Stability properties of neural networks Book Chapter

Fuller, EJ, Yerramalla, SK, Cukic, B. (2006). Stability properties of neural networks . 97-108. 10.1007/0-387-29485-6_5

cited authors

  • Fuller, EJ; Yerramalla, SK; Cukic, B



  • Lyapunov theory is a powerful tool for understanding the stability of neural networks. Once a measure of error is defined for the system, suitable Lyapunov functions can be described which then provide a rigorous characterization of network behavior across all of the possible set of states for the network for which the Lyapunov function is defined. In combination with probabilistic methods such as those of Schumann and Gupta [Schumann 2003], a high degree of reliability may be obtained for systems that integrate neural networks into their structure. © 2006 Springer Science+Business Media, Inc.

publication date

  • December 1, 2006

Digital Object Identifier (DOI)

start page

  • 97

end page

  • 108