Applications and training sets of machine learning potentials

Recently, machine learning potentials (MLPs) have been attracting interest as an alternative to the computationally expensive density-functional theory (DFT) calculations. The data-driven approach in MLPs requires carefully curated training datasets, which define the valid domain of simulations. The...

Full description

Bibliographic Details
Main Authors: Changho Hong, Jaehoon Kim, Jaesun Kim, Jisu Jung, Suyeon Ju, Jeong Min Choi, Seungwu Han
Format: Article
Language:English
Published: Taylor & Francis Group 2023-10-01
Series:Science and Technology of Advanced Materials: Methods
Subjects:
Online Access:http://dx.doi.org/10.1080/27660400.2023.2269948
_version_ 1797641042885869568
author Changho Hong
Jaehoon Kim
Jaesun Kim
Jisu Jung
Suyeon Ju
Jeong Min Choi
Seungwu Han
author_facet Changho Hong
Jaehoon Kim
Jaesun Kim
Jisu Jung
Suyeon Ju
Jeong Min Choi
Seungwu Han
author_sort Changho Hong
collection DOAJ
description Recently, machine learning potentials (MLPs) have been attracting interest as an alternative to the computationally expensive density-functional theory (DFT) calculations. The data-driven approach in MLPs requires carefully curated training datasets, which define the valid domain of simulations. Therefore, acquiring training datasets that comprehensively span the domain of the desired simulations is important. In this review, we attempt to set guidelines for the systematic construction of training datasets according to target simulations. To this end, we extensively analyze the training sets in previous literature according to four application types: thermal properties, diffusion properties, structure prediction, and chemical reactions. In each application, we summarize characteristic reference structures and discuss specific parameters for DFT calculations such as MD conditions. We hope this review serves as a comprehensive guide for researchers and practitioners aiming to harness the capabilities of MLPs in material simulations.
first_indexed 2024-03-11T13:39:55Z
format Article
id doaj.art-f5ca5df4b3014789848f16b9b0972d47
institution Directory Open Access Journal
issn 2766-0400
language English
last_indexed 2024-03-11T13:39:55Z
publishDate 2023-10-01
publisher Taylor & Francis Group
record_format Article
series Science and Technology of Advanced Materials: Methods
spelling doaj.art-f5ca5df4b3014789848f16b9b0972d472023-11-02T13:48:31ZengTaylor & Francis GroupScience and Technology of Advanced Materials: Methods2766-04002023-10-010010.1080/27660400.2023.22699482269948Applications and training sets of machine learning potentialsChangho Hong0Jaehoon Kim1Jaesun Kim2Jisu Jung3Suyeon Ju4Jeong Min Choi5Seungwu Han6Seoul National UniversitySeoul National UniversitySeoul National UniversitySeoul National UniversitySeoul National UniversitySeoul National UniversitySeoul National UniversityRecently, machine learning potentials (MLPs) have been attracting interest as an alternative to the computationally expensive density-functional theory (DFT) calculations. The data-driven approach in MLPs requires carefully curated training datasets, which define the valid domain of simulations. Therefore, acquiring training datasets that comprehensively span the domain of the desired simulations is important. In this review, we attempt to set guidelines for the systematic construction of training datasets according to target simulations. To this end, we extensively analyze the training sets in previous literature according to four application types: thermal properties, diffusion properties, structure prediction, and chemical reactions. In each application, we summarize characteristic reference structures and discuss specific parameters for DFT calculations such as MD conditions. We hope this review serves as a comprehensive guide for researchers and practitioners aiming to harness the capabilities of MLPs in material simulations.http://dx.doi.org/10.1080/27660400.2023.2269948machine learning potentialtraining setpotential energy surfacedensity functional theorymolecular dynamics
spellingShingle Changho Hong
Jaehoon Kim
Jaesun Kim
Jisu Jung
Suyeon Ju
Jeong Min Choi
Seungwu Han
Applications and training sets of machine learning potentials
Science and Technology of Advanced Materials: Methods
machine learning potential
training set
potential energy surface
density functional theory
molecular dynamics
title Applications and training sets of machine learning potentials
title_full Applications and training sets of machine learning potentials
title_fullStr Applications and training sets of machine learning potentials
title_full_unstemmed Applications and training sets of machine learning potentials
title_short Applications and training sets of machine learning potentials
title_sort applications and training sets of machine learning potentials
topic machine learning potential
training set
potential energy surface
density functional theory
molecular dynamics
url http://dx.doi.org/10.1080/27660400.2023.2269948
work_keys_str_mv AT changhohong applicationsandtrainingsetsofmachinelearningpotentials
AT jaehoonkim applicationsandtrainingsetsofmachinelearningpotentials
AT jaesunkim applicationsandtrainingsetsofmachinelearningpotentials
AT jisujung applicationsandtrainingsetsofmachinelearningpotentials
AT suyeonju applicationsandtrainingsetsofmachinelearningpotentials
AT jeongminchoi applicationsandtrainingsetsofmachinelearningpotentials
AT seungwuhan applicationsandtrainingsetsofmachinelearningpotentials