Mini-batch stochastic gradient descent with dynamic sample sizes
Optimization and Control
2017-08-03 v1
Abstract
We focus on solving constrained convex optimization problems using mini-batch stochastic gradient descent. Dynamic sample size rules are presented which ensure a descent direction with high probability. Empirical results from two applications show superior convergence compared to fixed sample implementations.
Cite
@article{arxiv.1708.00555,
title = {Mini-batch stochastic gradient descent with dynamic sample sizes},
author = {Michael R. Metel},
journal= {arXiv preprint arXiv:1708.00555},
year = {2017}
}