CoopHance: Cooperative Enhancement for Robustness of Deep Learning Systems

Adversarial attacks have been a threat to Deep Learning (DL) systems to be reckoned with. By adding human-imperceptible perturbation to benign inputs, adversarial attacks can cause the incorrect behavior of DL systems. Considering the popularity of DL systems in the industry, it is critical and urgent for developers to enhance the robustness of DL systems against adversarial attacks.