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

With the excellent accuracy and feasibility, the Neural Networks have beenwidely applied into the novel intelligent applications and systems. However,with the appearance of the Adversarial Attack, the NN based system performancebecomes extremely vulnerable:the image classification results can bearbitrarily misled by the adversarial examples, which are crafted images withhuman unperceivable pixel-level perturbation. As this raised a significantsystem security issue, we implemented a series of investigations on theadversarial attack in this work: We first identify an image's pixelvulnerability to the adversarial attack based on the adversarial saliencyanalysis. By comparing the analyzed saliency map and the adversarialperturbation distribution, we proposed a new evaluation scheme tocomprehensively assess the adversarial attack precision and efficiency. Then,with a novel adversarial saliency prediction method, a fast adversarial examplegeneration framework, namely "ASP", is proposed with significant attackefficiency improvement and dramatic computation cost reduction. Compared to theprevious methods, experiments show that ASP has at most 12 times speed-up foradversarial example generation, 2 times lower perturbation rate, and highattack success rate of 87% on both MNIST and Cifar10. ASP can be also wellutilized to support the data-hungry NN adversarial training. By reducing theattack success rate as much as 90%, ASP can quickly and effectively enhance thedefense capability of NN based system to the adversarial attacks.

ASP:A Fast Adversarial Attack Example Generation Framework based on Adversarial Saliency Prediction