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

As a critical threat to deep neural networks (DNNs), backdoor attacks can becategorized into two types, i.e., source-agnostic backdoor attacks (SABAs) andsource-specific backdoor attacks (SSBAs). Compared to traditional SABAs, SSBAsare more advanced in that they have superior stealthier in bypassing mainstreamcountermeasures that are effective against SABAs. Nonetheless, existing SSBAssuffer from two major limitations. First, they can hardly achieve a goodtrade-off between ASR (attack success rate) and FPR (false positive rate).Besides, they can be effectively detected by the state-of-the-art (SOTA)countermeasures (e.g., SCAn). To address the limitations above, we propose anew class of viable source-specific backdoor attacks, coined as CASSOCK. Ourkey insight is that trigger designs when creating poisoned data and cover datain SSBAs play a crucial role in demonstrating a viable source-specific attack,which has not been considered by existing SSBAs. With this insight, we focus ontrigger transparency and content when crafting triggers for poisoned datasetwhere a sample has an attacker-targeted label and cover dataset where a samplehas a ground-truth label. Specifically, we implement $CASSOCK_{Trans}$ and$CASSOCK_{Cont}$. While both they are orthogonal, they are complementary toeach other, generating a more powerful attack, called $CASSOCK_{Comp}$, withfurther improved attack performance and stealthiness. We perform acomprehensive evaluation of the three $CASSOCK$-based attacks on four populardatasets and three SOTA defenses. Compared with a representative SSBA as abaseline ($SSBA_{Base}$), $CASSOCK$-based attacks have significantly advancedthe attack performance, i.e., higher ASR and lower FPR with comparable CDA(clean data accuracy). Besides, $CASSOCK$-based attacks have effectivelybypassed the SOTA defenses, and $SSBA_{Base}$ cannot.

CASSOCK: Viable Backdoor Attacks against DNN in The Wall of Source-Specific Backdoor Defences