Journal of Physical Chemistry A, volume 124, issue 4, pages 731-745
Performance and Cost Assessment of Machine Learning Interatomic Potentials
Zuo Yunxing
1
,
Chen Chi-Hua
1
,
Li Xiang Guo
1
,
Deng Zhi
1
,
Chen Yiming
1
,
Behler Jörg
2
,
Csányi G.
3
,
Shapeev Alexander V.
4
,
Thompson Aidan
5
,
Wood Mitchell A.
5
,
1
Department of NanoEngineering, University of California San Diego, 9500 Gilman Drive, Mail Code 0448, La Jolla, California 92093-0448, United States
|
2
Institut für Physikalische Chemie, Theoretische Chemie, Universität Göttingen, Tammannstraße 6, 37077 Göttingen, Germany
|
3
5
Center for Computing Research, Sandia National Laboratories, Albuquerque, New Mexico 87185, United States
|
Publication type: Journal Article
Publication date: 2020-01-09
Journal:
Journal of Physical Chemistry A
Quartile SCImago
Q2
Quartile WOS
Q2
Impact factor: 2.9
ISSN: 10895639, 15205215
Physical and Theoretical Chemistry
Abstract
Machine learning of the quantitative relationship between local environment descriptors and the potential energy surface of a system of atoms has emerged as a new frontier in the development of interatomic potentials (IAPs). Here, we present a comprehensive evaluation of ML-IAPs based on four local environment descriptors --- Behler-Parrinello symmetry functions, smooth overlap of atomic positions (SOAP), the Spectral Neighbor Analysis Potential (SNAP) bispectrum components, and moment tensors --- using a diverse data set generated using high-throughput density functional theory (DFT) calculations. The data set comprising bcc (Li, Mo) and fcc (Cu, Ni) metals and diamond group IV semiconductors (Si, Ge) is chosen to span a range of crystal structures and bonding. All descriptors studied show excellent performance in predicting energies and forces far surpassing that of classical IAPs, as well as predicting properties such as elastic constants and phonon dispersion curves. We observe a general trade-off between accuracy and the degrees of freedom of each model, and consequently computational cost. We will discuss these trade-offs in the context of model selection for molecular dynamics and other applications.
Citations by journals
5
10
15
20
25
30
35
40
|
|
Journal of Chemical Physics
|
Journal of Chemical Physics
40 publications, 9.62%
|
npj Computational Materials
|
npj Computational Materials
28 publications, 6.73%
|
Physical Review Materials
|
Physical Review Materials
25 publications, 6.01%
|
Computational Materials Science
|
Computational Materials Science
20 publications, 4.81%
|
Physical Review B
|
Physical Review B
19 publications, 4.57%
|
Physical Chemistry Chemical Physics
|
Physical Chemistry Chemical Physics
15 publications, 3.61%
|
Journal of Chemical Theory and Computation
|
Journal of Chemical Theory and Computation
14 publications, 3.37%
|
Journal of Physical Chemistry C
|
Journal of Physical Chemistry C
11 publications, 2.64%
|
Journal of Physical Chemistry Letters
|
Journal of Physical Chemistry Letters
8 publications, 1.92%
|
Journal of Applied Physics
|
Journal of Applied Physics
7 publications, 1.68%
|
Journal of Physical Chemistry A
|
Journal of Physical Chemistry A
7 publications, 1.68%
|
Acta Materialia
|
Acta Materialia
6 publications, 1.44%
|
Journal of Materials Research
|
Journal of Materials Research
6 publications, 1.44%
|
Machine Learning: Science and Technology
|
Machine Learning: Science and Technology
6 publications, 1.44%
|
Chemistry of Materials
|
Chemistry of Materials
5 publications, 1.2%
|
Nature Communications
|
Nature Communications
5 publications, 1.2%
|
Computer Physics Communications
|
Computer Physics Communications
5 publications, 1.2%
|
Chemical Reviews
|
Chemical Reviews
5 publications, 1.2%
|
Journal of Physics Condensed Matter
|
Journal of Physics Condensed Matter
4 publications, 0.96%
|
Journal of Molecular Liquids
|
Journal of Molecular Liquids
4 publications, 0.96%
|
Carbon
|
Carbon
4 publications, 0.96%
|
Journal of Nuclear Materials
|
Journal of Nuclear Materials
4 publications, 0.96%
|
Chemical Science
|
Chemical Science
4 publications, 0.96%
|
Nature Computational Science
|
Nature Computational Science
3 publications, 0.72%
|
Physical Review Letters
|
Physical Review Letters
3 publications, 0.72%
|
Nanomaterials
|
Nanomaterials
3 publications, 0.72%
|
Scientific Reports
|
Scientific Reports
3 publications, 0.72%
|
Materials Horizons
|
Materials Horizons
3 publications, 0.72%
|
Journal of Physics Energy
|
Journal of Physics Energy
3 publications, 0.72%
|
5
10
15
20
25
30
35
40
|
Citations by publishers
10
20
30
40
50
60
70
80
|
|
Elsevier
|
Elsevier
77 publications, 18.51%
|
American Chemical Society (ACS)
|
American Chemical Society (ACS)
68 publications, 16.35%
|
Springer Nature
|
Springer Nature
53 publications, 12.74%
|
American Institute of Physics (AIP)
|
American Institute of Physics (AIP)
51 publications, 12.26%
|
American Physical Society (APS)
|
American Physical Society (APS)
50 publications, 12.02%
|
Royal Society of Chemistry (RSC)
|
Royal Society of Chemistry (RSC)
36 publications, 8.65%
|
IOP Publishing
|
IOP Publishing
24 publications, 5.77%
|
Multidisciplinary Digital Publishing Institute (MDPI)
|
Multidisciplinary Digital Publishing Institute (MDPI)
13 publications, 3.13%
|
Wiley
|
Wiley
11 publications, 2.64%
|
Cambridge University Press
|
Cambridge University Press
9 publications, 2.16%
|
Taylor & Francis
|
Taylor & Francis
3 publications, 0.72%
|
American Society of Civil Engineers (ASCE)
|
American Society of Civil Engineers (ASCE)
2 publications, 0.48%
|
Annual Reviews
|
Annual Reviews
2 publications, 0.48%
|
ASME
|
ASME
1 publication, 0.24%
|
Frontiers Media S.A.
|
Frontiers Media S.A.
1 publication, 0.24%
|
Materials Research Society
|
Materials Research Society
1 publication, 0.24%
|
Chinese Ceramic Society
|
Chinese Ceramic Society, 1, 0.24%
Chinese Ceramic Society
1 publication, 0.24%
|
Chemical Industry Press
|
Chemical Industry Press, 1, 0.24%
Chemical Industry Press
1 publication, 0.24%
|
Beilstein-Institut
|
Beilstein-Institut
1 publication, 0.24%
|
Pleiades Publishing
|
Pleiades Publishing
1 publication, 0.24%
|
Walter de Gruyter
|
Walter de Gruyter
1 publication, 0.24%
|
Proceedings of the National Academy of Sciences (PNAS)
|
Proceedings of the National Academy of Sciences (PNAS)
1 publication, 0.24%
|
Oxford University Press
|
Oxford University Press
1 publication, 0.24%
|
10
20
30
40
50
60
70
80
|
- We do not take into account publications that without a DOI.
- Statistics recalculated only for publications connected to researchers, organizations and labs registered on the platform.
- Statistics recalculated weekly.
{"yearsCitations":{"type":"bar","data":{"show":true,"labels":[2020,2021,2022,2023,2024],"ids":[0,0,0,0,0],"codes":[0,0,0,0,0],"imageUrls":["","","","",""],"datasets":[{"label":"Citations number","data":[53,79,100,135,47],"backgroundColor":["#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6"],"percentage":["12.74","18.99","24.04","32.45","11.3"],"barThickness":null}]},"options":{"indexAxis":"x","maintainAspectRatio":true,"scales":{"y":{"ticks":{"precision":0,"autoSkip":false,"font":{"family":"Montserrat"},"color":"#000000"}},"x":{"ticks":{"stepSize":1,"precision":0,"font":{"family":"Montserrat"},"color":"#000000"}}},"plugins":{"legend":{"position":"top","labels":{"font":{"family":"Montserrat"},"color":"#000000"}},"title":{"display":true,"text":"Citations per year","font":{"size":24,"family":"Montserrat","weight":600},"color":"#000000"}}}},"journals":{"type":"bar","data":{"show":true,"labels":["Journal of Chemical Physics","npj Computational Materials","Physical Review Materials","Computational Materials Science","Physical Review B","Physical Chemistry Chemical Physics","Journal of Chemical Theory and Computation","Journal of Physical Chemistry C","Journal of Physical Chemistry Letters","Journal of Applied Physics","Journal of Physical Chemistry A","Acta Materialia","Journal of Materials Research","Machine Learning: Science and Technology","Chemistry of Materials","Nature Communications","Computer Physics Communications","Chemical Reviews","Journal of Physics Condensed Matter","Journal of Molecular Liquids","Carbon","Journal of Nuclear Materials","Chemical Science","Nature Computational Science","Physical Review Letters","Nanomaterials","Scientific Reports","Materials Horizons","Journal of Physics Energy"],"ids":[544,19672,7673,3669,25280,1773,58,8859,21963,11951,15255,6469,1394,26682,6428,3231,20556,13718,4579,13597,37,21537,9646,28387,4343,10051,13767,826,27516],"codes":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"imageUrls":["\/storage\/images\/resized\/ARM4e6URKRsbRZvIF0vFis9DjxGloBjnBYJXbHmZ_medium.webp","\/storage\/images\/resized\/voXLqlsvTwv5p3iMQ8Dhs95nqB4AXOG7Taj7G4ra_medium.webp","\/storage\/images\/resized\/nrK64iXHTzj43wMrfN1ZoUQ0vanswGzWPN45K3jA_medium.webp","\/storage\/images\/resized\/GDnYOu1UpMMfMMRV6Aqle4H0YLLsraeD9IP9qScG_medium.webp","\/storage\/images\/resized\/nrK64iXHTzj43wMrfN1ZoUQ0vanswGzWPN45K3jA_medium.webp","\/storage\/images\/resized\/leiAYcRDGTSl5B1eCnwpSGqmDEUEfDPPoYisFGhT_medium.webp","\/storage\/images\/resized\/iLiQsFqFaSEx6chlGQ5fbAwF6VYU3WWa08hkss0g_medium.webp","\/storage\/images\/resized\/iLiQsFqFaSEx6chlGQ5fbAwF6VYU3WWa08hkss0g_medium.webp","\/storage\/images\/resized\/iLiQsFqFaSEx6chlGQ5fbAwF6VYU3WWa08hkss0g_medium.webp","\/storage\/images\/resized\/ARM4e6URKRsbRZvIF0vFis9DjxGloBjnBYJXbHmZ_medium.webp","\/storage\/images\/resized\/iLiQsFqFaSEx6chlGQ5fbAwF6VYU3WWa08hkss0g_medium.webp","\/storage\/images\/resized\/GDnYOu1UpMMfMMRV6Aqle4H0YLLsraeD9IP9qScG_medium.webp","\/storage\/images\/resized\/cF81zWlLdkmYekymGuDcSdvgBGNKaIDoMoeJtHS1_medium.webp","\/storage\/images\/resized\/LsKy6OnmmmRGcAU6CZgWQvNiP1polbaSLNrN7zqj_medium.webp","\/storage\/images\/resized\/iLiQsFqFaSEx6chlGQ5fbAwF6VYU3WWa08hkss0g_medium.webp","\/storage\/images\/resized\/voXLqlsvTwv5p3iMQ8Dhs95nqB4AXOG7Taj7G4ra_medium.webp","\/storage\/images\/resized\/GDnYOu1UpMMfMMRV6Aqle4H0YLLsraeD9IP9qScG_medium.webp","\/storage\/images\/resized\/iLiQsFqFaSEx6chlGQ5fbAwF6VYU3WWa08hkss0g_medium.webp","\/storage\/images\/resized\/LsKy6OnmmmRGcAU6CZgWQvNiP1polbaSLNrN7zqj_medium.webp","\/storage\/images\/resized\/GDnYOu1UpMMfMMRV6Aqle4H0YLLsraeD9IP9qScG_medium.webp","\/storage\/images\/resized\/GDnYOu1UpMMfMMRV6Aqle4H0YLLsraeD9IP9qScG_medium.webp","\/storage\/images\/resized\/GDnYOu1UpMMfMMRV6Aqle4H0YLLsraeD9IP9qScG_medium.webp","\/storage\/images\/resized\/leiAYcRDGTSl5B1eCnwpSGqmDEUEfDPPoYisFGhT_medium.webp","\/storage\/images\/resized\/voXLqlsvTwv5p3iMQ8Dhs95nqB4AXOG7Taj7G4ra_medium.webp","\/storage\/images\/resized\/nrK64iXHTzj43wMrfN1ZoUQ0vanswGzWPN45K3jA_medium.webp","\/storage\/images\/resized\/MjH1ITP7lMYGxeqUZfkt2BnVLgjkk413jwBV97XX_medium.webp","\/storage\/images\/resized\/voXLqlsvTwv5p3iMQ8Dhs95nqB4AXOG7Taj7G4ra_medium.webp","\/storage\/images\/resized\/leiAYcRDGTSl5B1eCnwpSGqmDEUEfDPPoYisFGhT_medium.webp","\/storage\/images\/resized\/LsKy6OnmmmRGcAU6CZgWQvNiP1polbaSLNrN7zqj_medium.webp"],"datasets":[{"label":"","data":[40,28,25,20,19,15,14,11,8,7,7,6,6,6,5,5,5,5,4,4,4,4,4,3,3,3,3,3,3],"backgroundColor":["#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6"],"percentage":[9.62,6.73,6.01,4.81,4.57,3.61,3.37,2.64,1.92,1.68,1.68,1.44,1.44,1.44,1.2,1.2,1.2,1.2,0.96,0.96,0.96,0.96,0.96,0.72,0.72,0.72,0.72,0.72,0.72],"barThickness":13}]},"options":{"indexAxis":"y","maintainAspectRatio":false,"scales":{"y":{"ticks":{"precision":0,"autoSkip":false,"font":{"family":"Montserrat"},"color":"#000000"}},"x":{"ticks":{"stepSize":null,"precision":0,"font":{"family":"Montserrat"},"color":"#000000"}}},"plugins":{"legend":{"position":"top","labels":{"font":{"family":"Montserrat"},"color":"#000000"}},"title":{"display":true,"text":"Journals","font":{"size":24,"family":"Montserrat","weight":600},"color":"#000000"}}}},"publishers":{"type":"bar","data":{"show":true,"labels":["Elsevier","American Chemical Society (ACS)","Springer Nature","American Institute of Physics (AIP)","American Physical Society (APS)","Royal Society of Chemistry (RSC)","IOP Publishing","Multidisciplinary Digital Publishing Institute (MDPI)","Wiley","Cambridge University Press","Taylor & Francis","American Society of Civil Engineers (ASCE)","Annual Reviews","ASME","Frontiers Media S.A.","Materials Research Society","Chinese Ceramic Society","Chemical Industry Press","Beilstein-Institut","Pleiades Publishing","Walter de Gruyter","Proceedings of the National Academy of Sciences (PNAS)","Oxford University Press"],"ids":[17,40,8,250,1539,123,2075,202,11,1,18,7061,6941,6977,208,6898,2953,287,6139,101,4,162,19],"codes":[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"imageUrls":["\/storage\/images\/resized\/GDnYOu1UpMMfMMRV6Aqle4H0YLLsraeD9IP9qScG_medium.webp","\/storage\/images\/resized\/iLiQsFqFaSEx6chlGQ5fbAwF6VYU3WWa08hkss0g_medium.webp","\/storage\/images\/resized\/voXLqlsvTwv5p3iMQ8Dhs95nqB4AXOG7Taj7G4ra_medium.webp","\/storage\/images\/resized\/ARM4e6URKRsbRZvIF0vFis9DjxGloBjnBYJXbHmZ_medium.webp","\/storage\/images\/resized\/nrK64iXHTzj43wMrfN1ZoUQ0vanswGzWPN45K3jA_medium.webp","\/storage\/images\/resized\/leiAYcRDGTSl5B1eCnwpSGqmDEUEfDPPoYisFGhT_medium.webp","\/storage\/images\/resized\/LsKy6OnmmmRGcAU6CZgWQvNiP1polbaSLNrN7zqj_medium.webp","\/storage\/images\/resized\/MjH1ITP7lMYGxeqUZfkt2BnVLgjkk413jwBV97XX_medium.webp","\/storage\/images\/resized\/bRyGpdm98BkAUYiK1YFNpl5Z7hPu6Gd87gbIeuG3_medium.webp","\/storage\/images\/resized\/cF81zWlLdkmYekymGuDcSdvgBGNKaIDoMoeJtHS1_medium.webp","\/storage\/images\/resized\/5YZtvLvkPZuc2JHOaZsjCvGSHFCuC3drUwN3YAc5_medium.webp","\/storage\/images\/resized\/GEGgoAuWiHS47loDJGWtCTstUblya6eHzuSGTSXU_medium.webp","\/storage\/images\/resized\/A7oGMwx1m3fFJlKJvQUVra8h0itmEa7JZW0q3Q0b_medium.webp","\/storage\/images\/resized\/yKbGpvfNleUTiwBfKt4XcIFlZMo3HgWv1KKVqn4g_medium.webp","\/storage\/images\/resized\/4QWA67eqfcfyOiA8Wk7YnqroHFqQbTsmDJUYTCTg_medium.webp","\/storage\/images\/resized\/uIgdLKrr5j2RcUZqeCI5v4KcXsLbz9m8NVJHai5O_medium.webp","","","\/storage\/images\/resized\/ex6KJoZujZOZFZh7jGfeHauiftuB3CI7iwJVFRDg_medium.webp","\/storage\/images\/resized\/oZgeErrVFhuDksyqFURLvYS1wtVSBWczh001igGo_medium.webp","\/storage\/images\/resized\/3SpVxcYL33bOvPq4sHxJLH2NeKNeDloahSUpNiO4_medium.webp","\/storage\/images\/resized\/mxFdPe9qujsfvfYfcN0QOclAiYORFb0xrRlwV8gs_medium.webp","\/storage\/images\/resized\/yNSijlgQghQF53VZuyFLA30CKDe4j3HK74Vtpnxa_medium.webp"],"datasets":[{"label":"","data":[77,68,53,51,50,36,24,13,11,9,3,2,2,1,1,1,1,1,1,1,1,1,1],"backgroundColor":["#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6","#3B82F6"],"percentage":[18.51,16.35,12.74,12.26,12.02,8.65,5.77,3.13,2.64,2.16,0.72,0.48,0.48,0.24,0.24,0.24,0.24,0.24,0.24,0.24,0.24,0.24,0.24],"barThickness":13}]},"options":{"indexAxis":"y","maintainAspectRatio":false,"scales":{"y":{"ticks":{"precision":0,"autoSkip":false,"font":{"family":"Montserrat"},"color":"#000000"}},"x":{"ticks":{"stepSize":null,"precision":0,"font":{"family":"Montserrat"},"color":"#000000"}}},"plugins":{"legend":{"position":"top","labels":{"font":{"family":"Montserrat"},"color":"#000000"}},"title":{"display":true,"text":"Publishers","font":{"size":24,"family":"Montserrat","weight":600},"color":"#000000"}}}}}
Metrics
Cite this
GOST |
RIS |
BibTex |
MLA
Cite this
GOST
Copy
Zuo Y. et al. Performance and Cost Assessment of Machine Learning Interatomic Potentials // Journal of Physical Chemistry A. 2020. Vol. 124. No. 4. pp. 731-745.
GOST all authors (up to 50)
Copy
Zuo Y., Chen C., Li X. G., Deng Z., Chen Y., Behler J., Csányi G., Shapeev A. V., Thompson A., Wood M. A., Ong S. P. Performance and Cost Assessment of Machine Learning Interatomic Potentials // Journal of Physical Chemistry A. 2020. Vol. 124. No. 4. pp. 731-745.
Cite this
RIS
Copy
TY - JOUR
DO - 10.1021/acs.jpca.9b08723
UR - https://doi.org/10.1021%2Facs.jpca.9b08723
TI - Performance and Cost Assessment of Machine Learning Interatomic Potentials
T2 - Journal of Physical Chemistry A
AU - Wood, Mitchell A.
AU - Zuo, Yunxing
AU - Chen, Chi-Hua
AU - Chen, Yiming
AU - Behler, Jörg
AU - Ong, Shyue Ping
AU - Li, Xiang Guo
AU - Deng, Zhi
AU - Csányi, G.
AU - Shapeev, Alexander V.
AU - Thompson, Aidan
PY - 2020
DA - 2020/01/09 00:00:00
PB - American Chemical Society (ACS)
SP - 731-745
IS - 4
VL - 124
SN - 1089-5639
SN - 1520-5215
ER -
Cite this
BibTex
Copy
@article{2020_Zuo,
author = {Mitchell A. Wood and Yunxing Zuo and Chi-Hua Chen and Yiming Chen and Jörg Behler and Shyue Ping Ong and Xiang Guo Li and Zhi Deng and G. Csányi and Alexander V. Shapeev and Aidan Thompson},
title = {Performance and Cost Assessment of Machine Learning Interatomic Potentials},
journal = {Journal of Physical Chemistry A},
year = {2020},
volume = {124},
publisher = {American Chemical Society (ACS)},
month = {jan},
url = {https://doi.org/10.1021%2Facs.jpca.9b08723},
number = {4},
pages = {731--745},
doi = {10.1021/acs.jpca.9b08723}
}
Cite this
MLA
Copy
Zuo, Yunxing, et al. “Performance and Cost Assessment of Machine Learning Interatomic Potentials.” Journal of Physical Chemistry A, vol. 124, no. 4, Jan. 2020, pp. 731-745. https://doi.org/10.1021%2Facs.jpca.9b08723.