Research

My Google Scholar page is up to date on publications.

I enjoy studying data through the lens of structures and geometry, including low-dimensional embeddings learned from generative ML models. Right now, I am most interested in:

• Learning on graphs, networks
• Learning on manifolds
• Optimal transport, diffusion-based models
• Nonconvex optimization
• Federated learning, privacy-preserving machine learning
• Learning in low-label settings
• Self-supervised learning, representation learning
• Applications: science of science, medical and bioinformatics

I use tools in signal processing, machine learning, optimization, information theory, network science, and (high-dimensional) statistics.

Mentored students

Current

• Amartya Banerjee. UNC CS PhD. Co-advised by Caroline Moosmueller in Math.
• Saurav Raj Pandey. UNC CS PhD.

Previous

• Olawumi Olasunkanmi. UNC CS PhD. Co-advised by Stan Ahalt from SDSS and RENCI.
• Scott Ye. UNC Biostatistics MS. Won department thesis award. First: Data Scientist at UCSF.
• Yidan Mei. UNC Math BS. First: MS student at Yale.
• Lu Cheng. UCLA Applied Math and Statistics BS. Advised honor’s thesis. First: Data Scientist at Meta.

Funding

My research is generously supported by the external funding sources listed below. I am also grateful for the many travel awards and internal grants that have supported my work over the years but are not listed here.

• Lead PI NSF DMS-2603388. Collaborative Research: Machine Learning on Stratified Matrix Manifolds Under Group Actions – Foundations, Algorithms and Applications. 2026-2029.