NeurIPS 2019
Sun Dec 8th through Sat the 14th, 2019 at Vancouver Convention Center
Paper ID:5520
Title:Nearly Tight Bounds for Robust Proper Learning of Halfspaces with a Margin


		
This work makes significant progress on a relatively natural theoretical problem of proper agnostic learning halfspaces with a margin. The paper is technically strong and is a clear accept. A downside is that the problem and proposed algorithms appear to be largely irrelevant to practice.