A Homogeneous Second-Order Descent Method for Nonconvex Optimization

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Zhang, Chuwen, Ge, Dongdong, He, Chang, Jiang, Bo, Jiang, Yuntian, Xue, Chenyu, Ye, Yinyu
Formato: Preprint
Publicado: 2022
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866916674249162752
author Zhang, Chuwen
Ge, Dongdong
He, Chang
Jiang, Bo
Jiang, Yuntian
Xue, Chenyu
Ye, Yinyu
author_facet Zhang, Chuwen
Ge, Dongdong
He, Chang
Jiang, Bo
Jiang, Yuntian
Xue, Chenyu
Ye, Yinyu
contents In this paper, we introduce a Homogeneous Second-Order Descent Method (HSODM) using the homogenized quadratic approximation to the original function. The merit of homogenization is that only the leftmost eigenvector of a gradient-Hessian integrated matrix is computed at each iteration. Therefore, the algorithm is a single-loop method that does not need to switch to other sophisticated algorithms and is easy to implement. We show that HSODM has a global convergence rate of $O(ε^{-3/2})$ to find an $ε$-approximate second-order stationary point, and has a local quadratic convergence rate under the standard assumptions. The numerical results demonstrate the advantage of the proposed method over other second-order methods.
format Preprint
id arxiv_https___arxiv_org_abs_2211_08212
institution arXiv
publishDate 2022
record_format arxiv
spellingShingle A Homogeneous Second-Order Descent Method for Nonconvex Optimization
Zhang, Chuwen
Ge, Dongdong
He, Chang
Jiang, Bo
Jiang, Yuntian
Xue, Chenyu
Ye, Yinyu
Optimization and Control
In this paper, we introduce a Homogeneous Second-Order Descent Method (HSODM) using the homogenized quadratic approximation to the original function. The merit of homogenization is that only the leftmost eigenvector of a gradient-Hessian integrated matrix is computed at each iteration. Therefore, the algorithm is a single-loop method that does not need to switch to other sophisticated algorithms and is easy to implement. We show that HSODM has a global convergence rate of $O(ε^{-3/2})$ to find an $ε$-approximate second-order stationary point, and has a local quadratic convergence rate under the standard assumptions. The numerical results demonstrate the advantage of the proposed method over other second-order methods.
title A Homogeneous Second-Order Descent Method for Nonconvex Optimization
topic Optimization and Control
url https://arxiv.org/abs/2211.08212