Interpolating Nonadiabatic Molecular Dynamics Hamiltonian with Bidirectional Long Short-Term Memory Networks

被引:7
|
作者
Wang, Bipeng [1 ]
Winkler, Ludwig [2 ]
Wu, Yifan [3 ]
Muller, Klaus-Robert [2 ,4 ,5 ,6 ,7 ]
Sauceda, Huziel E. [8 ,9 ]
Prezhdo, Oleg V. [1 ,3 ]
机构
[1] Univ Southern Calif, Dept Chem Engn, Los Angeles, CA 90089 USA
[2] Tech Univ Berlin, Machine Learning Grp, D-10587 Berlin, Germany
[3] Univ Southern Calif, Dept Chem, Los Angeles, CA 90089 USA
[4] Berlin Inst Fdn Learning & Data, BIFOLD, D-10587 Berlin, Germany
[5] Korea Univ, Dept Artificial Intelligence, Seoul 136713, South Korea
[6] Max Planck Inst Informat, D-66123 Saarbrucken, Germany
[7] Google Deepmind, D-10587 Berlin, Germany
[8] Tech Univ Berlin, BASF TU joint Lab, BASLEARN, D-10587 Berlin, Germany
[9] Univ Nacl Autonoma Mexico, Mexico City 01000, DF, Mexico
来源
JOURNAL OF PHYSICAL CHEMISTRY LETTERS | 2023年 / 14卷 / 31期
基金
美国国家科学基金会;
关键词
TOTAL-ENERGY CALCULATIONS; PYXAID PROGRAM; LOCALIZATION; SCHEMES;
D O I
10.1021/acs.jpclett.3c01723
中图分类号
O64 [物理化学(理论化学)、化学物理学];
学科分类号
070304 ; 081704 ;
摘要
Essential for understandingfar-from-equilibrium processes,nonadiabatic(NA) molecular dynamics (MD) requires expensive calculations of theexcitation energies and NA couplings. Machine learning (ML) can simplify computation; however, the NA Hamiltonian requires complex ML modelsdue to its intricate relationship to atomic geometry. Working directlyin the time domain, we employ bidirectional long short-term memorynetworks (Bi-LSTM) to interpolate the Hamiltonian. Applying this multiscaleapproach to three metal-halide perovskite systems, we achieve twoorders of magnitude computational savings compared to direct ab initiocalculation. Reasonable charge trapping and recombination times areobtained with NA Hamiltonian sampling every half a picosecond. TheBi-LSTM-NAMD method outperforms earlier models and captures both slowand fast time scales. In combination with ML force fields, the methodologyextends NAMD simulation times from picoseconds to nanoseconds, comparableto charge carrier lifetimes in many materials. Nanosecond samplingis particularly important in systems containing defects, boundaries,interfaces, etc. that can undergo slow rearrangements.
引用
收藏
页码:7092 / 7099
页数:8
相关论文
共 50 条
  • [21] A Novel Word Spotting Algorithm Using Bidirectional Long Short-Term Memory Neural Networks
    Frinken, Volkmar
    Fischer, Andreas
    Bunke, Horst
    ARTIFICIAL NEURAL NETWORKS IN PATTERN RECOGNITION, PROCEEDINGS, 2010, 5998 : 185 - 196
  • [22] Ground target classification using mmWave radar with bidirectional long short-term memory networks
    Gunes, Oytun
    Ege, Mert
    Morgul, Omer
    JOURNAL OF APPLIED REMOTE SENSING, 2022, 16 (04)
  • [23] Bidirectional Long Short-Term Memory Networks for Rapid Fault Detection in Marine Hydrokinetic Turbines
    Wilson, David
    Passmore, Sean
    Tang, Yufei
    VanZwieten, James
    2018 17TH IEEE INTERNATIONAL CONFERENCE ON MACHINE LEARNING AND APPLICATIONS (ICMLA), 2018, : 495 - 500
  • [24] Layered Multistep Bidirectional Long Short-Term Memory Networks for Biomedical Word Sense Disambiguation
    Bis, Daniel
    Zhang, Canlin
    Liu, Xiuwen
    He, Zhe
    PROCEEDINGS 2018 IEEE INTERNATIONAL CONFERENCE ON BIOINFORMATICS AND BIOMEDICINE (BIBM), 2018, : 313 - 320
  • [25] Assessment of stacked unidirectional and bidirectional long short-term memory networks for electricity load forecasting
    Atef, Sara
    Eltawil, Amr B.
    ELECTRIC POWER SYSTEMS RESEARCH, 2020, 187 (187)
  • [26] Sleep Stage Classification using Convolutional Neural Networks and Bidirectional Long Short-Term Memory
    Yulita, Intan Nurma
    Fanany, Mohamad Ivan
    Arymurthy, Aniati Murni
    2017 INTERNATIONAL CONFERENCE ON ADVANCED COMPUTER SCIENCE AND INFORMATION SYSTEMS (ICACSIS), 2017, : 303 - 307
  • [27] Kannada Named Entity Recognition and Classification using Bidirectional Long Short-Term Memory Networks
    Sathyanarayanan, Dinesh
    Ashok, Ashwin
    Mishra, Debanik
    Chimalamarri, Santwana
    Sitaram, Dinkar
    2018 3RD INTERNATIONAL CONFERENCE ON ELECTRICAL, ELECTRONICS, COMMUNICATION, COMPUTER, AND OPTIMIZATION TECHNIQUES (ICEECCOT - 2018), 2018, : 65 - 71
  • [28] Time Series-based Spoof Speech Detection Using Long Short-term Memory and Bidirectional Long Short-term Memory
    Mirza, Arsalan R.
    Al-Talabani, Abdulbasit K.
    ARO-THE SCIENTIFIC JOURNAL OF KOYA UNIVERSITY, 2024, 12 (02): : 119 - 129
  • [29] Hybrid Short-term Load Forecasting Method Based on Empirical Wavelet Transform and Bidirectional Long Short-term Memory Neural Networks
    Zhang, Xiaoyu
    Kuenzel, Stefanie
    Colombo, Nicolo
    Watkins, Chris
    JOURNAL OF MODERN POWER SYSTEMS AND CLEAN ENERGY, 2022, 10 (05) : 1216 - 1228
  • [30] Hybrid Short-term Load Forecasting Method Based on Empirical Wavelet Transform and Bidirectional Long Short-term Memory Neural Networks
    Xiaoyu Zhang
    Stefanie Kuenzel
    Nicolo Colombo
    Chris Watkins
    JournalofModernPowerSystemsandCleanEnergy, 2022, 10 (05) : 1216 - 1228