Stochastic Reweighted Gradient Descent

Abstract

Despite the strong theoretical guarantees that variance-reduced finite-sum optimization algorithms enjoy, their applicability remains limited to cases where the memory overhead they introduce (SAG/SAGA), or the periodic full gradient computation they require (SVRG/SARAH) are manageable. A promising approach to achieving variance reduction while avoiding these drawbacks is the use of importance sampling instead of control variates. While many such methods have been proposed in the literature, directly proving that they improve the convergence of the resulting optimization algorithm has remained elusive. In this work, we propose an importance-sampling-based algorithm we call SRG (stochastic reweighted gradient). We analyze the convergence of SRG in the strongly-convex case and show that, while it does not recover the linear rate of control variates methods, it provably outperforms SGD. We pay particular attention to the time and memory overhead of our proposed method, and design a specialized red-black tree allowing its efficient implementation. Finally, we present empirical results to support our findings.

References in corpus (7)

• SAGA: A Fast Incremental Gradient Method With Support for Non-Strongly Convex Composite Objectives
• A Stochastic Gradient Method with an Exponential Convergence Rate for Finite Training Sets
• Stochastic Optimization with Bandit Sampling
• On Convergence-Diagnostic based Step Sizes for Stochastic Gradient Descent
• Online Variance Reduction with Mixtures
• Adaptive Importance Sampling for Finite-Sum Optimization and Sampling with Decreasing Step-Sizes
• Limitations on Variance-Reduction and Acceleration Schemes for Finite Sum Optimization