About this document
E(n) Equivariant Graph Neural Networks by 124187371 is a document available to read on EtoBox.
The paper presents E(n)-Equivariant Graph Neural Networks (EGNNs), a model designed to learn graph neural networks that are equivariant to various transformations such as rotations and translations, without the need for expensive higher-order representations. EGNNs can be scaled to higher-dimensional spaces and demonstrate competitive performance in tasks like dynamical systems modeling and molecular property prediction. The authors provide a detailed analysis of the model
- Author
- 124187371
- Language
- EN