Trajectory of Mini-Batch Momentum: Batch Size Saturation and Convergence in High Dimensions

Kiwon Lee, Andrew Nicholas Cheng, Elliot Paquette, Courtney Paquette

Keywords: Stochastic Gradient Descent with Momentum, Random Matrix Theory, High Dimensional Probability

TL;DR: We provide exact dynamics for SGD with Momentum (SGD+M) in large scale and we show that SGD+M converges faster than SGD in large batch setting.

Abstract: We analyze the dynamics of large batch stochastic gradient descent with momentum (SGD+M) on the least squares problem when both the number of samples and dimensions are large. In this setting, we show that the dynamics of SGD+M converge to a deterministic discrete Volterra equation as dimension increases, which we analyze. We identify a stability measurement, the implicit conditioning ratio (ICR), which regulates the ability of SGD+M to accelerate the algorithm. When the batch size exceeds this ICR, SGD+M converges linearly at a rate of $\mathcal{O}(1/\sqrt{\kappa})$, matching optimal full-batch momentum (in particular performing as well as a full-batch but with a fraction of the size). For batch sizes smaller than the ICR, in contrast, SGD+M has rates that scale like a multiple of the single batch SGD rate. We give explicit choices for the learning rate and momentum parameter in terms of the Hessian spectra that achieve this performance.