Last-Iterate Complexity of SGD for Convex and Smooth Stochastic Problems

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Garrigos, Guillaume, Cortild, Daniel, Ketels, Lucas, Peypouquet, Juan
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915398215008256
author Garrigos, Guillaume
Cortild, Daniel
Ketels, Lucas
Peypouquet, Juan
author_facet Garrigos, Guillaume
Cortild, Daniel
Ketels, Lucas
Peypouquet, Juan
contents Most results on Stochastic Gradient Descent (SGD) in the convex and smooth setting are presented under the form of bounds on the ergodic function value gap. It is an open question whether bounds can be derived directly on the last iterate of SGD in this context. Recent advances suggest that it should be possible. For instance, it can be achieved by making the additional, yet unverifiable, assumption that the variance of the stochastic gradients is uniformly bounded. In this paper, we show that there is no need of such an assumption, and that SGD enjoys a $\tilde O \left( T^{-1/2} \right)$ last-iterate complexity rate for convex smooth stochastic problems.
format Preprint
id arxiv_https___arxiv_org_abs_2507_14122
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Last-Iterate Complexity of SGD for Convex and Smooth Stochastic Problems
Garrigos, Guillaume
Cortild, Daniel
Ketels, Lucas
Peypouquet, Juan
Optimization and Control
Most results on Stochastic Gradient Descent (SGD) in the convex and smooth setting are presented under the form of bounds on the ergodic function value gap. It is an open question whether bounds can be derived directly on the last iterate of SGD in this context. Recent advances suggest that it should be possible. For instance, it can be achieved by making the additional, yet unverifiable, assumption that the variance of the stochastic gradients is uniformly bounded. In this paper, we show that there is no need of such an assumption, and that SGD enjoys a $\tilde O \left( T^{-1/2} \right)$ last-iterate complexity rate for convex smooth stochastic problems.
title Last-Iterate Complexity of SGD for Convex and Smooth Stochastic Problems
topic Optimization and Control
url https://arxiv.org/abs/2507.14122