Presentation 2001/3/10
An experimental study on the basin of attraction in associative neural networks with a weighted pattern storage
Masaaki KANAKUBO, Masaharu ADACHI,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In this paper, we numerically analyse the dynamics of associative neural networks in which stored patterns are embedded with a weighted auto-associative matrix. The networks are composed of static model neurons and random model neurons. As the weight to a stored pattern is larger, the basin of attraction to the pattern is wider in the static neural network, and the retrieval frequency of the pattern is higher in the random neural network. Moreover, it is shown that the weighting to the stored pattern in the higher strage capacity is more effective to the enlargement of the basin of attraction than that in the lower capacity.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) associative memory / neural network / weighted auto-associative matrix
Paper # NLP2000-175
Date of Issue

Conference Information
Committee NLP
Conference Date 2001/3/10(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 JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) An experimental study on the basin of attraction in associative neural networks with a weighted pattern storage
Sub Title (in English)
Keyword(1) associative memory
Keyword(2) neural network
Keyword(3) weighted auto-associative matrix
1st Author's Name Masaaki KANAKUBO
1st Author's Affiliation Dept. of Electronic Engineering, Graduate School of Engineering, Tokyo Denki Univ.()
2nd Author's Name Masaharu ADACHI
2nd Author's Affiliation Dept. of Electronic Engineering, Graduate School of Engineering, Tokyo Denki Univ.:Dept. of Electronic Engineering, College of Engineering, Tokyo Denki University
Date 2001/3/10
Paper # NLP2000-175
Volume (vol) vol.100
Number (no) 681
Page pp.pp.-
#Pages 6
Date of Issue