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

The Universal Trigger (UniTrigger) is a recently-proposed powerfuladversarial textual attack method. Utilizing a learning-based mechanism,UniTrigger generates a fixed phrase that, when added to any benign inputs, candrop the prediction accuracy of a textual neural network (NN) model to nearzero on a target class. To defend against this attack that can causesignificant harm, in this paper, we borrow the "honeypot" concept from thecybersecurity community and propose DARCY, a honeypot-based defense frameworkagainst UniTrigger. DARCY greedily searches and injects multiple trapdoors intoan NN model to "bait and catch" potential attacks. Through comprehensiveexperiments across four public datasets, we show that DARCY detectsUniTrigger's adversarial attacks with up to 99% TPR and less than 2% FPR inmost cases, while maintaining the prediction accuracy (in F1) for clean inputswithin a 1% margin. We also demonstrate that DARCY with multiple trapdoors isalso robust to a diverse set of attack scenarios with attackers' varying levelsof knowledge and skills. Source code will be released upon the acceptance ofthis paper.

A Sweet Rabbit Hole by DARCY: Using Honeypots to Detect Universal Trigger's Adversarial Attacks