Enregistré dans:
| Auteurs principaux: | , , , , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2025
|
| Sujets: | |
| Accès en ligne: | https://arxiv.org/abs/2505.23086 |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866918149479202816 |
|---|---|
| author | An, Junyi Lu, Xinyu Qu, Chao Shi, Yunfei Lin, Peijia Tang, Qianwei Xu, Licheng Cao, Fenglei Qi, Yuan |
| author_facet | An, Junyi Lu, Xinyu Qu, Chao Shi, Yunfei Lin, Peijia Tang, Qianwei Xu, Licheng Cao, Fenglei Qi, Yuan |
| contents | Equivariant Graph Neural Networks (GNNs) have significantly advanced the modeling of 3D molecular structure by leveraging group representations. However, their message passing, heavily relying on Clebsch-Gordan tensor product convolutions, suffers from restricted expressiveness due to the limited non-linearity and low degree of group representations. To overcome this, we introduce the Equivariant Spherical Transformer (EST), a novel plug-and-play framework that applies a Transformer-like architecture to the Fourier spatial domain of group representations. EST achieves higher expressiveness than conventional models while preserving the crucial equivariant inductive bias through a uniform sampling strategy of spherical Fourier transforms. As demonstrated by our experiments on challenging benchmarks like OC20 and QM9, EST-based models achieve state-of-the-art performance. For the complex molecular systems within OC20, small models empowered by EST can outperform some larger models and those using additional data. In addition to demonstrating such strong expressiveness,we provide both theoretical and experimental validation of EST's equivariance as well, paving the way for new research in this area. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2505_23086 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Equivariant Spherical Transformer for Efficient Molecular Modeling An, Junyi Lu, Xinyu Qu, Chao Shi, Yunfei Lin, Peijia Tang, Qianwei Xu, Licheng Cao, Fenglei Qi, Yuan Machine Learning Artificial Intelligence Equivariant Graph Neural Networks (GNNs) have significantly advanced the modeling of 3D molecular structure by leveraging group representations. However, their message passing, heavily relying on Clebsch-Gordan tensor product convolutions, suffers from restricted expressiveness due to the limited non-linearity and low degree of group representations. To overcome this, we introduce the Equivariant Spherical Transformer (EST), a novel plug-and-play framework that applies a Transformer-like architecture to the Fourier spatial domain of group representations. EST achieves higher expressiveness than conventional models while preserving the crucial equivariant inductive bias through a uniform sampling strategy of spherical Fourier transforms. As demonstrated by our experiments on challenging benchmarks like OC20 and QM9, EST-based models achieve state-of-the-art performance. For the complex molecular systems within OC20, small models empowered by EST can outperform some larger models and those using additional data. In addition to demonstrating such strong expressiveness,we provide both theoretical and experimental validation of EST's equivariance as well, paving the way for new research in this area. |
| title | Equivariant Spherical Transformer for Efficient Molecular Modeling |
| topic | Machine Learning Artificial Intelligence |
| url | https://arxiv.org/abs/2505.23086 |