A random-sampling high dimensional model representation neural network for building potential energy surfaces.

We combine the high dimensional model representation (HDMR) idea of Rabitz and co-workers [J. Phys. Chem. 110, 2474 (2006)] with neural network (NN) fits to obtain an effective means of building multidimensional potentials. We verify that it is possible to determine an accurate many-dimensional potential by doing low dimensional fits. The final potential is a sum of terms each of which depends on a subset of the coordinates. This form facilitates quantum dynamics calculations. We use NNs to represent HDMR component functions that minimize error mode term by mode term. This NN procedure makes it possible to construct high-order component functions which in turn enable us to determine a good potential. It is shown that the number of available potential points determines the order of the HDMR which should be used.

[1]  Akio Kawano,et al.  Interpolating moving least-squares methods for fitting potential energy surfaces: Analysis of an application to a six-dimensional system. , 2004, The Journal of chemical physics.

[2]  James B Witkoskie,et al.  Neural Network Models of Potential Energy Surfaces:  Prototypical Examples. , 2005, Journal of chemical theory and computation.

[3]  H. Meyer,et al.  Reactive scattering using the multiconfiguration time‐dependent Hartree approximation: General aspects and application to the collinear H+H2→H2+H reaction , 1995 .

[4]  H. Rabitz,et al.  Reproducing kernel Hilbert space interpolation methods as a paradigm of high dimensional model representations: Application to multidimensional potential energy surface construction , 2003 .

[5]  W. B. Cook,et al.  Torsion-Rotation Energy Levels, Hindering Potential, and Inertial Parameters of the First Excited Vibrational State of the Antisymmetric O-H Stretch in Hydrogen Peroxide , 1995 .

[6]  Akio Kawano,et al.  Interpolating moving least-squares methods for fitting potential energy surfaces: applications to classical dynamics calculations. , 2004, The Journal of chemical physics.

[7]  David A. Sprecher,et al.  A representation theorem for continuous functions of several variables , 1965 .

[8]  Frank C. De Lucia,et al.  A study of the rotational-torsional spectrum of hydrogen peroxide between 80 and 700 GHz , 1981 .

[9]  S. Carter,et al.  Approximate single-valued representations of multivalued potential energy surfaces , 1984 .

[10]  Ah Chung Tsoi,et al.  Universal Approximation Using Feedforward Neural Networks: A Survey of Some Existing Methods, and Some New Results , 1998, Neural Networks.

[11]  Allan Pinkus,et al.  Lower bounds for approximation by MLP neural networks , 1999, Neurocomputing.

[12]  Peter L. Bartlett,et al.  Neural Network Learning - Theoretical Foundations , 1999 .

[13]  R. Friesner.  Ab initio quantum chemistry: methodology and applications. , 2005, Proceedings of the National Academy of Sciences of the United States of America.

[14]  Bernard Widrow,et al.  30 years of adaptive neural networks: perceptron, Madaline, and backpropagation , 1990, Proc. IEEE.

[15]  Ranga Komanduri,et al.  Molecular dynamics investigations of the dissociation of SiO2 on an ab initio potential energy surface obtained using neural network methods. , 2006, The Journal of chemical physics.

[16]  L. Schriver,et al.  The 7.9-μm Band of Hydrogen Peroxide: Line Positions and Intensities , 1995 .

[17]  H. Rabitz,et al.  Random Sampling-High Dimensional Model Representation (RS-HDMR) with Nonuniformly Distributed Variables: Application to an Integrated Multimedia/ Multipathway Exposure and Dose Model for Trichloroethylene , 2003 .

[18]  Xinhou Liu,et al.  Many-body expansion of the potential-energy surface for BeHF and Dynamical calculations for the reaction, Be + HF(v, J)→ BeF(v′, J′)+ H , 1991 .

[19]  Donald J. Kouri,et al.  Distributed approximating functional approach to fitting multi-dimensional surfaces , 1996 .

[20]  C. David Sherrill,et al.  An assessment of the accuracy of multireference configuration interaction (MRCI) and complete-active-space second-order perturbation theory (CASPT2) for breaking bonds to hydrogen , 2003 .

[21]  Herschel Rabitz,et al.  Constructing global functional maps between molecular potentials and quantum observables , 2001 .

[22]  Lawrence B. Harding,et al.  Ab initio calculations of force fields for H2CN and C1HCN and vibrational energies of H2CN , 1997 .

[23]  Herschel Rabitz,et al.  The Ar–HCl potential energy surface from a global map-facilitated inversion of state-to-state rotationally resolved differential scattering cross sections and rovibrational spectral data , 2001 .

[24]  David C. Clary,et al.  Potential optimized discrete variable representation , 1992 .

[25]  Antara Dutta,et al.  Full configuration interaction potential energy curves for breaking bonds to hydrogen: An assessment of single-reference correlation methods , 2003 .

[26]  Martin Quack,et al.  Spectra and Dynamics of Coupled Vibrations in Polyatomic Molecules , 1990 .

[27]  Kurt Hornik,et al.  Approximation capabilities of multilayer feedforward networks , 1991, Neural Networks.

[28]  Herschel Rabitz,et al.  High‐dimensional model representations generated from low order terms—lp‐RS‐HDMR , 2003, J. Comput. Chem..

[29]  Thomas B. Blank,et al.  Adaptive, global, extended Kalman filters for training feedforward neural networks , 1994 .

[30]  J. W. C. Johns,et al.  The far infrared spectrum of H2O2. First observation of the staggering of the levels and determination of the cis barrier , 1989 .

[31]  Vera Kurková,et al.  Kolmogorov's theorem and multilayer neural networks , 1992, Neural Networks.

[32]  Tucker Carrington,et al.  Efficient calculation of highly excited vibrational energy levels of floppy molecules: The band origins of H+3 up to 35 000 cm−1 , 1994 .

[33]  Frank Jensen,et al.  Using force fields methods for locating transition structures , 2003 .

[34]  William H. Press,et al.  Numerical Recipes: FORTRAN , 1988 .

[35]  J. W. C. Johns,et al.  Torsion-vibration interaction in H2O2: First high-resolution observation of ν3☆ , 1992 .

[36]  R Komanduri,et al.  Ab initio potential-energy surfaces for complex, multichannel systems using modified novelty sampling and feedforward neural networks. , 2005, The Journal of chemical physics.

[37]  J. J. Soares Neto,et al.  The fitting of potential energy surfaces using neural networks: Application to the study of vibrational levels of H3+ , 1998 .

[38]  Herschel Rabitz,et al.  Global, nonlinear algorithm for inverting quantum-mechanical observations , 2001 .

[39]  J. J. Soares Neto,et al.  The fitting of potential energy surfaces using neural networks. Application to the study of the photodissociation processes , 1998 .

[40]  Kersti Hermansson,et al.  Representation of Intermolecular Potential Functions by Neural Networks , 1998 .

[41]  António J. C. Varandas,et al.  Analytical potentials for triatomic molecules , 1982 .

[42]  C. Micchelli,et al.  Approximation by superposition of sigmoidal and radial basis functions , 1992 .

[43]  S. Ergezinger,et al.  An accelerated learning algorithm for multilayer perceptrons: optimization layer by layer , 1995, IEEE Trans. Neural Networks.

[44]  Chih-Cheng Chen,et al.  A fast multilayer neural-network training algorithm based on the layer-by-layer optimizing procedures , 1996, IEEE Trans. Neural Networks.

[45]  Ilya M. Sobol,et al.  Theorems and examples on high dimensional model representation , 2003, Reliab. Eng. Syst. Saf..

[46]  H. Rabitz,et al.  High Dimensional Model Representations , 2001 .

[47]  Michael A. Collins,et al.  Application of interpolated potential energy surfaces to quantum reactive scattering , 1999 .

[48]  Neil Shenvi,et al.  Efficient chemical kinetic modeling through neural network maps. , 2004, The Journal of chemical physics.

[49]  G. Rauhut.  Efficient calculation of potential energy surfaces for the generation of vibrational wave functions. , 2004, The Journal of chemical physics.

[50]  Alexander Dalgarno,et al.  Theoretical study of HCl+ : Potential curves, radiative lifetimes, and photodissociation cross sections , 1991 .

[51]  C. J. Stone,et al.  Additive Regression and Other Nonparametric Models , 1985 .

[52]  Michael A. Collins,et al.  Molecular potential-energy surfaces by interpolation: Furtherrefinements , 1997 .

[53]  R Komanduri,et al.  Theoretical investigation of the dissociation dynamics of vibrationally excited vinyl bromide on an ab initio potential-energy surface obtained using modified novelty sampling and feed-forward neural networks. , 2006, The Journal of chemical physics.

[54]  D. Sprecher.  On the structure of continuous functions of several variables , 1965 .

[55]  Donald L. Thompson,et al.  Interpolating moving least-squares methods for fitting potential energy surfaces: Detailed analysis of one-dimensional applications , 2003 .

[56]  Joel M. Bowman,et al.  Tests of MULTIMODE calculations of rovibrational energies of CH4 , 2006 .

[57]  T. Carrington,et al.  A general discrete variable method to calculate vibrational energy levels of three‐ and four‐atom molecules , 1993 .

[58]  Herschel Rabitz,et al.  Multicut‐HDMR with an application to an ionospheric model , 2004, J. Comput. Chem..

[59]  T. Carrington,et al.  A nested molecule-independent neural network approach for high-quality potential fits. , 2006, The journal of physical chemistry. A.

[60]  Tucker Carrington,et al.  The discrete variable representation of a triatomic Hamiltonian in bond length–bond angle coordinates , 1992 .

[61]  J. Light,et al.  Quasirandom distributed Gaussian bases for bound problems , 2001 .

[62]  Akio Kawano,et al.  Improving the accuracy of interpolated potential energy surfaces by using an analytical zeroth-order potential function. , 2004, The Journal of chemical physics.

[63]  H. Rabitz,et al.  Practical Approaches To Construct RS-HDMR Component Functions , 2002 .

[64]  H. Rabitz,et al.  General foundations of high‐dimensional model representations , 1999 .

[65]  Kurt Hornik,et al.  Multilayer feedforward networks are universal approximators , 1989, Neural Networks.

[66]  Kurt Hornik,et al.  Universal approximation of an unknown mapping and its derivatives using multilayer feedforward networks , 1990, Neural Networks.

[67]  H. Rabitz,et al.  High Dimensional Model Representations Generated from Low Dimensional Data Samples. I. mp-Cut-HDMR , 2001 .

[68]  Claude Leforestier,et al.  Grid method for the Wigner functions. Application to the van der Waals system Ar–H2O , 1994 .

[69]  Emily Weiss,et al.  Achieving the laboratory control of quantum dynamics phenomena using nonlinear functional maps , 2001 .

[70]  Donald G Truhlar,et al.  Calculation of converged rovibrational energies and partition function for methane using vibrational-rotational configuration interaction. , 2004, The Journal of chemical physics.

[71]  Lawrence B. Harding,et al.  State-to-state chemistry with fast hydrogen atoms. Reaction and collisional excitation in H + CO2 , 1987 .

[72]  Charles A. Micchelli,et al.  Dimension-independent bounds on the degree of approximation by neural networks , 1994, IBM J. Res. Dev..

[73]  Ian M. Mills,et al.  Analytical potentials for triatomic molecules: VIII. A two-valued surface for the lowest 1A′ states of H2O , 1981 .

[74]  D. W. Noid,et al.  Potential energy surfaces for macromolecules. A neural network technique , 1992 .

[75]  A. Gross,et al.  Representing high-dimensional potential-energy surfaces for reactions at surfaces by neural networks , 2004 .

[76]  Ken-ichi Funahashi,et al.  On the approximate realization of continuous mappings by neural networks , 1989, Neural Networks.

[77]  C. Chui,et al.  Approximation by ridge functions and neural networks with one hidden layer , 1992 .

[78]  Andrew R. Barron,et al.  Universal approximation bounds for superpositions of a sigmoidal function , 1993, IEEE Trans. Inf. Theory.

[79]  Hans-Dieter Meyer,et al.  Product representation of potential energy surfaces. II , 1996 .

[80]  T Hollebeek,et al.  Constructing multidimensional molecular potential energy surfaces from ab initio data. , 2003, Annual review of physical chemistry.

[81]  H. Rabitz,et al.  Random sampling-high dimensional model representation (RS-HDMR) and orthogonality of its different order component functions. , 2006, The journal of physical chemistry. A.

[82]  J. Murrell,et al.  Molecular Potential Energy Functions , 1985 .

[83]  Deborah L Crittenden,et al.  Efficiency considerations in the construction of interpolated potential energy surfaces for the calculation of quantum observables by diffusion Monte Carlo. , 2004, The Journal of chemical physics.

[84]  George C. Schatz,et al.  A quasi-classical trajectory study of product vibrational distributions in the OH + H2 → H2O + H reaction , 1980 .

[85]  G. G. Maisuradze,et al.  Interpolating moving least-squares methods for fitting potential energy surfaces: Illustrative approaches and applications , 2003 .

[86]  Herschel Rabitz,et al.  Estimation of molecular properties by high-dimensional model representation. , 2006, The journal of physical chemistry. A.

[87]  Herschel Rabitz,et al.  Efficient Implementation of High Dimensional Model Representations , 2001 .

[88]  Mark N. Gibbs,et al.  Combining ab initio computations, neural networks, and diffusion Monte Carlo: An efficient method to treat weakly bound molecules , 1996 .

[89]  Tucker Carrington,et al.  Adiabatic approach to the coupling between local‐ and normal‐like modes , 1987 .

[90]  David Zhang,et al.  Parameter by Parameter Algorithm for Multilayer Perceptrons , 2006, Neural Processing Letters.

[91]  Steven D. Brown,et al.  Neural network models of potential energy surfaces , 1995 .

[92]  H. Rabitz,et al.  Efficient input-output model representations , 1999 .

[93]  M.H. Hassoun,et al.  Fundamentals of Artificial Neural Networks , 1996, Proceedings of the IEEE.

[94]  B. Kuhn,et al.  A new six-dimensional analytical potential up to chemically significant energies for the electronic ground state of hydrogen peroxide , 1999 .

[95]  Joel M. Bowman,et al.  Variational Calculations of Rotational−Vibrational Energies of CH4 and Isotopomers Using an Adjusted ab Initio Potential , 2000 .

[96]  N. Handy,et al.  Extensions and tests of “multimode”: a code to obtain accurate vibration/rotation energies of many-mode molecules , 1998 .

[97]  H. Rabitz,et al.  A general method for constructing multidimensional molecular potential energy surfaces from ab initio calculations , 1996 .

[98]  J. Zupan,et al.  Neural networks: A new method for solving chemical problems or just a passing phase? , 1991 .

[99]  Joel M. Bowman,et al.  Vibrational self-consistent field method for many-mode systems: A new approach and application to the vibrations of CO adsorbed on Cu(100) , 1997 .

[100]  Nicholas C. Handy,et al.  On the representation of potential energy surfaces of polyatomic molecules in normal coordinates , 2002 .

[101]  Hua Guo,et al.  Six-dimensional quantum calculations of highly excited vibrational energy levels of hydrogen peroxide and its deuterated isotopomers , 2001 .

[102]  Donald E. Jennings,et al.  High-resolution infrared spectrum of hydrogen peroxide: The ν6 fundamental band , 1986 .