Detecting adversarial examples by positive and negative representations

Wenjian Luo, Chenwang Wu, Li Ni, Nan Zhou, Zhenya Zhang