Accelerating crystal structure prediction by machine-learning interatomic potentials with active learning

Evgeny V. Podryabinkin, Evgeny V. Tikhonov, Alexander V. Shapeev, Artem R. Oganov
(Submitted on 21 Feb 2018)

In this letter we propose a new methodology for crystal structure prediction, which is based on the evolutionary algorithm USPEX and the machine-learning interatomic potentials (MLIPs) actively learning on-the-fly. Our methodology allows for an automated construction of an interatomic interaction model from scratch replacing the expensive DFT with a speedup of several orders of magnitude. Those structures are then tested on DFT, ensuring that our machine-learning model does not introduce any prediction error. We tested our methodology on a very challenging problem of prediction of boron allotropes including those which have more than 100 atoms in the primitive cell. All the the main allotropes have been reproduced and a new 54-atom structure have been found at very modest computational efforts

https://arxiv.org/abs/1802.07605