UAdam: Unified Adam-Type Algorithmic Framework for Non-Convex Stochastic Optimization

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Jiang, Yiming, Liu, Jinlan, Xu, Dongpo, Mandic, Danilo P.
Format: Preprint
Publié: 2023
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866917780607991808
author Jiang, Yiming
Liu, Jinlan
Xu, Dongpo
Mandic, Danilo P.
author_facet Jiang, Yiming
Liu, Jinlan
Xu, Dongpo
Mandic, Danilo P.
contents Adam-type algorithms have become a preferred choice for optimisation in the deep learning setting, however, despite success, their convergence is still not well understood. To this end, we introduce a unified framework for Adam-type algorithms (called UAdam). This is equipped with a general form of the second-order moment, which makes it possible to include Adam and its variants as special cases, such as NAdam, AMSGrad, AdaBound, AdaFom, and Adan. This is supported by a rigorous convergence analysis of UAdam in the non-convex stochastic setting, showing that UAdam converges to the neighborhood of stationary points with the rate of $\mathcal{O}(1/T)$. Furthermore, the size of neighborhood decreases as $β$ increases. Importantly, our analysis only requires the first-order momentum factor to be close enough to 1, without any restrictions on the second-order momentum factor. Theoretical results also show that vanilla Adam can converge by selecting appropriate hyperparameters, which provides a theoretical guarantee for the analysis, applications, and further developments of the whole class of Adam-type algorithms.
format Preprint
id arxiv_https___arxiv_org_abs_2305_05675
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle UAdam: Unified Adam-Type Algorithmic Framework for Non-Convex Stochastic Optimization
Jiang, Yiming
Liu, Jinlan
Xu, Dongpo
Mandic, Danilo P.
Machine Learning
Numerical Analysis
Optimization and Control
Adam-type algorithms have become a preferred choice for optimisation in the deep learning setting, however, despite success, their convergence is still not well understood. To this end, we introduce a unified framework for Adam-type algorithms (called UAdam). This is equipped with a general form of the second-order moment, which makes it possible to include Adam and its variants as special cases, such as NAdam, AMSGrad, AdaBound, AdaFom, and Adan. This is supported by a rigorous convergence analysis of UAdam in the non-convex stochastic setting, showing that UAdam converges to the neighborhood of stationary points with the rate of $\mathcal{O}(1/T)$. Furthermore, the size of neighborhood decreases as $β$ increases. Importantly, our analysis only requires the first-order momentum factor to be close enough to 1, without any restrictions on the second-order momentum factor. Theoretical results also show that vanilla Adam can converge by selecting appropriate hyperparameters, which provides a theoretical guarantee for the analysis, applications, and further developments of the whole class of Adam-type algorithms.
title UAdam: Unified Adam-Type Algorithmic Framework for Non-Convex Stochastic Optimization
topic Machine Learning
Numerical Analysis
Optimization and Control
url https://arxiv.org/abs/2305.05675