Improving adversarial attacks on deep neural networks via constricted gradient-based perturbations

Yatie Xiao, Chi-Man Pun