Neural Functional Theory for Inhomogeneous (Non-)Equilibrium Fluids

Explore a 23-minute Lennard-Jones Centre discussion group seminar presented by Prof. Florian Sammüller from the University of Bayreuth, Germany, focusing on neural functional theory for inhomogeneous fluids. Learn how machine learning can be effectively applied to classical density functional theory and power functional theory to characterize inhomogeneous fluids in and out of equilibrium. Discover how neural networks trained with simulation data provide precise and flexible representations of central functional maps, enabling predictions that surpass analytic treatments in accuracy. The seminar covers multiscale applications including colloidal sedimentation-diffusion equilibrium under gravity and the inverse design of nonequilibrium flow. Additional reading materials from Journal of Physics: Condensed Matter and Proceedings of the National Academy of Sciences are provided for deeper understanding of this cutting-edge research presented on February 5th, 2024.