Geometry-enhanced pretraining on interatomic potentials

Machine learning interatomic potentials (MLIPs) describe the interactions between atoms in materials and molecules by learning them from a reference database generated by ab initio calculations. MLIPs can accurately and efficiently predict such interactions and have been applied to various fields of physical science. However, high-performance MLIPs rely on a large amount of labelled data, which are costly to obtain by ab initio calculations. Here we propose a geometric structure learning framework that leverages unlabelled configurations to improve the performance of MLIPs. Our framework consists of two stages: first, using classical molecular dynamics simulations to generate unlabelled configurations of the target molecular system; and second, applying geometry-enhanced self-supervised learning techniques, including masking, denoising and contrastive learning, to capture structural information. We evaluate our framework on various benchmarks ranging from small molecule datasets to complex periodic molecular systems with more types of elements. We show that our method significantly improves the accuracy and generalization of MLIPs with only a few additional computational costs and is compatible with different invariant or equivariant graph neural network architectures. Our method enhances MLIPs and advances the simulations of molecular systems. Using machine learning methods to model interatomic potentials enables molecular dynamics simulations with ab initio level accuracy at a relatively low computational cost, but requires a large number of labelled training data obtained through expensive ab initio computations. Cui and colleagues propose a geometric learning framework that leverages self-supervised learning pretraining to enhance existing machine learning based interatomic potential models at a negligible additional computational cost.