Scalable data-driven basis selection for linear machine learning interatomic potentials

Abstract Machine learning interatomic potentials provide an effective approach for accurately and efficiently modeling atomic interactions, expanding the capabilities of atomistic simulations to complex systems. However, a priori feature selection leads to high complexity, which can be detrimental to both computational cost and generalization, resulting in a need for hyperparameter tuning. We demonstrate the benefits of active set algorithms for automated data-driven feature selection. The proposed methods are implemented within the atomic cluster expansion (ACE) framework. Computational tests conducted on a variety of benchmark datasets indicate that sparse ACE models consistently enhance computational efficiency, generalization accuracy and interpretability over dense ACE models. An added benefit of the proposed algorithms is that they produce entire paths of models with varying cost/accuracy ratio.