Researchers develop recursively embedded atom neural network model

Researchers develop recursively embedded atom neural network model
Schematic diagram of the REANN model showing how the density descriptor is recursively embedded. Credit: ZHANG Yaolong et al.

In a recent study published in Physical Review Letters, a research team led by Prof. Jiang Bin from University of Science and Technology of China (USTC) of the Chinese Academy of Sciences proposed a recursively embedded atom neural network (REANN) model based on their previous work on creating high-precision machine-learning (ML) potential surface methods.

With the advancement of machine learning technologies, a common method to build potential functions is atom neural networks (ANNs) under which the is the sum of each atomic energy dependent on the local environment. Three-body descriptors have long been considered complete to describe the local environment.

Recent work, however, has found that three-body (or even four-body) descriptors could lead to the local structural degeneracy and thus fail to fully describe the local environment. This problem has posed difficulties to improve the precision of ANNs' potential surface training.

The REANN , using a recursively embedded density descriptor, shares the same nature with the less physically intuitive message-passing (MPNNs). The researchers proved that iteratively passing messages (namely updating orbital coefficients) to introduce many-body correlations can achieve a complete description of the without explicitly computing high-order features.

By testing the dataset of CH4 and bulk water, the researchers revealed the local completeness and nonlocality of this new model and showed that compared with current ML models, it has better accuracy.

The study provides a general way to easily improve existing ML potential surface frameworks to include more complicated many-body descriptors without changing their basic structures.


Explore further

Mesoscale neural plasticity helps in AI learning

More information: Yaolong Zhang et al, Physically Motivated Recursively Embedded Atom Neural Networks: Incorporating Local Completeness and Nonlocality, Physical Review Letters (2021). DOI: 10.1103/PhysRevLett.127.156002
Journal information: Physical Review Letters

Citation: Researchers develop recursively embedded atom neural network model (2021, October 26) retrieved 1 December 2021 from https://phys.org/news/2021-10-recursively-embedded-atom-neural-network.html
This document is subject to copyright. Apart from any fair dealing for the purpose of private study or research, no part may be reproduced without the written permission. The content is provided for information purposes only.
48 shares

Feedback to editors