郑仕欣

郑仕欣

Postdoc Associate, Department of Mathematics

About me

I am currently a Postdoctoral Associate at the University of Maryland, College Park, working under the mentorship of Professor Haizhao Yang. I earned my Ph.D. in Mathematics from Purdue University, where I was advised by Professor Xiangxiong Zhang.

My research interests lie in computational optimization, with a focus on both the development of efficient algorithms and the analysis of their theoretical guarantees. I am particularly interested in applications across machine learning, data science, partial differential equation (PDE) solvers, and deep learning. Currently, my work centers on convergence theory for neural-network training and the Riemannian geometry of modern training algorithms.

Professional Experience

  • Postdoctoral Associate, Aug 2025 - Present
    University of Maryland, College Park, MD, USA

Education

  • Ph.D. in Applied Mathematics, Sep 2018 – July 2025
    Purdue University, West Lafayette, IN, USA
    Research Area: Optimization, Riemannian Optimization, Riemannian Sampling, Machine Learning, Eigenvalue Problems

  • B.Sc. in Mathematics and Applied Mathematics, Sep 2014 – May 2018
    Shandong University, Jinan, Shandong, China

Publications & Preprints

Journal articles

  1. S. Zheng, W. Huang, B. Vandereycken, and X. Zhang. Riemannian optimization using three different metrics for Hermitian PSD fixed-rank constraints.
    Computational Optimization and Applications, 91(3):1135–1184, 2025.
    journal / arXiv:2204.07830

  2. S. Zheng, H. Yang, and X. Zhang. On the convergence of orthogonalization-free conjugate gradient method for extreme eigenvalues of Hermitian matrices: A Riemannian optimization interpretation.
    Journal of Computational and Applied Mathematics, 451:116053, 2024.
    journal / arXiv:2302.04974

Preprints

  1. J. Li, S. Zheng, and X. Zhang. Global convergence of an efficient splitting method for the defocusing Gross–Pitaevskii ground state problem.
    arXiv:2609.13509, 2026.

  2. S. Zheng, Y. Wang, and H. Yang. Global Convergence and Error Propagation in Neural Gradient Flows: A Riemannian Optimization Framework.
    arXiv:2605.27779, 2026. Under review.

  3. T. Yu, S. Zheng, J. Lu, G. Menon, and X. Zhang. Riemannian Langevin Monte Carlo schemes for sampling PSD matrices with fixed rank.
    arXiv:2309.04072, 2023.