Saved in:
Bibliographic Details
Main Authors: Zhong, Jindi, Yin, Congyaohui, Zhang, Zhaorong, Zhang, Huanshui
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2512.24552
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918491700854784
author Zhong, Jindi
Yin, Congyaohui
Zhang, Zhaorong
Zhang, Huanshui
author_facet Zhong, Jindi
Yin, Congyaohui
Zhang, Zhaorong
Zhang, Huanshui
contents This paper proposes a novel second-order optimization algorithm based on the Optimal Control Principle (OCP), applicable to large-scale optimization problems in neural network training. The algorithm has a computational complexity of O(d) and strong robustness. Extensive experiments on multiple benchmarks demonstrate the significant superiority of the proposed method.
format Preprint
id arxiv_https___arxiv_org_abs_2512_24552
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle OCP-GN: A Scalable Second-order Optimizer for Stochastic Optimization
Zhong, Jindi
Yin, Congyaohui
Zhang, Zhaorong
Zhang, Huanshui
Computer Vision and Pattern Recognition
Optimization and Control
This paper proposes a novel second-order optimization algorithm based on the Optimal Control Principle (OCP), applicable to large-scale optimization problems in neural network training. The algorithm has a computational complexity of O(d) and strong robustness. Extensive experiments on multiple benchmarks demonstrate the significant superiority of the proposed method.
title OCP-GN: A Scalable Second-order Optimizer for Stochastic Optimization
topic Computer Vision and Pattern Recognition
Optimization and Control
url https://arxiv.org/abs/2512.24552