Presentation 2007-03-05
Retrieval Properties of a Hopfield Type Associative Neural Network with Hysteretic Transfer Function
Erik OBERG, Yoshihiro HAYAKAWA, Koji NAKAJIMA,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) A Hopfield type associative neural network for storing and recalling patterns is investigated. The use of a hysteretic transfer function is shown to improve retrieval performance. This can be explained by considering the motion on the u-x plane when the neuron output approaches the nonlinear saturation region of the hysteretic transfer function. The switching between branches of the hysteretic transfer function introduces a discontinuity of the neuron output, which gives the network the ability to escape some local minima. This discontinuity also increases the convergence time, and thus speeds up memory recall.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Associative memory / Hopfield model / retrieval dynamics / hysteretic transfer function
Paper # NLP2006-153
Date of Issue

Conference Information
Committee NLP
Conference Date 2007/2/26(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 Nonlinear Problems (NLP)
Language ENG
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Retrieval Properties of a Hopfield Type Associative Neural Network with Hysteretic Transfer Function
Sub Title (in English)
Keyword(1) Associative memory
Keyword(2) Hopfield model
Keyword(3) retrieval dynamics
Keyword(4) hysteretic transfer function
1st Author's Name Erik OBERG
1st Author's Affiliation Laboratory for Brainware Systems()
2nd Author's Name Yoshihiro HAYAKAWA
2nd Author's Affiliation Nanoelectronics and Spintronics, Research Institute of Electrical Communication, Tohoku University
3rd Author's Name Koji NAKAJIMA
3rd Author's Affiliation Laboratory for Brainware Systems
Date 2007-03-05
Paper # NLP2006-153
Volume (vol) vol.106
Number (no) 573
Page pp.pp.-
#Pages 6
Date of Issue