Shotgun crystal structure prediction using machine-learned formation energies

Stable or metastable crystal structures of assembled atoms can be predicted by finding the global or local minima of the energy surface defined on the space of the atomic configurations. Generally, this requires repeated first-principles energy calculations that are impractical for large systems, such as those containing more than 30 atoms in the unit cell. Here, we have made significant progress in solving the crystal structure prediction problem with a simple but powerful machine-learning workflow; using a machine-learning surrogate for first-principles energy calculations, we performed non-iterative, single-shot screening using a large library of virtually created crystal structures. The present method relies on two key technical components: transfer learning, which enables a highly accurate energy prediction of pre-relaxed crystalline states given only a small set of training samples from first-principles calculations, and generative models to create promising and diverse crystal structures for screening. Here, first-principles calculations were performed only to generate the training samples, and for the optimization of a dozen or fewer finally narrowed-down crystal structures. Our shotgun method proved to be computationally less demanding compared to conventional methods, which heavily rely on iterations of first-principles calculations, and achieved an exceptional prediction accuracy, reaching 92.2% in a benchmark task involving the prediction of 90 different crystal structures.