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

Existing works have identified the limitation of top-$1$ attack success rate(ASR) as a metric to evaluate the attack strength but exclusively investigatedit in the white-box setting, while our work extends it to a more practicalblack-box setting: transferable attack. It is widely reported that strongerI-FGSM transfers worse than simple FGSM, leading to a popular belief thattransferability is at odds with the white-box attack strength. Our workchallenges this belief with empirical finding that stronger attack actuallytransfers better for the general top-$k$ ASR indicated by the interest classrank (ICR) after attack. For increasing the attack strength, with an intuitiveinterpretation of the logit gradient from the geometric perspective, weidentify that the weakness of the commonly used losses lie in prioritizing thespeed to fool the network instead of maximizing its strength. To this end, wepropose a new normalized CE loss that guides the logit to be updated in thedirection of implicitly maximizing its rank distance from the ground-truthclass. Extensive results in various settings have verified that our proposednew loss is simple yet effective for top-$k$ attack. Code is available at:\url{https://bit.ly/3uCiomP}

Investigating Top-$k$ White-Box and Transferable Black-box Attack