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

In this study, we propose a new methodology to control how user's data isrecognized and used by AI via exploiting the properties of adversarialexamples. For this purpose, we propose reversible adversarial example (RAE), anew type of adversarial example. A remarkable feature of RAE is that the imagecan be correctly recognized and used by the AI model specified by the userbecause the authorized AI can recover the original image from the RAE exactlyby eliminating adversarial perturbation. On the other hand, other unauthorizedAI models cannot recognize it correctly because it functions as an adversarialexample. Moreover, RAE can be considered as one type of encryption to computervision since reversibility guarantees the decryption. To realize RAE, wecombine three technologies, adversarial example, reversible data hiding forexact recovery of adversarial perturbation, and encryption for selectivecontrol of AIs who can remove adversarial perturbation. Experimental resultsshow that the proposed method can achieve comparable attack ability with thecorresponding adversarial attack method and similar visual quality with theoriginal image, including white-box attacks and black-box attacks.

Unauthorized AI cannot Recognize Me: Reversible Adversarial Example