Shapes of nonmonotonic activation functions in a chaotic neural network associative memory model and its evaluation

Abstract

AbstractThe purpose of this paper is to investigate the performance of the associative memory model using Aihara's chaotic neural network with different activation functions. Sigmoid function, a monotonic function, was used in Aihara's original model. However, in the static associative memory, it is reported that the storage capacity of the network is improved when a nonmonotonic function is used as the activation function. To improve the associative ability of chaotic neural network, kinds of nonmonotonic functions have been proposed to serve as activation function. This paper investigates their difference as to retrieval ability, and proposes an advanced nonmonotonic function. By computer simulation, we discuss what kind of shape is good regarding improving the associative ability of chaotic neural network. © 2008 Wiley Periodicals, Inc. Electron Comm Jpn, 91(3): 22– 27, 2008; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecj.10070