SeMi-Supervised Adaptive Resonance Theory (SMART2)

Christopher J. Merz; William E. Bond; Daniel C. St. Clair

doi:10.1109/IJCNN.1992.227046

SeMi-Supervised Adaptive Resonance Theory (SMART2)

Christopher J. Merz
, William E. Bond
, Daniel C. St. Clair

Missouri University of Science and Technology

Missouri University of Science and Technology

Research output: Contribution to journal › Article › peer-review

Abstract

Adaptive resonance theory (ART) algorithms represent a class of neural network architectures which self-organize stable recognition categories in response to arbitrary sequences of input patterns. The authors discuss incorporation of supervision into one of these architectures, ART2. Results of numerical experiments indicate that this new semi-supervised version of ART2 (SMART2) outperformed ART for classification problems. The results and analysis of runs on several data sets by SMART2, ART2, and backpropagation are analyzed. The test accuracy of SMART2 was similar to that of backpropagation. However, SMART2 network structures are easier to interpret than the corresponding structures produced by backpropagation.

Original language	American English
Journal	Proceedings of the International Joint Conference on Neural Networks, 1992
DOIs	https://doi.org/10.1109/IJCNN.1992.227046
State	Published - Jan 1 1992

Keywords

ART2
Adaptive Resonance Theory
Classification Problems
Neural Nets
Neural Network Architectures
Pattern Recognition
SMART2
Self Organizing
Stable Recognition Categories
Unsupervised Learning

Disciplines

Computer Sciences

Access to Document

10.1109/IJCNN.1992.227046License: Unspecified

SeMi-Supervised Adaptive Resonance Theory (SMART2)Final published version, 527 KBLicense: Unspecified

Cite this

@article{622ef4f2a7324c90965cdbc64e6fc4bd,

title = "SeMi-Supervised Adaptive Resonance Theory (SMART2)",

abstract = " Adaptive resonance theory (ART) algorithms represent a class of neural network architectures which self-organize stable recognition categories in response to arbitrary sequences of input patterns. The authors discuss incorporation of supervision into one of these architectures, ART2. Results of numerical experiments indicate that this new semi-supervised version of ART2 (SMART2) outperformed ART for classification problems. The results and analysis of runs on several data sets by SMART2, ART2, and backpropagation are analyzed. The test accuracy of SMART2 was similar to that of backpropagation. However, SMART2 network structures are easier to interpret than the corresponding structures produced by backpropagation.",

keywords = "ART2, Adaptive Resonance Theory, Classification Problems, Neural Nets, Neural Network Architectures, Pattern Recognition, SMART2, Self Organizing, Stable Recognition Categories, Unsupervised Learning",

author = "Merz, \{Christopher J.\} and Bond, \{William E.\} and \{St. Clair\}, \{Daniel C.\}",

year = "1992",

month = jan,

day = "1",

doi = "10.1109/IJCNN.1992.227046",

language = "American English",

journal = "Proceedings of the International Joint Conference on Neural Networks, 1992",

}

TY - JOUR

T1 - SeMi-Supervised Adaptive Resonance Theory (SMART2)

AU - Merz, Christopher J.

AU - Bond, William E.

AU - St. Clair, Daniel C.

PY - 1992/1/1

Y1 - 1992/1/1

N2 - Adaptive resonance theory (ART) algorithms represent a class of neural network architectures which self-organize stable recognition categories in response to arbitrary sequences of input patterns. The authors discuss incorporation of supervision into one of these architectures, ART2. Results of numerical experiments indicate that this new semi-supervised version of ART2 (SMART2) outperformed ART for classification problems. The results and analysis of runs on several data sets by SMART2, ART2, and backpropagation are analyzed. The test accuracy of SMART2 was similar to that of backpropagation. However, SMART2 network structures are easier to interpret than the corresponding structures produced by backpropagation.

AB - Adaptive resonance theory (ART) algorithms represent a class of neural network architectures which self-organize stable recognition categories in response to arbitrary sequences of input patterns. The authors discuss incorporation of supervision into one of these architectures, ART2. Results of numerical experiments indicate that this new semi-supervised version of ART2 (SMART2) outperformed ART for classification problems. The results and analysis of runs on several data sets by SMART2, ART2, and backpropagation are analyzed. The test accuracy of SMART2 was similar to that of backpropagation. However, SMART2 network structures are easier to interpret than the corresponding structures produced by backpropagation.

KW - ART2

KW - Adaptive Resonance Theory

KW - Classification Problems

KW - Neural Nets

KW - Neural Network Architectures

KW - Pattern Recognition

KW - SMART2

KW - Self Organizing

KW - Stable Recognition Categories

KW - Unsupervised Learning

UR - https://scholarsmine.mst.edu/comsci_facwork/277

U2 - 10.1109/IJCNN.1992.227046

DO - 10.1109/IJCNN.1992.227046

M3 - Article

JO - Proceedings of the International Joint Conference on Neural Networks, 1992

JF - Proceedings of the International Joint Conference on Neural Networks, 1992

ER -

SeMi-Supervised Adaptive Resonance Theory (SMART2)

Abstract

Keywords

Disciplines

Access to Document

Other files and links

Fingerprint

Cite this