Neural density functionals: Local learning and pair-correlation matching

Abstract

Recently Dijkman et al. (arXiv:2403.15007) proposed training classical neural density functionals via bulk pair-correlation matching. We show their method to be an efficient regularizer for neural functionals based on local learning of inhomogeneous one-body direct correlations [Sammüller et al., Proc. Natl. Acad. Sci. 120, e2312484120 (2023), 10.1073/pnas.2312484120]. While Dijkman et al. demonstrated pair-correlation matching of a global neural free energy functional, we argue in favor of local one-body learning for flexible neural modelling of the full Mermin-Evans density functional map. Using spatial localization gives access to accurate neural free energy functionals, including convolutional neural networks, that transcend the training box.

6 pages, 2 figures + supplementary material (7 pages, 6 figures)

References in corpus (9)

• Density functional theory for hard-sphere mixtures: the White-Bear version Mark II
• Machine learning and density functional theory
• A unified description of hydrophilic and superhydrophobic surfaces in terms of the wetting and drying transitions of liquids
• Machine-learning free-energy functionals using density profiles from simulations
• Analytical classical density functionals from an equation learning network
• Density depletion and enhanced fluctuations in water near hydrophobic solutes: identifying the underlying physics
• Physics-informed Bayesian inference of external potentials in classical density-functional theory
• Universality in Driven and Equilibrium Hard Sphere Liquid Dynamics
• Dynamic Decay and Superadiabatic Forces in the van Hove Dynamics of Bulk Hard Sphere Fluids