Skip to Main content Skip to Navigation
Conference papers

A Generic Acceleration Framework for Stochastic Composite Optimization

Andrei Kulunchakov 1 Julien Mairal 1
1 Thoth - Apprentissage de modèles à partir de données massives
Inria Grenoble - Rhône-Alpes, LJK - Laboratoire Jean Kuntzmann
Abstract : In this paper, we introduce various mechanisms to obtain accelerated first-order stochastic optimization algorithms when the objective function is convex or strongly convex. Specifically, we extend the Catalyst approach originally designed for deterministic objectives to the stochastic setting. Given an optimization method with mild convergence guarantees for strongly convex problems, the challenge is to accelerate convergence to a noise-dominated region, and then achieve convergence with an optimal worst-case complexity depending on the noise variance of the gradients. A side contribution of our work is also a generic analysis that can handle inexact proximal operators, providing new insights about the robustness of stochastic algorithms when the proximal operator cannot be exactly computed.
Complete list of metadata
Contributor : Julien Mairal Connect in order to contact the contributor
Submitted on : Monday, October 7, 2019 - 5:34:13 PM
Last modification on : Wednesday, November 3, 2021 - 8:41:38 AM


Files produced by the author(s)


  • HAL Id : hal-02139489, version 3
  • ARXIV : 1906.01164



Andrei Kulunchakov, Julien Mairal. A Generic Acceleration Framework for Stochastic Composite Optimization. NeurIPS 2019 - Thirty-third Conference Neural Information Processing Systems, Dec 2019, Vancouver, Canada. pp.12556-12567. ⟨hal-02139489v3⟩



Les métriques sont temporairement indisponibles