DeBackdoor: A Deductive Framework for Detecting Backdoor Attacks on Deep Models with Limited Data

Name: DeBackdoor: A Deductive Framework for Detecting Backdoor Attacks on Deep Models with Limited Data
Duration: 15 min
Description: Talk from 34th USENIX Security Symposium (USENIX Security '25).

Dorde Popovic

34th USENIX Security Symposium (USENIX Security '25) · Day 3 · ML and AI Security 3: Backdoors, Poisoning, Unlearning