An Associative Memory Neural Network to Recall Nearest Pattern from Input

Isao YAMADA; Satoshi IINO; Kohichi SAKANIWA

An Associative Memory Neural Network to Recall Nearest Pattern from Input

Isao YAMADA, Satoshi IINO, Kohichi SAKANIWA

Full Text Views

0

Cite this

Summary :

This paper proposes an associative memory neural network whose limiting state is the nearest point in a polyhedron from a given input. Two implementations of the proposed associative memory network are presented based on Dykstra's algorithm and a fixed point theorem for nonexpansive mappings. By these implementations, the set of all correctable errors by the network is characterized as a dual cone of the polyhedron at each pattern to be memorized, which leads to a simple amplifying technique to improve the error correction capability. It is shown by numerical examples that the proposed associative memory realizes much better error correction performance than the conventional one based on POCS at the expense of the increase of necessary number of iterations in the recalling stage.

Publication: IEICE TRANSACTIONS on Fundamentals Vol.E82-A No.12 pp.2811-2817

Publication Date: 1999/12/25

Publicized

Online ISSN

DOI

Type of Manuscript: PAPER

Category: Neural Networks

Cite this

Copy

Isao YAMADA, Satoshi IINO, Kohichi SAKANIWA, "An Associative Memory Neural Network to Recall Nearest Pattern from Input" in IEICE TRANSACTIONS on Fundamentals, vol. E82-A, no. 12, pp. 2811-2817, December 1999, doi: .
Abstract: This paper proposes an associative memory neural network whose limiting state is the nearest point in a polyhedron from a given input. Two implementations of the proposed associative memory network are presented based on Dykstra's algorithm and a fixed point theorem for nonexpansive mappings. By these implementations, the set of all correctable errors by the network is characterized as a dual cone of the polyhedron at each pattern to be memorized, which leads to a simple amplifying technique to improve the error correction capability. It is shown by numerical examples that the proposed associative memory realizes much better error correction performance than the conventional one based on POCS at the expense of the increase of necessary number of iterations in the recalling stage.
URL: https://global.ieice.org/en_transactions/fundamentals/10.1587/e82-a_12_2811/_p

Copy

@ARTICLE{e82-a_12_2811,
author={Isao YAMADA, Satoshi IINO, Kohichi SAKANIWA, },
journal={IEICE TRANSACTIONS on Fundamentals},
title={An Associative Memory Neural Network to Recall Nearest Pattern from Input},
year={1999},
volume={E82-A},
number={12},
pages={2811-2817},
abstract={This paper proposes an associative memory neural network whose limiting state is the nearest point in a polyhedron from a given input. Two implementations of the proposed associative memory network are presented based on Dykstra's algorithm and a fixed point theorem for nonexpansive mappings. By these implementations, the set of all correctable errors by the network is characterized as a dual cone of the polyhedron at each pattern to be memorized, which leads to a simple amplifying technique to improve the error correction capability. It is shown by numerical examples that the proposed associative memory realizes much better error correction performance than the conventional one based on POCS at the expense of the increase of necessary number of iterations in the recalling stage.},
keywords={},
doi={},
ISSN={},
month={December},}

Copy

TY - JOUR
TI - An Associative Memory Neural Network to Recall Nearest Pattern from Input
T2 - IEICE TRANSACTIONS on Fundamentals
SP - 2811
EP - 2817
AU - Isao YAMADA
AU - Satoshi IINO
AU - Kohichi SAKANIWA
PY - 1999
DO -
JO - IEICE TRANSACTIONS on Fundamentals
SN -
VL - E82-A
IS - 12
JA - IEICE TRANSACTIONS on Fundamentals
Y1 - December 1999
AB - This paper proposes an associative memory neural network whose limiting state is the nearest point in a polyhedron from a given input. Two implementations of the proposed associative memory network are presented based on Dykstra's algorithm and a fixed point theorem for nonexpansive mappings. By these implementations, the set of all correctable errors by the network is characterized as a dual cone of the polyhedron at each pattern to be memorized, which leads to a simple amplifying technique to improve the error correction capability. It is shown by numerical examples that the proposed associative memory realizes much better error correction performance than the conventional one based on POCS at the expense of the increase of necessary number of iterations in the recalling stage.
ER -