Latent Ewald summation for machine learning of long-range interactions

Fuente: arXiv
Saved in:
Bibliographic Details
Main Author: Cheng, Bingqing
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910753776205824
author Cheng, Bingqing
author_facet Cheng, Bingqing
contents Machine learning interatomic potentials (MLIPs) often neglect long-range interactions, such as electrostatic and dispersion forces. In this work, we introduce a straightforward and efficient method to account for long-range interactions by learning a latent variable from local atomic descriptors and applying an Ewald summation to this variable. We demonstrate that in systems including charged and polar molecular dimers, bulk water, and water-vapor interface, standard short-ranged MLIPs can lead to unphysical predictions even when employing message passing. The long-range models effectively eliminate these artifacts, with only about twice the computational cost of short-range MLIPs.
format Preprint
id arxiv_https___arxiv_org_abs_2408_15165
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Latent Ewald summation for machine learning of long-range interactions
Cheng, Bingqing
Machine Learning
Materials Science
Chemical Physics
Computational Physics
Machine learning interatomic potentials (MLIPs) often neglect long-range interactions, such as electrostatic and dispersion forces. In this work, we introduce a straightforward and efficient method to account for long-range interactions by learning a latent variable from local atomic descriptors and applying an Ewald summation to this variable. We demonstrate that in systems including charged and polar molecular dimers, bulk water, and water-vapor interface, standard short-ranged MLIPs can lead to unphysical predictions even when employing message passing. The long-range models effectively eliminate these artifacts, with only about twice the computational cost of short-range MLIPs.
title Latent Ewald summation for machine learning of long-range interactions
topic Machine Learning
Materials Science
Chemical Physics
Computational Physics
url https://arxiv.org/abs/2408.15165