Linked Presentation: PELICAN: Exploiting Backdoors of Naturally Trained Deep Learning Models In Binary Code AnalysisHard-label Black-box Universal Adversarial Patch Attack
