Parallel batch pattern training algorithm for deep neural network

Author

Turchenko, Volodymyr ; Golovko, Vladimir

Author_Institution

Res. Inst. for Intell. Comput. Syst., Ternopil Nat. Econ. Univ., Ternopil, Ukraine

fYear

2014

fDate

21-25 July 2014

Firstpage

697

Lastpage

702

Abstract

The development of parallel batch pattern training algorithm for deep multilayered neural network architecture and its parallelization efficiency research on many-core system are presented in this paper. The model of a deep neural network and batch pattern training algorithm are theoretically described. The algorithmic description of the parallel batch pattern training method is presented. Our results show high parallelization efficiency of the developed algorithm on many-core parallel system with 48 CPUs with the use of message passing parallelization technology.

Keywords

message passing; multilayer perceptrons; multiprocessing systems; parallel algorithms; deep multilayered neural network architecture; deep neural network; many-core parallel system; message passing parallelization technology; parallel batch pattern training algorithm; parallelization efficiency; parallelization efficiency research; Algorithm design and analysis; Artificial neural networks; Computer architecture; Neurons; Parallel algorithms; Training; Vectors;

fLanguage

English

Publisher

ieee

Conference_Titel

High Performance Computing & Simulation (HPCS), 2014 International Conference on

Conference_Location

Bologna

Print_ISBN

978-1-4799-5312-7

Type

conf

DOI

10.1109/HPCSim.2014.6903757

Filename

6903757