Adaptive coupling of a deep neural network potential to a classical force field
Abstract
An adaptive modeling method (AMM) that couples a deep neural network potential and a classical force field is introduced to address the accuracy-efficiency dilemma faced by the molecular simulation community. The AMM simulated system is decomposed into three types of regions. The first type captures the important phenomena in the system and requires high accuracy, for which we use the Deep Potential Molecular Dynamics (DeePMD) model in this work. The DeePMD model is trained to accurately reproduce the statistical properties of the ab initio molecular dynamics. The second type does not require high accuracy, and a classical force field is used to describe it in an efficient way. The third type is used for a smooth transition between the first and the second types of regions. By using a force interpolation scheme and imposing a thermodynamics force in the transition region, we make the DeePMD region embedded in the AMM simulated system as if it were embedded in a system that is fully described by the accurate potential. A representative example of the liquid water system is used to show the feasibility and promise of this method.
- Authors:
-
[1];
[2];
- Princeton Univ., NJ (United States)
- Lab. of Computational Physics, Inst. of Applied Physics and Computational Mathematics, Beijing (China)
- Princeton Univ., NJ (United States); Peking Univ., Beijing (China). Center for Data Science and Beijing International Center for Mathematical Research; Beijing Inst. of Big Data Research, Beijing (China)
- Publication Date:
- Research Org.:
- Princeton Univ., NJ (United States)
- Sponsoring Org.:
- USDOE Office of Science (SC); US Department of the Navy, Office of Naval Research (ONR); National Science Foundation of China (NSFC); National Key Research and Development Program of China
- OSTI Identifier:
- 1611098
- Alternate Identifier(s):
- OSTI ID: 1477902
- Grant/Contract Number:
- SC0008626; SC0009248; N00014-13-1-0338; U1430237; 91530322; 11501039; 11871110; 2016YFB0201200; 2016YFB0201203
- Resource Type:
- Accepted Manuscript
- Journal Name:
- Journal of Chemical Physics
- Additional Journal Information:
- Journal Volume: 149; Journal Issue: 15; Journal ID: ISSN 0021-9606
- Publisher:
- American Institute of Physics (AIP)
- Country of Publication:
- United States
- Language:
- English
- Subject:
- 71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSICS; 37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CHEMISTRY; Chemistry; Physics
Citation Formats
Zhang, Linfeng, Wang, Han, and E, Weinan. Adaptive coupling of a deep neural network potential to a classical force field. United States: N. p., 2018.
Web. doi:10.1063/1.5042714.
Zhang, Linfeng, Wang, Han, & E, Weinan. Adaptive coupling of a deep neural network potential to a classical force field. United States. https://doi.org/10.1063/1.5042714
Zhang, Linfeng, Wang, Han, and E, Weinan. Wed .
"Adaptive coupling of a deep neural network potential to a classical force field". United States. https://doi.org/10.1063/1.5042714. https://www.osti.gov/servlets/purl/1611098.
@article{osti_1611098,
title = {Adaptive coupling of a deep neural network potential to a classical force field},
author = {Zhang, Linfeng and Wang, Han and E, Weinan},
abstractNote = {An adaptive modeling method (AMM) that couples a deep neural network potential and a classical force field is introduced to address the accuracy-efficiency dilemma faced by the molecular simulation community. The AMM simulated system is decomposed into three types of regions. The first type captures the important phenomena in the system and requires high accuracy, for which we use the Deep Potential Molecular Dynamics (DeePMD) model in this work. The DeePMD model is trained to accurately reproduce the statistical properties of the ab initio molecular dynamics. The second type does not require high accuracy, and a classical force field is used to describe it in an efficient way. The third type is used for a smooth transition between the first and the second types of regions. By using a force interpolation scheme and imposing a thermodynamics force in the transition region, we make the DeePMD region embedded in the AMM simulated system as if it were embedded in a system that is fully described by the accurate potential. A representative example of the liquid water system is used to show the feasibility and promise of this method.},
doi = {10.1063/1.5042714},
journal = {Journal of Chemical Physics},
number = 15,
volume = 149,
place = {United States},
year = {Wed Oct 17 00:00:00 EDT 2018},
month = {Wed Oct 17 00:00:00 EDT 2018}
}
Free Publicly Available Full Text
Publisher's Version of Record
Other availability
Search WorldCat to find libraries that may hold this journal
Cited by: 10 works
Citation information provided by
Web of Science
Web of Science
Save to My Library
You must Sign In or Create an Account in order to save documents to your library.
Works referenced in this record:
The adaptive buffered force QM/MM method in the CP2K and AMBER software packages
journal, February 2015
- Mones, Letif; Jones, Andrew; Götz, Andreas W.
- Journal of Computational Chemistry, Vol. 36, Issue 9
Formulation of Liouville's theorem for grand ensemble molecular simulations
journal, February 2016
- Delle Site, Luigi
- Physical Review E, Vol. 93, Issue 2
Theoretical studies of enzymic reactions: Dielectric, electrostatic and steric stabilization of the carbonium ion in the reaction of lysozyme
journal, May 1976
- Warshel, A.; Levitt, M.
- Journal of Molecular Biology, Vol. 103, Issue 2
Adaptive Resolution Simulation (AdResS): A Smooth Thermodynamic and Structural Transition from Atomistic to Coarse Grained Resolution and Vice Versa in a Grand Canonical Fashion
journal, July 2012
- Wang, Han; Schütte, Christof; Delle Site, Luigi
- Journal of Chemical Theory and Computation, Vol. 8, Issue 8
Sampling the isothermal-isobaric ensemble by Langevin dynamics
journal, March 2016
- Gao, Xingyu; Fang, Jun; Wang, Han
- The Journal of Chemical Physics, Vol. 144, Issue 12
Comparison of simple potential functions for simulating liquid water
journal, July 1983
- Jorgensen, William L.; Chandrasekhar, Jayaraman; Madura, Jeffry D.
- The Journal of Chemical Physics, Vol. 79, Issue 2
Adaptive resolution molecular-dynamics simulation: Changing the degrees of freedom on the fly
journal, December 2005
- Praprotnik, Matej; Delle Site, Luigi; Kremer, Kurt
- The Journal of Chemical Physics, Vol. 123, Issue 22
GROMACS: High performance molecular simulations through multi-level parallelism from laptops to supercomputers
journal, September 2015
- Abraham, Mark James; Murtola, Teemu; Schulz, Roland
- SoftwareX, Vol. 1-2
GROMACS 4.5: a high-throughput and highly parallel open source molecular simulation toolkit
journal, February 2013
- Pronk, Sander; Páll, Szilárd; Schulz, Roland
- Bioinformatics, Vol. 29, Issue 7
DeePMD-kit: A deep learning package for many-body potential energy representation and molecular dynamics
journal, July 2018
- Wang, Han; Zhang, Linfeng; Han, Jiequn
- Computer Physics Communications, Vol. 228
Some fundamental problems for an energy-conserving adaptive-resolution molecular dynamics scheme
journal, October 2007
- Delle Site, Luigi
- Physical Review E, Vol. 76, Issue 4
Deep Potential: A General Representation of a Many-Body Potential Energy Surface
journal, January 2018
- Han, Jiequn; Zhang, Linfeng; Car, Roberto
- Communications in Computational Physics, Vol. 23, Issue 3
Deep Potential Molecular Dynamics: A Scalable Model with the Accuracy of Quantum Mechanics
journal, April 2018
- Zhang, Linfeng; Han, Jiequn; Wang, Han
- Physical Review Letters, Vol. 120, Issue 14
Molecular systems with open boundaries: Theory and simulation
journal, June 2017
- Delle Site, Luigi; Praprotnik, Matej
- Physics Reports, Vol. 693
QM/MM: what have we learned, where are we, and where do we go from here?
journal, July 2006
- Lin, Hai; Truhlar, Donald G.
- Theoretical Chemistry Accounts, Vol. 117, Issue 2
Path integral molecular dynamics within the grand canonical-like adaptive resolution technique: Simulation of liquid water
journal, September 2015
- Agarwal, Animesh; Delle Site, Luigi
- The Journal of Chemical Physics, Vol. 143, Issue 9
Adaptive resolution simulation in equilibrium and beyond
journal, June 2015
- Wang, H.; Agarwal, A.
- The European Physical Journal Special Topics, Vol. 224, Issue 12
QM/MM: What Have We Learned, Where Are We, and Where Do We Go from Here?
journal, May 2007
- Lin, Hai; Truhlar, Donald G.
- ChemInform, Vol. 38, Issue 22
How van der Waals interactions determine the unique properties of water
journal, July 2016
- Morawietz, Tobias; Singraber, Andreas; Dellago, Christoph
- Proceedings of the National Academy of Sciences, Vol. 113, Issue 30
Particle mesh Ewald: An N ⋅log( N ) method for Ewald sums in large systems
journal, June 1993
- Darden, Tom; York, Darrin; Pedersen, Lee
- The Journal of Chemical Physics, Vol. 98, Issue 12
Positive Definiteness of the Blended Force-Based Quasicontinuum Method
journal, January 2012
- Li, Xingjie Helen; Luskin, Mitchell; Ortner, Christoph
- Multiscale Modeling & Simulation, Vol. 10, Issue 3
Grand-Canonical-like Molecular-Dynamics Simulations by Using an Adaptive-Resolution Technique
journal, March 2013
- Wang, Han; Hartmann, Carsten; Schütte, Christof
- Physical Review X, Vol. 3, Issue 1
ANI-1: an extensible neural network potential with DFT accuracy at force field computational cost
text, January 2017
- ,
- The University of North Carolina at Chapel Hill University Libraries
Hybrid atomistic simulation methods for materials systems
journal, January 2009
- Bernstein, N.; Kermode, J. R.; Csányi, G.
- Reports on Progress in Physics, Vol. 72, Issue 2
Rational identification and characterisation of peptide ligands for targeting polysialic acid
journal, May 2020
- Shastry, Divya G.; Irudayanathan, Flaviyan Jerome; Williams, Asher
- Scientific Reports, Vol. 10, Issue 1
ANI-1: an extensible neural network potential with DFT accuracy at force field computational cost
journal, January 2017
- Smith, J. S.; Isayev, O.; Roitberg, A. E.
- Chemical Science, Vol. 8, Issue 4
Molecular dynamics in a grand ensemble: Bergmann–Lebowitz model and adaptive resolution simulation
journal, August 2015
- Agarwal, Animesh; Zhu, Jinglong; Hartmann, Carsten
- New Journal of Physics, Vol. 17, Issue 8
A Posteriori Error Control for Three Typical Force-Based Atomistic-to-Continuum Coupling Methods for an Atomistic Chain
journal, February 2019
- Wang, Hao
- Numerical Mathematics: Theory, Methods and Applications, Vol. 12, Issue 1
Quantum-chemical insights from deep tensor neural networks
journal, January 2017
- Schütt, Kristof T.; Arbabzadah, Farhad; Chmiela, Stefan
- Nature Communications, Vol. 8, Issue 1
Machine learning of accurate energy-conserving molecular force fields
text, January 2017
- Chmiela, Stefan; Tkatchenko, Alexandre; Sauceda, Huziel E.
- Technische Universität Berlin
Machine learning of accurate energy-conserving molecular force fields
journal, May 2017
- Chmiela, Stefan; Tkatchenko, Alexandre; Sauceda, Huziel E.
- Science Advances, Vol. 3, Issue 5
Improving the specificity of organophosphorus hydrolase to acephate by mutagenesis at its binding site: a computational study
journal, May 2021
- Badakhshan, Reza; Mohammadi, Mozafar; Farnoosh, Gholamreza
- Journal of Molecular Modeling, Vol. 27, Issue 6
Convergence of a Force-Based Hybrid Method in Three Dimensions
journal, October 2012
- Lu, Jianfeng; Ming, Pingbing
- Communications on Pure and Applied Mathematics, Vol. 66, Issue 1
The missing term in effective pair potentials
journal, November 1987
- Berendsen, H. J. C.; Grigera, J. R.; Straatsma, T. P.
- The Journal of Physical Chemistry, Vol. 91, Issue 24
Tests of an Adaptive QM/MM Calculation on Free Energy Profiles of Chemical Reactions in Solution
journal, September 2013
- Várnai, Csilla; Bernstein, Noam; Mones, Letif
- The Journal of Physical Chemistry B, Vol. 117, Issue 40
Theory-based benchmarking of the blended force-based quasicontinuum method
journal, January 2014
- Li, Xingjie Helen; Luskin, Mitchell; Ortner, Christoph
- Computer Methods in Applied Mechanics and Engineering, Vol. 268
Fast and accurate modeling of molecular atomization energies with machine learning
text, January 2012
- Rupp, Matthias; Tkatchenko, Alexandre; Müller, Klaus-Robert
- American Physical Society
DeePMD-kit: A deep learning package for many-body potential energy representation and molecular dynamics
dataset, January 2018
- Wang, Han
- Mendeley
Some fundamental problems for an energy conserving adaptive resolution molecular dynamics scheme
text, January 2007
- Site, L. Delle
- arXiv
Gaussian Approximation Potentials: the accuracy of quantum mechanics, without the electrons
text, January 2009
- Bartók, Albert P.; Payne, Mike C.; Kondor, Risi
- arXiv
Fast and Accurate Modeling of Molecular Atomization Energies with Machine Learning
text, January 2011
- Rupp, Matthias; Tkatchenko, Alexandre; Müller, Klaus-Robert
- arXiv
Adaptive resolution molecular dynamics simulation through coupling to an internal particle reservoir
text, January 2011
- Fritsch, Sebastian; Poblete, Simon; Junghans, Christoph
- arXiv
Grand-canonical-like molecular-dynamics simulations by using an adaptive-resolution technique
text, January 2013
- Wang, Han; Hartmann, Carsten; Schütte, Christof
- arXiv
Path Integral Molecular Dynamics within the Grand Canonical-like Adaptive Resolution Technique: Simulation of Liquid Water
text, January 2015
- Agarwal, Animesh; Site, Luigi Delle
- arXiv
Sampling the isothermal-isobaric ensemble by Langevin dynamics
text, January 2016
- Gao, Xingyu; Fang, Jun; Wang, Han
- arXiv
Formulation of Liouville's Theorem for Grand Ensemble Molecular Simulations
text, January 2016
- Site, Luigi Delle
- arXiv
Quantum-Chemical Insights from Deep Tensor Neural Networks
text, January 2016
- Schütt, Kristof T.; Arbabzadah, Farhad; Chmiela, Stefan
- arXiv
Machine Learning of Accurate Energy-Conserving Molecular Force Fields
text, January 2016
- Chmiela, Stefan; Tkatchenko, Alexandre; Sauceda, Huziel E.
- arXiv
Deep Potential: a general representation of a many-body potential energy surface
text, January 2017
- Han, Jiequn; Zhang, Linfeng; Car, Roberto
- arXiv
DeePMD-kit: A deep learning package for many-body potential energy representation and molecular dynamics
text, January 2017
- Wang, Han; Zhang, Linfeng; Han, Jiequn
- arXiv
The missing term in effective pair potentials
journal, November 1987
- Berendsen, H. J. C.; Grigera, J. R.; Straatsma, T. P.
- The Journal of Physical Chemistry, Vol. 91, Issue 24
Adaptive Resolution Molecular Dynamics Simulation: Changing the Degrees of Freedom on the Fly
text, January 2005
- Praprotnik, Matej; Site, Luigi Delle; Kremer, Kurt
- arXiv
Works referencing / citing this record:
Deep learning inter-atomic potential model for accurate irradiation damage simulations
journal, June 2019
- Wang, Hao; Guo, Xun; Zhang, Linfeng
- Applied Physics Letters, Vol. 114, Issue 24
Deep learning inter-atomic potential model for accurate irradiation damage simulations
text, January 2019
- Wang, Hao; Guo, Xun; Zhang, Linfeng
- arXiv