Adversarial Vector Machines

Proposed and Implemented a Learning framework for optimal prediction confidence and theoretical guarantees for complete robustness to norm-bounded adversarial perturbations on the training data, and adversarial sample detection at test time.