E(n) Equivariant Graph Neural Networks.¶
Satorras, V. G., Hoogeboom, & Welling, M. (2021). E(n) Equivariant Graph Neural Networks. Proceedings of the 38th International Conference on Machine Learning (ICML), 9323-9332.
Cited by¶
1 citation across 1 artifact.
Each citation links to the sentence it supports in the citing article.
Primes¶
- Equivariance
- Machine learning: Convolutional layers are translation-equivariant (shift the image, the feature map shifts identically), the founding insight of geometric deep learning; group-equivariant CNNs extend this to rotations and reflections; equivariant graph neural networks respect permutation symmetry of nodes; equivariant transformers and message-passing networks for molecules respect the rotation/translation symmetry of 3D space, with Satorras and colleagues (2021) showing that E(n)-equivariant graph networks predict molecular properties more sample-efficiently than unconstrained models.
This sourceShows E(n)-equivariant graph networks (EGNNs) predict molecular and physical-system properties (QM9) competitively or better than unconstrained models without expensive higher-order representations, by building rotation/translation/reflection/permutation equivariance into the architecture.
- Machine learning: Convolutional layers are translation-equivariant (shift the image, the feature map shifts identically), the founding insight of geometric deep learning; group-equivariant CNNs extend this to rotations and reflections; equivariant graph neural networks respect permutation symmetry of nodes; equivariant transformers and message-passing networks for molecules respect the rotation/translation symmetry of 3D space, with Satorras and colleagues (2021) showing that E(n)-equivariant graph networks predict molecular properties more sample-efficiently than unconstrained models.
Verification¶
This reference passed the adversarial substantiation pipeline: it was checked to exist and to support the claim it is attached to. See how references were verified.
Registry ID ref:f0a000d9136d · see in the full table