Zhonglin Xie

prof_pic.jpg

Welcome to my personal homepage! I am a fourth-year Ph.D. student at the Beijing International Center for Mathematical Research, Peking University, under the guidance of Prof. Zaiwen Wen. I am a member of the Elite Ph.D. Program in Applied Mathematics.

My current research focus is on Large Language Models (LLMs), where I am particularly interested in addressing key challenges and advancing the state-of-the-art in this rapidly evolving field. My work centers around the following directions:

  • Data selection strategies: Investigating how to optimize training and fine-tuning processes by selecting high-quality and diverse datasets to improve model performance and efficiency.
  • More efficient and scalable Mixture-of-Experts (MoE) architectures: Designing innovative MoE structures that balance computational cost with model capacity, enabling better scalability for large-scale models.
  • L2O techniques in LLM scenarios: Exploring how Learning to Optimize (L2O) can be applied to enhance optimization in LLM training and inference pipelines, such as adaptive learning rate scheduling and resource allocation.

I am actively seeking opportunities to transition into the LLM industry and contribute to cutting-edge advancements in this domain. If you are interested in my background or any of the research directions mentioned above, please feel free to contact me immediately—I would love to connect and explore potential collaborations or opportunities!

Previously, my research primarily focused on machine learning and optimization, with an emphasis on theoretical studies in Learning to Optimize (L2O). I worked on designing novel optimization algorithms using machine learning techniques and analyzing optimization methods from the perspective of ordinary differential equations (ODEs). Additionally, I explored how tools and insights from optimization could address challenges in deep learning.

I received my bachelor’s degree in 2021 from the School of Mathematical Sciences at Peking University, majoring in computational mathematics. During my undergraduate studies, I was also a member of the Elite Program of Applied Mathematics and Statistics for Undergraduates. Additionally, I earned a second bachelor’s degree in economics from the National School of Development between 2018 and 2021.

Looking forward to connecting with like-minded researchers and professionals!

selected publications

* denotes co-first authors
  1. arXiv
    OptMATH: A Scalable Bidirectional Data Synthesis Framework for Optimization Modeling
    Hongliang Lu*Zhonglin Xie*, Yaoyu Wu, Can Ren, Yuxuan Chen, and Zaiwen Wen
    arXiv preprint, arXiv:2502.11102, 2025
  2. arXiv
    ODE-based Learning to Optimize
    Zhonglin Xie, Wotao Yin, and Zaiwen Wen
    arXiv preprint, arXiv:2406.02006, 2024

news

Apr 03, 2025 I will give two talks titled ‘‘OptMATH: A Scalable Bidirectional Data Synthesis Framework for Optimization Modeling ‘’ and ‘‘ODE-based Learning to Optimize’’ at MOS2025.
Sep 26, 2024 I will give a talk titled ‘‘ODE-based Learning to Optimize’’ at The Applied Math PhD Seminar, Fudan University.
Jun 05, 2024 I will present the “ODE-based Learning to Optimize” at the poster session of 2024 International Workshop on Modern Optimization and Applications. The poster can be found here.
May 26, 2024 I will give a talk at The China conference on Scientific Machine Learning 2024 on the session of recent advances on learning to optimize.

latest posts