Multi-fidelity statistical machine learning for molecular crystal structure prediction
Multi-fidelity statistical machine learning for molecular crystal structure prediction
The prediction of crystal structures from first-principles requires highly accurate energies for large numbers of putative crystal structures. High accuracy of solid state density functional theory (DFT) calculations is often required, but hundreds or more structures can be present in the low energy region of interest, so that the associated computational costs are prohibitive. Here, we apply statistical machine learning to predict expensive hybrid functional DFT (PBE0) calculations using a multifidelity approach to re-evaluate the energies of crystal structures predicted with an inexpensive force field. The method uses an autoregressive Gaussian process, making use of less expensive GGA DFT (PBE) calculations to bridge the gap between the force field and PBE0 energies. The method is benchmarked on the crystal structure landscapes of three small, hydrogen-bonded organic molecules and shown to produce accurate predictions of energies and crystal structure ranking using small numbers of the most expensive calculations; the PBE0 energies can be predicted with errors of less than 1 kJ mol-1 with between 4.2 and 6.8% of the cost of the full calculations. As the model that we have developed is probabilistic, we discuss how the uncertainties in predicted energies impact the assessment of the energetic ranking of crystal structures.
computational chemistry, crystal structure prediction, machine learning, statistical learning
8065-8078
Egorova, Olga
49e91576-5a13-476b-8dbd-6091d82ce907
Hafizi, Roohollah
bdf707e3-cfc0-4c9b-8daa-d1acc5123632
Woods, David
ae21f7e2-29d9-4f55-98a2-639c5e44c79c
Day, Graeme M.
e3be79ba-ad12-4461-b735-74d5c4355636
1 October 2020
Egorova, Olga
49e91576-5a13-476b-8dbd-6091d82ce907
Hafizi, Roohollah
bdf707e3-cfc0-4c9b-8daa-d1acc5123632
Woods, David
ae21f7e2-29d9-4f55-98a2-639c5e44c79c
Day, Graeme M.
e3be79ba-ad12-4461-b735-74d5c4355636
Egorova, Olga, Hafizi, Roohollah, Woods, David and Day, Graeme M.
(2020)
Multi-fidelity statistical machine learning for molecular crystal structure prediction.
Journal of Physical Chemistry A, 124 (39), .
(doi:10.1021/acs.jpca.0c05006).
Abstract
The prediction of crystal structures from first-principles requires highly accurate energies for large numbers of putative crystal structures. High accuracy of solid state density functional theory (DFT) calculations is often required, but hundreds or more structures can be present in the low energy region of interest, so that the associated computational costs are prohibitive. Here, we apply statistical machine learning to predict expensive hybrid functional DFT (PBE0) calculations using a multifidelity approach to re-evaluate the energies of crystal structures predicted with an inexpensive force field. The method uses an autoregressive Gaussian process, making use of less expensive GGA DFT (PBE) calculations to bridge the gap between the force field and PBE0 energies. The method is benchmarked on the crystal structure landscapes of three small, hydrogen-bonded organic molecules and shown to produce accurate predictions of energies and crystal structure ranking using small numbers of the most expensive calculations; the PBE0 energies can be predicted with errors of less than 1 kJ mol-1 with between 4.2 and 6.8% of the cost of the full calculations. As the model that we have developed is probabilistic, we discuss how the uncertainties in predicted energies impact the assessment of the energetic ranking of crystal structures.
Text
multifidelity_modelling_CSP_revision
- Accepted Manuscript
Text
Supporting information
Restricted to Repository staff only
Request a copy
More information
Accepted/In Press date: 3 September 2020
e-pub ahead of print date: 3 September 2020
Published date: 1 October 2020
Additional Information:
Funding Information:
We thank the EPSRC for funding, via Grant EP/S015418/1. We acknowledge the use of the IRIDIS High Performance Computing Facility and associated support services at the University of Southampton in the completion of this work. We are also grateful to the U.K. Materials and Molecular Modelling Hub for computational resources, which is partially funded by the EPSRC (EP/P020194/1). We thank David McDonagh for sharing his experiences during the first stages of the project. R.H. thanks Barry Searle for his helpful advice on convergence of crystal17 calculations.
Publisher Copyright:
© 2020 American Chemical Society.
Keywords:
computational chemistry, crystal structure prediction, machine learning, statistical learning
Identifiers
Local EPrints ID: 443809
URI: http://eprints.soton.ac.uk/id/eprint/443809
ISSN: 1089-5639
PURE UUID: 94c2b526-f649-4f18-9de5-dd5de3861d32
Catalogue record
Date deposited: 14 Sep 2020 16:31
Last modified: 17 Mar 2024 05:52
Export record
Altmetrics
Contributors
Author:
Olga Egorova
Author:
Roohollah Hafizi
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics