English

Convergence of Adam for Non-convex Objectives: Relaxed Hyperparameters and Non-ergodic Case

Optimization and Control 2025-02-12 v1 Machine Learning Numerical Analysis Numerical Analysis

Abstract

Adam is a commonly used stochastic optimization algorithm in machine learning. However, its convergence is still not fully understood, especially in the non-convex setting. This paper focuses on exploring hyperparameter settings for the convergence of vanilla Adam and tackling the challenges of non-ergodic convergence related to practical application. The primary contributions are summarized as follows: firstly, we introduce precise definitions of ergodic and non-ergodic convergence, which cover nearly all forms of convergence for stochastic optimization algorithms. Meanwhile, we emphasize the superiority of non-ergodic convergence over ergodic convergence. Secondly, we establish a weaker sufficient condition for the ergodic convergence guarantee of Adam, allowing a more relaxed choice of hyperparameters. On this basis, we achieve the almost sure ergodic convergence rate of Adam, which is arbitrarily close to o(1/K)o(1/\sqrt{K}). More importantly, we prove, for the first time, that the last iterate of Adam converges to a stationary point for non-convex objectives. Finally, we obtain the non-ergodic convergence rate of O(1/K)O(1/K) for function values under the Polyak-Lojasiewicz (PL) condition. These findings build a solid theoretical foundation for Adam to solve non-convex stochastic optimization problems.

Keywords

Cite

@article{arxiv.2307.11782,
  title  = {Convergence of Adam for Non-convex Objectives: Relaxed Hyperparameters and Non-ergodic Case},
  author = {Meixuan He and Yuqing Liang and Jinlan Liu and Dongpo Xu},
  journal= {arXiv preprint arXiv:2307.11782},
  year   = {2025}
}