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 are highly vulnerable to adversarial examples, whichimposes severe security issues for these state-of-the-art models. Many defensemethods have been proposed to mitigate this problem. However, a lot of themdepend on modification or additional training of the target model. In thiswork, we analytically investigate each layer's representation of non-perturbedand perturbed images and show the effect of perturbations on each of theserepresentations. Accordingly, a method based on whitening coloring transform isproposed in order to diminish the misrepresentation of any desirable layercaused by adversaries. Our method can be applied to any layer of any arbitrarymodel without the need of any modification or additional training. Due to thefact that the full whitening of the layer's representation is not easilydifferentiable, our proposed method is superbly robust against white-boxattacks. Furthermore, we demonstrate the strength of our method against somestate-of-the-art black-box attacks.

Diminishing the Effect of Adversarial Perturbations via Refining Feature Representation