Training a digital model of a deep spiking neural network using backpropagation

Deep spiking neural networks are one of the promising eventbased sensor signal processing concepts. However, the practical application of such networks is difficult with standard deep neural network training packages. In this paper, we propose a vector-matrix description of a spike neural network th...

Full description

Bibliographic Details
Main Author:	Bondarev V
Format:	Article
Language:	English
Published:	EDP Sciences 2020-01-01
Series:	E3S Web of Conferences
Online Access:	https://www.e3s-conferences.org/articles/e3sconf/pdf/2020/84/e3sconf_TPACEE2020_01026.pdf

id	doaj-9fdf6a9b96cb47cebe581114de81ae88
record_format	Article
spelling	doaj-9fdf6a9b96cb47cebe581114de81ae882021-04-02T18:59:32ZengEDP SciencesE3S Web of Conferences2267-12422020-01-012240102610.1051/e3sconf/202022401026e3sconf_TPACEE2020_01026Training a digital model of a deep spiking neural network using backpropagationBondarev V0Sevastopol State UniversityDeep spiking neural networks are one of the promising eventbased sensor signal processing concepts. However, the practical application of such networks is difficult with standard deep neural network training packages. In this paper, we propose a vector-matrix description of a spike neural network that allows us to adapt the traditional backpropagation algorithm for signals represented as spike time sequences. We represent spike sequences as binary vectors. This enables us to derive expressions for the forward propagation of spikes and the corresponding spike training algorithm based on the back propagation of the loss function sensitivities. The capabilities of the proposed vector-matrix model are demonstrated on the problem of handwritten digit recognition on the MNIST data set. The classification accuracy on test data for spiking neural network with 3 hidden layers is equal to 98.14%.https://www.e3s-conferences.org/articles/e3sconf/pdf/2020/84/e3sconf_TPACEE2020_01026.pdf
collection	DOAJ
language	English
format	Article
sources	DOAJ
author	Bondarev V
spellingShingle	Bondarev V Training a digital model of a deep spiking neural network using backpropagation E3S Web of Conferences
author_facet	Bondarev V
author_sort	Bondarev V
title	Training a digital model of a deep spiking neural network using backpropagation
title_short	Training a digital model of a deep spiking neural network using backpropagation
title_full	Training a digital model of a deep spiking neural network using backpropagation
title_fullStr	Training a digital model of a deep spiking neural network using backpropagation
title_full_unstemmed	Training a digital model of a deep spiking neural network using backpropagation
title_sort	training a digital model of a deep spiking neural network using backpropagation
publisher	EDP Sciences
series	E3S Web of Conferences
issn	2267-1242
publishDate	2020-01-01
description	Deep spiking neural networks are one of the promising eventbased sensor signal processing concepts. However, the practical application of such networks is difficult with standard deep neural network training packages. In this paper, we propose a vector-matrix description of a spike neural network that allows us to adapt the traditional backpropagation algorithm for signals represented as spike time sequences. We represent spike sequences as binary vectors. This enables us to derive expressions for the forward propagation of spikes and the corresponding spike training algorithm based on the back propagation of the loss function sensitivities. The capabilities of the proposed vector-matrix model are demonstrated on the problem of handwritten digit recognition on the MNIST data set. The classification accuracy on test data for spiking neural network with 3 hidden layers is equal to 98.14%.
url	https://www.e3s-conferences.org/articles/e3sconf/pdf/2020/84/e3sconf_TPACEE2020_01026.pdf
work_keys_str_mv	AT bondarevv trainingadigitalmodelofadeepspikingneuralnetworkusingbackpropagation
_version_	1721550160321839104

Training a digital model of a deep spiking neural network using backpropagation

Similar Items