Presentation 2001/6/21
A model of context-dependent association using selective desensitization of nonmonotonic neural elements
Kouhei MATSUZAWA, Kazuhiko MURATA, Masahiko MORITA,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) Existing neural network models with distributed information representation have difficulty in associating the same input with various targets. We consider the reason of this difficulty and present a model of context-dependent association using a novel method of contextual modification. We also show by computer simulations that this model can simulate a large-scale finite state machine with a relatively small number of neural elements and limited repetitions of learning.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Neural network / Associative memory / Selective desensitization / Nonmonotone dynamics / Trajectory attractor / Finite state machine
Paper # NC2001-21
Date of Issue

Conference Information
Committee NC
Conference Date 2001/6/21(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Neurocomputing (NC)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) A model of context-dependent association using selective desensitization of nonmonotonic neural elements
Sub Title (in English)
Keyword(1) Neural network
Keyword(2) Associative memory
Keyword(3) Selective desensitization
Keyword(4) Nonmonotone dynamics
Keyword(5) Trajectory attractor
Keyword(6) Finite state machine
1st Author's Name Kouhei MATSUZAWA
1st Author's Affiliation Doctral Program in Systems and Information Engineering, University of Tsukuba()
2nd Author's Name Kazuhiko MURATA
2nd Author's Affiliation College of Engineering Systems, University of Tsukuba
3rd Author's Name Masahiko MORITA
3rd Author's Affiliation Institute of Engineering Mechanics and Systems, University of Tsukuba
Date 2001/6/21
Paper # NC2001-21
Volume (vol) vol.101
Number (no) 153
Page pp.pp.-
#Pages 8
Date of Issue