Memory trace in spiking neural networks

Autor(en): Castellano, M.
Pipa, G. 
Stichwörter: Continuous delayed systems; delayed-dynamical systems; Dynamical systems; History dependencies; Information exchanges; Intelligent agents, Neural networks; Limited memory; memory trace; Neural computations; reservoir computing; spiking neural networks; Spiking neuron, Computer architecture
Erscheinungsdatum: 2013
Journal: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volumen: 8131 LNCS
Startseite: 264
Seitenende: 271
Zusammenfassung: 
Spiking neural networks have a limited memory capacity, such that a stimulus arriving at time t would vanish over a timescale of 200-300 milliseconds [1]. Therefore, only neural computations that require history dependencies within this short range can be accomplished. In this paper, the limited memory capacity of a spiking neural network is extended by coupling it to an delayed-dynamical system. This presents the possibility of information exchange between spiking neurons and continuous delayed systems. © 2013 Springer-Verlag Berlin Heidelberg.
Beschreibung: 
Conference of 23rd International Conference on Artificial Neural Networks, ICANN 2013 ; Conference Date: 10 September 2013 Through 13 September 2013; Conference Code:99717
ISBN: 9783642407277
ISSN: 03029743
DOI: 10.1007/978-3-642-40728-4_33
Externe URL: https://www.scopus.com/inward/record.uri?eid=2-s2.0-84884922381&doi=10.1007%2f978-3-642-40728-4_33&partnerID=40&md5=6aeeb94a998000c70fd7bddddd2ae370

Show full item record

Page view(s)

7
Last Week
0
Last month
0
checked on Apr 23, 2024

Google ScholarTM

Check

Altmetric