Optimal mini-batch and step sizes for SAGA

Abstract

Recently it has been shown that the step sizes of a family of variance reduced gradient methods called the JacSketch methods depend on the expected smoothness constant. In particular, if this expected smoothness constant could be calculated a priori, then one could safely set much larger step sizes which would result in a much faster convergence rate. We fill in this gap, and provide simple closed form expressions for the expected smoothness constant and careful numerical experiments verifying these bounds. Using these bounds, and since the SAGA algorithm is part of this JacSketch family, we suggest a new standard practice for setting the step sizes and mini-batch size for SAGA that are competitive with a numerical grid search. Furthermore, we can now show that the total complexity of the SAGA algorithm decreases linearly in the mini-batch size up to a pre-defined value: the optimal mini-batch size. This is a rare result in the stochastic variance reduced literature, only previously shown for the Katyusha algorithm. Finally we conjecture that this is the case for many other stochastic variance reduced methods and that our bounds and analysis of the expected smoothness constant is key to extending these results.

34 pages, 27 figures

References in corpus (3)

• SAGA: A Fast Incremental Gradient Method With Support for Non-Strongly Convex Composite Objectives
• Sharp analysis of low-rank kernel matrix approximations
• Note on sampling without replacing from a finite collection of matrices

Cited by in corpus (9)

• SGD: General Analysis and Improved Rates
• The Practicality of Stochastic Optimization in Imaging Inverse Problems
• Adaptive Gradient Descent for Convex and Non-Convex Stochastic Optimization
• One Method to Rule Them All: Variance Reduction for Data, Parameters and Many New Methods
• MISO is Making a Comeback With Better Proofs and Rates
• Leveraging Sparse Linear Layers for Debuggable Deep Networks
• SGD with Coordinate Sampling: Theory and Practice
• A Survey on Large-scale Machine Learning
• On Adaptive Stochastic Optimization for Streaming Data: A Newton's Method with O(dN) Operations