Random iterated neural networks: Asymptotic behavior

Authors

German Hernandez, University of Memphis
Fernando Nino, University of Memphis
Fernanda Botelho, University of MemphisFollow
Anthony Quas, University of Memphis

Abstract

The random iterated neural networks, a class of stochastic discrete recurrent neural networks, are introduced. These networks implement random discrete dynamical systems that have very special asymptotic properties, compact global attractor and an ergodic measure supported on the attractor.

Publication Title

Intelligent Engineering Systems Through Artificial Neural Networks

Recommended Citation

Hernandez, G., Nino, F., Botelho, F., & Quas, A. (1999). Random iterated neural networks: Asymptotic behavior. Intelligent Engineering Systems Through Artificial Neural Networks, 9, 653-658. Retrieved from https://digitalcommons.memphis.edu/facpubs/5567