Research

My research interests are machine learning, statistical computation, and their applications. My Google scholar page can be found here.

My statistics and machine learning research group at HKUST investigates the fundamental theory of machine learning. Based on theoretical understanding, we also design effective machine learning algorithms for practical problems. We also apply machine learning methods to various applications such as computer vision and natural language processing. Our core machine learning research currently focuses on the following topics.

  • image

    Theoretical Foundation of Machine Learning

    This research topic is to study the mathematical models and statistical convergence behavior of machine learning algorithms. For example, the mathematical models for deep neural networks, and statistical analysis of various learning algorithms.

  • image

    Efficient Computational Algorithms

    This research topic is concerned with efficient convex and nonconvex optimization, large scale and distributed training, automatic tuning of machine learning models.

  • image

    Robust and Adaptive Methods

    This research topic is concerned with the generalization of machine learning procedures to new scenarios, and related issues of distribution shift. We consider problems such as effective adaptation of ML models to new domains, unsupervised pretraining and fine-tuning, and exploration issues in reinforcement learning.

  • image

    Applications

    This research topic is to develop tools and algorithms in applications such as natural language processing, computer vision, and science. We are especially interested in statistically effective models, computationally efficient algorithms, and robust procedures under distribution shift.