Liang will be developing new theoretical models of properties of practical data and providing frameworks for proving performance guarantees, including for learning in the presence of adversarial attacks or limited labeled data. He will also design new learning methods that are provably more robust and labeled-data efficient. This direction is still largely unexplored, despite significant recent research activities. The proposed research can be transformational for modern intelligent systems by laying the foundations for further development. It will also help to solve new theoretical problems from practice that are not adequately addressed by current theory and will have lasting impacts on machine learning and optimization.
Liang has been a UW-Madison CS faculty member since 2017. His primary area of research interest is machine learning, in particular focusing on optimization and generalization in deep learning, robust machine learning, and their applications. He has done studies in many important areas, such as deep learning theory and applications, non-convex optimization, theoretical analysis of natural language models, clustering, and large-scale and distributed machine learning. Related to the award, he is one of the first to study the generalization and optimization of deep learning on data with practical structures and one of the first to study deep learning in the teacher-student setting.