Characterization of potential energy surfaces using machine-learning techniques

PN 3-4

Project description

Machine-learned surrogate models are used to find minima, saddle points and reaction paths on potential energy surfaces of materials. These are used to characterize the energetics of chemical processes or the structure of materials on an atomistic scale. We build machine-learned models on-the-fly as more and more data from electronic structure calculations become available. Optimization algorithms are developed and adjusted to different ML approaches to deal with the inherent numerical noise in the data. The re-use of data as well as the combination of data sources with different cost and accuracy increases the computational efficiency. The goal is to gain as much knowledge about stationary points and optimal paths on the potential energy surface with as few electronic structure calculations as possible, since the latter are the computational bottleneck.

Project information

Project title Characterization of potential energy surfaces using machinge-learning techniques
Project leaders Johannes Kästner (Bernard Haasdonk)
Project partners

Christian Holm (PN 3-3): Atomic Descriptors for Machine Learning.
PN 5-7: Gaussian Processes
Ingo Steinwart (PN 6-3): Implementing Physical Constraints in Machine Learning
Project duration June 2019 - November 2022
Project number PN 3-4

Publications PN 3-4 and PN 3-4 (II)

  1. 2024

    1. J. C. del Valle, P. Redondo, J. Kästner, and G. Molpeceres, “Formation of the Interstellar Sugar Precursor, (Z)-1,2-Ethenediol, through Radical Reactions on Dust Grains,” Astrophys. J., vol. 974, p. 129, 2024, doi: 10.3847/1538-4357/ad6f9a.
    2. S. V. Klostermann, J. Kappler, A. Waigum, M. R. Buchmeiser, A. Köhn, and J. Kästner, “The reduction behavior of sulfurized polyacrylonitrile (SPAN) in lithium–sulfur batteries using a carbonate electrolyte: a computational study,” Phys. Chem. Chem. Phys., vol. 26, pp. 9998–10007, 2024, doi: 10.1039/D3CP06248A.
    3. P. M. Becker, K. Heinze, B. Sarkar, and J. Kästner, “Redox−Acid/Base Phase Diagrams as an Entry to Computational Redox Chemistry,” ChemElectroChem, vol. 11, no. 20, Art. no. 20, 2024, doi: https://doi.org/10.1002/celc.202400301.

  2. 2023

    1. Molpeceres, G., Zaverkin, V., Furuya, K., Aikawa, Y., and Kästner, J., “Reaction dynamics on amorphous solid water surfaces using interatomic machine-learned potentials - Microscopic energy partition revealed from the P + H → PH reaction,” Astronomy & Astrophysics, vol. 673, p. A51, 2023, doi: 10.1051/0004-6361/202346073.
    2. V. Zaverkin, D. Holzmüller, L. Bonfirraro, and J. Kästner, “Transfer learning for chemically accurate interatomic neural network potentials,” Physical Chemistry Chemical Physics, vol. 25, no. 7, Art. no. 7, 2023, doi: 10.1039/D2CP05793J.
    3. K. Gubaev, V. Zaverkin, P. Srinivasan, A. I. Duff, J. Kästner, and B. Grabowski, “Performance of two complementary machine-learned potentials in modelling chemically complex systems,” NPJ Computational Materials, vol. 9, p. 129, 2023, doi: 10.1038/s41524-023-01073-w.

  3. 2022

    1. C. Kessler et al., “Influence of layer slipping on adsorption of light gases in covalent organic frameworks: A combined experimental and computational study,” Microporous and Mesoporous Materials, vol. 336, p. 111796, May 2022, doi: 10.1016/j.micromeso.2022.111796.
    2. V. Zaverkin, D. Holzmüller, R. Schuldt, and J. Kästner, “Predicting properties of periodic systems from cluster data: A case study of liquid water,” The Journal of Chemical Physics, vol. 156, no. 11, Art. no. 11, 2022, doi: 10.1063/5.0078983.
    3. V. Zaverkin, J. Netz, F. Zills, A. Köhn, and J. Kästner, “Thermally Averaged Magnetic Anisotropy Tensors via Machine Learning Based on Gaussian Moments,” Journal of Chemical Theory and Computation, vol. 18, pp. 1–12, 2022, doi: 10.1021/acs.jctc.1c00853.

  4. 2021

    1. D. Born and J. Kästner, “Geometry Optimization in Internal Coordinates Based on Gaussian Process Regression: Comparison of Two Approaches,” Journal of Chemical Theory and Computation, vol. 17, no. 9, Art. no. 9, 2021, doi: 10.1021/acs.jctc.1c00517.
    2. G. Molpeceres, V. Zaverkin, N. Watanabe, and J. Kästner, “Binding energies and sticking coefficients of H₂ on crystalline and amorphous CO ice,” Astronomy & Astrophysics, vol. 648, p. A84, 2021, doi: 10.1051/0004-6361/202040023.

  5. 2020

    1. G. Molpeceres, V. Zaverkin, and J. Kästner, “Neural-network assisted study of nitrogen atom dynamics on amorphous solid water – I. adsorption and desorption,” Mon. Not. R. Astron. Soc., vol. 499, pp. 1373–1384, 2020, doi: 10.1093/mnras/staa2891.
    2. V. Zaverkin and J. Kästner, “Gaussian Moments as Physically Inspired Molecular Descriptors for Accurate and Scalable Machine Learning Potentials,” Journal of Chemical Theory and Computation, vol. 16, pp. 5410–5421, 2020, doi: 10.1021/acs.jctc.0c00347.
    3. A. Denzel and J. Kästner, “Hessian Matrix Update Scheme for Transition State Search Based on Gaussian Process Regression,” Journal of Chemical Theory and Computation, vol. 16, no. 8, Art. no. 8, Jul. 2020, doi: 10.1021/acs.jctc.0c00348.

  6. 2019

    1. A. Denzel, B. Haasdonk, and J. Kästner, “Gaussian Process Regression for Minimum Energy Path Optimization and Transition State Search,” The Journal of Physical Chemistry A, vol. 123, no. 44, Art. no. 44, 2019, doi: 10.1021/acs.jpca.9b08239.

  7. 2018

    1. A. Denzel and J. Kästner, “Gaussian Process Regression for Transition State Search,” Journal of Chemical Theory and Computation, vol. 14, no. 11, Art. no. 11, 2018, doi: 10.1021/acs.jctc.8b00708.
    2. A. Denzel and J. Kästner, “Gaussian process regression for geometry optimization,” Journal of Chemical Physics, vol. 148, no. 9, Art. no. 9, 2018, doi: 10.1063/1.5017103.
To the top of the page