- TL;DR: Directional message passing incorporates spatial directional information to improve graph neural networks.
- Abstract: Graph neural networks have recently achieved great successes in predicting quantum mechanical properties of molecules. These models represent a molecule as a graph using only the distance between atoms (nodes) and not the spatial direction from one atom to another. However, directional information plays a central role in empirical potentials for molecules, e.g. in angular potentials. To alleviate this limitation we propose directional message passing, in which we embed the messages passed between atoms instead of the atoms themselves. Each message is associated with a direction in coordinate space. These directional message embeddings are rotationally equivariant since the associated directions rotate with the molecule. We propose a message passing scheme analogous to belief propagation, which uses the directional information by transforming messages based on the angle between them. Additionally, we use spherical Bessel functions to construct a theoretically well-founded, orthogonal radial basis that achieves better performance than the currently prevalent Gaussian radial basis functions while using more than 4x fewer parameters. We leverage these innovations to construct the directional message passing neural network (DimeNet). DimeNet outperforms previous GNNs on average by 77% on MD17 and by 41% on QM9.
- Code: https://figshare.com/s/8f5355542c7ed98ffb2d
- Keywords: GNN, Graph neural network, message passing, graphs, equivariance, molecules