Accelerating wavepacket propagation with machine learning

Research output: Contribution to journalArticleResearchpeer review

Authors

  • Kanishka Singh
  • Ka Hei Lee
  • Daniel Peláez
  • Annika Bande

External Research Organisations

  • Helmholtz-Zentrum Berlin für Materialien und Energie (HZB)
  • Freie Universität Berlin (FU Berlin)
  • Universite Paris-Sud XI
View graph of relations

Details

Original languageEnglish
Pages (from-to)2360-2373
Number of pages14
JournalJournal of computational chemistry
Volume45
Issue number28
Publication statusPublished - 2 Sept 2024

Abstract

In this work, we discuss the use of a recently introduced machine learning (ML) technique known as Fourier neural operators (FNO) as an efficient alternative to the traditional solution of the time-dependent Schrödinger equation (TDSE). FNOs are ML models which are employed in the approximated solution of partial differential equations. For a wavepacket propagating in an anharmonic potential and for a tunneling system, we show that the FNO approach can accurately and faithfully model wavepacket propagation via the density. Additionally, we demonstrate that FNOs can be a suitable replacement for traditional TDSE solvers in cases where the results of the quantum dynamical simulation are required repeatedly such as in the case of parameter optimization problems (e.g., control). The speed-up from the FNO method allows for its combination with the Markov-chain Monte Carlo approach in applications that involve solving inverse problems such as optimal and coherent laser control of the outcome of dynamical processes.

Keywords

    Fourier neural operators, machine learning, quantum dynamics

ASJC Scopus subject areas

Cite this

Accelerating wavepacket propagation with machine learning. / Singh, Kanishka; Lee, Ka Hei; Peláez, Daniel et al.
In: Journal of computational chemistry, Vol. 45, No. 28, 02.09.2024, p. 2360-2373.

Research output: Contribution to journalArticleResearchpeer review

Singh, K, Lee, KH, Peláez, D & Bande, A 2024, 'Accelerating wavepacket propagation with machine learning', Journal of computational chemistry, vol. 45, no. 28, pp. 2360-2373. https://doi.org/10.1002/jcc.27443
Singh K, Lee KH, Peláez D, Bande A. Accelerating wavepacket propagation with machine learning. Journal of computational chemistry. 2024 Sept 2;45(28):2360-2373. doi: 10.1002/jcc.27443
Singh, Kanishka ; Lee, Ka Hei ; Peláez, Daniel et al. / Accelerating wavepacket propagation with machine learning. In: Journal of computational chemistry. 2024 ; Vol. 45, No. 28. pp. 2360-2373.
Download
@article{941ea78ca4cf403882b1a89a8fd8d0e1,
title = "Accelerating wavepacket propagation with machine learning",
abstract = "In this work, we discuss the use of a recently introduced machine learning (ML) technique known as Fourier neural operators (FNO) as an efficient alternative to the traditional solution of the time-dependent Schr{\"o}dinger equation (TDSE). FNOs are ML models which are employed in the approximated solution of partial differential equations. For a wavepacket propagating in an anharmonic potential and for a tunneling system, we show that the FNO approach can accurately and faithfully model wavepacket propagation via the density. Additionally, we demonstrate that FNOs can be a suitable replacement for traditional TDSE solvers in cases where the results of the quantum dynamical simulation are required repeatedly such as in the case of parameter optimization problems (e.g., control). The speed-up from the FNO method allows for its combination with the Markov-chain Monte Carlo approach in applications that involve solving inverse problems such as optimal and coherent laser control of the outcome of dynamical processes.",
keywords = "Fourier neural operators, machine learning, quantum dynamics",
author = "Kanishka Singh and Lee, {Ka Hei} and Daniel Pel{\'a}ez and Annika Bande",
note = "Publisher Copyright: {\textcopyright} 2024 The Author(s). Journal of Computational Chemistry published by Wiley Periodicals LLC.",
year = "2024",
month = sep,
day = "2",
doi = "10.1002/jcc.27443",
language = "English",
volume = "45",
pages = "2360--2373",
journal = "Journal of computational chemistry",
issn = "0192-8651",
publisher = "John Wiley and Sons Inc.",
number = "28",

}

Download

TY - JOUR

T1 - Accelerating wavepacket propagation with machine learning

AU - Singh, Kanishka

AU - Lee, Ka Hei

AU - Peláez, Daniel

AU - Bande, Annika

N1 - Publisher Copyright: © 2024 The Author(s). Journal of Computational Chemistry published by Wiley Periodicals LLC.

PY - 2024/9/2

Y1 - 2024/9/2

N2 - In this work, we discuss the use of a recently introduced machine learning (ML) technique known as Fourier neural operators (FNO) as an efficient alternative to the traditional solution of the time-dependent Schrödinger equation (TDSE). FNOs are ML models which are employed in the approximated solution of partial differential equations. For a wavepacket propagating in an anharmonic potential and for a tunneling system, we show that the FNO approach can accurately and faithfully model wavepacket propagation via the density. Additionally, we demonstrate that FNOs can be a suitable replacement for traditional TDSE solvers in cases where the results of the quantum dynamical simulation are required repeatedly such as in the case of parameter optimization problems (e.g., control). The speed-up from the FNO method allows for its combination with the Markov-chain Monte Carlo approach in applications that involve solving inverse problems such as optimal and coherent laser control of the outcome of dynamical processes.

AB - In this work, we discuss the use of a recently introduced machine learning (ML) technique known as Fourier neural operators (FNO) as an efficient alternative to the traditional solution of the time-dependent Schrödinger equation (TDSE). FNOs are ML models which are employed in the approximated solution of partial differential equations. For a wavepacket propagating in an anharmonic potential and for a tunneling system, we show that the FNO approach can accurately and faithfully model wavepacket propagation via the density. Additionally, we demonstrate that FNOs can be a suitable replacement for traditional TDSE solvers in cases where the results of the quantum dynamical simulation are required repeatedly such as in the case of parameter optimization problems (e.g., control). The speed-up from the FNO method allows for its combination with the Markov-chain Monte Carlo approach in applications that involve solving inverse problems such as optimal and coherent laser control of the outcome of dynamical processes.

KW - Fourier neural operators

KW - machine learning

KW - quantum dynamics

UR - http://www.scopus.com/inward/record.url?scp=85196514834&partnerID=8YFLogxK

U2 - 10.1002/jcc.27443

DO - 10.1002/jcc.27443

M3 - Article

AN - SCOPUS:85196514834

VL - 45

SP - 2360

EP - 2373

JO - Journal of computational chemistry

JF - Journal of computational chemistry

SN - 0192-8651

IS - 28

ER -