Sparse and Imperceptible Adversarial Attack via a Homotopy Algorithm
Mingkang Zhu 1 Tianlong Chen 1 Zhangyang Wang 1
Abstract
Sparse adversarial attacks can fool deep neural
networks (DNNs) by only perturbing a few pixels (a) otterhound (b) stingray (c) `0 = 1803 (d) `0 = 12264
(regularized by `0 norm). Recent efforts combine
it with another `∞ imperceptible on the pertur-
bation magnitudes. The resultant sparse and im-
perceptibl ...


雷达卡




京公网安备 11010802022788号







