Profile

Zhao DING (丁钊). Computational Mathematics PhD candidate at Wuhan University. Expect to enter the job market in 2025. Working on solving PDE by deep learning & diffusion model based generation. With 4 years of experience in scientific computing and 2 years in generative learning.

[简历]       [Curriculum Vitae]

Contact

E-mail: dingz489@gmail.com, zd1998@whu.edu.cn

Phone: +86 157 2701 5212

Recent Work

Working on new one-step generation scheme derived from diffusion models, exploiting the deterministic nature of ODE, achieving SOTA results among methods of the same kind. Participated in algorithm design and numerical implementation. Post.

Education

  • PhD, Computational Mathematics, Wuhan University, 2020-present.
  • BSc, Mathematics and Applied Mathematics, Wuhan University, 2016-2020.

Publication

  • Zhao Ding, Chenguang Duan, Yuling Jiao, Ruoxuan Li, Jerry Zhijian Yang and Pingwen Zhang (2024). Characteristic Learning for Provable One Step Generation.
    arXiv:2405.05512. [Project Page].
  • Zhao Ding, Chenguang Duan, Yuling Jiao and Jerry Zhijian Yang (2024). Semi-Supervised Deep Sobolev Regression: Estimation, Variable Selection and Beyond.
    arXiv:2401.04535.
  • Jinyuan Chang, Zhao Ding, Yuling Jiao, Ruoxuan Li and Jerry Zhijian Yang (2024). Deep conditional distribution learning via conditional Föllmer flow.
    arXiv:2402.01460. [Project Page].
  • Zhao Ding, Yuling Jiao, Xiliang Lu, Jerry Zhijian Yang and Cheng Yuan (2023). Sampling via Föllmer Flow.
    arXiv:2311.03660. [Project Page].
  • Zhao Ding, Junjun Huang, Yuling Jiao, Xiliang Lu and Jerry Zhijian Yang (2020). Robust decoding from binary measurements with cardinality constraint least squares.
    In press with Communications in Computational Physics.
  • Mo Chen, Zhao Ding, Yuling Jiao, Xiliang Lu, Peiying Wu and Jerry Zhijian Yang (2023). Convergence analysis of PINNs with Over-parametrization.
    In press with Communications in Computational Physics.

Experience

  • Teaching Assistant, Wuhan University, 2021

    Numerical tutorials in MATLAB for “Numerical Linear Algebra”. [Project Page]

  • Invited Talks

    • “Sampling via Föllmer Flow” at the student chapter of the 21st China Society for Industrial and Applied Mathematics (CSIAM) conference, 2023.

    • “ODE-based Sampling and Generative Models” at the student chapter of the 17th East Asia Section for Industrial and Applied Mathematics (EASIAM) conference, 2024.

  • Member of Projects, Wuhan University & Huawei, 2020-2024

    1. Machine Learning Library

      Develop SVD, PCA, LDA and ARMA, with performance comparable to (partially better than) scikit-learn and statsmodels. (C++)

    2. Vector Statistical Library

      Develop 1st-4th original & central SUM and MOM statistical functions. (C)

    3. MindX Models

      Implement of FNO & PINO (operator networks) in native MindSpore. (Python)

    4. Math Function Library with High Precision

      Develop interval algorithms with arbitrary precision, built upon mpfr library, in object oriented style, for testing basic math functions at any precision. (C++)

    5. Derivative Constrained Path Fitting

      Develop B-spline fitting algorithm under 1st-3rd derivative constraints. (MATLAB,C++)

    6. Sparse Matrix Solver

      Design the architecture for sparse Cholesky decomposition and develop the symbol factorization part. Reach comparable performance compared to C library CHOLMOD. (C++)

Skill

  • Scientific Computing

    in Python (numpy, scipy, matplotlib, seaborn, …), MATLAB, C/C++

  • Deep Learning

    with PyTorch, experienced in data parallel training/inference of diffusion models

  • Language

    English: CET4 (620), CET6 (625), IELTS (6.5)

  • Other

    bash script, git, make, LaTex

Award

  • SIAM Award for Student Chapter at Wuhan University, 2024
  • First Prize of China Undergraduate Mathematical Contest in Modeling, 2018

Hobby

  • Sudoku (middle level player)
  • Movies (Horror, Suspense, Sci-Fi)

Post