Equivariant Graph Neural Networks for Prediction of Tensor Material Properties of Crystals
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2024
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866929604154884096 |
|---|---|
| author | Heilman, Alex Schlesinger, Claire Yan, Qimin |
| author_facet | Heilman, Alex Schlesinger, Claire Yan, Qimin |
| contents | Modern E(3)-Equivariant networks may be used to predict rotationally equivariant properties, including tensorial quantities. Three such quantities: the dielectric, piezoelectric, and elasticity tensors, are computationally expensive to produce ab initio for crystalline systems; however, with greater availability of such data in large material property databases, we now have a sufficient target space to begin training equivariant models in the prediction of such properties. Here we explicitly develop spherical harmonic decompositions of these tensorial properties using their general symmetries. We then apply three distinct E(3)-equivariant convolutional structures to the prediction of the components of these decompositions, allowing us to predict the aforementioned tensorial quantities in an equivariant manner and compare performance. We further report results testing the transferability of these predictive models between different tensorial target sets. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2406_03563 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Equivariant Graph Neural Networks for Prediction of Tensor Material Properties of Crystals Heilman, Alex Schlesinger, Claire Yan, Qimin Computational Physics Materials Science Modern E(3)-Equivariant networks may be used to predict rotationally equivariant properties, including tensorial quantities. Three such quantities: the dielectric, piezoelectric, and elasticity tensors, are computationally expensive to produce ab initio for crystalline systems; however, with greater availability of such data in large material property databases, we now have a sufficient target space to begin training equivariant models in the prediction of such properties. Here we explicitly develop spherical harmonic decompositions of these tensorial properties using their general symmetries. We then apply three distinct E(3)-equivariant convolutional structures to the prediction of the components of these decompositions, allowing us to predict the aforementioned tensorial quantities in an equivariant manner and compare performance. We further report results testing the transferability of these predictive models between different tensorial target sets. |
| title | Equivariant Graph Neural Networks for Prediction of Tensor Material Properties of Crystals |
| topic | Computational Physics Materials Science |
| url | https://arxiv.org/abs/2406.03563 |