COSIRES.pdf (812.71 kB)
A new approach to potential fitting using neural networks
journal contribution
posted on 2006-10-16, 12:33 authored by Ajeevsing Bholoa, Steven KennySteven Kenny, Roger SmithA methodology is presented for developing transferable empirical
potential functions without following the usual procedure of postulating
a functional form. Instead, a neural network (NN) is employed
to learn the functional relationships of potential energy surfaces from
the local geometric arrangement of atoms. The methodology is illustrated
by training the NN model on tens of thousands of individual
data points derived from the tight-binding (TB) method for a wide
range of silicon systems including both small clusters and bulk structures.
Comparisons of the potential’s properties with experimental
data, quantum methods and other Si potentials have been made.
The NN model successfully fitted energy variations of the different
test cases as a function of bond distances, bond angles, lattice constants
and elastic properties for both equilibrium and non-equilibrium
small cluster and bulk structures. This indicates a robust and consistent
methodology for fitting empirical potentials which can be applied
to a wide range of materials independent of the type of bonding or
their crystal structure.
History
School
- Science
Department
- Mathematical Sciences
Pages
832214 bytesCitation
BHOLOA, A., KENNY, S.D. and SMITH, R., 2007. A new approach to potential fitting using neural networks. Nuclear instruments and methods in physics research section B: Beam interactions with materials and atoms, 255 (1), pp. 1-7 [doi:10.1016/j.nimb.2006.11.040]Publisher
© ElsevierPublication date
2007Notes
This article has been accepted for publication in the journal, Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms [© Elsevier] and the definitive version is available at: http://www.sciencedirect.com/science/journal/0168583XISSN
0168-583XLanguage
- en