Towards Federated Learning at Scale: System Design

  • 2019-02-04 06:27:41
  • Keith Bonawitz, Hubert Eichner, Wolfgang Grieskamp, Dzmitry Huba, Alex Ingerman, Vladimir Ivanov, Chloe Kiddon, Jakub Konecny, Stefano Mazzocchi, H. Brendan McMahan, Timon Van Overveldt, David Petrou, Daniel Ramage, Jason Roselander
  • 59

Abstract

Federated Learning is a distributed machine learning approach which enablesmodel training on a large corpus of decentralized data. We have built ascalable production system for Federated Learning in the domain of mobiledevices, based on TensorFlow. In this paper, we describe the resultinghigh-level design, sketch some of the challenges and their solutions, and touchupon the open problems and future directions.

 

Quick Read (beta)

loading the full paper ...