TY - JOUR
T1 - Transfer-Learning-Based Coarse-Graining Method for Simple Fluids
T2 - Toward Deep Inverse Liquid-State Theory
AU - Moradzadeh, Alireza
AU - Aluru, Narayana R.
N1 - Funding Information:
We thank Professor R. Srikant from the Electrical and Computer Engineering Department at the University of Illinois at Urbana− Champaign for helpful discussions. This work was supported by the National Science Foundation under grants 1506619, 1545907, 1708852, 1720633, and 1720701. We acknowledge the use of Blue Waters supercomputing resources at the University of Illinois at Urbana−Champaign. This work partially used the Extreme Science and Engineering Discovery Environment (XSEDE) Stampede2 at the Texas Advanced Computing Center through allocation TG-CDA100010.
Publisher Copyright:
© 2019 American Chemical Society.
PY - 2019/3/21
Y1 - 2019/3/21
N2 - Machine learning is an attractive paradigm to circumvent difficulties associated with the development and optimization of force-field parameters. In this study, a deep neural network (DNN) is used to study the inverse problem of the liquid-state theory, in particular, to obtain the relation between the radial distribution function (RDF) and the Lennard-Jones (LJ) potential parameters at various thermodynamic states. Using molecular dynamics (MD), many observables, including RDF, are determined once the interatomic potential is specified. However, the inverse problem (parametrization of the potential for a specific RDF) is not straightforward. Here we present a framework integrating DNN with big data from 1.5 TB of MD trajectories with a cumulative simulation time of 52 μs for 26 000 distinct systems to predict LJ potential parameters. Our results show that DNN is successful not only in the parametrization of the atomic LJ liquids but also in parametrizing the LJ potential for coarse-grained models of simple multiatom molecules.
AB - Machine learning is an attractive paradigm to circumvent difficulties associated with the development and optimization of force-field parameters. In this study, a deep neural network (DNN) is used to study the inverse problem of the liquid-state theory, in particular, to obtain the relation between the radial distribution function (RDF) and the Lennard-Jones (LJ) potential parameters at various thermodynamic states. Using molecular dynamics (MD), many observables, including RDF, are determined once the interatomic potential is specified. However, the inverse problem (parametrization of the potential for a specific RDF) is not straightforward. Here we present a framework integrating DNN with big data from 1.5 TB of MD trajectories with a cumulative simulation time of 52 μs for 26 000 distinct systems to predict LJ potential parameters. Our results show that DNN is successful not only in the parametrization of the atomic LJ liquids but also in parametrizing the LJ potential for coarse-grained models of simple multiatom molecules.
UR - http://www.scopus.com/inward/record.url?scp=85062839077&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85062839077&partnerID=8YFLogxK
U2 - 10.1021/acs.jpclett.8b03872
DO - 10.1021/acs.jpclett.8b03872
M3 - Article
C2 - 30818949
AN - SCOPUS:85062839077
SN - 1948-7185
VL - 10
SP - 1242
EP - 1250
JO - Journal of Physical Chemistry Letters
JF - Journal of Physical Chemistry Letters
IS - 6
ER -