Computation with spiking neural networks
Keywords
No Thumbnail Available
Authors
Issue Date
2014-10-30
Language
en
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Unlike generic spiking neural networks, networks with stable irregular dynamics are stable against small perturbations. Furthermore, big networks were shown have a state space fractured in flux tubes. We describe the novel effect of flux tube merging in big networks and expanded the concept of flux tubes also to very small networks. Using intuition from simple examples, we present an approach for an analytica! description of a state space with flux tubes. This allows to formally understand flux tubes in both small and big networks. Then, we identify a number of computationally beneficia! properties of such state space structures and propose setups that exploit these to perform real spike-time-dependent computation. With the analytica! understanding it may become possible to design network connectivities for specific flux tube dynamics.
Description
Citation
Supervisor
Faculty
Faculteit der Sociale Wetenschappen