SpikeComp: An Evolving Spiking Neural Network with Adaptive Compact Structure for Pattern Classification

Jinling Wang, A Belatreche, Liam Maguire, T.Martin McGinnity

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

2 Citations (Scopus)

Abstract

This paper presents a new supervised learning algorithm (SpikeComp) with an adaptive compact structure for Spiking Neural Networks (SNNs). SpikeComp consists of two layers of spiking neurons: an encoding layer which temporally encodes real valued features into spatio-temporal spike patterns, and an output layer of dynamically grown neurons which perform spatio-temporal pattern classification. The weights between the neurons in the encoding layer and the new added neuron in the output layer are initialised based on the precise spiking times in the encoding layer. New strategies are proposed to either add a new neuron, or update the network parameters when a new sample is presented to the network. The proposed learning algorithm was demonstrated on several benchmark classification datasets and the obtained results show that SpikeComp can perform pattern classification with a comparable performance and a much compact network structure compared with other existing SNN training algorithm.
Original languageEnglish
Title of host publicationConference Proceedings ICONIP: International Conference on Neural Information Processing - Neural Information Processing
Pages259-267
Number of pages8
ISBN (Electronic)978-3-319-26535-3
DOIs
Publication statusE-pub ahead of print - 10 Nov 2015

Publication series

NameLecture Notes in Computer Science

Fingerprint Dive into the research topics of 'SpikeComp: An Evolving Spiking Neural Network with Adaptive Compact Structure for Pattern Classification'. Together they form a unique fingerprint.

Cite this