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

We study the realistic potential of conducting backdoor attack against deepneural networks (DNNs) during deployment stage. Specifically, our goal is todesign a deployment-stage backdoor attack algorithm that is both threateningand realistically implementable. To this end, we propose Subnet ReplacementAttack (SRA), which is capable of embedding backdoor into DNNs by directlymodifying a limited number of model parameters. Considering the realisticpracticability, we abandon the strong white-box assumption widely adopted inexisting studies, instead, our algorithm works in a gray-box setting, wherearchitecture information of the victim model is available but the adversariesdo not have any knowledge of parameter values. The key philosophy underlyingour approach is -- given any neural network instance (regardless of itsspecific parameter values) of a certain architecture, we can always embed abackdoor into that model instance, by replacing a very narrow subnet of abenign model (without backdoor) with a malicious backdoor subnet, which isdesigned to be sensitive (fire large activation value) to a particular backdoortrigger pattern.

Subnet Replacement: Deployment-stage backdoor attack against deep neural networks in gray-box setting