1 paper
Roie Kazoom, Alon Goldberg, Hodaya Cohen +1
Adversarial patch attacks pose a severe threat to deep neural networks, yet most existing approaches rely on unrealistic white-box assumptions, untargeted objectives, or produce vi…