Deep Learning for Nonadiabatic Excited-State Dynamics.

In this work we show that deep learning (DL) can be used for exploring complex and highly nonlinear multistate potential energy surfaces of polyatomic molecules and related nonadiabatic dynamics. Our DL is based on deep neural networks (DNNs), which are used as accurate representations of the CASSCF ground- and excited-state potential energy surfaces (PESs) of CH2NH. After geometries near conical intersection are included in the training set, the DNN models accurately reproduce excited-state topological structures; photoisomerization paths; and, importantly, conical intersections. We have also demonstrated that the results from nonadiabatic dynamics run with the DNN models are very close to those from the dynamics run with the pure ab initio method. The present work should encourage further studies of using machine learning methods to explore excited-state potential energy surfaces and nonadiabatic dynamics of polyatomic molecules.

[1]  Klaus-Robert Müller,et al.  Machine learning of accurate energy-conserving molecular force fields , 2016, Science Advances.

[2]  Lihong Hu,et al.  A generalized exchange-correlation functional: the Neural-Networks approach ☆ , 2003, physics/0311024.

[3]  S. Hammes-Schiffer,et al.  Proton transfer in solution: Molecular dynamics with quantum transitions , 1994 .

[4]  Kipton Barros,et al.  Discovering a Transferable Charge Assignment Model Using Machine Learning. , 2018, The journal of physical chemistry letters.

[5]  Bernard R Brooks,et al.  Machine Learning Force Field Parameters from Ab Initio Data. , 2017, Journal of chemical theory and computation.

[6]  Lihong Hu,et al.  Improving the Performance of Long-Range-Corrected Exchange-Correlation Functional with an Embedded Neural Network. , 2017, The journal of physical chemistry. A.

[7]  J. Behler.  Neural network potential-energy surfaces in chemistry: a tool for large-scale simulations. , 2011, Physical chemistry chemical physics : PCCP.

[8]  U. Manthe,et al.  Neural network based coupled diabatic potential energy surfaces for reactive scattering. , 2017, The Journal of chemical physics.

[9]  W. Thiel,et al.  Implementation of surface hopping molecular dynamics using semiempirical methods , 2008 .

[10]  Hua Guo,et al.  Permutation invariant polynomial neural network approach to fitting potential energy surfaces. IV. Coupled diabatic potential energy matrices. , 2018, The Journal of chemical physics.

[11]  Ryan P. Adams,et al.  Design of efficient molecular organic light-emitting diodes by a high-throughput virtual screening and experimental approach. , 2016, Nature materials.

[12]  Lihong Hu,et al.  Combined first-principles calculation and neural-network correction approach for heat of formation , 2003 .

[13]  M. Rupp,et al.  Machine Learning for Quantum Mechanical Properties of Atoms in Molecules , 2015, 1505.00350.

[14]  Dong H. Zhang,et al.  Methane dissociation on Ni(111): A fifteen-dimensional potential energy surface using neural network method. , 2015, The Journal of chemical physics.

[15]  Deping Hu,et al.  Inclusion of Machine Learning Kernel Ridge Regression Potential Energy Surfaces in On-the-Fly Nonadiabatic Molecular Dynamics Simulation. , 2018, The journal of physical chemistry letters.

[16]  Walter Thiel,et al.  Nonadiabatic Excited-State Dynamics with Machine Learning , 2018, The journal of physical chemistry letters.

[17]  Leroy Cronin,et al.  Controlling an organic synthesis robot with machine learning to search for new reactivity , 2018, Nature.

[18]  Dong H. Zhang,et al.  Construction of diabatic energy surfaces for LiFH with artificial neural networks. , 2017, The Journal of chemical physics.

[19]  Justin S. Smith,et al.  Transferable Dynamic Molecular Charge Assignment Using Deep Neural Networks. , 2018, Journal of chemical theory and computation.

[20]  Lihong Hu,et al.  Alternative approach to chemical accuracy: a neural networks-based first-principles method for heat of formation of molecules made of H, C, N, O, F, S, and Cl. , 2014, The journal of physical chemistry. A.

[21]  John E Herr,et al.  The many-body expansion combined with neural networks. , 2016, The Journal of chemical physics.

[22]  Weitao Yang,et al.  Force Field for Water Based on Neural Network. , 2018, The journal of physical chemistry letters.

[23]  Linfeng Zhang,et al.  DeePCG: Constructing coarse-grained models via deep neural networks. , 2018, The Journal of chemical physics.

[24]  D. Yarkony,et al.  An improved quasi-diabatic representation of the 1, 2, 3(1)A coupled adiabatic potential energy surfaces of phenol in the full 33 internal coordinates. , 2016, The Journal of chemical physics.

[25]  Jinlan Wang,et al.  Accelerated discovery of stable lead-free hybrid organic-inorganic perovskites via machine learning , 2018, Nature Communications.

[26]  J. Behler.  First Principles Neural Network Potentials for Reactive Simulations of Large Molecular and Condensed Systems. , 2017, Angewandte Chemie.

[27]  Hua Guo,et al.  High-Dimensional Atomistic Neural Network Potentials for Molecule-Surface Interactions: HCl Scattering from Au(111). , 2017, The journal of physical chemistry letters.

[28]  Matthias Rupp,et al.  Big Data Meets Quantum Chemistry Approximations: The Δ-Machine Learning Approach. , 2015, Journal of chemical theory and computation.

[29]  Walter Thiel,et al.  Machine Learning of Parameters for Accurate Semiempirical Quantum Chemical Calculations , 2015, Journal of chemical theory and computation.

[30]  Paul Raccuglia,et al.  Machine-learning-assisted materials discovery using failed experiments , 2016, Nature.

[31]  GuanHua Chen,et al.  Neural network correction for heats of formation with a larger experimental training set and new descriptors , 2005 .

[32]  Weitao Yang,et al.  Multiscale Quantum Mechanics/Molecular Mechanics Simulations with Neural Networks. , 2016, Journal of chemical theory and computation.

[33]  S. Habershon,et al.  Direct Quantum Dynamics Using Grid-Based Wave Function Propagation and Machine-Learned Potential Energy Surfaces. , 2017, Journal of chemical theory and computation.

[34]  M. Parrinello,et al.  Molecular dynamics in low-spin excited states , 1998 .

[35]  Min Zhang,et al.  Improving the accuracy of density-functional theory calculation: the genetic algorithm and neural network approach. , 2007, The Journal of chemical physics.

[36]  Jörg Behler,et al.  Constructing high‐dimensional neural network potentials: A tutorial review , 2015 .

[37]  Mike Preuss,et al.  Planning chemical syntheses with deep neural networks and symbolic AI , 2017, Nature.

[38]  Alireza Khorshidi,et al.  Amp: A modular approach to machine learning in atomistic simulations , 2016, Comput. Phys. Commun..

[39]  E Weinan,et al.  DeePMD-kit: A deep learning package for many-body potential energy representation and molecular dynamics , 2017, Comput. Phys. Commun..

[40]  Alán Aspuru-Guzik,et al.  Inverse molecular design using machine learning: Generative models for matter engineering , 2018, Science.

[41]  Hua Guo,et al.  Constructing High-Dimensional Neural Network Potential Energy Surfaces for Gas–Surface Scattering and Reactions , 2018 .

[42]  Tristan Bereau,et al.  Transferable Atomic Multipole Machine Learning Models for Small Organic Molecules. , 2015, Journal of chemical theory and computation.

[43]  Hiroki Nakamura,et al.  The two‐state linear curve crossing problems revisited. III. Analytical approximations for Stokes constant and scattering matrix: Nonadiabatic tunneling case , 1993 .

[44]  J. Tully,et al.  Trajectory Surface Hopping Approach to Nonadiabatic Molecular Collisions: The Reaction of H+ with D2 , 1971 .

[45]  Li Li,et al.  Bypassing the Kohn-Sham equations with machine learning , 2016, Nature Communications.