Comment on "First-order methods almost always avoid strict saddle points"
Optimization and Control
2022-04-04 v1
Abstract
The analysis on the global stability of Riemannian gradient descent method in manifold optimization (i.e., it avoids strict saddle points for almost all initializations) due to Lee et al. (Math. Program. 176:311-337) is corrected. Moreover, an explicit bound on the step-size is presented by the newly introduced retraction L-smooth property.
Keywords
Cite
@article{arxiv.2204.00521,
title = {Comment on "First-order methods almost always avoid strict saddle points"},
author = {Jinyang Zheng and Yong Xia},
journal= {arXiv preprint arXiv:2204.00521},
year = {2022}
}
Comments
7 pages