Abstract
Modular neural networks have the possibility of overcoming common scalability and interference problems experienced by fully connected neural networks when applied to large databases. In this paper we trial an approach to constructing modular ANN’s for a very large problem from CEDAR for the classification of handwritten characters. In our approach, we apply progressive task decomposition methods based upon clustering and regression techniques to find modules. We then test methods for combining the modules into ensembles and compare their structural characteristics and classification performance with that of an ANN having a fully connected topology. The results reveal improvements to classification rates as well as network topologies for this problem.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Auda, G., Kamel, M.: Modular Neural Networks: A Survey. International Journal of NeuralSystems 9, 129–151 (1999)
Boers, E.J.W., Kuipers, H.: Biological Metaphors and the Design of Modular Artificial Neural Networks, Masters Thesis, Leiden Univesity, Netherlands (1992)
Boers, E.J.W., Kuipers, H., Happel, B.L.M., Sprinkhuizen-Kuyper, I.G.: Designing Modular Artificial Neural Networks, Computing Science in the Netherlands. In: Wijshoff, H.A. (ed.) Proc (CSN 1993), Strichting Mathematisch Centrum, Amsterdam, pp. 87–96 (1993)
Collier, P.A., Waugh, S.G.: Characteristics of Data Suitable For Learning With Connectionist and Symbolic Methods, Dept. of Computer Science, University of Tasmania (1994)
Craven, M., Shavlik, J.: Rule extraction: Where do we go to from here?, Working paper 99-1, University of Wisconsin Machine Learning Group (1999)
Fayyad, U.M.: Data Mining and Knowledge Discovery: Making sense out of Data. IEEE Expert Intelligent Systems and Their Applications 11, 20–26 (1996)
Golea, M., Tickle, A., Andrews, R., Diederich, J.: The truth will come to light. IEEE Transactions on Neural Newtorks 9, 1057–1068 (1998)
Happel, B., Murre, J.M.: Design and evolution of modular neural network architectures. Neural Networks 7, 985–1004 (1994)
Jacobs, R.A., Jordan, M.I., Barto, A.G.: Task decomposition through competition in a modular connectionist architecture: The what and where vision tasks. Cognitive Science 15, 219–250 (1989)
Nayak, R.: Intelligent Data Analysis: Issues and Challenges. In: Proc. of the 6th. World Multi Conferences on Systematics, Cybernetics and Informatics, Florida USA, July 14-18 (2002)
Pal, S., Mitra, S.: Rough Fuzzy MLP: Modular evolution, rule generation and evaluation. IEEE Transactions Knowledge and Data Engineering 15, 14–25 (2003)
Quinlan, J.R.: Comparing Connectionist and Symbolic Learning Methods, Dept. of Computer Science, University of Sydney (1993)
Rueckl, J., Cave, K.R., Kosslyn, S.M.: Why are ‘what’ and ‘where’ processed by separate cortical visual systems? a computational investigation. Journal of Cognitive Neuroscience 1, 171–186 (1989)
Schmidt, A.: A modular neural network architecture with additional generalisation abilities for high dimensional input vectors. Masters Thesis, Manchester Metropolitan University (1996)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2004 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Ferguson, B., Ghosh, R., Yearwood, J. (2004). An Experiment in Task Decomposition and Ensembling for a Modular Artificial Neural Network. In: Orchard, B., Yang, C., Ali, M. (eds) Innovations in Applied Artificial Intelligence. IEA/AIE 2004. Lecture Notes in Computer Science(), vol 3029. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-540-24677-0_11
Download citation
DOI: https://doi.org/10.1007/978-3-540-24677-0_11
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-22007-7
Online ISBN: 978-3-540-24677-0
eBook Packages: Springer Book Archive