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

The black-box adversarial attack has attracted impressive attention for itspractical use in the field of deep learning security. Meanwhile, it is verychallenging as there is no access to the network architecture or internalweights of the target model. Based on the hypothesis that if an example remainsadversarial for multiple models, then it is more likely to transfer the attackcapability to other models, the ensemble-based adversarial attack methods areefficient and widely used for black-box attacks. However, ways of ensembleattack are rather less investigated, and existing ensemble attacks simply fusethe outputs of all the models evenly. In this work, we treat the iterativeensemble attack as a stochastic gradient descent optimization process, in whichthe variance of the gradients on different models may lead to poor localoptima. To this end, we propose a novel attack method called the stochasticvariance reduced ensemble (SVRE) attack, which could reduce the gradientvariance of the ensemble models and take full advantage of the ensemble attack.Empirical results on the standard ImageNet dataset demonstrate that theproposed method could boost the adversarial transferability and outperformsexisting ensemble attacks significantly. Code is available athttps://github.com/JHL-HUST/SVRE.

Stochastic Variance Reduced Ensemble Adversarial Attack for Boosting the Adversarial Transferability