Orbital-free bond breaking via machine learning

2013/06/07 by John Snyder, John C. Snyder, Matthias Rupp +5 · 145 citations
Chemistry · Computer Science · Engineering · Materials Science · Mathematics · Physics and Astronomy · #Ab initio #Atomic physics #Chemistry #Classical mechanics #Computational Drug Discovery Methods #Computational chemistry #Density functional theory #Diatomic molecule #Energy (signal processing) #Fuel Cells and Related Materials #Kinetic energy #Machine Learning in Materials Science #Molecular dynamics #Molecule #Physics #Quantum mechanics #Statistical physics #cond-mat.mtrl-sci #physics.chem-ph #stat.ML

paper · pdf · doi:10.1063/1.4834075

published in The Journal of Chemical Physics 139(22), 224104 (American Institute of Physics)

arxiv created 2013/06/07 · openalex publication_date 2013/12/10 · arxiv updated 2015/06/16 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/05

Abstract

Using a one-dimensional model, we explore the ability of machine learning to approximate the non-interacting kinetic energy density functional of diatomics. This nonlinear interpolation between Kohn-Sham reference calculations can (i) accurately dissociate a diatomic, (ii) be systematically improved with increased reference data and (iii) generate accurate self-consistent densities via a projection method that avoids directions with no data. With relatively few densities, the error due to the interpolation is smaller than typical errors in standard exchange-correlation functionals.

Citations

• DFT in a nutshell
• Condition on the Kohn–Sham kinetic energy and modern parametrization of the Thomas–Fermi density
• Fast and Accurate Modeling of Molecular Atomization Energies with Machine Learning
• Reference electronic structure calculations in one dimension
• Self-Consistent Equations Including Exchange and Correlation Effects
• Zur Theorie der Kernmassen

Cited by

• Mitigating error cancellation in density functional approximations via machine learning correction
• Pure density functional for strong correlation and the thermodynamic limit from machine learning
• Predicting fermionic densities using a Projected Quantum Kernel method