Neural networks offer an unbiased and numerically very accurate approach to represent high-dimensional ab initio potential-energy surfaces. Once constructed, neural network potentials can provide the energies and forces many orders of magnitude faster than electronic structure calculations, and thus enable molecular dynamics simulations of large systems. However, Cartesian coordinates are not a good choice to represent the atomic positions, and a transformation to symmetry functions is required. Using simple benchmark systems, the properties of several types of symmetry functions suitable for the construction of high-dimensional neural network potential-energy surfaces are discussed in detail. The symmetry functions are general and can be applied to all types of systems such as molecules, crystalline and amorphous solids, and liquids.
REFERENCES
1.
R.
Car
and M.
Parrinello
, Phys. Rev. Lett.
55
, 2471
(1985
).2.
D.
Marx
and J.
Hutter
, Ab initio Molecular Dynamics: Basic Theory and Advanced Methods
(Cambridge University Press
, Cambridge
, 2009
).3.
W. H.
Press
, S. A.
Teukolsky
, W. T.
Vetterling
, and P.
Flannert
, Numerical Recipes in Fortran
, 2nd ed. (Cambridge University Press
, Cambridge
, 1992
).4.
J.
Ischtwan
and M. A.
Collins
, J. Chem. Phys.
100
, 8080
(1994
).5.
M. J. T.
Jordan
, K. C.
Thompson
, and M. A.
Collins
, J. Chem.Phys.
102
, 5647
(1995
).6.
M. A.
Collins
, Theor. Chem. Acc.
108
, 313
(2002
).7.
A. P.
Bartok
, M. C.
Payne
, R.
Kondor
, and G.
Csanyi
, Phys. Rev. Lett.
104
, 136403
(2010
).8.
C. M.
Bishop
, Neural Networks for Pattern Recognition
(Oxford University Press
, Oxford
, 1995
).9.
G.
Cybenko
, Math. Control, Signals Syst.
2
, 303
(1989
).10.
K.
Hornik
, M.
Stinchcombe
, and H.
White
, Neural Networks
2
, 359
(1989
).11.
C. M.
Handley
and P. L. A.
Popelier
, J. Phys. Chem. A
114
, 3371
(2010
).12.
J.
Behler
, Chem. Modelling
7
, 1
(2010
).13.
J.
Behler
and M.
Parrinello
, Phys. Rev. Lett.
98
, 146401
(2007
).14.
J.
Behler
, R.
Martoňák
, D.
Donadio
, and M.
Parrinello
, Phys. Status Solidi (b)
245
, 2618
(2008
).15.
J.
Behler
, R.
Martoňák
, D.
Donadio
, and M.
Parrinello
, Phys. Rev. Lett.
100
, 185501
(2008
).16.
H.
Eshet
, R. Z.
Khaliullin
, T.
Kühne
, J.
Behler
, and M.
Parrinello
, Phys. Rev. B
81
, 184107
(2010
).17.
R. Z.
Khaliullin
, H.
Eshet
, T.
Kühne
, J.
Behler
, and M.
Parrinello
, Phys. Rev. B
81
, 100103
(2010
).18.
RuNNer—A Neural Network Code for High-Dimensional Potential-Energy Surfaces,
Jörg
Behler
, Lehrstuhl für Theoretische Chemie, Ruhr-Universität Bochum
, Germany.19.
J. B.
Witkoskie
and D. J.
Doren
, J. Chem. Theory Comput.
1
, 14
(2005
).20.
H. M.
Le
and L. M.
Raff
, J. Phys. Chem. A
114
, 45
(2010
).21.
A.
Pukrittayakamee
, M.
Malshe
, M.
Hagan
, L. M.
Raff
, R.
Narulkar
, S.
Bukkapatnum
, and R.
Komanduri
, J. Chem. Phys.
130
, 134101
(2009
).22.
F. V.
Prudente
and J. J. Soares
Neto
, Chem. Phys. Lett.
287
, 585
(1998
).23.
F. V.
Prudente
, P. H.
Acioli
, and J. J. Soares
Neto
, J. Chem. Phys.
109
, 8801
(1998
).24.
A. C. P.
Bittencourt
, F. V.
Prudente
, and J. D. M.
Vianna
, Chem. Phys.
297
, 153
(2004
).25.
T. M. Rocha
Filho
, Z. T.
Oliveira
, Jr., L. A. C.
Malbouisson
, R.
Gargano
, and J. J.
Soares Neto
, Int. J. Quantum. Chem.
95
, 281
(2003
).26.
S.
Manzhos
, X.
Wang
, R.
Dawes
, and T.
Carrington
, Jr., J. Phys. Chem. A
110
, 5295
(2006
).27.
S.
Manzhos
and T.
Carrington
, Jr., J. Chem. Phys.
125
, 84109
(2006
).28.
S.
Manzhos
and T.
Carrington
, Jr., J. Chem. Phys.
125
, 194105
(2006
).29.
S.
Manzhos
and T.
Carrington
, Jr., J. Chem. Phys.
127
, 014103
(2007
).30.
S.
Manzhos
and T.
Carrington
, Jr., J. Chem. Phys.
129
, 224104
(2008
).31.
M.
Malshe
, R.
Narulkar
, L. M.
Raff
, M.
Hagan
, S.
Bukkapatnam
, P. M.
Agrawal
, and R.
Komanduri
, J. Chem. Phys.
130
, 184102
(2009
).32.
M.
Malshe
, R.
Narulkar
, L. M.
Raff
, M.
Hagan
, S.
Bukkapatnam
, and R.
Komanduri
, J. Chem. Phys.
129
, 044111
(2008
).33.
T. B.
Blank
, S. D.
Brown
, A. W.
Calhoun
, and D. J.
Doren
, J. Chem. Phys.
103
, 4129
(1995
).34.
S.
Lorenz
, A.
Groß
, and M.
Scheffler
, Chem. Phys. Lett.
395
, 210
(2004
).35.
S.
Lorenz
, M.
Scheffler
, and A.
Groß
, Phys. Rev. B
73
, 115431
(2006
).36.
J.
Behler
, S.
Lorenz
, and K.
Reuter
, J. Chem. Phys.
127
, 014705
(2007
).37.
J.
Behler
, K.
Reuter
, and M.
Scheffler
, Phys. Rev. B
77
, 115421
(2008
).38.
J.
Ludwig
and D. G.
Vlachos
, J. Chem. Phys.
127
, 154716
(2007
).39.
D. A. R. S.
Latino
, R. P. S.
Fartaria
, F. F. M.
Freitas
, J.
Aires-de-Sousa
, and F. M. S.
Silva Fernandes
, J. Electroanal. Chem.
624
, 109
(2008
). 40.
C.
Carbogno
, J.
Behler
, K.
Reuter
, and A.
Groß
, Phys. Rev. B
81
, 035410
(2010
).41.
P. M.
Agrawal
, L. M.
Raff
, M. T.
Hagan
, and R.
Komanduri
, J. Chem. Phys.
124
, 124306
(2006
).42.
J.
Tersoff
, Phys. Rev. B
39
, 5566
(1989
).43.
We thank an unknown referee for suggesting the following cutoff function with continuous first and second derivatives:
\begin{equation*}f_{\mathrm{c}}(R_{ij})=\Bigg \lbrace \begin{array}{lc@{\hskip-6pt}c}\tanh ^3\left[1-\frac{R_{ij}}{R_{\mathrm{c}}}\right] &\quad \mbox{for} &\quad R_{ij}\le R_{\mathrm{c}}\\[6pt]0 &\quad \mbox{for} &\quad R_{ij} > R_{\mathrm{c}}. \end{array}\end{equation*}
44.
S.
Hobday
, R.
Smith
, and J.
Belbruno
, Modell. Simul. Mater. Sci. Eng.
7
, 397
(1999
).45.
S.
Hobday
, R.
Smith
, and J.
BelBruno
, Nucl. Instrum. Methods Phys. Res. B
153
, 247
(1999
).46.
A.
Bholoa
, S. D.
Kenny
, and R.
Smith
, Nucl. Instrum. Methods Phys. Res. B
255
, 1
(2007
).47.
E.
Sanville
, A.
Bholoa
, R.
Smith
, and S. D.
Kenny
, J. Phys. Condens. Matter
20
, 285219
(2008
).48.
H.
Gassner
, M.
Probst
, A.
Lauenstein
, and K.
Hermansson
, J. Phys. Chem. A
102
, 4596
(1998
).49.
M.
Parrinello
and A.
Rahman
, Phys. Rev. Lett.
45
, 1196
(1980
).50.
M.
Parrinello
and A.
Rahman
, J. Appl. Phys.
52
, 7182
(1981
).51.
R.
Martoňák
, A.
Laio
, and M.
Parrinello
, Phys. Rev. Lett.
90
, 75503
(2003
).© 2011 American Institute of Physics.
2011
American Institute of Physics
You do not currently have access to this content.
