Files in this item



application/pdfECE499-Sp2011-wang.pdf (2MB)Restricted to U of Illinois
(no description provided)PDF


Title:Implementation and Analysis of an AB Initio Multi-Scale Model of Associative Memory
Author(s):Wang, Felix
Contributor(s):Levinson, Stephen
Subject(s):artificial intelligence
associative memory
multiscale modeling
Abstract:In this work, I explore the behaviors and metrics relating to a nonlinear dynamical multi-scale model of associative memory developed by Alex Duda. The model utilizes an ab initio approach with the classic Hodgkin-Huxley neuron serving as the basis, and it aims to capture the functionality of associative memory as it works in humans. Particular to my analysis are the first two scales of the model, extending from the individual Hodgkin-Huxley neuron to a population of interconnected neurons. At the first scale, the neurocomputational behaviors exhibited that are intrinsic to the neuron model, as well as those requiring modifications to the original model parameters, are studied. A discussion on the effects of these parameters is also given. At the second scale, we pay attention to metrics relating to the network as a whole: phase synchrony among neurons and the role topological structure plays. To this end, the effects of various methods of network initialization as well as application of external input are explored and discussed.
Issue Date:2011-05
Publication Status:unpublished
Peer Reviewed:not peer reviewed
Date Available in IDEALS:2014-01-17

This item appears in the following Collection(s)

Item Statistics