Federated Optimization: Distributed Machine Learning for On-Device Intelligence

TL;DR

This paper introduces Federated Optimization, a new distributed learning setting where data is unevenly spread across many devices, emphasizing communication efficiency and proposing a novel algorithm suitable for sparse convex problems.

Contribution

It defines the federated optimization setting, highlights the limitations of existing algorithms, and proposes a new algorithm tailored for this scenario with promising experimental results.

Findings

Existing algorithms are unsuitable for federated optimization.

The proposed algorithm performs well on sparse convex problems.

Federated optimization sets a foundation for future research in distributed on-device learning.

Abstract

We introduce a new and increasingly relevant setting for distributed optimization in machine learning, where the data defining the optimization are unevenly distributed over an extremely large number of nodes. The goal is to train a high-quality centralized model. We refer to this setting as Federated Optimization. In this setting, communication efficiency is of the utmost importance and minimizing the number of rounds of communication is the principal goal. A motivating example arises when we keep the training data locally on users' mobile devices instead of logging it to a data center for training. In federated optimziation, the devices are used as compute nodes performing computation on their local data in order to update a global model. We suppose that we have extremely large number of devices in the network --- as many as the number of users of a given service, each of which has only a tiny fraction of the total data available. In particular, we expect the number of data points available locally to be much smaller than the number of devices. Additionally, since different users generate data with different patterns, it is reasonable to assume that no device has a representative sample of the overall distribution. We show that existing algorithms are not suitable for this setting, and propose a new algorithm which shows encouraging experimental results for sparse convex problems. This work also sets a path for future research needed in the context of \federated optimization.