Wanli is a second-year PhD student under the data science program working with Professor Shuyang Ling. His research interest lies in the interdisciplinary area of mathematics and data science such as optimal transport, theoretical deep learning, trajectory inference, etc. He is currently working with Professor Shuyang Ling on using mean field techniques to explain neural network behaviors during the terminal phase of training. He is also cooperating with faculties and students from Center for Data Science, NYU. He also holds an M.S. degree in applied mathematics from Columbia University.
Publications
- Neural collapse for unconstrained feature model under cross-entropy loss with imbalanced data. W. Hong, S. Ling, submitted, 2023.
- Fourier-Based Bounds for Wasserstein Distances and Their Implications in Computational Inversion. W. Hong, V. Kobzar, K. Ren. Optimal Transport and Machine Learning workshop, NeurIPS 2023.
Research Interests
- Theory of deep learning, mean field modeling of neural networks, neural collapse
- Computational optimal transport
- Messaging passing algorithms