INDEEDopt: a deep learning-based ReaxFF parameterization framework

Abstract Empirical interatomic potentials require optimization of force field parameters to tune interatomic interactions to mimic ones obtained by quantum chemistry-based methods. The optimization of the parameters is complex and requires the development of new techniques. Here, we propose an INiti...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Mert Y. Sengul, Yao Song, Nadire Nayir, Yawei Gao, Ying Hung, Tirthankar Dasgupta, Adri C. T. van Duin
Formato: article
Lenguaje:EN
Publicado: Nature Portfolio 2021
Materias:
Acceso en línea:https://doaj.org/article/e38e18c02dd2449aa860d83c0501a7d3
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
id oai:doaj.org-article:e38e18c02dd2449aa860d83c0501a7d3
record_format dspace
spelling oai:doaj.org-article:e38e18c02dd2449aa860d83c0501a7d32021-12-02T15:45:30ZINDEEDopt: a deep learning-based ReaxFF parameterization framework10.1038/s41524-021-00534-42057-3960https://doaj.org/article/e38e18c02dd2449aa860d83c0501a7d32021-05-01T00:00:00Zhttps://doi.org/10.1038/s41524-021-00534-4https://doaj.org/toc/2057-3960Abstract Empirical interatomic potentials require optimization of force field parameters to tune interatomic interactions to mimic ones obtained by quantum chemistry-based methods. The optimization of the parameters is complex and requires the development of new techniques. Here, we propose an INitial-DEsign Enhanced Deep learning-based OPTimization (INDEEDopt) framework to accelerate and improve the quality of the ReaxFF parameterization. The procedure starts with a Latin Hypercube Design (LHD) algorithm that is used to explore the parameter landscape extensively. The LHD passes the information about explored regions to a deep learning model, which finds the minimum discrepancy regions and eliminates unfeasible regions, and constructs a more comprehensive understanding of physically meaningful parameter space. We demonstrate the procedure here for the parameterization of a nickel–chromium binary force field and a tungsten–sulfide–carbon–oxygen–hydrogen quinary force field. We show that INDEEDopt produces improved accuracies in shorter development time compared to the conventional optimization method.Mert Y. SengulYao SongNadire NayirYawei GaoYing HungTirthankar DasguptaAdri C. T. van DuinNature PortfolioarticleMaterials of engineering and construction. Mechanics of materialsTA401-492Computer softwareQA76.75-76.765ENnpj Computational Materials, Vol 7, Iss 1, Pp 1-9 (2021)
institution DOAJ
collection DOAJ
language EN
topic Materials of engineering and construction. Mechanics of materials
TA401-492
Computer software
QA76.75-76.765
spellingShingle Materials of engineering and construction. Mechanics of materials
TA401-492
Computer software
QA76.75-76.765
Mert Y. Sengul
Yao Song
Nadire Nayir
Yawei Gao
Ying Hung
Tirthankar Dasgupta
Adri C. T. van Duin
INDEEDopt: a deep learning-based ReaxFF parameterization framework
description Abstract Empirical interatomic potentials require optimization of force field parameters to tune interatomic interactions to mimic ones obtained by quantum chemistry-based methods. The optimization of the parameters is complex and requires the development of new techniques. Here, we propose an INitial-DEsign Enhanced Deep learning-based OPTimization (INDEEDopt) framework to accelerate and improve the quality of the ReaxFF parameterization. The procedure starts with a Latin Hypercube Design (LHD) algorithm that is used to explore the parameter landscape extensively. The LHD passes the information about explored regions to a deep learning model, which finds the minimum discrepancy regions and eliminates unfeasible regions, and constructs a more comprehensive understanding of physically meaningful parameter space. We demonstrate the procedure here for the parameterization of a nickel–chromium binary force field and a tungsten–sulfide–carbon–oxygen–hydrogen quinary force field. We show that INDEEDopt produces improved accuracies in shorter development time compared to the conventional optimization method.
format article
author Mert Y. Sengul
Yao Song
Nadire Nayir
Yawei Gao
Ying Hung
Tirthankar Dasgupta
Adri C. T. van Duin
author_facet Mert Y. Sengul
Yao Song
Nadire Nayir
Yawei Gao
Ying Hung
Tirthankar Dasgupta
Adri C. T. van Duin
author_sort Mert Y. Sengul
title INDEEDopt: a deep learning-based ReaxFF parameterization framework
title_short INDEEDopt: a deep learning-based ReaxFF parameterization framework
title_full INDEEDopt: a deep learning-based ReaxFF parameterization framework
title_fullStr INDEEDopt: a deep learning-based ReaxFF parameterization framework
title_full_unstemmed INDEEDopt: a deep learning-based ReaxFF parameterization framework
title_sort indeedopt: a deep learning-based reaxff parameterization framework
publisher Nature Portfolio
publishDate 2021
url https://doaj.org/article/e38e18c02dd2449aa860d83c0501a7d3
work_keys_str_mv AT mertysengul indeedoptadeeplearningbasedreaxffparameterizationframework
AT yaosong indeedoptadeeplearningbasedreaxffparameterizationframework
AT nadirenayir indeedoptadeeplearningbasedreaxffparameterizationframework
AT yaweigao indeedoptadeeplearningbasedreaxffparameterizationframework
AT yinghung indeedoptadeeplearningbasedreaxffparameterizationframework
AT tirthankardasgupta indeedoptadeeplearningbasedreaxffparameterizationframework
AT adrictvanduin indeedoptadeeplearningbasedreaxffparameterizationframework
_version_ 1718385781096054784