{"ID":2823575,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2512.24552","arxiv_id":"2512.24552","title":"OCP-GN: A Scalable Second-order Optimizer for Stochastic Optimization","abstract":"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.","short_abstract":"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 t...","url_abs":"https://arxiv.org/abs/2512.24552","url_pdf":"https://arxiv.org/pdf/2512.24552v2","authors":"[\"Jindi Zhong\",\"Congyaohui Yin\",\"Zhaorong Zhang\",\"Huanshui Zhang\"]","published":"2025-12-31T01:21:08Z","proceeding":"cs.CV","tasks":"[\"cs.CV\",\"math.OC\"]","methods":"[]","has_code":false}