Equation Learning for Statistical Physics
While simple classical fluids are reasonably well understood, theories about the emerging physical behaviour of fluid particles that interact through an anisotropic potential are still quite limited. We are using machine learning methods, especially in the context of symbolic regression, to help us find an effective description (density functional theory) of these systems.
Publications
2024
Machine learning of a density functional for anisotropic patchy particles
Journal of Chemical Theory and Computation
·
2024
2020
Analytical classical density functionals from an equation learning network
The Journal of Chemical Physics
·
2020