Abstract
Feedforward neural network structures have extensively been considered in the literature. In a significant volume of research and development studies hyperbolic tangent type of a neuronal nonlinearity has been utilized. This paper dwells on the widely used neuronal activation functions as well as two new ones composed of sines and cosines, and a sinc function characterizing the firing of a neuron. The viewpoint here is to consider the hidden layer(s) as transforming blocks composed of nonlinear basis functions, which may assume different forms. This paper considers 8 different activation functions which are differentiable and utilizes Levenberg-Marquardt algorithm for parameter tuning purposes. The studies carried out have a guiding quality based on empirical results on several training data sets.
Similar content being viewed by others
References
Kamruzzaman J, Aziz SM (2002) A note on activation function in multilayer feedforward learning. In: Proc. of the 2002 int. joint. conf. on Neural Networks (IJCNN ’02), 12–17 May, vol 1, pp 519–523
Hara K ,Nakayamma K (1994) Comparison of activation functions in multilayer neural network for pattern classification. In: IEEE world congress on computational intelligence, 1994 IEEE in conf. on neural networks, 27 June–2 July, vol 5, pp 2997–3002
Wu Y, Zhao M, Ding X (1997) Beyond weights adaptation: a new neuron model with trainable activation function and its supervised learning. In: Int. conf. on neural networks, 9–12 June, vol 2, pp 1152–1157
Karaköse M, Akın E (2004) Type-2 fuzzy activation function for multilayer feedforward neural networks. In: IEEE int. conf. on systems, man and cybernetics, 10–13 Oct., vol 4, pp 3762–3767
Chiang C-C, Fu H-C (1992) A variant of second-order multilayer perceptron and its application to function approximations. In: Int. joint. conf. on neural networks (IJCNN ’92), 7–11 June, vol 3, pp 887–892
Jang J-SR, Sun C-T, Mizutani E (1997) Neuro-fuzzy and soft computing. PTR Prentice-Hall
Hagan MT, Menhaj MB (1994) Training feedforward networks with the Marquardt algorithm. IEEE Trans Neural Netw NN-5: 989–993
Kretzschmar R, Büeler R, Karayiannis NB, Eggimann F (2000) Quantum neural networks versus conventional feedforward neural networks: an experimental study. In: Proc. of the 2000 IEEE signal processing society workshop on neural networks for signal processing X, Dec. 11–13, vol 1, pp 328–337
Efe MÖ, Abadoğlu E, Kaynak O (1999) A novel analysis and design of a neural network assisted nonlinear controller for a bioreactor. Int J Robust Nonlinear Control 9(11): 799–815
Puskorius GV, Feldkamp LA (1990) Neurocontrol of nonlinear dynamical systems with Kalman filter trained recurrent networks. IEEE Trans Neural Netw 5(2): 279–297
Ungar LH (1997) A bioreactor benchmark for adaptive-network based process control. In: Miller WT III, Sutton RS, Werbos PJ (eds) Neural networks for control. MIT Press, pp 387–402
Narendra KS, Parthasarathy K (1990) Identification and control of dynamical systems using neural networks. IEEE Trans Neural Netw 1: 4–27
Efe MÖ, Debiasi M, Yan P, Özbay H, Samimy M (2008) Neural network based generalization and modeling of subsonic cavity flow regimes. Int J Syst Sci 39(2): 105–117
Samimy M, Debiasi M, Caraballo E, Malone J, Little J, Özbay H, Efe MÖ, Yan P, Yuan X, DeBonis J, Myatt JH, Camphouse RC (2004) Exploring strategies for closed-loop cavity flow control. In: 42nd AIAA Aerospace Sciences meeting and exhibit, January 5–8, Reno, Nevada, U.S.A., 2004 (Paper # AIAA 2004-0576)
Debiasi M, Samimy M (2004) Logic-based active control of subsonic cavity-flow resonance. AIAA J 42: 1901–1909
Asuncion A, Newman DJ (2007) UCI machine learning Repository [http://www.ics.uci.edu/~mlearn/MLRepository.html]. University of California, School of Information and Computer Science, Irvine, CA
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Efe, M.Ö. Novel Neuronal Activation Functions for Feedforward Neural Networks. Neural Process Lett 28, 63–79 (2008). https://doi.org/10.1007/s11063-008-9082-0
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11063-008-9082-0