Dissertation available on cognitive modeling with spiking neurons

[J.Sougne@ulg.ac.be]

Dear Connectionists,

The following dissertation is now available on-line at
   http://www.fapse.ulg.ac.be/Lab/cogsci/jsougne/jspapers.html


http://www.fapse.ulg.ac.be/Lab/cogsci/jsougne/JSougneThesis.ps.Z
(compressed postscript)
http://www.fapse.ulg.ac.be/Lab/cogsci/jsougne/JSougneThesis.pdf (pdf)




     INFERNET: A Neurocomputational Model of Binding and Inference

	                            by

                             Jacques P. Sougné



ABSTRACT

An implementation of a network of integrate-and-fire neuron-like elements
is presented.  Integrate-and-fire nodes fire at a precise moment and
transmit their activation, with a particular strength and delay, to nodes
connected to them.  The receiving nodes accumulate potential but also
slowly loose their potential through decay.  When the potential of the node
reaches a particular threshold, it emits a spike.  Thereafter, the
potential is reset to a resting value.  As with real neurons, there is a
short refractory period during which this node will be completely
insensitive to incoming signals, after which its sensitivity will slowly
increase.  Precise timing properties have been used to represent symbols in
a distributed manner, and to solve the problems of binding and multiple
instantiation.  This architecture produced several predictions about human
short-term memory, predicate processing, complex reasoning, and multiple
instantiation.  These predictions have been tested by empirical studies on
humans.  This network shows symbolic processing abilities using
neurologically and psychologically plausible mechanisms that maintain the
advantages of generalization and noise tolerance found in connectionist
networks.