Tong Ding
Postdoctoral Research Associate
Siebel School of Computing and Data Science
University of Illinois Urbana-Champaign
📧 | 🏢 Office 4217
My research lies in numerical linear algebra and tensor methods. My current work focuses on tensor decompositions and graph algorithms, with earlier work on matrix structures arising in shallow neural network training.
Research Interests
Numerical Linear Algebra
Tensor Computations
Graph Algorithms
Scientific Computing
Education
Ph.D. in Mathematics, Purdue University, 2019 – 2026
Thesis: Matrix Computations for Training Shallow Neural Networks; Advisor: Jianlin Xia
B.S. in Mathematics, Purdue University, 2015 – 2019
Publications [CV]
Matrix analysis for shallow ReLU neural network least-squares approximations
A Structure-Guided Gauss-Newton Method for Shallow ReLU Neural Network
Presentations
Talks
Why Neural Network Optimization Matrices Are Ill-Conditioned and Why That's Okay
(Lightning talk)
Matrix Analysis and Fast Solvers in Neural Networks
(Minisymposium talk)
Matrix Analysis for Shallow ReLU Neural Network Least-Squares Approximations
(Session talk)
A Structure-Guided Gauss-Newton Method for Shallow ReLU Neural Networks
(Minisymposium talk)
Posters
ICERM Workshop on Randomized Numerical Linear Algebra (RNLA)
CBMS Conference on Research at the Interface of Applied Mathematics and Machine Learning
NSF Computational Mathematics PI Meeting
NSF Computational Mathematics PI Meeting
Purdue Undergraduate Research Exhibition
Professional Activities
January 2024 – May 2026
President, SIAM Student Chapter, Purdue University
- Organized annual student conferences (2024, 2025, and 2026)
Teaching
MA 303: Differential Equations for Engineering
Delivered lectures for undergraduate students, prepared course materials, and facilitated class discussions on differential equations topics.
MA 16600: Calculus II
Led recitation sessions and held office hours to support students in mastering calculus concepts.
Graduate Grader
Graded coursework for MA 52700 (Advanced Mathematics for Engineers), MA 35100 (Linear Algebra), and MA 45000 (Abstract Algebra).