Analysis of SpikeProp convergence with alternative spike response functions

Thiruvarudchelvan, Vaenthan, Crane, James W. and Bossomaier, Terry (2013). Analysis of SpikeProp convergence with alternative spike response functions. In: Proceedings of the 2013 IEEE Symposium on Foundations of Computational Intelligence (FOCI). FOCI 2013: IEEE Symposium on Foundations of Computational Intelligence, Singapore, (98-105). 16-19 April, 2013. doi:10.1109/FOCI.2013.6602461


Author Thiruvarudchelvan, Vaenthan
Crane, James W.
Bossomaier, Terry
Title of paper Analysis of SpikeProp convergence with alternative spike response functions
Conference name FOCI 2013: IEEE Symposium on Foundations of Computational Intelligence
Conference location Singapore
Conference dates 16-19 April, 2013
Proceedings title Proceedings of the 2013 IEEE Symposium on Foundations of Computational Intelligence (FOCI)
Journal name IEEE Symposium Series on Computational Intelligence (SSCI)
Place of Publication Piscataway, NJ, United States
Publisher Institute of Electrical and Electronics Engineers
Publication Year 2013
Sub-type Fully published paper
DOI 10.1109/FOCI.2013.6602461
Open Access Status
ISBN 9781467359009
9781467359016
Start page 98
End page 105
Total pages 8
Collection year 2014
Formatted Abstract/Summary
SpikeProp is a supervised learning algorithm for spiking neural networks analogous to backpropagation. Like backpropagation, it may fail to converge for particular networks, parameters and datasets. However there are several behaviours and additional failure modes unique to SpikeProp which have not been explicitly outlined in the literature. These factors hinder the adoption of SpikeProp for general machine learning use. In this paper we examine the mathematics of SpikeProp in detail and identify the various causes of failure therein. The analysis implies that applying certain constraints on parameters like initial weights can improve the rates of convergence. It also suggests that alternative spike response functions could improve the learning rate and reduce the number of convergence failures. We tested two alternative functions and found these predictions to be true.
Subjects 1702 Cognitive Sciences
Q-Index Code E1
Q-Index Status Confirmed Code
Institutional Status UQ

Document type: Conference Paper
Collections: Queensland Brain Institute Publications
Official 2014 Collection
 
Versions
Version Filter Type
Citation counts: Scopus Citation Count Cited 2 times in Scopus Article | Citations
Google Scholar Search Google Scholar
Created: Thu, 28 Nov 2013, 20:17:43 EST by System User on behalf of Queensland Brain Institute