Recent Advances in Robust Estimation for the Huber's \epsilon contamination model

Nov 02, 3pm NSH 3305

Speaker: Simon Du

Abstract: Huber's \epsilon contamination model is a half-century-old problem in statistics. Classical approaches based on Tukey's give optimal statistical rate but is computationally infeasible. In this talk, I'll first describe polynomial time robust estimators for Gaussian mean proposed last year by theoretical computer science researchers. Next, I'll talk about our work on how to generalize the estimators to the high-dimensional setting. Last, I'll list several open problems.

References: https://arxiv.org/abs/1604.06443; https://arxiv.org/abs/1604.06968; https://arxiv.org/abs/1702.07709; https://arxiv.org/abs/1703.00893