Astrid Klipfel

Unified Model for Crystalline Material Generation

In this paper we have proposed an equivariant message passing neural network for periodic materials. Our proposed model is based on GemNet and can act on both atomic position and lattice shape.

commutative diagram

Vector fields of the lattice parameters

How to cite

@inproceedings{klipfel_gemsnet_ijcai_2023,
  author       = {Astrid Klipfel and Ya{\"{e}}l Fr{\'{e}}gier and Adlane Sayede and Zied Bouraoui},
  title        = {Unified Model for Crystalline Material Generation},
  booktitle    = {Proceedings of the Thirty-Second International Joint Conference on Artificial Intelligence, {IJCAI} 2023, 19th-25th August 2023, Macao, SAR, China},
  pages        = {6031--6039},
  year         = {2023},
  url          = {https://doi.org/10.24963/ijcai.2023/669},
  doi          = {10.24963/IJCAI.2023/669},
}