Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2023 Dec 1;28(23):7900.
doi: 10.3390/molecules28237900.

Intermolecular Non-Bonded Interactions from Machine Learning Datasets

Jia-An Chen  1 , Sheng D Chao  1   2
Affiliations

Intermolecular Non-Bonded Interactions from Machine Learning Datasets

Jia-An Chen et al. Molecules. .

Abstract

Accurate determination of intermolecular non-covalent-bonded or non-bonded interactions is the key to potentially useful molecular dynamics simulations of polymer systems. However, it is challenging to balance both the accuracy and computational cost in force field modelling. One of the main difficulties is properly representing the calculated energy data as a continuous force function. In this paper, we employ well-developed machine learning techniques to construct a general purpose intermolecular non-bonded interaction force field for organic polymers. The original ab initio dataset SOFG-31 was calculated by us and has been well documented, and here we use it as our training set. The CLIFF kernel type machine learning scheme is used for predicting the interaction energies of heterodimers selected from the SOFG-31 dataset. Our test results show that the overall errors are well below the chemical accuracy of about 1 kcal/mol, thus demonstrating the promising feasibility of machine learning techniques in force field modelling.

Keywords: artificial intelligence; machine learning potentials; non-bonded interactions; quantum chemistry datasets; symmetry adapted perturbation theory.

PubMed Disclaimer

Conflict of interest statement

The authors declare no conflict of interest.

Figures

Figure 1
Figure 1
Convergence of the loss function during the fitting process.
Figure 2
Figure 2
Correlation plots for calculating the energies of the SOFG-31-heterodimer using the SOFG-31 dataset as the training set. The blue line is the reference line for the correlation.
Figure 2
Figure 2
Correlation plots for calculating the energies of the SOFG-31-heterodimer using the SOFG-31 dataset as the training set. The blue line is the reference line for the correlation.
Figure 3
Figure 3
Convergence of loss function in the fitting process for the Dimer 31 + 47 dataset.
Figure 4
Figure 4
Correlation plots for calculating SOFG-31-heterodimer using Dimer 31 + 47 as the fitting dataset. The blue line is the reference line for the correlation.
Figure 4
Figure 4
Correlation plots for calculating SOFG-31-heterodimer using Dimer 31 + 47 as the fitting dataset. The blue line is the reference line for the correlation.
Figure 5
Figure 5
Correlation plot for calculating Des370k at k = 0 using Dimer 31 + 47 as the fitting dataset. The blue line is the reference line for the correlation.
Figure 6
Figure 6
Correlation plot for calculating Des370k at k = 0 using CLIFF0 as the fitting dataset. The blue line is the reference line for the correlation.
See this image and copyright information in PMC

References

    1. Behler J., Parrinello M. Generalized Neural-Network Representation of High-Dimensional Potential-Energy Surfaces. Phys. Rev. Lett. 2007;98:146401. doi: 10.1103/PhysRevLett.98.146401. - DOI - PubMed
    1. Behler J. Atom-centered symmetry functions for constructing high-dimensional neural network potentials. J. Chem. Phys. 2011;134:074106. doi: 10.1063/1.3553717. - DOI - PubMed
    1. Schütt K.T., Arbabzadah F., Chmiela S., Müller K.R., Tkatchenko A. Quantum-chemical insights from deep tensor neural networks Schütt. Nat. Commun. 2017;8:13890. doi: 10.1038/ncomms13890. - DOI - PMC - PubMed
    1. Smith J.S., Isayev O., Roitberg A.E. ANI-1: An extensible neural network potential with DFT accuracy at force field computational cost Smith. Chem. Sci. 2017;8:3192. doi: 10.1039/C6SC05720A. - DOI - PMC - PubMed
    1. Schütt K.T., Sauceda H.E., Kindermans P.-J., Tkatchenko A., Müller K.-R. SchNet—A deep learning architecture for molecules and materials. Chem. Phys. 2018;148:241722. doi: 10.1063/1.5019779. - DOI - PubMed

LinkOut - more resources