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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Deep speech classification has achieved tremendous success and greatlypromoted the emergence of many real-world applications. However, backdoorattacks present a new security threat to it, particularly with untrustworthythird-party platforms, as pre-defined triggers set by the attacker can activatethe backdoor. Most of the triggers in existing speech backdoor attacks aresample-agnostic, and even if the triggers are designed to be unnoticeable, theycan still be audible. This work explores a backdoor attack that utilizessample-specific triggers based on voice conversion. Specifically, we adopt apre-trained voice conversion model to generate the trigger, ensuring that thepoisoned samples does not introduce any additional audible noise. Extensiveexperiments on two speech classification tasks demonstrate the effectiveness ofour attack. Furthermore, we analyzed the specific scenarios that activated theproposed backdoor and verified its resistance against fine-tuning.

Fake the Real: Backdoor Attack on Deep Speech Classification via Voice Conversion