ElemNet: Deep Learning the Chemistry of Materials From Only Elemental Composition

Abstract Conventional machine learning approaches for predicting material properties from elemental compositions have emphasized the importance of leveraging domain knowledge when designing model inputs. Here, we demonstrate that by using a deep learning approach, we can bypass such manual feature e...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Dipendra Jha, Logan Ward, Arindam Paul, Wei-keng Liao, Alok Choudhary, Chris Wolverton, Ankit Agrawal
Formato:	article
Lenguaje:	EN
Publicado:	Nature Portfolio 2018
Materias:	Manual Feature Engineering Deep Neural Networks (DNN) Inorganic Crystal Structure Database (ICSD) Positive Formation Enthalpy Deep Learning Models Medicine R Science Q
Acceso en línea:	https://doaj.org/article/91d0cbe550334e66bd4c6aaf25de24d9
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:91d0cbe550334e66bd4c6aaf25de24d9
record_format	dspace
spelling	oai:doaj.org-article:91d0cbe550334e66bd4c6aaf25de24d92021-12-02T15:08:26ZElemNet: Deep Learning the Chemistry of Materials From Only Elemental Composition10.1038/s41598-018-35934-y2045-2322https://doaj.org/article/91d0cbe550334e66bd4c6aaf25de24d92018-12-01T00:00:00Zhttps://doi.org/10.1038/s41598-018-35934-yhttps://doaj.org/toc/2045-2322Abstract Conventional machine learning approaches for predicting material properties from elemental compositions have emphasized the importance of leveraging domain knowledge when designing model inputs. Here, we demonstrate that by using a deep learning approach, we can bypass such manual feature engineering requiring domain knowledge and achieve much better results, even with only a few thousand training samples. We present the design and implementation of a deep neural network model referred to as ElemNet; it automatically captures the physical and chemical interactions and similarities between different elements using artificial intelligence which allows it to predict the materials properties with better accuracy and speed. The speed and best-in-class accuracy of ElemNet enable us to perform a fast and robust screening for new material candidates in a huge combinatorial space; where we predict hundreds of thousands of chemical systems that could contain yet-undiscovered compounds.Dipendra JhaLogan WardArindam PaulWei-keng LiaoAlok ChoudharyChris WolvertonAnkit AgrawalNature PortfolioarticleManual Feature EngineeringDeep Neural Networks (DNN)Inorganic Crystal Structure Database (ICSD)Positive Formation EnthalpyDeep Learning ModelsMedicineRScienceQENScientific Reports, Vol 8, Iss 1, Pp 1-13 (2018)
institution	DOAJ
collection	DOAJ
language	EN
topic	Manual Feature Engineering Deep Neural Networks (DNN) Inorganic Crystal Structure Database (ICSD) Positive Formation Enthalpy Deep Learning Models Medicine R Science Q
spellingShingle	Manual Feature Engineering Deep Neural Networks (DNN) Inorganic Crystal Structure Database (ICSD) Positive Formation Enthalpy Deep Learning Models Medicine R Science Q Dipendra Jha Logan Ward Arindam Paul Wei-keng Liao Alok Choudhary Chris Wolverton Ankit Agrawal ElemNet: Deep Learning the Chemistry of Materials From Only Elemental Composition
description	Abstract Conventional machine learning approaches for predicting material properties from elemental compositions have emphasized the importance of leveraging domain knowledge when designing model inputs. Here, we demonstrate that by using a deep learning approach, we can bypass such manual feature engineering requiring domain knowledge and achieve much better results, even with only a few thousand training samples. We present the design and implementation of a deep neural network model referred to as ElemNet; it automatically captures the physical and chemical interactions and similarities between different elements using artificial intelligence which allows it to predict the materials properties with better accuracy and speed. The speed and best-in-class accuracy of ElemNet enable us to perform a fast and robust screening for new material candidates in a huge combinatorial space; where we predict hundreds of thousands of chemical systems that could contain yet-undiscovered compounds.
format	article
author	Dipendra Jha Logan Ward Arindam Paul Wei-keng Liao Alok Choudhary Chris Wolverton Ankit Agrawal
author_facet	Dipendra Jha Logan Ward Arindam Paul Wei-keng Liao Alok Choudhary Chris Wolverton Ankit Agrawal
author_sort	Dipendra Jha
title	ElemNet: Deep Learning the Chemistry of Materials From Only Elemental Composition
title_short	ElemNet: Deep Learning the Chemistry of Materials From Only Elemental Composition
title_full	ElemNet: Deep Learning the Chemistry of Materials From Only Elemental Composition
title_fullStr	ElemNet: Deep Learning the Chemistry of Materials From Only Elemental Composition
title_full_unstemmed	ElemNet: Deep Learning the Chemistry of Materials From Only Elemental Composition
title_sort	elemnet: deep learning the chemistry of materials from only elemental composition
publisher	Nature Portfolio
publishDate	2018
url	https://doaj.org/article/91d0cbe550334e66bd4c6aaf25de24d9
work_keys_str_mv	AT dipendrajha elemnetdeeplearningthechemistryofmaterialsfromonlyelementalcomposition AT loganward elemnetdeeplearningthechemistryofmaterialsfromonlyelementalcomposition AT arindampaul elemnetdeeplearningthechemistryofmaterialsfromonlyelementalcomposition AT weikengliao elemnetdeeplearningthechemistryofmaterialsfromonlyelementalcomposition AT alokchoudhary elemnetdeeplearningthechemistryofmaterialsfromonlyelementalcomposition AT chriswolverton elemnetdeeplearningthechemistryofmaterialsfromonlyelementalcomposition AT ankitagrawal elemnetdeeplearningthechemistryofmaterialsfromonlyelementalcomposition
_version_	1718388135268712448

ElemNet: Deep Learning the Chemistry of Materials From Only Elemental Composition

Ejemplares similares