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) are susceptible to backdoor attacks, wheremalicious functionality is embedded to allow attackers to trigger incorrectclassifications. Old-school backdoor attacks use strong trigger features thatcan easily be learned by victim models. Despite robustness against inputvariation, the robustness however increases the likelihood of unintentionaltrigger activations. This leaves traces to existing defenses, which findapproximate replacements for the original triggers that can activate thebackdoor without being identical to the original trigger via, e.g., reverseengineering and sample overlay.  In this paper, we propose and investigate a new characteristic of backdoorattacks, namely, backdoor exclusivity, which measures the ability of backdoortriggers to remain effective in the presence of input variation. Building uponthe concept of backdoor exclusivity, we propose Backdoor Exclusivity LifTing(BELT), a novel technique which suppresses the association between the backdoorand fuzzy triggers to enhance backdoor exclusivity for defense evasion.Extensive evaluation on three popular backdoor benchmarks validate, ourapproach substantially enhances the stealthiness of four old-school backdoorattacks, which, after backdoor exclusivity lifting, is able to evade sixstate-of-the-art backdoor countermeasures, at almost no cost of the attacksuccess rate and normal utility. For example, one of the earliest backdoorattacks BadNet, enhanced by BELT, evades most of the state-of-the-art defensesincluding ABS and MOTH which would otherwise recognize the backdoored model.

BELT: Old-School Backdoor Attacks can Evade the State-of-the-Art Defense with Backdoor Exclusivity Lifting