A Parameter-Free and Near-Optimal Zeroth-Order Algorithm for Stochastic Convex Optimization
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866918008033640448 |
|---|---|
| author | Ren, Kunjie Luo, Luo |
| author_facet | Ren, Kunjie Luo, Luo |
| contents | This paper considers zeroth-order optimization for stochastic convex minimization problem. We propose a parameter-free stochastic zeroth-order method (POEM) by introducing a step-size scheme based on the distance over finite difference and an adaptive smoothing parameter. We provide the theoretical analysis to show that POEM achieves the near-optimal stochastic zeroth-order oracle complexity. We further conduct the numerical experiments to demonstrate POEM outperforms existing zeroth-order methods in practice. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2502_05600 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | A Parameter-Free and Near-Optimal Zeroth-Order Algorithm for Stochastic Convex Optimization Ren, Kunjie Luo, Luo Optimization and Control This paper considers zeroth-order optimization for stochastic convex minimization problem. We propose a parameter-free stochastic zeroth-order method (POEM) by introducing a step-size scheme based on the distance over finite difference and an adaptive smoothing parameter. We provide the theoretical analysis to show that POEM achieves the near-optimal stochastic zeroth-order oracle complexity. We further conduct the numerical experiments to demonstrate POEM outperforms existing zeroth-order methods in practice. |
| title | A Parameter-Free and Near-Optimal Zeroth-Order Algorithm for Stochastic Convex Optimization |
| topic | Optimization and Control |
| url | https://arxiv.org/abs/2502.05600 |