Safe Machine Learning and Defeating Adversarial Attacks

Adversarial attacks have exposed the unreliability of machine learning models for decision making in autonomous agents. This article discusses recent research for deep learning model assurance in face of adversarial attacks.