Complexity of normalized stochastic first-order methods with momentum under heavy-tailed noise

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: He, Chuan, Lu, Zhaosong, Sun, Defeng, Deng, Zhanwang
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910018248376320
author He, Chuan
Lu, Zhaosong
Sun, Defeng
Deng, Zhanwang
author_facet He, Chuan
Lu, Zhaosong
Sun, Defeng
Deng, Zhanwang
contents In this paper, we propose practical normalized stochastic first-order methods with Polyak momentum, multi-extrapolated momentum, and recursive momentum for solving unconstrained optimization problems. These methods employ dynamically updated algorithmic parameters and do not require explicit knowledge of problem-dependent quantities such as the Lipschitz constant or noise bound. We establish first-order oracle complexity results for finding approximate stochastic stationary points under heavy-tailed noise and weakly average smoothness conditions -- both of which are weaker than the commonly used bounded variance and mean-squared smoothness assumptions. Our complexity bounds either improve upon or match the best-known results in the literature. Numerical experiments are presented to demonstrate the practical effectiveness of the proposed methods.
format Preprint
id arxiv_https___arxiv_org_abs_2506_11214
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Complexity of normalized stochastic first-order methods with momentum under heavy-tailed noise
He, Chuan
Lu, Zhaosong
Sun, Defeng
Deng, Zhanwang
Optimization and Control
Artificial Intelligence
Computational Complexity
Machine Learning
49M05, 49M37, 90C25, 90C30
In this paper, we propose practical normalized stochastic first-order methods with Polyak momentum, multi-extrapolated momentum, and recursive momentum for solving unconstrained optimization problems. These methods employ dynamically updated algorithmic parameters and do not require explicit knowledge of problem-dependent quantities such as the Lipschitz constant or noise bound. We establish first-order oracle complexity results for finding approximate stochastic stationary points under heavy-tailed noise and weakly average smoothness conditions -- both of which are weaker than the commonly used bounded variance and mean-squared smoothness assumptions. Our complexity bounds either improve upon or match the best-known results in the literature. Numerical experiments are presented to demonstrate the practical effectiveness of the proposed methods.
title Complexity of normalized stochastic first-order methods with momentum under heavy-tailed noise
topic Optimization and Control
Artificial Intelligence
Computational Complexity
Machine Learning
49M05, 49M37, 90C25, 90C30
url https://arxiv.org/abs/2506.11214