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 advancements in CNNs have shown remarkable achievements in variousCV/AI applications. Though CNNs show near human or better than humanperformance in many critical tasks, they are quite vulnerable to adversarialattacks. These attacks are potentially dangerous in real-life deployments.Though there have been many adversarial attacks proposed in recent years, thereis no proper way of quantifying the effectiveness of these attacks. As oftoday, mere fooling rate is used for measuring the susceptibility of themodels, or the effectiveness of adversarial attacks. Fooling rate justconsiders label flipping and does not consider the cost of such flipping, forinstance, in some deployments, flipping between two species of dogs may not beas severe as confusing a dog category with that of a vehicle. Therefore, themetric to quantify the vulnerability of the models should capture the severityof the flipping as well. In this work we first bring out the drawbacks of theexisting evaluation and propose novel metrics to capture various aspects of thefooling. Further, for the first time, we present a comprehensive analysis ofseveral important adversarial attacks over a set of distinct CNN architectures.We believe that the presented analysis brings valuable insights about thecurrent adversarial attacks and the CNN models.

Adversarial Fooling Beyond "Flipping the Label"