Catalyst Acceleration for First-order Convex Optimization: from Theory to Practice
Abstract
We introduce a generic scheme for accelerating gradient-based optimization methods in the sense of Nesterov. The approach, called Catalyst, builds upon the inexact accelerated proximal point algorithm for minimizing a convex objective function, and consists of approximately solving a sequence of well-chosen auxiliary problems, leading to faster convergence. One of the keys to achieve acceleration in theory and in practice is to solve these sub-problems with appropriate accuracy by using the right stopping criterion and the right warm-start strategy. We give practical guidelines to use Catalyst and present a comprehensive analysis of its global complexity. We show that Catalyst applies to a large class of algorithms, including gradient descent, block coordinate descent, incremental algorithms such as SAG, SAGA, SDCA, SVRG, MISO/Finito, and their proximal variants. For all of these methods, we establish faster rates using the Catalyst acceleration, for strongly convex and non-strongly convex objectives. We conclude with extensive experiments showing that acceleration is useful in practice, especially for ill-conditioned problems.
Keywords
Cite
@article{arxiv.1712.05654,
title = {Catalyst Acceleration for First-order Convex Optimization: from Theory to Practice},
author = {Hongzhou Lin and Julien Mairal and Zaid Harchaoui},
journal= {arXiv preprint arXiv:1712.05654},
year = {2018}
}
Comments
link to publisher website: http://jmlr.org/papers/volume18/17-748/17-748.pdf