The relative permittivity of a crystal is a fundamental property that links microscopic chemical bonding to macroscopic electromagnetic response. Multiple models, including analytical, numerical, and statistical descriptions, have been made to understand and predict dielectric behavior. Analytical models are often limited to a particular type of compound, whereas machine learning (ML) models often lack interpretability. Here, we combine supervised ML, density functional perturbation theory, and analysis based on game theory to predict and explain the physical trends in optical dielectric constants of crystals. Two ML models, support vector regression and deep neural networks, were trained on a dataset of 1364 dielectric constants. Analysis of Shapley additive explanations of the ML models reveals that they recover correlations described by textbook Clausius–Mossotti and Penn models, which gives confidence in their ability to describe physical behavior, while providing superior predictive power.
Skip Nav Destination
,
,
,
Article navigation
14 July 2020
Research Article|
July 10 2020
Modeling the dielectric constants of crystals using machine learning
Special Collection:
Machine Learning Meets Chemical Physics
Kazuki Morita
;
Kazuki Morita
1
Department of Materials, Imperial College London
, London SW7 2AZ, United Kingdom
Search for other works by this author on:
Daniel W. Davies
;
Daniel W. Davies
2
Department of Chemistry, University College London
, London WC1H 0AJ, United Kingdom
Search for other works by this author on:
Keith T. Butler
;
Keith T. Butler
a)
3
SciML, Scientific Computer Division, Rutherford Appleton Laboratory
, Harwell OX11 0QX, United Kingdom
Search for other works by this author on:
Aron Walsh
Aron Walsh
a)
1
Department of Materials, Imperial College London
, London SW7 2AZ, United Kingdom
4
Department of Materials Science and Engineering, Yonsei University
, Seoul 03722, South Korea
Search for other works by this author on:
Kazuki Morita
1
Daniel W. Davies
2
Keith T. Butler
3,a)
Aron Walsh
1,4,a)
1
Department of Materials, Imperial College London
, London SW7 2AZ, United Kingdom
2
Department of Chemistry, University College London
, London WC1H 0AJ, United Kingdom
3
SciML, Scientific Computer Division, Rutherford Appleton Laboratory
, Harwell OX11 0QX, United Kingdom
4
Department of Materials Science and Engineering, Yonsei University
, Seoul 03722, South Korea
Note: This paper is part of the JCP Special Topic on Machine Learning Meets Chemical Physics.
J. Chem. Phys. 153, 024503 (2020)
Article history
Received:
May 08 2020
Accepted:
June 22 2020
Citation
Kazuki Morita, Daniel W. Davies, Keith T. Butler, Aron Walsh; Modeling the dielectric constants of crystals using machine learning. J. Chem. Phys. 14 July 2020; 153 (2): 024503. https://doi.org/10.1063/5.0013136
Download citation file:
Pay-Per-View Access
$40.00
Sign In
You could not be signed in. Please check your credentials and make sure you have an active account and try again.
Citing articles via
DeePMD-kit v2: A software package for deep potential models
Jinzhe Zeng, Duo Zhang, et al.
CREST—A program for the exploration of low-energy molecular chemical space
Philipp Pracht, Stefan Grimme, et al.
Related Content
First-principles calculation of the bulk photovoltaic effect in the polar compounds LiAsS2, LiAsSe2, and NaAsSe2
J. Chem. Phys. (November 2014)