Relaxed Equivariant Graph Neural Networks

Elyssa Hofgard, Rui Wang, Robin Walters, Tess Smidt
7/30/2024
cs.LG

No official code found. Best implementation: atomicarchitects/RelaxedE3NN (4 stars, MIT, last commit 2024-11-08).

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.

View on arXivView PDF

Code Implementations(1)

atomicarchitects/RelaxedE3NNCommunity100%
MIT

[GRaM at ICML'24] Relaxed Equivariant Graph Neural Networks

41Python, Jupyter NotebookNov 7, 20241 years agoMIT

Similar Papers

Improving Equivariant Networks with Probabilistic Symmetry Breaking

86% match2025

On Universality Classes of Equivariant Networks

85% match2025

Approximately Piecewise E(3) Equivariant Point Networks

85% match2024

Symmetry-Based Structured Matrices for Efficient Approximately Equivariant Networks

85% match1 repo62024

Symmetry-preserving neural networks in lattice field theories

84% match2025

Graph Automorphism Group Equivariant Neural Networks

84% match2023

Cite this paper

@article{hofgard2024relaxed,
  title  = {Relaxed Equivariant Graph Neural Networks},
  author = {Elyssa Hofgard and Rui Wang and Robin Walters and Tess Smidt},
  year   = {2024},
  eprint = {2407.20471},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2407.20471v2}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2407.20471.svg)](https://codeofpaper.com/paper/2407.20471)

Discussion