Learning algorithm for global fault immunization of supervised ANN

Fault immunization techniques are techniques to further enhance fault tolerance of a neural network. The technique of Chun and McNamee [1990] is based on the trial-and-error training which requires a high computational-time. Lursinsap and Tanprasert [1997] proposed a mathematical model to capture the characteristic of the fault immunization. However this model is performed locally to each neuron after training which may deteriorate the target error and increase the computational time. In this paper, we investigate the capability of two random optimization techniques for the the fault immunization improvement. A new cost function which combines the target error function and the immunization function is also proposed