Predicting the Young’s Modulus of Silicate Glasses using High Throughput Molecular Dynamics Simulations and Machine Learning

Kai Yang, Xinyi Xu, Benjamin Yang, Brian Cook, Herbert Ramos, N. M. Anoop Krishnan, Morten Mattrup Smedskjær, Christian Hoover, Mathieu Bauchy

Publikation: Bidrag til tidsskriftTidsskriftartikelForskningpeer review

1 Citation (Scopus)
6 Downloads (Pure)

Resumé

The application of machine learning to predict materials’ properties usually requires a large number of consistent data for training. However, experimental datasets of high quality are not always available or self-consistent. Here, as an alternative route, we combine machine learning with high-throughput molecular dynamics simulations to predict the Young’s modulus of silicate glasses. We demonstrate that this combined approach offers good and reliable predictions over the entire compositional domain. By comparing the performances of select machine learning algorithms, we discuss the nature of the balance between accuracy, simplicity, and interpretability in machine learning.

OriginalsprogEngelsk
Artikelnummer8739
TidsskriftScientific Reports
Vol/bind9
Antal sider11
ISSN2045-2322
DOI
StatusUdgivet - 19 jun. 2019

Fingerprint

Silicates
Learning systems
Molecular dynamics
Elastic moduli
Throughput
Glass
Computer simulation
Learning algorithms
Materials properties

Citer dette

Yang, Kai ; Xu, Xinyi ; Yang, Benjamin ; Cook, Brian ; Ramos, Herbert ; Krishnan, N. M. Anoop ; Smedskjær, Morten Mattrup ; Hoover, Christian ; Bauchy, Mathieu. / Predicting the Young’s Modulus of Silicate Glasses using High Throughput Molecular Dynamics Simulations and Machine Learning. I: Scientific Reports. 2019 ; Bind 9.
@article{c6a4bbf3e84d4a63860d884d886da7a2,
title = "Predicting the Young’s Modulus of Silicate Glasses using High Throughput Molecular Dynamics Simulations and Machine Learning",
abstract = "The application of machine learning to predict materials’ properties usually requires a large number of consistent data for training. However, experimental datasets of high quality are not always available or self-consistent. Here, as an alternative route, we combine machine learning with high-throughput molecular dynamics simulations to predict the Young’s modulus of silicate glasses. We demonstrate that this combined approach offers good and reliable predictions over the entire compositional domain. By comparing the performances of select machine learning algorithms, we discuss the nature of the balance between accuracy, simplicity, and interpretability in machine learning.",
author = "Kai Yang and Xinyi Xu and Benjamin Yang and Brian Cook and Herbert Ramos and Krishnan, {N. M. Anoop} and Smedskj{\ae}r, {Morten Mattrup} and Christian Hoover and Mathieu Bauchy",
year = "2019",
month = "6",
day = "19",
doi = "10.1038/s41598-019-45344-3",
language = "English",
volume = "9",
journal = "Scientific Reports",
issn = "2045-2322",
publisher = "Nature Publishing Group",

}

Predicting the Young’s Modulus of Silicate Glasses using High Throughput Molecular Dynamics Simulations and Machine Learning. / Yang, Kai; Xu, Xinyi; Yang, Benjamin; Cook, Brian; Ramos, Herbert; Krishnan, N. M. Anoop; Smedskjær, Morten Mattrup; Hoover, Christian; Bauchy, Mathieu.

I: Scientific Reports, Bind 9, 8739, 19.06.2019.

Publikation: Bidrag til tidsskriftTidsskriftartikelForskningpeer review

TY - JOUR

T1 - Predicting the Young’s Modulus of Silicate Glasses using High Throughput Molecular Dynamics Simulations and Machine Learning

AU - Yang, Kai

AU - Xu, Xinyi

AU - Yang, Benjamin

AU - Cook, Brian

AU - Ramos, Herbert

AU - Krishnan, N. M. Anoop

AU - Smedskjær, Morten Mattrup

AU - Hoover, Christian

AU - Bauchy, Mathieu

PY - 2019/6/19

Y1 - 2019/6/19

N2 - The application of machine learning to predict materials’ properties usually requires a large number of consistent data for training. However, experimental datasets of high quality are not always available or self-consistent. Here, as an alternative route, we combine machine learning with high-throughput molecular dynamics simulations to predict the Young’s modulus of silicate glasses. We demonstrate that this combined approach offers good and reliable predictions over the entire compositional domain. By comparing the performances of select machine learning algorithms, we discuss the nature of the balance between accuracy, simplicity, and interpretability in machine learning.

AB - The application of machine learning to predict materials’ properties usually requires a large number of consistent data for training. However, experimental datasets of high quality are not always available or self-consistent. Here, as an alternative route, we combine machine learning with high-throughput molecular dynamics simulations to predict the Young’s modulus of silicate glasses. We demonstrate that this combined approach offers good and reliable predictions over the entire compositional domain. By comparing the performances of select machine learning algorithms, we discuss the nature of the balance between accuracy, simplicity, and interpretability in machine learning.

U2 - 10.1038/s41598-019-45344-3

DO - 10.1038/s41598-019-45344-3

M3 - Journal article

VL - 9

JO - Scientific Reports

JF - Scientific Reports

SN - 2045-2322

M1 - 8739

ER -