A Riemannian Accelerated Proximal Gradient Method

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Feng, Shuailing, Jiang, Yuhang, Huang, Wen, Ying, Shihui
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915516052930560
author Feng, Shuailing
Jiang, Yuhang
Huang, Wen
Ying, Shihui
author_facet Feng, Shuailing
Jiang, Yuhang
Huang, Wen
Ying, Shihui
contents Riemannian accelerated gradient methods have been well studied for smooth optimization, typically treating geodesically convex and geodesically strongly convex cases separately. However, their extension to nonsmooth problems on manifolds with theoretical acceleration remains underexplored. To address this issue, we propose a unified Riemannian accelerated proximal gradient method for problems of the form $F(x) = f(x) + h(x)$ on manifolds, where $f$ can be either geodesically convex or geodesically strongly convex, and $h$ is $ρ$-retraction-convex, possibly nonsmooth. We rigorously establish accelerated convergence rate under reasonable conditions. Additionally, we introduce a safeguard mechanism to ensure global convergence in non-convex settings. Numerical results validate the theoretical acceleration of the proposed method.
format Preprint
id arxiv_https___arxiv_org_abs_2509_21897
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle A Riemannian Accelerated Proximal Gradient Method
Feng, Shuailing
Jiang, Yuhang
Huang, Wen
Ying, Shihui
Optimization and Control
90C25, 90C26, 90C48
Riemannian accelerated gradient methods have been well studied for smooth optimization, typically treating geodesically convex and geodesically strongly convex cases separately. However, their extension to nonsmooth problems on manifolds with theoretical acceleration remains underexplored. To address this issue, we propose a unified Riemannian accelerated proximal gradient method for problems of the form $F(x) = f(x) + h(x)$ on manifolds, where $f$ can be either geodesically convex or geodesically strongly convex, and $h$ is $ρ$-retraction-convex, possibly nonsmooth. We rigorously establish accelerated convergence rate under reasonable conditions. Additionally, we introduce a safeguard mechanism to ensure global convergence in non-convex settings. Numerical results validate the theoretical acceleration of the proposed method.
title A Riemannian Accelerated Proximal Gradient Method
topic Optimization and Control
90C25, 90C26, 90C48
url https://arxiv.org/abs/2509.21897