Forced Formation of a Geometrical Feature Space by a Neural Network Model with Supervised Learning

Toshiaki TAKEDA; Hiroki MIZOE; Koichiro KISHI; Takahide MATSUOKA

Forced Formation of a Geometrical Feature Space by a Neural Network Model with Supervised Learning

Toshiaki TAKEDA, Hiroki MIZOE, Koichiro KISHI, Takahide MATSUOKA

Full Text Views

0

Share
Cite this

Summary :

To investigate necessary conditions for the object recognition by simulations using neural network models is one of ways to acquire suggestions for understanding the neuronal representation of objects in the brain. In the present study, we trained a three layered neural network to form a geometrical feature representation in its output layer using back-propagation algorithm. After training using 73 learning examples, 65 testing patterns made by various combinations of above features could be recognized with the network at a rate of 95.3% appropriate response. We could classify four types of hidden layer units on the basis of effects on the output layer.

Publication: IEICE TRANSACTIONS on Fundamentals Vol.E76-A No.7 pp.1129-1132

Publication Date: 1993/07/25

Publicized

Online ISSN

DOI

Type of Manuscript: Special Section LETTER (Special Section of Letters Selected from the 1993 IEICE Spring Conference)

Category

Cite this

Copy

Toshiaki TAKEDA, Hiroki MIZOE, Koichiro KISHI, Takahide MATSUOKA, "Forced Formation of a Geometrical Feature Space by a Neural Network Model with Supervised Learning" in IEICE TRANSACTIONS on Fundamentals, vol. E76-A, no. 7, pp. 1129-1132, July 1993, doi: .
Abstract: To investigate necessary conditions for the object recognition by simulations using neural network models is one of ways to acquire suggestions for understanding the neuronal representation of objects in the brain. In the present study, we trained a three layered neural network to form a geometrical feature representation in its output layer using back-propagation algorithm. After training using 73 learning examples, 65 testing patterns made by various combinations of above features could be recognized with the network at a rate of 95.3% appropriate response. We could classify four types of hidden layer units on the basis of effects on the output layer.
URL: https://globals.ieice.org/en_transactions/fundamentals/10.1587/e76-a_7_1129/_p

Copy

@ARTICLE{e76-a_7_1129,
author={Toshiaki TAKEDA, Hiroki MIZOE, Koichiro KISHI, Takahide MATSUOKA, },
journal={IEICE TRANSACTIONS on Fundamentals},
title={Forced Formation of a Geometrical Feature Space by a Neural Network Model with Supervised Learning},
year={1993},
volume={E76-A},
number={7},
pages={1129-1132},
abstract={To investigate necessary conditions for the object recognition by simulations using neural network models is one of ways to acquire suggestions for understanding the neuronal representation of objects in the brain. In the present study, we trained a three layered neural network to form a geometrical feature representation in its output layer using back-propagation algorithm. After training using 73 learning examples, 65 testing patterns made by various combinations of above features could be recognized with the network at a rate of 95.3% appropriate response. We could classify four types of hidden layer units on the basis of effects on the output layer.},
keywords={},
doi={},
ISSN={},
month={July},}

Copy

TY - JOUR
TI - Forced Formation of a Geometrical Feature Space by a Neural Network Model with Supervised Learning
T2 - IEICE TRANSACTIONS on Fundamentals
SP - 1129
EP - 1132
AU - Toshiaki TAKEDA
AU - Hiroki MIZOE
AU - Koichiro KISHI
AU - Takahide MATSUOKA
PY - 1993
DO -
JO - IEICE TRANSACTIONS on Fundamentals
SN -
VL - E76-A
IS - 7
JA - IEICE TRANSACTIONS on Fundamentals
Y1 - July 1993
AB - To investigate necessary conditions for the object recognition by simulations using neural network models is one of ways to acquire suggestions for understanding the neuronal representation of objects in the brain. In the present study, we trained a three layered neural network to form a geometrical feature representation in its output layer using back-propagation algorithm. After training using 73 learning examples, 65 testing patterns made by various combinations of above features could be recognized with the network at a rate of 95.3% appropriate response. We could classify four types of hidden layer units on the basis of effects on the output layer.
ER -