Spherical Channels for Modeling Atomic Interactions

TL;DR

The paper introduces the Spherical Channel Network, a graph neural network leveraging spherical harmonics for modeling atomic energies and forces, achieving state-of-the-art accuracy and efficiency in computational chemistry tasks.

Contribution

The novel SCN architecture uses spherical channels and relaxed equivariance to improve atomic interaction modeling over existing methods.

Findings

Achieves state-of-the-art results on Open Catalyst dataset.

Utilizes spherical harmonics for rotationally equivariant embeddings.

Relaxing equivariance improves prediction accuracy.

Abstract

Modeling the energy and forces of atomic systems is a fundamental problem in computational chemistry with the potential to help address many of the world's most pressing problems, including those related to energy scarcity and climate change. These calculations are traditionally performed using Density Functional Theory, which is computationally very expensive. Machine learning has the potential to dramatically improve the efficiency of these calculations from days or hours to seconds. We propose the Spherical Channel Network (SCN) to model atomic energies and forces. The SCN is a graph neural network where nodes represent atoms and edges their neighboring atoms. The atom embeddings are a set of spherical functions, called spherical channels, represented using spherical harmonics. We demonstrate, that by rotating the embeddings based on the 3D edge orientation, more information may be utilized while maintaining the rotational equivariance of the messages. While equivariance is a desirable property, we find that by relaxing this constraint in both message passing and aggregation, improved accuracy may be achieved. We demonstrate state-of-the-art results on the large-scale Open Catalyst dataset in both energy and force prediction for numerous tasks and metrics.