FFLUX is a new force field that combines the accuracy of quantum mechanics with the speed of force fields, without any link to the architecture of classical force fields. This force field is atom-focused and adopts the parameter-free topological atom from Quantum Chemical Topology (QCT). FFLUX uses Gaussian process regression (also known as kriging) models to make predictions of atomic properties, which in this work are atomic energies according to QCT’s interacting quantum atom approach. Here, we report the adaptive sampling technique maximum expected prediction error to create data-compact, efficient, and accurate kriging models (sub-kJ mol−1 for water, ammonia, methane, and methanol and sub-kcal mol−1 for N-methylacetamide). The models cope with large molecular distortions and are ready for use in molecular simulation. A brand new press-one-button Python pipeline, called ICHOR, carries out the training.
Skip Nav Destination
Article navigation
7 August 2020
Research Article|
August 04 2020
Creating Gaussian process regression models for molecular simulations using adaptive sampling
Matthew J. Burn
;
Matthew J. Burn
Manchester Institute of Biotechnology, The University of Manchester
, Manchester M1 7DN, United Kingdom
and Department of Chemistry, The University of Manchester
, Manchester M13 9PL, United Kingdom
Search for other works by this author on:
Paul L. A. Popelier
Manchester Institute of Biotechnology, The University of Manchester
, Manchester M1 7DN, United Kingdom
and Department of Chemistry, The University of Manchester
, Manchester M13 9PL, United Kingdom
a)Author to whom correspondence should be addressed: pla@manchester.ac.uk. Telephone: +44 161 3064511
Search for other works by this author on:
a)Author to whom correspondence should be addressed: pla@manchester.ac.uk. Telephone: +44 161 3064511
b)
Present address: Manchester Institute of Biotechnology, The University of Manchester, Manchester M1 7DN, United Kingdom.
Note: This paper is part of the JCP Special Topic on Machine Learning Meets Chemical Physics.
J. Chem. Phys. 153, 054111 (2020)
Article history
Received:
June 12 2020
Accepted:
July 13 2020
Citation
Matthew J. Burn, Paul L. A. Popelier; Creating Gaussian process regression models for molecular simulations using adaptive sampling. J. Chem. Phys. 7 August 2020; 153 (5): 054111. https://doi.org/10.1063/5.0017887
Download citation file:
Sign in
Don't already have an account? Register
Sign In
You could not be signed in. Please check your credentials and make sure you have an active account and try again.
Pay-Per-View Access
$40.00
Citing articles via
DeePMD-kit v2: A software package for deep potential models
Jinzhe Zeng, Duo Zhang, et al.
Related Content
Toward a simple yet efficient cost function for the optimization of Gaussian process regression model hyperparameters
AIP Advances (September 2023)
The accuracy of ab initio calculations without ab initio calculations for charged systems: Kriging predictions of atomistic properties for ions in aqueous solutions
J. Chem. Phys. (April 2018)
Kriging atomic properties with a variable number of inputs
J. Chem. Phys. (September 2016)
Gaussian process models of potential energy surfaces with boundary optimization
J. Chem. Phys. (October 2021)
Flexible multipole moments in smooth particle mesh Ewald
J. Chem. Phys. (June 2022)