Attractor neural networks with patchy connectivity
2007 (English)In: ESANN 2005 Proceedings - 13th European Symposium on Artificial Neural Networks, 2007, 429-434 p.Conference paper (Refereed)
We investigate the effects of patchy (clustered) connectivity in sparsely connected attractor neural networks (NNs). This study is motivated by the fact that the connectivity of pyramidal neurons in layer II/III of the mammalian visual cortex is patchy and sparse. The storage capacity of hypercolumnar attractor NNs that use the Hopfield and Willshaw learning rules with this kind of connectivity is investigated analytically as well as by simulation experiments. We find that patchy connectivity gives a higher storage capacity, given an overall sparse connectivity and regardless of learning rule.
Place, publisher, year, edition, pages
2007. 429-434 p.
Artificial intelligence, Cognitive systems
Other Computer and Information Science
IdentifiersURN: urn:nbn:se:kth:diva-154667ScopusID: 2-s2.0-84886997164ISBN: 2930307056ISBN: 978-293030705-3OAI: oai:DiVA.org:kth-154667DiVA: diva2:758596
13th European Symposium on Artificial Neural Networks, ESANN 2005, 27 April 2005 through 29 April 2005, Bruges, Belgium
QC 201410272014-10-272014-10-272014-10-27Bibliographically approved