Abstract:Density functional theory is routinely applied to predict crystal structures. The most common exchange-correlation functionals used to this end are the Perdew–Burke–Ernzerhof (PBE) approximation and its variant PBEsol. We investigate the performance of these functionals for the prediction of lattice parameters and show how to enhance their accuracy using machine learning. Our data set is constituted by experimental crystal structures of the Inorganic Crystal Structure Database matched with PBE-optimized structures stored in the materials project database. We complement these data with PBEsol calculations. We demonstrate that the accuracy and precision of PBE/PBEsol volume predictions can be noticeably improved a posteriori by employing simple, explainable machine learning models. These models can improve PBE unit cell volumes to match the accuracy of PBEsol calculations, and reduce the error of the latter with respect to experiment by 35 percent. Further, the error of PBE lattice constants is reduced by a factor of 3–5. A further benefit of our approach is the implicit correction of finite temperature effects without performing phonon calculations. Knowledge about the crystal structure of solids is essential for describing their elastic and electronic properties. In particular, their accurate prediction is essential to predict the electronic properties of not-yet-synthesized materials. Lattice parameters are most commonly calculated by density functional theory using the Perdew–Burke–Ernzerhof (PBE) approximation and its variant PBEsol as exchange-correlation functional. They are successful in describing materials properties but do, however, not always achieve the desired accuracy in comparison with experiments. We propose a computationally efficient scheme based on interpretable machine learning to optimize crystal structures. We demonstrate that the accuracy of PBE- and PBEsol-structures can be, therewith, enhanced noticeably. In particular, the PBE unit cells, available in materials databases, can be improved to the level of the more accurate PBEsol calculations and the error of the latter with respect to the experiment can be reduced by 35 percent. An additional advantage of our scheme is the implicit inclusion of finite temperature corrections, which makes expensive phonon calculations unnecessary.

Machine learning for accuracy in density functional approximations

Machine learning for accuracy in density functional approximations

Machine Learning approaches to classical density functional theory

Deepks: A Comprehensive Data-Driven Approach Toward Chemically Accurate Density Functional Theory

A Deep Dive into Machine Learning Density Functional Theory for Materials Science and Chemistry

Machine learning density functionals from the random-phase approximation

Using Machine Learning to Find New Density Functionals

Density Functionals with Quantum Chemical Accuracy: From Machine Learning to Molecular Dynamics

A semilocal machine-learning correction to density functional approximations

Deep dive into machine learning density functional theory for materials science and chemistry

Improving Density Functional Prediction of Molecular Thermochemical Properties with a Machine-Learning-Corrected Generalized Gradient Approximation.

Completing density functional theory by machine learning hidden messages from molecules

Machine-learned approximations to Density Functional Theory Hamiltonians

Bridging electronic and classical density-functional theory using universal machine-learned functional approximations

Machine-learning correction to density-functional crystal structure optimization

Construct exchange-correlation functional via machine learning

Constructing Accurate and Efficient General-Purpose Atomistic Machine Learning Model with Transferable Accuracy for Quantum Chemistry

Development of a machine learning finite-range nonlocal density functional

By-passing the Kohn-Sham equations with machine learning

Learning electron densities in the condensed phase