The best DFT functional is the ensemble of functionals
Preprint
- 19 July 2024
- preprint
- Published by American Chemical Society (ACS) in ChemRxiv
Abstract
The development of better density functional theory (DFT) methods is one of the most active research areas, given the importance of DFT for ubiquitous molecular and materials simulations. However, this research primarily focuses on improving a specific exchange-correlation Kohn–Sham density functional. Here, we propose a robust procedure for constructing transferable ensembles of density functionals that perform superior to any constituent individual density functional. We show that such ensembles built only with the density functionals predating the GMTKN55 benchmark can reach a record-low weighted error of 1.69 kcal/mol on this benchmark compared to 3.08 kcal/mol of the best constituent density functional. We also introduce the DENS24 density functional ensembles as practical DFT methods with consistently accurate performance for various simulations at affordable cost. DENS24 ensembles will be available open-source and for simulations online as described at https://xacs.xmu.edu.cn/docs/mlatom/tutorial_dens.html.Keywords
All Related Versions
- Version 1, 2024-07-19, ChemRxiv
- Published version: Advanced Science, 11 (47), e2408239.
Funding Information
- National Natural Science Foundation of China
This publication has 33 references indexed in Scilit:
- MLatom 3: A Platform for Machine Learning-Enhanced Computational Chemistry Simulations and WorkflowsJournal of Chemical Theory and Computation, 2024
- A Survey on ensemble learning under the era of deep learningArtificial Intelligence Review, 2022
- Approximately Self-Consistent Ensemble Density Functional Theory: Toward Inclusion of All CorrelationsThe Journal of Physical Chemistry Letters, 2020
- Ensemble learning: A surveyWIREs Data Mining and Knowledge Discovery, 2018
- Exchange–Correlation Functional with Good Accuracy for Both Structural and Energetic Properties while Depending Only on the Density and Its GradientJournal of Chemical Theory and Computation, 2012
- W4-11: A high-confidence benchmark dataset for computational thermochemistry derived from first-principles W4 dataChemical Physics Letters, 2011
- Effect of the damping function in dispersion corrected density functional theoryJournal of Computational Chemistry, 2011
- Density Functionals with Broad Applicability in ChemistryAccounts of Chemical Research, 2008
- Progress in the development of exchange-correlation functionals2005
- Toward reliable adiabatic connection models free from adjustable parametersChemical Physics Letters, 1997