You are here

MODULAR AND HIERARCHICAL SPECIALIZATION IN NEURAL NETWORKS

Journal Name:

Publication Year:

Abstract (2. Language): 
Modularity and hierarchy are fundamental notions in structured system design. By subdividing a large and unstructured problem into smaller and tractable chunks, design automation becomes possible. In this paper we discuss the use of modularity and hierarchy for functional specialization during the development of neural networks. We study the behavioral differences and requirements for back-propagation training of feed-forward networks. Further we illustrate that a deliberate mix of hierarchically imposed evaluation functions will improve network accuracy and learning speed.
719-726

REFERENCES

References: 

[1] G. Auda and M. Kamel, “Modular neuraNetworks: A survey”, Int. Journal of Neural Systems, Vol. 9, No. 2, (April 1999), pp. 129-151.
l [1] [6] A. Sharkey, “Multi-Net Systems” in “Combining Artificial Neural Nets” Ed. A Sharkey, Springer-Verlag, London, ISBN 1-85233-004-X.,1999.
[2] T. Caelli, L.Guan and W.Wan, “Modularitin Neural Computin”, Proceedings of the IEEE 87, No.9, September 1999, pp. 1497-1518.
y [2] [7] M.H. ter Brugge et al., “On the representation of data for optimal learning”, Proceedings ICNN’95 Perth, pp. 3180-3184, 1995.
[3] E. Barakova, “Learning Reliability: a studon indecisiveness in sample selection” PhD thesis, Rijksuniversiteit Groningen, ISBN 90-367-09873, 1999.
y [3] [8] H. Keegstra et alieni, “Exploiting Network Redundancy for Low-Cost Neural Network Realization”, Digest ICNN’96,Washington, USA, pp. 951-955, 1996.
[4] L. Spaanenburg, W.J. Jansen, J.A.G. Nijhuis“Over Multiple Rule-blocks to Modular Net”, Proceedings Euromicro’97, (Budapest, 1997), pp. 698-705, 1997.
, [4] [9] M. Staley, “Learning with Piece-Wise Linear Networks”, Int. Journal of Neural Systems, Vol. 6, No. 1, pp. 43-59, March 1995.

Thank you for copying data from http://www.arastirmax.com