Machine Learning for Parameter Auto-tuning in Molecular Dynamics Simulations: Efficient Dynamics of Ions near Polarizable Nanoparticles. Kadupitiya, J., Fox, G., C., & Jadhao, V. Technical Report 2018. ![Machine Learning for Parameter Auto-tuning in Molecular Dynamics Simulations: Efficient Dynamics of Ions near Polarizable Nanoparticles [pdf]](https://bibbase.org/img/filetypes/pdf.svg "pdf")
Paper Website doi abstract bibtex Simulating the dynamics of ions near polarizable nanoparticles (NPs) is extremely challenging due to the need to solve the Poisson equation at every simulation timestep. Recently, a molecular dynamics (MD) method based on a dynamical optimization framework bypassed this obstacle by representing the polarization charge density as virtual dynamic variables, and evolving them in parallel with the physical dynamics of ions. We highlight the computational gains accessible with the integration of machine learning (ML) methods for parameter prediction in MD simulations by demonstrating how they were realized in MD simulations of ions near polarizable NPs. An artificial neural network based regression model was integrated with MD and predicted the optimal simulation timestep and critical parameters characterizing the virtual system on-the-fly with 94.3% success. The integration of ML method with hybrid OpenMP/MPI parallelized MD simulations generated accurate dynamics of thousands of ions in the presence of polarizable NPs for over 10 million steps (with a maximum simulated physical time over 30 ns) while reducing the computational time from thousands of hours to tens of hours yielding a maximum speedup of ≈ 3 from ML-only acceleration and a maximum overall speedup of ≈ 600 from ML-hybrid Open/MPI combined method. Extraction of ionic structure in concentrated electrolytes near oil-water emulsions demonstrates the success of the method. The approach can be generalized to select optimal parameters in other molecular dynamics applications and energy minimization problems.
@techreport{
title = {Machine Learning for Parameter Auto-tuning in Molecular Dynamics Simulations: Efficient Dynamics of Ions near Polarizable Nanoparticles},
type = {techreport},
year = {2018},
keywords = {Auto-tuning,Energy Minimization,Hybrid MPI/OpenMP,Machine Learning,Nanoscale Simulations,Parallel Computing},
pages = {15},
websites = {www.sagepub.com/,http://dsc.soic.indiana.edu/publications/Manuscript.IJHPCA.Nov2018.pdf},
id = {93e36f16-de3c-38ac-8569-20194aa0e1dd},
created = {2019-10-01T17:20:54.076Z},
file_attached = {true},
profile_id = {42d295c0-0737-38d6-8b43-508cab6ea85d},
last_modified = {2020-05-11T14:43:32.762Z},
read = {false},
starred = {false},
authored = {true},
confirmed = {true},
hidden = {false},
citation_key = {Kadupitiya2018},
private_publication = {false},
abstract = {Simulating the dynamics of ions near polarizable nanoparticles (NPs) is extremely challenging due to the need to solve the Poisson equation at every simulation timestep. Recently, a molecular dynamics (MD) method based on a dynamical optimization framework bypassed this obstacle by representing the polarization charge density as virtual dynamic variables, and evolving them in parallel with the physical dynamics of ions. We highlight the computational gains accessible with the integration of machine learning (ML) methods for parameter prediction in MD simulations by demonstrating how they were realized in MD simulations of ions near polarizable NPs. An artificial neural network based regression model was integrated with MD and predicted the optimal simulation timestep and critical parameters characterizing the virtual system on-the-fly with 94.3% success. The integration of ML method with hybrid OpenMP/MPI parallelized MD simulations generated accurate dynamics of thousands of ions in the presence of polarizable NPs for over 10 million steps (with a maximum simulated physical time over 30 ns) while reducing the computational time from thousands of hours to tens of hours yielding a maximum speedup of ≈ 3 from ML-only acceleration and a maximum overall speedup of ≈ 600 from ML-hybrid Open/MPI combined method. Extraction of ionic structure in concentrated electrolytes near oil-water emulsions demonstrates the success of the method. The approach can be generalized to select optimal parameters in other molecular dynamics applications and energy minimization problems.},
bibtype = {techreport},
author = {Kadupitiya, Jcs and Fox, Geoffrey C and Jadhao, Vikram},
doi = {10.1177/ToBeAssigned}
}
Downloads: 0
{"_id":"fnBH8nDxSypkkFY2K","bibbaseid":"kadupitiya-fox-jadhao-machinelearningforparameterautotuninginmoleculardynamicssimulationsefficientdynamicsofionsnearpolarizablenanoparticles-2018","authorIDs":[],"author_short":["Kadupitiya, J.","Fox, G., C.","Jadhao, V."],"bibdata":{"title":"Machine Learning for Parameter Auto-tuning in Molecular Dynamics Simulations: Efficient Dynamics of Ions near Polarizable Nanoparticles","type":"techreport","year":"2018","keywords":"Auto-tuning,Energy Minimization,Hybrid MPI/OpenMP,Machine Learning,Nanoscale Simulations,Parallel Computing","pages":"15","websites":"www.sagepub.com/,http://dsc.soic.indiana.edu/publications/Manuscript.IJHPCA.Nov2018.pdf","id":"93e36f16-de3c-38ac-8569-20194aa0e1dd","created":"2019-10-01T17:20:54.076Z","file_attached":"true","profile_id":"42d295c0-0737-38d6-8b43-508cab6ea85d","last_modified":"2020-05-11T14:43:32.762Z","read":false,"starred":false,"authored":"true","confirmed":"true","hidden":false,"citation_key":"Kadupitiya2018","private_publication":false,"abstract":"Simulating the dynamics of ions near polarizable nanoparticles (NPs) is extremely challenging due to the need to solve the Poisson equation at every simulation timestep. Recently, a molecular dynamics (MD) method based on a dynamical optimization framework bypassed this obstacle by representing the polarization charge density as virtual dynamic variables, and evolving them in parallel with the physical dynamics of ions. We highlight the computational gains accessible with the integration of machine learning (ML) methods for parameter prediction in MD simulations by demonstrating how they were realized in MD simulations of ions near polarizable NPs. An artificial neural network based regression model was integrated with MD and predicted the optimal simulation timestep and critical parameters characterizing the virtual system on-the-fly with 94.3% success. The integration of ML method with hybrid OpenMP/MPI parallelized MD simulations generated accurate dynamics of thousands of ions in the presence of polarizable NPs for over 10 million steps (with a maximum simulated physical time over 30 ns) while reducing the computational time from thousands of hours to tens of hours yielding a maximum speedup of ≈ 3 from ML-only acceleration and a maximum overall speedup of ≈ 600 from ML-hybrid Open/MPI combined method. Extraction of ionic structure in concentrated electrolytes near oil-water emulsions demonstrates the success of the method. The approach can be generalized to select optimal parameters in other molecular dynamics applications and energy minimization problems.","bibtype":"techreport","author":"Kadupitiya, Jcs and Fox, Geoffrey C and Jadhao, Vikram","doi":"10.1177/ToBeAssigned","bibtex":"@techreport{\n title = {Machine Learning for Parameter Auto-tuning in Molecular Dynamics Simulations: Efficient Dynamics of Ions near Polarizable Nanoparticles},\n type = {techreport},\n year = {2018},\n keywords = {Auto-tuning,Energy Minimization,Hybrid MPI/OpenMP,Machine Learning,Nanoscale Simulations,Parallel Computing},\n pages = {15},\n websites = {www.sagepub.com/,http://dsc.soic.indiana.edu/publications/Manuscript.IJHPCA.Nov2018.pdf},\n id = {93e36f16-de3c-38ac-8569-20194aa0e1dd},\n created = {2019-10-01T17:20:54.076Z},\n file_attached = {true},\n profile_id = {42d295c0-0737-38d6-8b43-508cab6ea85d},\n last_modified = {2020-05-11T14:43:32.762Z},\n read = {false},\n starred = {false},\n authored = {true},\n confirmed = {true},\n hidden = {false},\n citation_key = {Kadupitiya2018},\n private_publication = {false},\n abstract = {Simulating the dynamics of ions near polarizable nanoparticles (NPs) is extremely challenging due to the need to solve the Poisson equation at every simulation timestep. Recently, a molecular dynamics (MD) method based on a dynamical optimization framework bypassed this obstacle by representing the polarization charge density as virtual dynamic variables, and evolving them in parallel with the physical dynamics of ions. We highlight the computational gains accessible with the integration of machine learning (ML) methods for parameter prediction in MD simulations by demonstrating how they were realized in MD simulations of ions near polarizable NPs. An artificial neural network based regression model was integrated with MD and predicted the optimal simulation timestep and critical parameters characterizing the virtual system on-the-fly with 94.3% success. The integration of ML method with hybrid OpenMP/MPI parallelized MD simulations generated accurate dynamics of thousands of ions in the presence of polarizable NPs for over 10 million steps (with a maximum simulated physical time over 30 ns) while reducing the computational time from thousands of hours to tens of hours yielding a maximum speedup of ≈ 3 from ML-only acceleration and a maximum overall speedup of ≈ 600 from ML-hybrid Open/MPI combined method. Extraction of ionic structure in concentrated electrolytes near oil-water emulsions demonstrates the success of the method. The approach can be generalized to select optimal parameters in other molecular dynamics applications and energy minimization problems.},\n bibtype = {techreport},\n author = {Kadupitiya, Jcs and Fox, Geoffrey C and Jadhao, Vikram},\n doi = {10.1177/ToBeAssigned}\n}","author_short":["Kadupitiya, J.","Fox, G., C.","Jadhao, V."],"urls":{"Paper":"https://bibbase.org/service/mendeley/42d295c0-0737-38d6-8b43-508cab6ea85d/file/1ac4e463-14e8-697f-727c-8e1825063f6e/Kadupitiya_Fox_Jadhao___2018___Machine_Learning_for_Parameter_Auto_tuning_in_Molecular_Dynamics_Simulations_Efficient_Dy.pdf.pdf","Website":"www.sagepub.com/,http://dsc.soic.indiana.edu/publications/Manuscript.IJHPCA.Nov2018.pdf"},"biburl":"https://bibbase.org/service/mendeley/42d295c0-0737-38d6-8b43-508cab6ea85d","bibbaseid":"kadupitiya-fox-jadhao-machinelearningforparameterautotuninginmoleculardynamicssimulationsefficientdynamicsofionsnearpolarizablenanoparticles-2018","role":"author","keyword":["Auto-tuning","Energy Minimization","Hybrid MPI/OpenMP","Machine Learning","Nanoscale Simulations","Parallel Computing"],"metadata":{"authorlinks":{}},"downloads":0},"bibtype":"techreport","creationDate":"2019-09-12T13:19:08.409Z","downloads":0,"keywords":["auto-tuning","energy minimization","hybrid mpi/openmp","machine learning","nanoscale simulations","parallel computing"],"search_terms":["machine","learning","parameter","auto","tuning","molecular","dynamics","simulations","efficient","dynamics","ions","near","polarizable","nanoparticles","kadupitiya","fox","jadhao"],"title":"Machine Learning for Parameter Auto-tuning in Molecular Dynamics Simulations: Efficient Dynamics of Ions near Polarizable Nanoparticles","year":2018,"biburl":"https://bibbase.org/service/mendeley/42d295c0-0737-38d6-8b43-508cab6ea85d","dataSources":["zgahneP4uAjKbudrQ","ya2CyA73rpZseyrZ8","2252seNhipfTmjEBQ"]}