Speaker: Johannes Kästner, Universität Stuttgart

The development of machine-learned interatomic potentials requires generating sufficiently expressive atomistic data sets. Active learning algorithms select data points on which labels, i.e., energies and forces, are calculated for inclusion in the training set. However, for batch mode active learning, i.e., when multiple data points are selected at once, conventional active-learning algorithms can perform poorly. Therefore, we investigate algorithms specifically designed for this setting and show that they can outperform traditional algorithms. We investigate selection based on the resulting training set’s informativeness, diversity, and representativeness. We propose using gradient features specific to atomistic neural networks to evaluate the informativeness of queried samples, including several approximations allowing for their efficient evaluation. To avoid selecting similar structures, we present several methods that enforce the diversity and representativeness of the selected batch. Furthermore, we use transfer learning to improve the quality of the resulting potential, use training data from cluster calculations to predict bulk properties, and present a scheme to learn tensorial quantities, like the magnetic anisotropy.