Fast Distributed k-Means with a Small Number of Rounds

Tom Hess, Ron Visbord, Sivan Sabato

We propose a new algorithm for k-means clustering in a distributed setting, where the data is distributed across many machines, and a coordinator communicates with these machines t…