Please use this identifier to cite or link to this item: http://dx.doi.org/10.14279/depositonce-10467
For citation please use:
Main Title: Combining SchNet and SHARC: The SchNarc Machine Learning Approach for Excited-State Dynamics
Author(s): Westermayr, Julia
Gastegger, Michael
Marquetand, Philipp
Type: Article
Language Code: en
Abstract: In recent years, deep learning has become a part of our everyday life and is revolutionizing quantum chemistry as well. In this work, we show how deep learning can be used to advance the research field of photochemistry by learning all important properties—multiple energies, forces, and different couplings—for photodynamics simulations. We simplify such simulations substantially by (i) a phase-free training skipping costly preprocessing of raw quantum chemistry data; (ii) rotationally covariant nonadiabatic couplings, which can either be trained or (iii) alternatively be approximated from only ML potentials, their gradients, and Hessians; and (iv) incorporating spin–orbit couplings. As the deep-learning method, we employ SchNet with its automatically determined representation of molecular structures and extend it for multiple electronic states. In combination with the molecular dynamics program SHARC, our approach termed SchNarc is tested on two polyatomic molecules and paves the way toward efficient photodynamics simulations of complex systems.
URI: https://depositonce.tu-berlin.de/handle/11303/11580
http://dx.doi.org/10.14279/depositonce-10467
Issue Date: 20-Apr-2020
Date Available: 19-Aug-2020
DDC Class: 530 Physik
Subject(s): coupling reactions
quantum mechanics
hessians
molecular dynamics simulations
molecules
Sponsor/Funder: EC/H2020/792572/EU/Machine Learning for Catalytic Carbon Dioxide Activation/MachineCat
EC/H2020/730897/EU/Transnational Access Programme for a Pan-European Network of HPC Research Infrastructures and Laboratories for scientific computing/HPC-EUROPA3
License: https://creativecommons.org/licenses/by/4.0/
Journal Title: Journal of Physical Chemistry Letters
Publisher: American Chemical Society (ACS)
Publisher Place: Washington, DC
Volume: 11
Publisher DOI: 10.1021/acs.jpclett.0c00527
Page Start: 3828
Page End: 3834
EISSN: 1948-7185
Appears in Collections:FG Maschinelles Lernen » Publications

Files in This Item:
acs.jpclett.0c00527.pdf
Format: Adobe PDF | Size: 1.05 MB
DownloadShow Preview
Thumbnail

Item Export Bar

This item is licensed under a Creative Commons License Creative Commons