Efficiently Learning Adversarially Robust Halfspaces with Noise

Part of Proceedings of the International Conference on Machine Learning 1 pre-proceedings (ICML 2020)

Bibtex »Metadata »Paper »

Bibtek download is not availble in the pre-proceeding


Omar Montasser, Surbhi Goel, Ilias Diakonikolas, Nati Srebro


We study the problem of learning adversarially robust halfspaces in the distribution-independent setting. We give the first computationally efficient algorithm for this problem in the realizable setting and in the presence of random label noise with respect to any $\ell_p$-perturbation (and, more generally, perturbations with respect to any norm).