Flexible SE(2) graph neural networks with applications to PDE surrogates

This repository contains code accompanying the paper Flexible SE(2) graph neural networks with applications to PDE surrogates by Maria Bånkestad, Olof Mogren and Aleksis Pirinen, all affiliated with the deep learning research group at RISE Research Institutes of Sweden. This repo is under construction.

Rollout predictions of our SE(2) equivariant GNN (middle), its non-equivariant counterpart (right), and the true simulation (left). The models have only seen the first 3 seconds of the simulation. The bottom simulation is the same as Figure 16 in the article.

Rollout predictions of our SE(2) equivariant GNN (middle), its non-equivariant counterpart (right), and the true simulation (left). The models have only seen the first 2 seconds of the simulation (Figure 9 in the article).

Citation

If you find our code and/or our paper interesting or helpful, please consider citing:

@article{baankestad2024flexible,
  title={Flexible SE (2) graph neural networks with applications to PDE surrogates},
  author={B{\aa}nkestad, Maria and Mogren, Olof and Pirinen, Aleksis},
  journal={arXiv preprint arXiv:2405.20287},
  year={2024}
}

Acknowledgements

This work was funded by Vinnova, grant number 2023-01398. The simulations were performed on the Luxembourg national supercomputer MeluXina. The authors gratefully acknowledge the LuxProvide teams for their expert support.