Transferring predictions of formation energy across lattices of increasing size
In this study, we show the transferability of graph convolutional neural network (GCNN) predictions of the formation energy of the nickel-platinum solid solution alloy across atomic structures of increasing sizes. The original dataset was generated with the large-scale atomic/molecular massively par...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IOP Publishing
2024-01-01
|
| Series: | Machine Learning: Science and Technology |
| Subjects: | |
| Online Access: | https://doi.org/10.1088/2632-2153/ad3d2c |
| _version_ | 1797201833778741248 |
|---|---|
| author | Massimiliano Lupo Pasini Mariia Karabin Markus Eisenbach |
| author_facet | Massimiliano Lupo Pasini Mariia Karabin Markus Eisenbach |
| author_sort | Massimiliano Lupo Pasini |
| collection | DOAJ |
| description | In this study, we show the transferability of graph convolutional neural network (GCNN) predictions of the formation energy of the nickel-platinum solid solution alloy across atomic structures of increasing sizes. The original dataset was generated with the large-scale atomic/molecular massively parallel simulator using the second nearest-neighbor modified embedded-atom method empirical interatomic potential. Geometry optimization was performed on the initially randomly generated face centered cubic crystal structures and the formation energy has been calculated at each step of the geometry optimization, with configurations spanning the whole compositional range. Using data from various steps of the geometry optimization, we first trained our open-source, scalable implementation of GCNN called HydraGNN on a lattice of 256 atoms, which accounts well for the short-range interactions. Using this data, we predicted the formation energy for lattices of 864 atoms and 2048 atoms, which resulted in lower-than-expected accuracy due to the long-range interactions present in these larger lattices. We accounted for the long-range interactions by including a small amount of training data representative for those two larger sizes, whereupon the predictions of HydraGNN scaled linearly with the size of the lattice. Therefore, our strategy ensured scalability while reducing significantly the computational cost of training on larger lattice sizes. |
| first_indexed | 2024-04-24T07:53:50Z |
| format | Article |
| id | doaj.art-de9263242f884da285a0926f80d5d6b0 |
| institution | Directory Open Access Journal |
| issn | 2632-2153 |
| language | English |
| last_indexed | 2024-04-24T07:53:50Z |
| publishDate | 2024-01-01 |
| publisher | IOP Publishing |
| record_format | Article |
| series | Machine Learning: Science and Technology |
| spelling | doaj.art-de9263242f884da285a0926f80d5d6b02024-04-18T07:32:32ZengIOP PublishingMachine Learning: Science and Technology2632-21532024-01-015202501510.1088/2632-2153/ad3d2cTransferring predictions of formation energy across lattices of increasing sizeMassimiliano Lupo Pasini0https://orcid.org/0000-0002-4980-6924Mariia Karabin1Markus Eisenbach2https://orcid.org/0000-0001-8805-8327Oak Ridge National Laboratory, Computational Sciences and Engineering Division , Oak Ridge, TN 37831, United States of AmericaOak Ridge National Laboratory, National Center for Computational Sciences Division , Oak Ridge, TN 37831, United States of AmericaOak Ridge National Laboratory, National Center for Computational Sciences Division , Oak Ridge, TN 37831, United States of AmericaIn this study, we show the transferability of graph convolutional neural network (GCNN) predictions of the formation energy of the nickel-platinum solid solution alloy across atomic structures of increasing sizes. The original dataset was generated with the large-scale atomic/molecular massively parallel simulator using the second nearest-neighbor modified embedded-atom method empirical interatomic potential. Geometry optimization was performed on the initially randomly generated face centered cubic crystal structures and the formation energy has been calculated at each step of the geometry optimization, with configurations spanning the whole compositional range. Using data from various steps of the geometry optimization, we first trained our open-source, scalable implementation of GCNN called HydraGNN on a lattice of 256 atoms, which accounts well for the short-range interactions. Using this data, we predicted the formation energy for lattices of 864 atoms and 2048 atoms, which resulted in lower-than-expected accuracy due to the long-range interactions present in these larger lattices. We accounted for the long-range interactions by including a small amount of training data representative for those two larger sizes, whereupon the predictions of HydraGNN scaled linearly with the size of the lattice. Therefore, our strategy ensured scalability while reducing significantly the computational cost of training on larger lattice sizes.https://doi.org/10.1088/2632-2153/ad3d2csolid solution alloysgraph neural networksatomistic materials modelingformation energy |
| spellingShingle | Massimiliano Lupo Pasini Mariia Karabin Markus Eisenbach Transferring predictions of formation energy across lattices of increasing size Machine Learning: Science and Technology solid solution alloys graph neural networks atomistic materials modeling formation energy |
| title | Transferring predictions of formation energy across lattices of increasing size |
| title_full | Transferring predictions of formation energy across lattices of increasing size |
| title_fullStr | Transferring predictions of formation energy across lattices of increasing size |
| title_full_unstemmed | Transferring predictions of formation energy across lattices of increasing size |
| title_short | Transferring predictions of formation energy across lattices of increasing size |
| title_sort | transferring predictions of formation energy across lattices of increasing size |
| topic | solid solution alloys graph neural networks atomistic materials modeling formation energy |
| url | https://doi.org/10.1088/2632-2153/ad3d2c |
| work_keys_str_mv | AT massimilianolupopasini transferringpredictionsofformationenergyacrosslatticesofincreasingsize AT mariiakarabin transferringpredictionsofformationenergyacrosslatticesofincreasingsize AT markuseisenbach transferringpredictionsofformationenergyacrosslatticesofincreasingsize |