Machine Learning the Physical Non-Local Exchange-Correlation Functional of Density-Functional Theory

Abstract

We train a neural network as the universal exchange-correlation functional of density-functional theory that simultaneously reproduces both the exact exchange-correlation energy and potential. This functional is extremely non-local, but retains the computational scaling of traditional local or semi-local approximations. It therefore holds the promise of solving some of the delocalization problems that plague density-functional theory, while maintaining the computational efficiency that characterizes the Kohn-Sham equations. Furthermore, by using automatic differentiation, a capability present in modern machine-learning frameworks, we impose the exact mathematical relation between the exchange-correlation energy and the potential, leading to a fully consistent method. We demonstrate the feasibility of our approach by looking at one-dimensional systems with two strongly-correlated electrons, where density-functional methods are known to fail, and investigate the behavior and performance of our functional by varying the degree of non-locality.

8 pages, 5 figures; accepted for publication in J. Phys. Chem. Lett

References in corpus (15)

• Fast and Accurate Deep Network Learning by Exponential Linear Units (ELUs)
• Big Data of Materials Science - Critical Role of the Descriptor
• By-passing the Kohn-Sham equations with machine learning
• Finding Density Functionals with Machine Learning
• Real-space grids and the Octopus code as tools for the development of new simulation approaches for electronic systems
• Self-consistent hybrid functional for condensed systems
• Revisiting Small Batch Training for Deep Neural Networks
• Fractional spins and static correlation error in density functional theory
• Orbital-free Bond Breaking via Machine Learning
• Pure density functional for strong correlations and the thermodynamic limit from machine learning
• Numerical Methods for the Inverse Problem of Density Functional Theory
• Density functional theory beyond the linear regime: Validating adiabatic LDA
• Reference electronic structure calculations in one dimension
• Neural-network Kohn-Sham exchange-correlation potential and its out-of-training transferability
• Towards a formal definition of static and dynamic electronic correlations

Cited by in corpus (36)

• Combining Machine Learning and Computational Chemistry for Predictive Insights Into Chemical Systems
• Perspective on integrating machine learning into computational chemistry and materials science
• Kohn-Sham equations as regularizer: building prior knowledge into machine-learned physics
• Quantum Machine Learning for Chemistry and Physics
• A Deep Dive into Machine Learning Density Functional Theory for Materials Science and Chemistry
• Deep learning the Hohenberg-Kohn maps of Density Functional Theory
• Neural functional theory for inhomogeneous fluids: Fundamentals and applications
• Drawing Phase Diagrams of Random Quantum Systems by Deep Learning the Wave Functions
• Hybrid convolutional neural network and PEPS wave functions for quantum many-particle states
• Reduced Density Matrix Functional Theory for Bosons
• Machine learning the derivative discontinuity of density-functional theory
• Density Functional Theory Transformed into a One-electron Reduced Density Matrix Functional Theory for the Capture of Static Correlation
• Machine Learning Exchange-Correlation Potential in Time Dependent Density Functional Theory
• Machine Learning Universal Bosonic Functionals
• Comparison of Density-Matrix Corrections to Density Functional Theory
• Why neural functionals suit statistical mechanics
• Machine learning density functionals from the random-phase approximation
• Hyper-density functional theory of soft matter
• Grad DFT: a software library for machine learning enhanced density functional theory
• Efficient Learning of a One-dimensional Density Functional Theory
• Completing density functional theory by machine-learning hidden messages from molecules
• Neural density functionals: Local learning and pair-correlation matching
• Neural density functional theory of liquid-gas phase coexistence
• KineticNet: Deep learning a transferable kinetic energy functional for orbital-free density functional theory
• Why hyperdensity functionals describe any equilibrium observable
• Exact and model exchange-correlation potentials for open-shell systems
• Training-free hyperparameter optimization of neural networks for electronic structures in matter
• Inverting the Kohn-Sham equations with physics-informed machine learning
• Homogeneous electron gas in arbitrary dimensions
• MOLPIPx: an end-to-end differentiable package for permutationally invariant polynomials in Python and Rust
• The complete inverse Kohn-Sham problem: from the density to the energy
• Leveraging Normalizing Flows for Orbital-Free Density Functional Theory
• Homogeneous Electron Liquid in Arbitrary Dimensions: Exchange and Correlation Using the Singwi-Tosi-Land-Sjölander Approach
• Machine-learning semi-local density functional theory for many-body lattice models at zero and finite temperature
• Quantum-Enhanced Neural Exchange-Correlation Functionals
• Machine learning based prediction of the electronic structure of quasi-one-dimensional materials under strain