Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

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

About

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.

Yaolong Zhang, Junfan Xia, Bin Jiang• 2021

Related benchmarks

TaskDatasetResultRank
Interatomic Potential PredictionLiquid water 32-water system revPBE0-D3 (test)
Energy RMSE (meV/atom)0.8
11
Energy and force predictionWater (test)
Force RMSE (meV/Å)53
9
Energy Predictionwater dataset (test)
Energy RMSE (meV/atom)0.8
9
Interatomic potential modelingGAP 17 (test)
RMSE (Energy)31
7
Polarizability predictionDimer water system (val)
RRMSE1.6
6
Polarizability predictionLiquid water system (val)
RRMSE (%)2.1
6
Polarizability predictionZundel water system (val)
RRMSE0.1
6
Dipole moment predictionMonomer water system (val)
RRMSE5
6
Polarizability predictionMonomer water system (val)
RRMSE0.06
6
Dipole moment predictionDimer water system (val)
RRMSE (%)3
6
Showing 10 of 12 rows

Other info

Follow for update