Shengchao LIU

Shengchao LIU is an Assistant Professor in the Department of Computer Science and Engineering at The Chinese University of Hong Kong. He earned his Ph.D. at the Quebec AI Institute (Mila) in 2023 and completed a two-year postdoctoral fellowship at the University of California, Berkeley in 2025. His research focuses on the intersection of AI and physics, developing methods to both accelerate existing scientific paradigms and enable novel ones, with applications in chemistry, materials science, biology, and geography. His work has appeared in top machine learning conferences and leading scientific journals, including Nature and Science sub-journals, PNAS, and ACS.

Shengchao Liu focuses on GenAI, PhysAI, Dynamics, and the integration of Physics with Machine Learning, particularly in Pretraining. His research explores molecular graph representation, multi-modal molecule structure-text models, and graph-text models for molecule zero-shot learning. He investigates methods for drug function prediction, molecular geometry pretraining, and symmetry-informed geometric representations for molecules, proteins, and crystalline materials. His work also addresses challenges in multi-task learning, negative transfer reduction, and self-supervised learning for molecular graph embeddings. Liu's contributions span drug discovery, chemical space navigation, and the development of machine learning platforms for molecular design.