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

Many existing adversarial attacks generate $L_p$-norm perturbations on imageRGB space. Despite some achievements in transferability and attack successrate, the crafted adversarial examples are easily perceived by human eyes.Towards visual imperceptibility, some recent works explore unrestricted attackswithout $L_p$-norm constraints, yet lacking transferability of attackingblack-box models. In this work, we propose a novel imperceptible andtransferable attack by leveraging both the generative and discriminative powerof diffusion models. Specifically, instead of direct manipulation in pixelspace, we craft perturbations in the latent space of diffusion models. Combinedwith well-designed content-preserving structures, we can generatehuman-insensitive perturbations embedded with semantic clues. For bettertransferability, we further "deceive" the diffusion model which can be viewedas an implicit recognition surrogate, by distracting its attention away fromthe target regions. To our knowledge, our proposed method, DiffAttack, is thefirst that introduces diffusion models into the adversarial attack field.Extensive experiments on various model structures, datasets, and defensemethods have demonstrated the superiority of our attack over the existingattack methods.

Diffusion Models for Imperceptible and Transferable Adversarial Attack