Learning local equivariant representations for large-scale atomistic dynamics
The paper presents a method that allows scaling machine learning interatomic potentials to extremely large systems, while at the same time retaining the remarkable accuracy and learning efficiency of deep equivariant models. This is obtained with an E(3)- equivariant neural network architecture that...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2023-02-01
|
| Series: | Nature Communications |
| Online Access: | https://doi.org/10.1038/s41467-023-36329-y |
| _version_ | 1811171659909955584 |
|---|---|
| author | Albert Musaelian Simon Batzner Anders Johansson Lixin Sun Cameron J. Owen Mordechai Kornbluth Boris Kozinsky |
| author_facet | Albert Musaelian Simon Batzner Anders Johansson Lixin Sun Cameron J. Owen Mordechai Kornbluth Boris Kozinsky |
| author_sort | Albert Musaelian |
| collection | DOAJ |
| description | The paper presents a method that allows scaling machine learning interatomic potentials to extremely large systems, while at the same time retaining the remarkable accuracy and learning efficiency of deep equivariant models. This is obtained with an E(3)- equivariant neural network architecture that combines the high accuracy of equivariant neural networks with the scalability of local methods. |
| first_indexed | 2024-04-10T17:17:38Z |
| format | Article |
| id | doaj.art-c834dacf32f9492789e8747321d5d7e6 |
| institution | Directory Open Access Journal |
| issn | 2041-1723 |
| language | English |
| last_indexed | 2024-04-10T17:17:38Z |
| publishDate | 2023-02-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Nature Communications |
| spelling | doaj.art-c834dacf32f9492789e8747321d5d7e62023-02-05T12:18:18ZengNature PortfolioNature Communications2041-17232023-02-0114111510.1038/s41467-023-36329-yLearning local equivariant representations for large-scale atomistic dynamicsAlbert Musaelian0Simon Batzner1Anders Johansson2Lixin Sun3Cameron J. Owen4Mordechai Kornbluth5Boris Kozinsky6Harvard UniversityHarvard UniversityHarvard UniversityHarvard UniversityHarvard UniversityRobert Bosch LLC Research and Technology CenterHarvard UniversityThe paper presents a method that allows scaling machine learning interatomic potentials to extremely large systems, while at the same time retaining the remarkable accuracy and learning efficiency of deep equivariant models. This is obtained with an E(3)- equivariant neural network architecture that combines the high accuracy of equivariant neural networks with the scalability of local methods.https://doi.org/10.1038/s41467-023-36329-y |
| spellingShingle | Albert Musaelian Simon Batzner Anders Johansson Lixin Sun Cameron J. Owen Mordechai Kornbluth Boris Kozinsky Learning local equivariant representations for large-scale atomistic dynamics Nature Communications |
| title | Learning local equivariant representations for large-scale atomistic dynamics |
| title_full | Learning local equivariant representations for large-scale atomistic dynamics |
| title_fullStr | Learning local equivariant representations for large-scale atomistic dynamics |
| title_full_unstemmed | Learning local equivariant representations for large-scale atomistic dynamics |
| title_short | Learning local equivariant representations for large-scale atomistic dynamics |
| title_sort | learning local equivariant representations for large scale atomistic dynamics |
| url | https://doi.org/10.1038/s41467-023-36329-y |
| work_keys_str_mv | AT albertmusaelian learninglocalequivariantrepresentationsforlargescaleatomisticdynamics AT simonbatzner learninglocalequivariantrepresentationsforlargescaleatomisticdynamics AT andersjohansson learninglocalequivariantrepresentationsforlargescaleatomisticdynamics AT lixinsun learninglocalequivariantrepresentationsforlargescaleatomisticdynamics AT cameronjowen learninglocalequivariantrepresentationsforlargescaleatomisticdynamics AT mordechaikornbluth learninglocalequivariantrepresentationsforlargescaleatomisticdynamics AT boriskozinsky learninglocalequivariantrepresentationsforlargescaleatomisticdynamics |