Jiannan Jiang

Wean Hall 6213
Mathematics Department
Mellon College of Science
Carnegie Mellon University
5000 Forbes Ave,
Pittsburgh, PA 15213

Email: jiannanj at andrew.cmu.edu
    Jiannan Jiang     Bio

Research interests

Teaching

Links

CV

Github

Bio

I am a third-year PhD Candidate at the analysis group in the Mathematics Department of Carnegie Mellon University, advised by Hayden Schaeffer. I received my Bachelor's degree in Applied Mathematics and Computer Science at University of California, Berkeley in 2019.

Research interests

My current research interest focuses on imposing mathematically-provable structured regularizations on data-driven models over dynamical systems. Specifically, modern machine learning techniques over dynamical systems often first impose a large hypothesis class and then regularize based on specific needs (SGD/dropout achieves statistically L2 normalization with low computational costs; L1/L0 nomalization enforces sparsity). Recent advances in deep learning focuses on imposing a pre-determined structure on neural networks based on the specific tasks, as a way to limit the hypothesis class. These regularization metohds are powerful but lack a well-structured theory to quantify the behaviors in convergence. On the other hand, the number of parameters used in these models scales exponentially as the processing and computing power grows, suggesting a potential improvement by truncating inactive computing units. I'm interested in how to scale down the number of parameters in large models while preserving its accuracy via mathematical tools. Particularly, I focus on linear/nonlinear models of dynamical systems.

Research Directions

Teaching Experience

Carnegie Mellon University

Relavant Links

This page was last updated on Sept. 30 2022. Still Working on updating