Molecular Models Empowering Data-Driven Approaches to Materials Discovery

J. Wu
University of California, Riverside,
United States

Keywords: statistical mechanics, machine learning, inverse design

Summary:

Physics-based modeling and data science are complementary but progressed almost in parallel until very recently. Combination of physics-based modeling with data science will accelerate scientific progress for better understanding and more reliable predictions of the physicochemical properties and phase behavior of multicomponent chemical systems that are essential for engineering design of chemical products and processes. In this talk, I explore opportunities that may promote mutual understanding and potential intercourses between these intrinsically multidisciplinary fields. Illustrative examples will be discussed on how combination of physics-based modeling and data science yields useful guidelines for materials discovery and applications.