Gavin (Jialun) Zhang

• Email: jialun2@illinois.edu
• Office: ECE Building, 306 N Wright St, Urbana, IL 61801
• Phone: (909) 575-7644
• Google Scholar

About Me

Hello! My name is Gavin (Jialun) Zhang. I am a third-year PhD student in the ECE department at the University of Illinois at Urbana-Champaign. My advisor is Richard Y. Zhang. I am also a member of the coordinated science laboratory at UIUC.

I’m mainly interested using optimization and machine learning to solve large-scale, data-driven problems with provable guarantees. Previously I also have worked in other areas of applied and computational mathematics, such as numerical PDEs and mathematical biology.

I received a bachelor’s degree in mathematics from Harvey Mudd College, where I worked with professors Chiu-Yen Kao and Andrew Bernoff on numerical PDEs and mathematical biology. I also received a master’s degree in computational mathematics from Stanford, where I worked with Mert Pilanci.

Publications and Preprints

• Fast and Minimax Optimal Estimation of Low-rank Matrices via Gradient Descent
  Gavin Zhang, Hong-ming Chiu, Richard Y. Zhang
  Submitted
• Simple Alternating Minimization Provably Solves Complete Dictionary Learning
  Geyu Liang, Gavin Zhang, Salar Fattahi, Richard Y. Zhang
  Submitted
• Preconditioned Gradient Descent for Overparameterized Nonconvex Burer-Monteiro Factorization with Global Optimality Certification Gavin Zhang, Salar Fattahi, Richard Y. Zhang
  Journal of Machine Learning Research (JMLR 2023)
• Accelerating SGD for Highly Ill-Conditioned Huge-Scale Online Matrix Completion
  Gavin Zhang, Hong-ming Chiu, Richard Y. Zhang
  Advances in Neural Information Processing Systems (NeurIPS 2022)
• Preconditioned Gradient Descent for Over-parameterized Nonconvex Matrix Factorization
  Gavin Zhang, Salar Fattahi, Richard Y. Zhang.
  Advances in Neural Information Processing Systems (NeurIPS 2021)
• How Many Samples is a Good Initial Point Worth in Low-rank Matrix Recovery?
  Gavin Zhang, Richard Y. Zhang.
  Spotlight Presentation (Top 4%). Advances in Neural Information Processing Systems (NeurIPS 2020)
• Weak Solution of a Doubly Degenerate Parabolic Equation
  Di Kang, Tharathep Sangsawang, Gavin Zhang
  (Submitted)
• Agent-based and Continuous Models of Locust Hopper Bands
  A. J. Bernoff, C. M. Topaz, M. R. D’Orsogna, L. Edelstein-Keshet, R. Jones, J. Zhang, S. J. DeVore and S. Schein (Submitted)