Momentum via Primal Averaging: Theoretical Insights and Learning Rate Schedules for Non-Convex Optimization
Machine Learning
2021-06-02 v4 Optimization and Control
Machine Learning
Abstract
Momentum methods are now used pervasively within the machine learning community for training non-convex models such as deep neural networks. Empirically, they out perform traditional stochastic gradient descent (SGD) approaches. In this work we develop a Lyapunov analysis of SGD with momentum (SGD+M), by utilizing a equivalent rewriting of the method known as the stochastic primal averaging (SPA) form. This analysis is much tighter than previous theory in the non-convex case, and due to this we are able to give precise insights into when SGD+M may out-perform SGD, and what hyper-parameter schedules will work and why.
Keywords
Cite
@article{arxiv.2010.00406,
title = {Momentum via Primal Averaging: Theoretical Insights and Learning Rate Schedules for Non-Convex Optimization},
author = {Aaron Defazio},
journal= {arXiv preprint arXiv:2010.00406},
year = {2021}
}