Connectivity Optimized Nested Graph Networks for Crystal Structures
This work addresses computational efficiency and performance in materials science GNNs, offering incremental improvements with broad applicability to crystal graph networks.
The authors tackled the problem of high computational cost in graph neural networks for crystalline materials by proposing an asymmetric unit cell representation that reduces atoms using symmetries, which cut training time without accuracy loss, and introduced a Nested Graph Network architecture that improved state-of-the-art results across all tasks in the MatBench benchmark.
Graph neural networks (GNNs) have been applied to a large variety of applications in materials science and chemistry. Here, we recapitulate the graph construction for crystalline (periodic) materials and investigate its impact on the GNNs model performance. We suggest the asymmetric unit cell as a representation to reduce the number of atoms by using all symmetries of the system. This substantially reduced the computational cost and thus time needed to train large graph neural networks without any loss in accuracy. Furthermore, with a simple but systematically built GNN architecture based on message passing and line graph templates, we introduce a general architecture (Nested Graph Network, NGN) that is applicable to a wide range of tasks. We show that our suggested models systematically improve state-of-the-art results across all tasks within the MatBench benchmark. Further analysis shows that optimized connectivity and deeper message functions are responsible for the improvement. Asymmetric unit cells and connectivity optimization can be generally applied to (crystal) graph networks, while our suggested nested graph framework will open new ways of systematic comparison of GNN architectures.