Communication-efficient distributed optimization of self-concordant empirical loss

Yuchen Zhang, Lin Xiao

Research output: Chapter in Book/Report/Conference proceedingChapterpeer-review

9 Scopus citations

Abstract

We consider distributed convex optimization problems originating from sample average approximation of stochastic optimization, or empirical risk minimization in machine learning. We assume that each machine in the distributed computing system has access to a local empirical loss function, constructed with i.i.d. data sampled from a common distribution. We propose a communication-efficient distributed algorithm to minimize the overall empirical loss, which is the average of the local empirical losses. The algorithm is based on an inexact damped Newton method, where the inexact Newton steps are computed by a distributed preconditioned conjugate gradient method. We analyze its iteration complexity and communication efficiency for minimizing self-concordant empirical loss functions, and discuss the results for ridge regression, logistic regression and binary classification with a smoothed hinge loss. In a standard setting for supervised learning where the condition number of the problem grows with square root of the sample size, the required number of communication rounds of the algorithm does not increase with the sample size, and only grows slowly with the number of machines.

Original languageEnglish
Title of host publicationLecture Notes in Mathematics
PublisherSpringer Verlag
Pages289-341
Number of pages53
DOIs
StatePublished - 2018
Externally publishedYes

Publication series

NameLecture Notes in Mathematics
Volume2227
ISSN (Print)0075-8434

Keywords

  • Distributed optimization
  • Empirical risk minimization
  • Inexact Newton methods
  • Self-concordant functions

Fingerprint

Dive into the research topics of 'Communication-efficient distributed optimization of self-concordant empirical loss'. Together they form a unique fingerprint.

Cite this