers