Postdoctoral Scholar
Istituto Italiano di Tecnologia (IIT) · Genova, Italy
My research focuses on the mathematical foundations of deep learning, examining how network architecture, compositional structure, and learned representations shape generalization and sample complexity.
At IIT, I work with Lorenzo Rosasco. During this appointment, I undertook a two-week research visit to MIT in October 2025, hosted by Tomaso Poggio. I received my PhD in Mathematics from City University of Hong Kong in 2023, advised by Han Feng and Ding-Xuan Zhou.
🌟 Sep. 2026 · Our paper Learning Sparse Compositional Functions with Norm-Constrained Neural Networks has been accepted at NeurIPS 2026.
April 2026 · New preprint on sparsity-aware neural networks for nonlinear functionals, with approximation guarantees that mitigate dimensional dependence.
March 2026 · Spherical Analysis of Learning Nonlinear Functionals was published in Constructive Approximation.