Efficient determination of the Hamiltonian and electronic properties using graph neural network with complete local coordinates
Despite the successes of machine learning methods in physical sciences, the prediction of the Hamiltonian, and thus the electronic properties, is still unsatisfactory. Based on graph neural network (NN) architecture, we present an extendable NN model to determine the Hamiltonian from ab initio data,...
Main Authors: | , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
IOP Publishing
2023-01-01
|
Series: | Machine Learning: Science and Technology |
Subjects: | |
Online Access: | https://doi.org/10.1088/2632-2153/accb26 |
_version_ | 1797773519174500352 |
---|---|
author | Mao Su Ji-Hui Yang Hong-Jun Xiang Xin-Gao Gong |
author_facet | Mao Su Ji-Hui Yang Hong-Jun Xiang Xin-Gao Gong |
author_sort | Mao Su |
collection | DOAJ |
description | Despite the successes of machine learning methods in physical sciences, the prediction of the Hamiltonian, and thus the electronic properties, is still unsatisfactory. Based on graph neural network (NN) architecture, we present an extendable NN model to determine the Hamiltonian from ab initio data, with only local atomic structures as inputs. The rotational equivariance of the Hamiltonian is achieved by our complete local coordinates (LCs). The LC information, encoded using a convolutional NN and designed to preserve Hermitian symmetry, is used to map hopping parameters onto local structures. We demonstrate the performance of our model using graphene and SiGe random alloys as examples. We show that our NN model, although trained using small-size systems, can predict the Hamiltonian, as well as electronic properties such as band structures and densities of states for large-size systems within the ab initio accuracy, justifying its extensibility. In combination with the high efficiency of our model, which takes only seconds to get the Hamiltonian of a 1728-atom system, the present work provides a general framework to predict electronic properties efficiently and accurately, which provides new insights into computational physics and will accelerate the research for large-scale materials. |
first_indexed | 2024-03-12T22:07:40Z |
format | Article |
id | doaj.art-0756982315e04c43962a76bf041ab767 |
institution | Directory Open Access Journal |
issn | 2632-2153 |
language | English |
last_indexed | 2024-03-12T22:07:40Z |
publishDate | 2023-01-01 |
publisher | IOP Publishing |
record_format | Article |
series | Machine Learning: Science and Technology |
spelling | doaj.art-0756982315e04c43962a76bf041ab7672023-07-24T09:56:12ZengIOP PublishingMachine Learning: Science and Technology2632-21532023-01-014303501010.1088/2632-2153/accb26Efficient determination of the Hamiltonian and electronic properties using graph neural network with complete local coordinatesMao Su0https://orcid.org/0000-0003-2066-881XJi-Hui Yang1Hong-Jun Xiang2https://orcid.org/0000-0002-9396-3214Xin-Gao Gong3https://orcid.org/0000-0001-7539-5471Key Laboratory for Computational Physical Sciences (MOE), State Key Laboratory of Surface Physics, Department of Physics, Fudan University , Shanghai 200433, People’s Republic of China; Shanghai AI Laboratory , Shanghai 200030, People’s Republic of ChinaKey Laboratory for Computational Physical Sciences (MOE), State Key Laboratory of Surface Physics, Department of Physics, Fudan University , Shanghai 200433, People’s Republic of China; Shanghai Qi Zhi Institute , Shanghai 200030, People’s Republic of ChinaKey Laboratory for Computational Physical Sciences (MOE), State Key Laboratory of Surface Physics, Department of Physics, Fudan University , Shanghai 200433, People’s Republic of China; Shanghai Qi Zhi Institute , Shanghai 200030, People’s Republic of ChinaKey Laboratory for Computational Physical Sciences (MOE), State Key Laboratory of Surface Physics, Department of Physics, Fudan University , Shanghai 200433, People’s Republic of China; Shanghai Qi Zhi Institute , Shanghai 200030, People’s Republic of ChinaDespite the successes of machine learning methods in physical sciences, the prediction of the Hamiltonian, and thus the electronic properties, is still unsatisfactory. Based on graph neural network (NN) architecture, we present an extendable NN model to determine the Hamiltonian from ab initio data, with only local atomic structures as inputs. The rotational equivariance of the Hamiltonian is achieved by our complete local coordinates (LCs). The LC information, encoded using a convolutional NN and designed to preserve Hermitian symmetry, is used to map hopping parameters onto local structures. We demonstrate the performance of our model using graphene and SiGe random alloys as examples. We show that our NN model, although trained using small-size systems, can predict the Hamiltonian, as well as electronic properties such as band structures and densities of states for large-size systems within the ab initio accuracy, justifying its extensibility. In combination with the high efficiency of our model, which takes only seconds to get the Hamiltonian of a 1728-atom system, the present work provides a general framework to predict electronic properties efficiently and accurately, which provides new insights into computational physics and will accelerate the research for large-scale materials.https://doi.org/10.1088/2632-2153/accb26Hamiltonianneural networklocal coordinate |
spellingShingle | Mao Su Ji-Hui Yang Hong-Jun Xiang Xin-Gao Gong Efficient determination of the Hamiltonian and electronic properties using graph neural network with complete local coordinates Machine Learning: Science and Technology Hamiltonian neural network local coordinate |
title | Efficient determination of the Hamiltonian and electronic properties using graph neural network with complete local coordinates |
title_full | Efficient determination of the Hamiltonian and electronic properties using graph neural network with complete local coordinates |
title_fullStr | Efficient determination of the Hamiltonian and electronic properties using graph neural network with complete local coordinates |
title_full_unstemmed | Efficient determination of the Hamiltonian and electronic properties using graph neural network with complete local coordinates |
title_short | Efficient determination of the Hamiltonian and electronic properties using graph neural network with complete local coordinates |
title_sort | efficient determination of the hamiltonian and electronic properties using graph neural network with complete local coordinates |
topic | Hamiltonian neural network local coordinate |
url | https://doi.org/10.1088/2632-2153/accb26 |
work_keys_str_mv | AT maosu efficientdeterminationofthehamiltonianandelectronicpropertiesusinggraphneuralnetworkwithcompletelocalcoordinates AT jihuiyang efficientdeterminationofthehamiltonianandelectronicpropertiesusinggraphneuralnetworkwithcompletelocalcoordinates AT hongjunxiang efficientdeterminationofthehamiltonianandelectronicpropertiesusinggraphneuralnetworkwithcompletelocalcoordinates AT xingaogong efficientdeterminationofthehamiltonianandelectronicpropertiesusinggraphneuralnetworkwithcompletelocalcoordinates |