Atomistic Line Graph Neural Network for Improved Materials Property Predictions

TL;DR

The paper introduces ALIGNN, a graph neural network that explicitly incorporates bond angles via line graph message passing, significantly enhancing atomistic property prediction accuracy.

Contribution

ALIGNN is a novel GNN architecture that explicitly models bond angles, improving prediction performance over existing models.

Findings

ALIGNN outperforms previous GNN models by up to 85% in accuracy.

ALIGNN achieves comparable or better training speed.

Effective inclusion of bond angles improves property prediction.

Abstract

Graph neural networks (GNN) have been shown to provide substantial performance improvements for atomistic material representation and modeling compared with descriptor-based machine learning models. While most existing GNN models for atomistic predictions are based on atomic distance information, they do not explicitly incorporate bond angles, which are critical for distinguishing many atomic structures. Furthermore, many material properties are known to be sensitive to slight changes in bond angles. We present an Atomistic Line Graph Neural Network (ALIGNN), a GNN architecture that performs message passing on both the interatomic bond graph and its line graph corresponding to bond angles. We demonstrate that angle information can be explicitly and efficiently included, leading to improved performance on multiple atomistic prediction tasks. We ALIGNN models for predicting 52 solid-state and molecular properties available in the JARVIS-DFT, Materials project, and QM9 databases. ALIGNN can outperform some previously reported GNN models on atomistic prediction tasks by up to 85% in accuracy with better or comparable model training speed.