Guardado en:
| Autores principales: | , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2024
|
| Materias: | |
| Acceso en línea: | https://arxiv.org/abs/2402.19327 |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866914930710544384 |
|---|---|
| author | Xie, Fankai Lu, Tenglong Meng, Sheng Liu, Miao |
| author_facet | Xie, Fankai Lu, Tenglong Meng, Sheng Liu, Miao |
| contents | This study introduces a novel AI force field, namely graph-based pre-trained transformer force field (GPTFF), which can simulate arbitrary inorganic systems with good precision and generalizability. Harnessing a large trove of the data and the attention mechanism of transformer algorithms, the model can accurately predict energy, atomic forces, and stress with Mean Absolute Error (MAE) values of 32 meV/atom, 71 meV/Å, and 0.365 GPa, respectively. The dataset used to train the model includes 37.8 million single-point energies, 11.7 billion force pairs, and 340.2 million stresses. We also demonstrated that GPTFF can be universally used to simulate various physical systems, such as crystal structure optimization, phase transition simulations, and mass transport. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2402_19327 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | GPTFF: A high-accuracy out-of-the-box universal AI force field for arbitrary inorganic materials Xie, Fankai Lu, Tenglong Meng, Sheng Liu, Miao Materials Science This study introduces a novel AI force field, namely graph-based pre-trained transformer force field (GPTFF), which can simulate arbitrary inorganic systems with good precision and generalizability. Harnessing a large trove of the data and the attention mechanism of transformer algorithms, the model can accurately predict energy, atomic forces, and stress with Mean Absolute Error (MAE) values of 32 meV/atom, 71 meV/Å, and 0.365 GPa, respectively. The dataset used to train the model includes 37.8 million single-point energies, 11.7 billion force pairs, and 340.2 million stresses. We also demonstrated that GPTFF can be universally used to simulate various physical systems, such as crystal structure optimization, phase transition simulations, and mass transport. |
| title | GPTFF: A high-accuracy out-of-the-box universal AI force field for arbitrary inorganic materials |
| topic | Materials Science |
| url | https://arxiv.org/abs/2402.19327 |