Title:Physically-motivated Recursively Embedded Atom Neural Networks: Incorporating Local Completeness and Nonlocality

Abstract:Recent advances in machine-learned interatomic potentials have largely benefited from the atomistic representation and locally invariant many-body descriptors. It was however recently argued that including three- (or even four-) body features is insufficient to distinguish some specific local environments. Utilizing an embedded density descriptor made by linear combinations of neighboring atomic orbitals and realizing that each orbital coefficient physically depends on its own local environment, we propose a recursively embedded atom neural network model. We formally prove that this model can efficiently incorporate complete many-body correlations without explicitly computing high-order terms. This model not only successfully addresses challenges of the local completeness and nonlocality in representative systems, but also provides a general way to adapt local descriptors to a message-passing form without changing their basic structures.