Clarkson Center for Complex Systems Science
Data-driven modeling with applications to science, engineering and computing
Roman Grigoriev, Georgia Tech
The recent decade saw a revolution in science and engineering where data-driven methods have in many cases become more popular and powerful than more traditional human-driven approaches. In particular, machine learning has demonstrated unprecedented capabilities for generating new knowledge and understanding by discovering easily interpretable governing equations describing noisy and occasionally incomplete data. This talk will introduce a general equation inference framework that can be used to synthesize complete mathematical models of continuous or discrete systems such as fluids, excitable systems and active matter. I will illustrate the power and flexibility of this framework using specific examples of learning new physics directly from experimental data and validation of direct numerical simulations.
Bio: Roman Grigoriev is a Professor in the School of Physics at Georgia Tech. He is a recipient of the Frenkiel Award from the APS Division of Fluid Dynamics. Dr. Grigoriev’s research interests cover a broad range of topics in dynamical systems and control with applications to fluid dynamics, excitable media, instabilities, and turbulence. A significant portion of his current research is focused on developing the foundations of machine learning with applications to scientific discovery and modeling.
Friday March 14, 12:00 pm SC166