Stochastic convergence of parallel asynchronous adaptive first-order methods
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866913178151026688 |
|---|---|
| author | Gratton, Serge Toint, Philippe L. |
| author_facet | Gratton, Serge Toint, Philippe L. |
| contents | A new class of asynchronous adaptive first-order optimization methods is introduced, comprising asynchronous variants of several popular algorithms. Versions of these methods using momentum and/or inexact normalization are also considered. The convergence of methods in the class on non-convex functions is analyzed in a fully stochastic setting, and is shown to be (up to logarithmic factors) of order O(1/sqrt{t}) under reasonable assumptions. Numerical experiments suggest that such asynchronous adaptive algorithms are very relevant in heterogeneous large-scale machine learning systems. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2606_01787 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Stochastic convergence of parallel asynchronous adaptive first-order methods Gratton, Serge Toint, Philippe L. Artificial Intelligence Optimization and Control 90C26, 90C06, 90C15, 90C30 I.2.6; F.2.1; G.1.6 A new class of asynchronous adaptive first-order optimization methods is introduced, comprising asynchronous variants of several popular algorithms. Versions of these methods using momentum and/or inexact normalization are also considered. The convergence of methods in the class on non-convex functions is analyzed in a fully stochastic setting, and is shown to be (up to logarithmic factors) of order O(1/sqrt{t}) under reasonable assumptions. Numerical experiments suggest that such asynchronous adaptive algorithms are very relevant in heterogeneous large-scale machine learning systems. |
| title | Stochastic convergence of parallel asynchronous adaptive first-order methods |
| topic | Artificial Intelligence Optimization and Control 90C26, 90C06, 90C15, 90C30 I.2.6; F.2.1; G.1.6 |
| url | https://arxiv.org/abs/2606.01787 |