Cargando…
Machine Learning Force Fields
[Image: see text] In recent years, the use of machine learning (ML) in computational chemistry has enabled numerous advances previously out of reach due to the computational complexity of traditional electronic-structure methods. One of the most promising applications is the construction of ML-based...
Autores principales: | , , , , , , , |
---|---|
Formato: | Online Artículo Texto |
Lenguaje: | English |
Publicado: |
American Chemical Society
2021
|
Acceso en línea: | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8391964/ https://www.ncbi.nlm.nih.gov/pubmed/33705118 http://dx.doi.org/10.1021/acs.chemrev.0c01111 |
_version_ | 1783743395653484544 |
---|---|
author | Unke, Oliver T. Chmiela, Stefan Sauceda, Huziel E. Gastegger, Michael Poltavsky, Igor Schütt, Kristof T. Tkatchenko, Alexandre Müller, Klaus-Robert |
author_facet | Unke, Oliver T. Chmiela, Stefan Sauceda, Huziel E. Gastegger, Michael Poltavsky, Igor Schütt, Kristof T. Tkatchenko, Alexandre Müller, Klaus-Robert |
author_sort | Unke, Oliver T. |
collection | PubMed |
description | [Image: see text] In recent years, the use of machine learning (ML) in computational chemistry has enabled numerous advances previously out of reach due to the computational complexity of traditional electronic-structure methods. One of the most promising applications is the construction of ML-based force fields (FFs), with the aim to narrow the gap between the accuracy of ab initio methods and the efficiency of classical FFs. The key idea is to learn the statistical relation between chemical structure and potential energy without relying on a preconceived notion of fixed chemical bonds or knowledge about the relevant interactions. Such universal ML approximations are in principle only limited by the quality and quantity of the reference data used to train them. This review gives an overview of applications of ML-FFs and the chemical insights that can be obtained from them. The core concepts underlying ML-FFs are described in detail, and a step-by-step guide for constructing and testing them from scratch is given. The text concludes with a discussion of the challenges that remain to be overcome by the next generation of ML-FFs. |
format | Online Article Text |
id | pubmed-8391964 |
institution | National Center for Biotechnology Information |
language | English |
publishDate | 2021 |
publisher | American Chemical Society |
record_format | MEDLINE/PubMed |
spelling | pubmed-83919642021-08-31 Machine Learning Force Fields Unke, Oliver T. Chmiela, Stefan Sauceda, Huziel E. Gastegger, Michael Poltavsky, Igor Schütt, Kristof T. Tkatchenko, Alexandre Müller, Klaus-Robert Chem Rev [Image: see text] In recent years, the use of machine learning (ML) in computational chemistry has enabled numerous advances previously out of reach due to the computational complexity of traditional electronic-structure methods. One of the most promising applications is the construction of ML-based force fields (FFs), with the aim to narrow the gap between the accuracy of ab initio methods and the efficiency of classical FFs. The key idea is to learn the statistical relation between chemical structure and potential energy without relying on a preconceived notion of fixed chemical bonds or knowledge about the relevant interactions. Such universal ML approximations are in principle only limited by the quality and quantity of the reference data used to train them. This review gives an overview of applications of ML-FFs and the chemical insights that can be obtained from them. The core concepts underlying ML-FFs are described in detail, and a step-by-step guide for constructing and testing them from scratch is given. The text concludes with a discussion of the challenges that remain to be overcome by the next generation of ML-FFs. American Chemical Society 2021-03-11 2021-08-25 /pmc/articles/PMC8391964/ /pubmed/33705118 http://dx.doi.org/10.1021/acs.chemrev.0c01111 Text en © 2021 The Authors. Published by American Chemical Society https://creativecommons.org/licenses/by-nc-nd/4.0/Permits non-commercial access and re-use, provided that author attribution and integrity are maintained; but does not permit creation of adaptations or other derivative works (https://creativecommons.org/licenses/by-nc-nd/4.0/). |
spellingShingle | Unke, Oliver T. Chmiela, Stefan Sauceda, Huziel E. Gastegger, Michael Poltavsky, Igor Schütt, Kristof T. Tkatchenko, Alexandre Müller, Klaus-Robert Machine Learning Force Fields |
title | Machine Learning Force Fields |
title_full | Machine Learning Force Fields |
title_fullStr | Machine Learning Force Fields |
title_full_unstemmed | Machine Learning Force Fields |
title_short | Machine Learning Force Fields |
title_sort | machine learning force fields |
url | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8391964/ https://www.ncbi.nlm.nih.gov/pubmed/33705118 http://dx.doi.org/10.1021/acs.chemrev.0c01111 |
work_keys_str_mv | AT unkeolivert machinelearningforcefields AT chmielastefan machinelearningforcefields AT saucedahuziele machinelearningforcefields AT gasteggermichael machinelearningforcefields AT poltavskyigor machinelearningforcefields AT schuttkristoft machinelearningforcefields AT tkatchenkoalexandre machinelearningforcefields AT mullerklausrobert machinelearningforcefields |