A Graph Neural Network for the Era of Large Atomistic Models
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866914274133147648 |
|---|---|
| author | Zhang, Duo Peng, Anyang Cai, Chun Li, Wentao Zhou, Yuanchang Zeng, Jinzhe Guo, Mingyu Zhang, Chengqian Li, Bowen Jiang, Hong Zhu, Tong Jia, Weile Zhang, Linfeng Wang, Han |
| author_facet | Zhang, Duo Peng, Anyang Cai, Chun Li, Wentao Zhou, Yuanchang Zeng, Jinzhe Guo, Mingyu Zhang, Chengqian Li, Bowen Jiang, Hong Zhu, Tong Jia, Weile Zhang, Linfeng Wang, Han |
| contents | Foundation models, or large atomistic models (LAMs), aim to universally represent the ground-state potential energy surface (PES) of atomistic systems as defined by density functional theory (DFT). The scaling law is pivotal in the development of large models, suggesting that their generalizability in downstream tasks consistently improves with increased model size, expanded training datasets, and larger computational budgets. In this study, we present DPA3, a multi-layer graph neural network founded on line graph series (LiGS), designed explicitly for the era of LAMs. We demonstrate that the generalization error of the DPA3 model adheres to the scaling law. The scalability in the number of model parameters is attained by stacking additional layers within DPA3. Additionally, the model employs a dataset encoding mechanism that decouples the scaling of training data size from the model size within its multi-task training framework. When trained as problem-oriented potential energy models, the DPA3 model exhibits superior accuracy in the majority of benchmark cases, encompassing systems with diverse features, including molecules, bulk materials, surface and cluster catalysts, two-dimensional materials, and battery materials. When trained as a LAM on the OpenLAM-v1 dataset, the DPA-3.1-3M model exhibits lowest overall zero-shot generalization error across 12 downstream tasks spanning a diverse array of research domains. This performance suggests superior accuracy as an out-of-the-box potential model, requiring minimal fine-tuning data for downstream scientific applications. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2506_01686 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | A Graph Neural Network for the Era of Large Atomistic Models Zhang, Duo Peng, Anyang Cai, Chun Li, Wentao Zhou, Yuanchang Zeng, Jinzhe Guo, Mingyu Zhang, Chengqian Li, Bowen Jiang, Hong Zhu, Tong Jia, Weile Zhang, Linfeng Wang, Han Computational Physics Foundation models, or large atomistic models (LAMs), aim to universally represent the ground-state potential energy surface (PES) of atomistic systems as defined by density functional theory (DFT). The scaling law is pivotal in the development of large models, suggesting that their generalizability in downstream tasks consistently improves with increased model size, expanded training datasets, and larger computational budgets. In this study, we present DPA3, a multi-layer graph neural network founded on line graph series (LiGS), designed explicitly for the era of LAMs. We demonstrate that the generalization error of the DPA3 model adheres to the scaling law. The scalability in the number of model parameters is attained by stacking additional layers within DPA3. Additionally, the model employs a dataset encoding mechanism that decouples the scaling of training data size from the model size within its multi-task training framework. When trained as problem-oriented potential energy models, the DPA3 model exhibits superior accuracy in the majority of benchmark cases, encompassing systems with diverse features, including molecules, bulk materials, surface and cluster catalysts, two-dimensional materials, and battery materials. When trained as a LAM on the OpenLAM-v1 dataset, the DPA-3.1-3M model exhibits lowest overall zero-shot generalization error across 12 downstream tasks spanning a diverse array of research domains. This performance suggests superior accuracy as an out-of-the-box potential model, requiring minimal fine-tuning data for downstream scientific applications. |
| title | A Graph Neural Network for the Era of Large Atomistic Models |
| topic | Computational Physics |
| url | https://arxiv.org/abs/2506.01686 |