Simple activation functions for neural and fuzzy neural networks

In this paper, we propose some activation functions designed to simplify the computational complexity and the hardware implementation of neural and fuzzy neural networks. These functions approximate the sigmoidal and the Gaussian shapes using only simple arithmetic operations. The simulation results show the good approximation capabilities of the resulting networks.

[1]  B. Mendil,et al.  Hardware oriented fuzzy neural network , 1996, 1996 IEEE International Symposium on Circuits and Systems. Circuits and Systems Connecting the World. ISCAS 96.

[2]  Takayuki Yamada,et al.  Neural network controller using autotuning method for nonlinear functions , 1992, IEEE Trans. Neural Networks.

[3]  Jan M. Rabaey,et al.  Digital Integrated Circuits: A Design Perspective , 1995 .

[4]  S Mangrulkar,et al.  Artificial neural systems. , 1990, ISA transactions.

[5]  Bimal K. Bose,et al.  Expert system, fuzzy logic, and neural network applications in power electronics and motion control , 1994, Proc. IEEE.

[6]  Peter J. Gawthrop,et al.  Neural networks for control systems - A survey , 1992, Autom..

[7]  Bart Kosko,et al.  Neural networks and fuzzy systems: a dynamical systems approach to machine intelligence , 1991 .

[8]  D.R. Hush,et al.  Progress in supervised neural networks , 1993, IEEE Signal Processing Magazine.