Presentation 1997/7/24
Characteristics of the complex-valued associative memory model having penalty term
Hiroyuki AOKI, Yukio KOSUGI,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In this paper, the multistate and complex-valued neural networks, which deal with variables expressed in complex number and have me energy function having penalty term, is focused on and the characteristics of the model are investigated by computer simulation. In the proposed model, some rules are embeded in making patterns to be memorized, and the rules are reflected in the energy function as the penalty term. The storage capacity and the retrieval ability of the stored patterns can be improved by using the rules as prior information. As an example in making patterns to be memorized, we halved each pattern into two string and imposed a rule between two strings that each element of a string should be the complex conjugate of me partner element, situated at the same position in the other string. Furthermore, we show that the memorized patters based on M-sequence improved the performance of the model significantly. The complex-valued associative memory can be applied to error correcting means in communication systems.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) complex associative memory / energy function / penalty term / redundancy / communication / error correction / apriori information
Paper # NC97-33
Date of Issue

Conference Information
Committee NC
Conference Date 1997/7/24(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) Characteristics of the complex-valued associative memory model having penalty term
Sub Title (in English)
Keyword(1) complex associative memory
Keyword(2) energy function
Keyword(3) penalty term
Keyword(4) redundancy
Keyword(5) communication
Keyword(6) error correction
Keyword(7) apriori information
1st Author's Name Hiroyuki AOKI
1st Author's Affiliation Department of Electronic Engineering, Tokyo National College of Technology()
2nd Author's Name Yukio KOSUGI
2nd Author's Affiliation Interdisciplinary Graduate School of Science and Engineering, Tokyo Institute of Technology
Date 1997/7/24
Paper # NC97-33
Volume (vol) vol.97
Number (no) 201
Page pp.pp.-
#Pages 8
Date of Issue