Back to Search Start Over

Statistically Preconditioned Accelerated Gradient Method for Distributed Optimization

Authors :
Hendrikx, Hadrien
Xiao, Lin
Bubeck, Sebastien
Bach, Francis
Massoulie, Laurent
Publication Year :
2020

Abstract

We consider the setting of distributed empirical risk minimization where multiple machines compute the gradients in parallel and a centralized server updates the model parameters. In order to reduce the number of communications required to reach a given accuracy, we propose a \emph{preconditioned} accelerated gradient method where the preconditioning is done by solving a local optimization problem over a subsampled dataset at the server. The convergence rate of the method depends on the square root of the relative condition number between the global and local loss functions. We estimate the relative condition number for linear prediction models by studying \emph{uniform} concentration of the Hessians over a bounded domain, which allows us to derive improved convergence rates for existing preconditioned gradient methods and our accelerated method. Experiments on real-world datasets illustrate the benefits of acceleration in the ill-conditioned regime.

Details

Database :
arXiv
Publication Type :
Report
Accession number :
edsarx.2002.10726
Document Type :
Working Paper