Accelerated Gradient Descent via Long Steps
Abstract
Recently Grimmer [1] showed for smooth convex optimization by utilizing longer steps periodically, gradient descent's textbook convergence guarantees can be improved by constant factors, conjecturing an accelerated rate strictly faster than could be possible. Here we prove such a big-O gain, establishing gradient descent's first accelerated convergence rate in this setting. Namely, we prove a rate for smooth convex minimization by utilizing a nonconstant nonperiodic sequence of increasingly large stepsizes. It remains open if one can achieve the rate conjectured by Das Gupta et. al. [2] or the optimal gradient method rate of . Big-O convergence rate accelerations from long steps follow from our theory for strongly convex optimization, similar to but somewhat weaker than those concurrently developed by Altschuler and Parrilo [3].
Cite
@article{arxiv.2309.09961,
title = {Accelerated Gradient Descent via Long Steps},
author = {Benjamin Grimmer and Kevin Shu and Alex L. Wang},
journal= {arXiv preprint arXiv:2309.09961},
year = {2023}
}