Relaxed Equivariant Graph Neural Networks
Abstract
3D Euclidean symmetry equivariant neural networks have demonstrated notable success in modeling complex physical systems. We introduce a framework for relaxed $E(3)$ graph equivariant neural networks that can learn and represent symmetry breaking within continuous groups. Building on the existing e3nn framework, we propose the use of relaxed weights to allow for controlled symmetry breaking. We show empirically that these relaxed weights learn the correct amount of symmetry breaking.
Cite
Text
Hofgard et al. "Relaxed Equivariant Graph Neural Networks." ICML 2024 Workshops: GRaM, 2024.Markdown
[Hofgard et al. "Relaxed Equivariant Graph Neural Networks." ICML 2024 Workshops: GRaM, 2024.](https://mlanthology.org/icmlw/2024/hofgard2024icmlw-relaxed/)BibTeX
@inproceedings{hofgard2024icmlw-relaxed,
title = {{Relaxed Equivariant Graph Neural Networks}},
author = {Hofgard, Elyssa and Wang, Rui and Walters, Robin and Smidt, Tess},
booktitle = {ICML 2024 Workshops: GRaM},
year = {2024},
url = {https://mlanthology.org/icmlw/2024/hofgard2024icmlw-relaxed/}
}