1 paper · 1 filter
Jiacheng Guo, Tianyun Zhang, Lei Li +3
Deep neural networks are highly vulnerable to adversarial perturbations, raising serious safety concerns in the real-world systems. While prior work mainly explores single-task att…