Statistics Papers


Making Gradient Descent Optimal for Strongly Convex Stochastic Optimization

Document Type

Conference Paper

Date of this Version


Publication Source

Proceedings of the 29th International Conference on Machine Learning


Stochastic gradient descent (SGD) is a simple and popular method to solve stochastic optimization problems which arise in machine learning. For strongly convex problems, its convergence rate was known to be O(log(T/T ), by running SGD for T iterations and returning the average point. How- ever, recent results showed that using a different algorithm, one can get an optimal O(1/T ) rate. This might lead one to believe that standard SGD is suboptimal, and maybe should even be replaced as a method of choice. In this paper, we investigate the optimality of SGD in a stochastic setting. We show that for smooth problems, the algorithm attains the optimal O(1/T) rate. However, for non-smooth problems, the convergence rate with averaging might really be Ω (log(T)/T ), and this is not just an artifact of the analysis. On the flip side, we show that a simple modification of the averaging step success to recover the O(1/T ) rate, and no other change of the algorithm is necessary. We also present experimental results which support our findings, and point out open problems.

This document is currently not available here.



Date Posted: 27 November 2017