Artificial neural network(ANN)potentials enable the efficient large-scale atomistic modeling of complex materials with near firstprinciples accuracy.For molecular dynamics simulations,accurate energies and interatomic...Artificial neural network(ANN)potentials enable the efficient large-scale atomistic modeling of complex materials with near firstprinciples accuracy.For molecular dynamics simulations,accurate energies and interatomic forces are a prerequisite,but training ANN potentials simultaneously on energies and forces from electronic structure calculations is computationally demanding.Here,we introduce an efficient alternative method for the training of ANN potentials on energy and force information,based on an extrapolation of the total energy via a Taylor expansion.By translating the force information to approximate energies,the quadratic scaling with the number of atoms exhibited by conventional force-training methods can be avoided,which enables the training on reference datasets containing complex atomic structures.展开更多
基金N.A.and A.U.acknowledge financial support by the U.S.Department of Energy(DOE)Office of Energy Efficiency and Renewable Energy,Vehicle Technologies Office,Contract No.DE-SC0012704A.C.and J.K.acknowledge financial support by the European Union’s Horizon 2020 research and innovation programme(Grant Agreement No.646717,TUNNELCHEM)+7 种基金the German Research Foundation(DFG)through the Cluster of Excellence in Simulation Technology(No.EXC 310/2)at the University of StuttgartComputational resources for the water cluster DFT calculations were provided by the state of Baden-Württemberg through bwHPC and the German Research Foundation(DFG)through grant no INST 40/467-1 FUGGDevelopment ofænet used the Extreme Science and Engineering Discovery Environment(XSEDE),which is supported by National Science Foundation grant number ACI-1053575The LMNTO DFT calculations used resources of the Center for Functional Nanomaterials,which is a U.S.DOE Office of Science Facility,at Brookhaven National Laboratory under Contract No.DE-SC0012704Development ofænet used the Extreme Science and Engineering Discovery Environment(XSEDE),which is supported by National Science Foundation grant number ACI-1053575The LMNTO DFT calculations used resources of the Center for Functional Nanomaterials,which is a U.S.DOE Office of Science Facility,at Brookhaven National Laboratory under Contract No.DE-SC0012704For the simulations of bulk water,we acknowledge computing resources from Columbia University’s Shared Research Computing Facility project,which is supported by NIH Research Facility Improvement Grant 1G20RR030893-01associated funds from the New York State Empire State Development,Division of Science Technology and Innovation(NYSTAR)Contract C090171,both awarded April 15,2010.
文摘Artificial neural network(ANN)potentials enable the efficient large-scale atomistic modeling of complex materials with near firstprinciples accuracy.For molecular dynamics simulations,accurate energies and interatomic forces are a prerequisite,but training ANN potentials simultaneously on energies and forces from electronic structure calculations is computationally demanding.Here,we introduce an efficient alternative method for the training of ANN potentials on energy and force information,based on an extrapolation of the total energy via a Taylor expansion.By translating the force information to approximate energies,the quadratic scaling with the number of atoms exhibited by conventional force-training methods can be avoided,which enables the training on reference datasets containing complex atomic structures.
