Convergence rates for the stochastic gradient descent method for non-convex objective functions
Numerical Analysis
2021-11-02 v2 Machine Learning
Numerical Analysis
Probability
Machine Learning
Abstract
We prove the local convergence to minima and estimates on the rate of convergence for the stochastic gradient descent method in the case of not necessarily globally convex nor contracting objective functions. In particular, the results are applicable to simple objective functions arising in machine learning.
Cite
@article{arxiv.1904.01517,
title = {Convergence rates for the stochastic gradient descent method for non-convex objective functions},
author = {Benjamin Fehrman and Benjamin Gess and Arnulf Jentzen},
journal= {arXiv preprint arXiv:1904.01517},
year = {2021}
}
Comments
59 pages