Fine-Grain Parallelization of Recurrent Neural Networks Training

Author

Turchenko, Volodymyr

Author_Institution

Dept. of Inf. Comput. Syst. & Control, Ternopil State Econ. Univ., Ternopil

fYear

2006

fDate

Feb. 28 2006-March 4 2006

Firstpage

208

Lastpage

211

Abstract

An approach to development of fine-grain parallel algorithm of artificial neural network training using parallelization of computational operations of each elementary neuron is presented in this paper. A training algorithm of back error propagation is described and parallel section of the algorithm is developed. The results of experimental research of the parallel algorithm are given using analysis of parallelization speedup and efficiency on parallel computer Origin 300.

Keywords

backpropagation; parallel algorithms; parallel machines; parallel processing; recurrent neural nets; Origin 300 parallel computer; artificial neural network training; back error propagation; computational parallelization; fine-grain parallel algorithm; fine-grain parallelization; parallelization efficiency; parallelization speedup; recurrent neural networks training; training algorithm; Artificial neural networks; Computer errors; Computer networks; Concurrent computing; Hardware; Neural networks; Neurons; Parallel algorithms; Parallel processing; Recurrent neural networks; fine-grain parallelization; parallel computer; recurrent neural network;

fLanguage

English

Publisher

ieee

Conference_Titel

Modern Problems of Radio Engineering, Telecommunications, and Computer Science, 2006. TCSET 2006. International Conference

Conference_Location

Lviv-Slavsko

Print_ISBN

966-553-507-2

Type

conf

DOI

10.1109/TCSET.2006.4404497

Filename

4404497