A software engineering approach to develop adaptive RBF neural networks

Authors:
A. Talevski;E. Chang;D. H. Wang;T. S. Dillon
Affiliations:
Department of Computer Science and Computer Engineering, La Trobe University, Bundoora, VIC, Australia;School of Information Systems, Cutrin University, WA, Australia;Department of Computer Science and Computer Engineering, La Trobe University, Bundoora, VIC, Australia;Faculty of Information Technology, University of Technology, Broadway, NSW, Australia
Venue:
Design and application of hybrid intelligent systems
Year:
2003

Citing 8
Cited 0

Adaptive plug-and-play components for evolutionary software development

Proceedings of the 13th ACM SIGPLAN conference on Object-oriented programming, systems, languages, and applications
Using adapters to reduce interaction complexity in reusable component-based software development

SSR '99 Proceedings of the 1999 symposium on Software reusability
Neural Networks: A Comprehensive Foundation

Neural Networks: A Comprehensive Foundation
Specification, implementation, and deployment of components

Communications of the ACM
A goal-driven approach to enterprise component identification and specification

Communications of the ACM
Applying model-integrated computing to component middleware and enterprise applications

Communications of the ACM
Theoretical Foundation for Nonlinear Edge-Preserving Regularized Learning Image Restoration

AI '02 Proceedings of the 15th Australian Joint Conference on Artificial Intelligence: Advances in Artificial Intelligence
Effects of moving the center's in an RBF network

IEEE Transactions on Neural Networks

Quantified Score

Hi-index	0.00

Visualization

Abstract

Component-based software engineering is a way of raising the level of abstraction of software development so that software can be built out of existing context-independent components that can be widely reused. Research has shown that component-based software engineering leads to software that is of higher quality, is developed in a shorter time and therefore results in lower cost. However, the lack of a framework for expressing component collaboration makes component-oriented programs more difficult to maintain, expand and widely reuse. This paper demonstrates the use of an adaptive component-based meta model driven framework that eases the integration of heterogeneous components into an application at runtime. By using the proposed framework, this paper introduces a new software engineering approach to the implementation of an adaptive RBF network. The dynamic RBF network is applied to an image restoration task in order to realize the functional mapping from a degraded image space to the original image space, where no prior knowledge and assumptions about the blurring process and the additive noise are required. The proposed RBF network can run in either sequential or parallel modes.