Momentum-Based Variance Reduction in Non-Convex SGD
Abstract
Variance reduction has emerged in recent years as a strong competitor to stochastic gradient descent in non-convex problems, providing the first algorithms to improve upon the converge rate of stochastic gradient descent for finding first-order critical points. However, variance reduction techniques typically require carefully tuned learning rates and willingness to use excessively large "mega-batches" in order to achieve their improved results. We present a new algorithm, STORM, that does not require any batches and makes use of adaptive learning rates, enabling simpler implementation and less hyperparameter tuning. Our technique for removing the batches uses a variant of momentum to achieve variance reduction in non-convex optimization. On smooth losses , STORM finds a point with in iterations with variance in the gradients, matching the optimal rate but without requiring knowledge of .
Keywords
Cite
@article{arxiv.1905.10018,
title = {Momentum-Based Variance Reduction in Non-Convex SGD},
author = {Ashok Cutkosky and Francesco Orabona},
journal= {arXiv preprint arXiv:1905.10018},
year = {2020}
}
Comments
Added Ack