Differentially Private Optimizers Can Learn Adversarially Robust Models

Yuan Zhang, Zhiqi Bu

Machine learning models have shone in a variety of domains and attracted increasing attention from both the security and the privacy communities. One important yet worrying questio…