Abstract
A particular group of neural network (NN) learning algorithms known as constructive algorithms (CoNN) congregates algorithms that dynamically combine two processes: (1) the definition of the NN architecture and (2) learning. Generally both processes alternate, depending on each others′ performance. During training CoNN algorithms incrementally add hidden neurons and connections to the network until some stopping criterion is satisfied. This paper describes an investigation into the semantic role played by the hidden neurons added into the NN, when learning Boolean functions. Five CoNN algorithms namely Tower, Pyramid, Tiling, Perceptron-Cascade and Shift are examined in that respect. Results show that hidden neurons represent Boolean sub-expressions whose combination represents a disjunction of prime implicants.
This is a preview of subscription content, log in via an institution to check access.
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
Gallant, S.I.: Neural Network Learning & Expert Systems. The MIT Press, Cambridge (1994)
do Carmo Nicoletti, M., Bertini Jr., J.R., Elizondo, D., Franco, L., Jerez, J.M.: Constructive neural network algorithms for feedforward architectures suitable for classification tasks. In: Franco, L., Elizondo, D.A., Jerez, J.M. (eds.) Constructive Neural Networks. SCI, vol. 258, pp. 1–23. Springer, Heidelberg (2009)
Bishop, C.M.: Neural networks for pattern recognition. Oxford University Press, Great Britain (1999)
Franco, L., Elizondo, D.A., Jérez, J.M. (eds.): Constructive Neural Networks. SCI, vol. 258. Springer, Germany (2009)
Setiono, R., Leow, W.K.: FERNN: an algorithm for fast extraction of rules from neural networks. Appl. Intell. 1, 15–25 (2000)
Garcez, A.S.A., Broda, K., Gavia, D.M.: Symbolic knowledge extraction from trained neural network: a sound approach. Artif. Intell. 125, 155–207 (2001)
Fu, L.: Rule generation from neural networks. IEEE Trans. Syst. Man Cybern. 24(8), 1114–1124 (1994)
Towell, G.G., Shavlik, J.W.: Extracting refined rules from knowledge-based neural networks. Mach. Learn. 13, 71–101 (1993)
Craven, M.W., Shavlik, J.W.: Learning symbolic rules using artificial neural networks. In: Proc. of the 10th ICML, pp. 73–80. Morgan Kaufmann, San Mateo (1993)
Andrews, R., Diederich, J., Tickle, A.B.: Survey and critique of techniques for extracting rules from trained artificial neural networks. Knowl. Base. Syst. 8(6), 373–389 (1995)
Pagallo, G., Haussler, D.: Boolean feature discovery in empirical learning. Mach. Learn. 5, 71–99 (1990)
Mézard, M., Nadal, J.: Learning feedforward networks: the tiling algorithm. J. Phys. A: Math. Gen. 22, 2191–2203 (1989)
Burgess, N.: A constructive algorithm that converges for real-valued input patterns. Int. J. Neural Syst. 5(1), 59–66 (1994)
Amaldi, E., Guenin, B.: Two constructive methods for designing compact feedfoward networks of threshold units. Int. J. Neural Syst. 8(5&6), 629–645 (1997)
Fahlman, S., Lebiere, C.: The cascade correction architecture. In: Advances in Neural Information Processing Systems, vol. 2, pp. 524–532. Morgan Kaufmann, San Mateo (1990)
Frean, M.: The upstart algorithm: a method for constructing and training feedforward neural networks. Neural Comput. 2, 198–209 (1990)
Berztiss, A.T.: Data structures – theory and practice, 2nd edn. Academic Press, N.Y (1975)
Quine, W.V.: The problem of simplifying truth tables. Amer. Math. Month. 59, 51–531 (1952)
McCluskey, E.J.: Minimization of Boolean functions. Bell System Tech. J. 35, 1417–1444 (1956)
Bertini Jr., J.R., Nicoletti, M.C.: A constructive neural network algorithm based on the geometric concept of barycenter of convext hull. In: Rutkoski, R.L., Tadeusiewicz, R., Zadeh, L.A., Zurada, J. (eds.) Computational Intelligence: Methods and Applications, pp. 1–12. Academic Publishing House Exit, Poland (2008)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2011 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
do Carmo Nicoletti, M., Bertini, J.R., de Oliveira, O.L. (2011). Interpreting Hidden Neurons in Boolean Constructive Neural Networks. In: Yin, H., Wang, W., Rayward-Smith, V. (eds) Intelligent Data Engineering and Automated Learning - IDEAL 2011. IDEAL 2011. Lecture Notes in Computer Science, vol 6936. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-23878-9_5
Download citation
DOI: https://doi.org/10.1007/978-3-642-23878-9_5
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-23877-2
Online ISBN: 978-3-642-23878-9
eBook Packages: Computer ScienceComputer Science (R0)