English

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

Machine Learning 2024-09-24 v1 Numerical Analysis Numerical Analysis Optimization and Control

Abstract

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 O(1/T)\mathcal{O}(1/T). Furthermore, the size of neighborhood decreases as β\beta 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.

Keywords

Cite

@article{arxiv.2305.05675,
  title  = {UAdam: Unified Adam-Type Algorithmic Framework for Non-Convex Stochastic Optimization},
  author = {Yiming Jiang and Jinlan Liu and Dongpo Xu and Danilo P. Mandic},
  journal= {arXiv preprint arXiv:2305.05675},
  year   = {2024}
}
R2 v1 2026-06-28T10:30:18.182Z