The Concept and Properties of Sigma-if Neural Network
Our recent works on artificial neural networks point to the possibility of extending the activation function of a standard artificial neuron model using the conditional signal accumulation technique, thus significantly enhancing the capabilities of neural networks. We present a new artificial neuron model, called Sigma-if, with the ability to dynamically tune the size of the decision space under consideration, resulting from a novel activation function. The paper discusses construction of the proposed neuron as well as training Sigma-if feedforward neural networks for well known sample classification problems.
KeywordsHide Layer Neural Information Processing System Decision Space Classic Neural Network Classic Artificial Neural Network
Unable to display preview. Download preview PDF.
- Duch, W., Jankowski, N. (1999) Survey of neural transfer functions. Neural Computing Surveys 2: 163–212Google Scholar
- Anthony, M., Bartlett, P.L. (1999) Neural networks learning: Theoretical foundations, 1st ed. Cambridge University Press, CambridgeGoogle Scholar
- Murray, A.F., Edwards, P.J. (1992) Synaptic weight noise during MLP learning enhances fault-tolerance, generalisation and learning trajectory. Advances in Neural Information Processing Systems 5: 491–498Google Scholar
- Durbin, R., Rumelhart, D. (1989) Product units: A computationally powerful and biologically plausible extension to backpropagation networks. Neural Computation 1: 133–142Google Scholar
- Duch, W., Jankowski, N. (2001) Transfer functions: hidden possibilities for better neural networks. In Verleysen, M. (ed.) 9th European Symposium on Artificial Neural Networks. D-Facto, Brugge, pp. 81–94Google Scholar
- Huk, M. (2004) The Sigma-if neural network as a method of dynamic selection of decision subspaces for medical reasoning systems. Journal of Medical Informatics & Technologies 7: 65–73Google Scholar
- Banarer, V., Perwass, C, Sommer, G. (2003) The hypersphere neuron. In: Verleysen, M. (ed.) 11th European Symposium on Artificial Neural Networks. D-Side, Brugge, pp. 469–474Google Scholar
- Huk, M. (2003) Determining the relevancy of attributes in classification tasks through nondestructive elimination of interneuronal connections in a neural network (in Polish). Pozyskiwanie Wiedzy z Baz Danych 975: 138–147Google Scholar
- Fahlman, S.E., Lebiere, C. (1990) The Cascade-Correlation learning architecture. Advances in Neural Information Processing Systems 2: 524–532Google Scholar
- Kaufman, L., (1999) Solving the quadratic pro-gramming problem arising in support vector classification. In: Scholpkopf, B., Burges, C.J.C., Smola, A.J. (eds.) Advances in Kernel Methods-Support Vector Learning. MIT Press, Boston, pp. 146–167Google Scholar
- Craven, M.W., Shavlik, J.W. (1996) Extracting tree-structured representations of trained networks. Advances in Neural Information Processing Systems 8: 24–30Google Scholar