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

Deep neural networks (DNNs) have been found to be vulnerable to adversarialexamples. Adversarial examples are malicious images with visually imperceptibleperturbations. While these carefully crafted perturbations restricted withtight $\Lp$ norm bounds are small, they are still easily perceivable by humans.These perturbations also have limited success rates when attacking black-boxmodels or models with defenses like noise reduction filters. To solve theseproblems, we propose Demiguise Attack, crafting ``unrestricted'' perturbationswith Perceptual Similarity. Specifically, we can create powerful andphotorealistic adversarial examples by manipulating semantic information basedon Perceptual Similarity. Adversarial examples we generate are friendly to thehuman visual system (HVS), although the perturbations are of large magnitudes.We extend widely-used attacks with our approach, enhancing adversarialeffectiveness impressively while contributing to imperceptibility. Extensiveexperiments show that the proposed method not only outperforms variousstate-of-the-art attacks in terms of fooling rate, transferability, androbustness against defenses but can also improve attacks effectively. Inaddition, we also notice that our implementation can simulate illumination andcontrast changes that occur in real-world scenarios, which will contribute toexposing the blind spots of DNNs.

Demiguise Attack: Crafting Invisible Semantic Adversarial Perturbations with Perceptual Similarity