Master Thesis / Project
Recent advances in machine learning and specifically deep learning enabled the deployment of automatic agents that interact with humans on a large scale. However, the vulnerability of deep learning algorithms to small adversarial perturbations remains an open problem. Thus, ...