Improved Learning Rates for Stochastic Optimization

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Li, Shaojie, Tang, Pengwei, Liu, Yong
Format: Preprint
Published: 2021
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915872972472320
author Li, Shaojie
Tang, Pengwei
Liu, Yong
author_facet Li, Shaojie
Tang, Pengwei
Liu, Yong
contents Stochastic optimization is a cornerstone of modern machine learning. This paper studies the generalization performance of two classical stochastic optimization algorithms: stochastic gradient descent (SGD) and Nesterov's accelerated gradient (NAG). We establish new learning rates for both algorithms, with improved guarantees in some settings or comparable rates under weaker assumptions in others. We also provide numerical experiments to support the theory.
format Preprint
id arxiv_https___arxiv_org_abs_2107_08686
institution arXiv
publishDate 2021
record_format arxiv
spellingShingle Improved Learning Rates for Stochastic Optimization
Li, Shaojie
Tang, Pengwei
Liu, Yong
Machine Learning
Optimization and Control
Stochastic optimization is a cornerstone of modern machine learning. This paper studies the generalization performance of two classical stochastic optimization algorithms: stochastic gradient descent (SGD) and Nesterov's accelerated gradient (NAG). We establish new learning rates for both algorithms, with improved guarantees in some settings or comparable rates under weaker assumptions in others. We also provide numerical experiments to support the theory.
title Improved Learning Rates for Stochastic Optimization
topic Machine Learning
Optimization and Control
url https://arxiv.org/abs/2107.08686