A New Type of Adversarial Examples

Xingyang Nie, Guojie Xiao, Caoliang Zhang +4

Most machine learning models are vulnerable to adversarial examples, which poses security concerns on these models. Adversarial examples are crafted by applying subtle but intentio…