My research is mostly problem-driven and has its roots from both scientific and engineering applications. These problems arise from astronomy, brain imaging, computer experiment and recommender system. And many of them involve modern data complications such as big data size, high dimensionality and manifold structures. Broadly speaking, I tackle them with nonparametric and semi-parametric modeling, combining with efficient computational techniques.
- Nonparametric and semi-parametric modeling
- Regularization methods (e.g. $\ell_1$, $\ell_2$ and nuclear-norm penalty)
- Statistical applications to astronomy, brain imaging, computer experiments and recommender systems
- Statistical learning
My research is currently supported by National Science Foundation grants: