A Combined Network Architecture Using Art2 and Back Propagation for Adaptive Estimation of Dynamic Processes
A neural network architecture called ART2/BP is proposed. Thc goal has been to construct an artificial neural network that learns incrementally an unknown mapping, and is motivated by the instability found in back propagation (BP) networks: after first learning pattern A and then pattern B, a BP net...
Main Author: | |
---|---|
Format: | Article |
Language: | English |
Published: |
Norwegian Society of Automatic Control
1990-10-01
|
Series: | Modeling, Identification and Control |
Subjects: | |
Online Access: | http://www.mic-journal.no/PDF/1990/MIC-1990-4-2.pdf |
id |
doaj-528c3dd2d35743d38ed73dabe99baad5 |
---|---|
record_format |
Article |
spelling |
doaj-528c3dd2d35743d38ed73dabe99baad52020-11-24T22:52:53ZengNorwegian Society of Automatic ControlModeling, Identification and Control0332-73531890-13281990-10-0111419119910.4173/mic.1990.4.2A Combined Network Architecture Using Art2 and Back Propagation for Adaptive Estimation of Dynamic ProcessesEinar SørheimA neural network architecture called ART2/BP is proposed. Thc goal has been to construct an artificial neural network that learns incrementally an unknown mapping, and is motivated by the instability found in back propagation (BP) networks: after first learning pattern A and then pattern B, a BP network often has completely 'forgotten' pattern A. A network using both supervised and unsupervised training is proposed, consisting of a combination of ART2 and BP. ART2 is used to build and focus a supervised backpropagation network consisting of many small subnetworks each specialized on a particular domain of the input space. The ART2/BP network has the advantage of being able to dynamically expand itself in response to input patterns containing new information. Simulation results show that the ART2/BP network outperforms a classical maximum likelihood method for the estimation of a discrete dynamic and nonlinear transfer function. http://www.mic-journal.no/PDF/1990/MIC-1990-4-2.pdfSystem identificationnonlinear systemsadaptive controlartificial neural networksback propagation |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Einar Sørheim |
spellingShingle |
Einar Sørheim A Combined Network Architecture Using Art2 and Back Propagation for Adaptive Estimation of Dynamic Processes Modeling, Identification and Control System identification nonlinear systems adaptive control artificial neural networks back propagation |
author_facet |
Einar Sørheim |
author_sort |
Einar Sørheim |
title |
A Combined Network Architecture Using Art2 and Back Propagation for Adaptive Estimation of Dynamic Processes |
title_short |
A Combined Network Architecture Using Art2 and Back Propagation for Adaptive Estimation of Dynamic Processes |
title_full |
A Combined Network Architecture Using Art2 and Back Propagation for Adaptive Estimation of Dynamic Processes |
title_fullStr |
A Combined Network Architecture Using Art2 and Back Propagation for Adaptive Estimation of Dynamic Processes |
title_full_unstemmed |
A Combined Network Architecture Using Art2 and Back Propagation for Adaptive Estimation of Dynamic Processes |
title_sort |
combined network architecture using art2 and back propagation for adaptive estimation of dynamic processes |
publisher |
Norwegian Society of Automatic Control |
series |
Modeling, Identification and Control |
issn |
0332-7353 1890-1328 |
publishDate |
1990-10-01 |
description |
A neural network architecture called ART2/BP is proposed. Thc goal has been to construct an artificial neural network that learns incrementally an unknown mapping, and is motivated by the instability found in back propagation (BP) networks: after first learning pattern A and then pattern B, a BP network often has completely 'forgotten' pattern A. A network using both supervised and unsupervised training is proposed, consisting of a combination of ART2 and BP. ART2 is used to build and focus a supervised backpropagation network consisting of many small subnetworks each specialized on a particular domain of the input space. The ART2/BP network has the advantage of being able to dynamically expand itself in response to input patterns containing new information. Simulation results show that the ART2/BP network outperforms a classical maximum likelihood method for the estimation of a discrete dynamic and nonlinear transfer function. |
topic |
System identification nonlinear systems adaptive control artificial neural networks back propagation |
url |
http://www.mic-journal.no/PDF/1990/MIC-1990-4-2.pdf |
work_keys_str_mv |
AT einarsørheim acombinednetworkarchitectureusingart2andbackpropagationforadaptiveestimationofdynamicprocesses AT einarsørheim combinednetworkarchitectureusingart2andbackpropagationforadaptiveestimationofdynamicprocesses |
_version_ |
1725664180527169536 |