Gradient Descent Only Converges to Minimizers
Abstract
We show that gradient descent converges to a local minimizer, almost surely with random initial- ization. This is proved by applying the Stable Manifold Theorem from dynamical systems theory.
Cite
Text
Lee et al. "Gradient Descent Only Converges to Minimizers." Annual Conference on Computational Learning Theory, 2016.Markdown
[Lee et al. "Gradient Descent Only Converges to Minimizers." Annual Conference on Computational Learning Theory, 2016.](https://mlanthology.org/colt/2016/lee2016colt-gradient/)BibTeX
@inproceedings{lee2016colt-gradient,
title = {{Gradient Descent Only Converges to Minimizers}},
author = {Lee, Jason D. and Simchowitz, Max and Jordan, Michael I. and Recht, Benjamin},
booktitle = {Annual Conference on Computational Learning Theory},
year = {2016},
pages = {1246-1257},
url = {https://mlanthology.org/colt/2016/lee2016colt-gradient/}
}