Machine-learning free-energy functionals using density profiles from simulations

Abstract

The formally exact framework of equilibrium Density Functional Theory (DFT) is capable of simultaneously and consistently describing thermodynamic and structural properties of interacting many-body systems in arbitrary external potentials. In practice, however, DFT hinges on approximate (free-)energy functionals from which density profiles (and hence the thermodynamic potential) follow via an Euler-Lagrange equation. Here we explore a relatively simple Machine Learning (ML) approach to improve the standard mean-field approximation of the excess Helmholtz free energy functional of a 3D Lennard-Jones system at a supercritical temperature. The learning set consists of density profiles from grand-canonical Monte Carlo simulations of this system at varying chemical potentials and external potentials in a planar geometry only. Using the DFT formalism we nevertheless can extract not only very accurate 3D bulk equations of state but also radial distribution functions using the Percus test-particle method. Unfortunately, our ML approach did not provide very reliable Ornstein-Zernike direct correlation functions for small distances.

arXiv comment12 pages, 10 figures

References in corpus (8)

• Adam: A Method for Stochastic Optimization
• Density functional theory for hard-sphere mixtures: the White-Bear version Mark II
• Tension and stiffness of the hard sphere crystal-fluid interface
• A model colloidal fluid with competing interactions: bulk and interfacial properties
• The standard mean-field treatment of inter-particle attraction in classical DFT is better than one might expect
• Fundamental measure theory for the electric double layer: implications for blue-energy harvesting and water desalination
• A classical density functional from machine learning and a convolutional neural network
• Analytical classical density functionals from an equation learning network

Cited by in corpus (24)

• Power functional theory for many-body dynamics
• Neural functional theory for inhomogeneous fluids: Fundamentals and applications
• Perspective: How to overcome dynamical density functional theory
• Force density functional theory in- and out-of-equilibrium
• Machine learning of a density functional for anisotropic patchy particles
• Learning classical density functionals for ionic fluids
• Perspective: New directions in dynamical density functional theory
• Reliable emulation of complex functionals by active learning with error control
• Why neural functionals suit statistical mechanics
• Physics-informed Bayesian inference of external potentials in classical density-functional theory
• Hyper-density functional theory of soft matter
• Learning Neural Free-Energy Functionals with Pair-Correlation Matching
• Hyperforce balance via thermal Noether invariance of any observable
• Neural density functionals: Local learning and pair-correlation matching
• Neural density functional theory of liquid-gas phase coexistence
• Neural force functional for non-equilibrium many-body colloidal systems
• Why hyperdensity functionals describe any equilibrium observable
• Accelerated calculation of configurational free energy using a combination of reverse Monte Carlo and neural network models: Adsorption isotherm for 2D square and triangular lattices
• Metadensity functional theory for classical fluids: Extracting the pair potential
• Learning the bulk and interfacial physics of liquid-liquid phase separation with neural density functionals
• Neural Density Functional Theory in Higher Dimensions with Convolutional Layers
• Determining the chemical potential via universal density functional learning
• The roles of bulk and surface thermodynamics in the selective adsorption of a confined azeotropic mixture
• Metadensity functional learning for classical fluids: Regularizing with pair correlations