Physically Motivated Recursively Embedded Atom Neural Networks: Incorporating Local Completeness and Nonlocality

Abstract

Recent advances in machine-learned interatomic potentials largely benefit from the atomistic representation and locally invariant many-body descriptors. It was however recently argued that including three- (or even four-) body features is incomplete to distinguish specific local structures. 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 regarding local completeness and nonlocality in representative systems, but also provides an easy and general way to update local many-body descriptors to have a message-passing form without changing their basic structures.

This work has been published in Phys. Rev. Lett

References in corpus (6)

• ANI-1: An extensible neural network potential with DFT accuracy at force field computational cost
• Machine Learning Unifies the Modelling of Materials and Molecules
• Recursive evaluation and iterative contraction of $N$-body equivariant features
• Symmetry-Adapted High Dimensional Neural Network Representation of Electronic Friction Tensor of Adsorbates on Metals
• Conceptual and practical bases for the high accuracy of machine learning interatomic potential
• A practical approach to Hohenberg-Kohn maps based on many-body correlations: learning the electronic density

Cited by in corpus (1)

• REANN: A PyTorch-based End-to-End Multi-functional Deep Neural Network Package for Molecular, Reactive and Periodic Systems