Biography

Shaocong Ma is a postdoctoral researcher at the University of Maryland, working under the supervision of Professor Heng Huang. He received his Ph.D. in Electrical and Computer Engineering from the University of Utah, where he was advised by Professor Yi Zhou. Previously, he completed an M.A. in Statistics at the University of California, Santa Barbara, and earned his B.S. in Statistics from Sichuan University.

My most-updated CV: PDF

Research summary: PDF

Research Interests:

  • Optimization and Reinforcement Learning: Driven by real-world challenges, my research is centered on optimization techniques applied in diverse fields such as (multi-agent) reinforcement learning and physic-informed machine learning. More specifically, I prioritize creating data-efficient and environment-robust algorithms that are substantiated with theoretical guarantees.

Professional Services:

  • Conference Reviewer/Program Committee: ICML; ICLR; NeurIPS; IEEE BigData; IJCAI; UAI; AAAI; AISTAT.
  • Journal Reviewer: Transactions on Machine Learning Research (TMLR); IEEE Transactions on Emerging Topics in Computational Intelligence (TETCI); European Journal of Control.
  • Workshop Reviewer: ICLR 2024 Blogpost.

Experiences

Postdoctoral Researcher

2024.10 - Present
University of Maryland, College Park

I joined the Heng Huang lab as a postdoctoral researcher. I led several Large Language Model (LLM) projects and contributed to more efficient and robust machine learning systems.

Research Intern (AI4Science)

2022.5 - 2022.8
Lawrence Livermore National Security

(1) Designed a hybrid model with external black-box PDE solvers, addressing the non-differentiability challenges in fluid flow predictions. (2) Rigorously assessed the Physics-Informed Graph Neural Network’s resilience in out-of-distribution scenarios, achieving comparable performance with differentiable solvers.

Research Assistant

2019 - 2024
University of Utah

I led several machine learning projects in Professor Yi Zhou’s lab. My role is to design fast and stable algorithms in the large-scale machine learning and reinforcement learning. Results developed during this period were published on top conferences including ICML, NeurIPS, and ICLR.

Teaching Assistant

2020 - 2021
University of Utah

I instructed the lab section of an electrical and computer engineering course during the Ph.D. program. Course title:

  • ECE 3500: Fundamentals of Signals and Systems

Teaching Assistant

2018 - 2019
University of California, Santa Barbara

I instructed sections/labs of statistics and data science courses during the Master program. Courses include:

  • PSTAT 5A: Statistics
  • PSTAT 5LS: Statistics for Life Science
  • PSTAT 109: Statistics for Economics
  • PSTAT 172A: Actuarial Statistics
  • PSTAT 175: Survival Analysis

Publications

  • Variance-Reduced Off-Policy TDC Learning: Non-Asymptotic Convergence Analysis
  • Shaocong Ma, Yi Zhou, Shaofeng Zou
    NeurIPS. 2020. (Acceptance rate: 20.1%)
  • Greedy-GQ with Variance Reduction: Finite-time Analysis and Improved Complexity
  • Shaocong Ma, Ziyi Chen, Yi Zhou, Shaofeng Zou
    ICLR. 2021. (Acceptance rate: 28.7%)
  • Sample Efficient Stochastic Policy Extragradient Algorithm for Zero-Sum Markov Game
  • Ziyi Chen, Shaocong Ma, Yi Zhou
    ICLR. 2022. (Acceptance rate: 32.3%)
  • Data Sampling Affects the Complexity of Online SGD over Dependent Data
  • Shaocong Ma, Ziyi Chen, Yi Zhou, Kaiyi Ji, Yingbin Liang
    UAI. 2022.
  • Finding Correlated Equilibrium of Constrained Markov Game: A Primal-Dual Approach
  • Ziyi Chen, Shaocong Ma, Yi Zhou
    NeurIPS. 2022. (Acceptance rate: 25.6% )
  • Decentralized Robust V-Learning for Solving Markov Games with Model Uncertainty.
  • Shaocong Ma, Ziyi Chen, Shaofeng Zou, Yi Zhou
    JMLR 2023.
  • End-to-End Mesh Optimization of a Hybrid Deep Learning Black-Box PDE Solver.
  • Shaocong Ma, James Diffenderfer, Bhavya Kailkhura, and Yi Zhou
    NeurIPS 2023 (ML4PS Workshop).
  • When Non-Differentiable PDE Solver Meets Deep Learning: Partially Differentiable Learning for Efficient Fluid Flow Prediction
  • Shaocong Ma, James Diffenderfer, Bhavya Kailkhura, and Yi Zhou
    Submitted.

    Others

    Session/Lab Notes, Presentation Slides, and Course Projects.