NeurIPS 2019
Sun Dec 8th through Sat the 14th, 2019 at Vancouver Convention Center
Paper ID:1283
Title:Order Optimal One-Shot Distributed Learning


		
This paper provides exactly what the title suggests. Theoretically, this is a solid contribution and would be of interest to researchers in statistical learning. The reviewers felt that the author response clarified most of their technical questions. However, a number of improvements could be made to the manuscript, particularly in the related work, some additional experimental validation, and a more expanded discussion of the implications of the work in terms of network resources. In the end, the story here is clean, compelling, and has some novel insight for a problem setting that has become and will remain important in the coming years.