Directional Message Passing for Molecular Graphs

TL;DR

This paper introduces DimeNet, a novel graph neural network that incorporates directional information in message passing, leading to significant improvements in predicting molecular properties.

Contribution

The paper proposes directional message passing with rotationally equivariant embeddings and spherical basis functions, enhancing molecular graph neural network performance.

Findings

DimeNet outperforms previous GNNs by 76% on MD17.

DimeNet achieves 31% better results on QM9.

Uses fewer parameters with better accuracy than Gaussian basis methods.

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). They do not, however, consider the spatial direction from one atom to another, despite directional information playing 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 and spherical harmonics to construct theoretically well-founded, orthogonal representations that achieve better performance than the currently prevalent Gaussian radial basis representations while using fewer than 1/4 of the parameters. We leverage these innovations to construct the directional message passing neural network (DimeNet). DimeNet outperforms previous GNNs on average by 76% on MD17 and by 31% on QM9. Our implementation is available online.