Efficiency research of batch and single pattern MLP parallel training algorithms

Author

Turchenko, Volodymyr ; Grandinetti, Lucio

Author_Institution

Dept. of Electron., Inf. & Syst., Univ. of Calabria, Rende, Italy

fYear

2009

fDate

21-23 Sept. 2009

Firstpage

218

Lastpage

224

Abstract

The development of parallel algorithms for batch and single pattern back propagation training of a multilayer perceptron and the research of their efficiency on a general-purpose parallel computer are presented in this paper. The multilayer perceptron model and the sequential batch and single pattern training algorithms are theoretically described. An algorithmic description of the parallel versions of the batch and single pattern training methods are specified. The efficiencies of the developed parallel algorithms are investigated by progressively increasing the dimension of the parallelized problem on a general-purpose parallel computer NEC TX-7.

Keywords

backpropagation; multilayer perceptrons; parallel algorithms; back propagation training; batch MLP parallel training algorithm; general-purpose parallel computer; multilayer perceptron model; single pattern MLP parallel training algorithm; Computational modeling; Computer networks; Concurrent computing; Grid computing; Hardware; High performance computing; Multilayer perceptrons; Neural networks; Neurons; Parallel algorithms; Parallel batch pattern training; multilayer perceptron; parallel single pattern training; parallelization efficiency;

fLanguage

English

Publisher

ieee

Conference_Titel

Intelligent Data Acquisition and Advanced Computing Systems: Technology and Applications, 2009. IDAACS 2009. IEEE International Workshop on

Conference_Location

Rende

Print_ISBN

978-1-4244-4901-9

Electronic_ISBN

978-1-4244-4882-1

Type

conf

DOI

10.1109/IDAACS.2009.5342990

Filename

5342990