Fault tolerance of feedforward neural nets for classification tasks

A method is proposed to estimate the fault tolerance of feedforward artificial neural nets (ANNs) and to synthesize robust nets. Fault models are presented, and a procedure is developed to build fault tolerant ANNs by replicating the hidden units. Based on this procedure, metrics are derived to quantify the fault tolerance as a function of redundancy. A significant amount of redundancy is shown to be necessary to achieve complete fault tolerance even if only single faults are considered. Furthermore, lower bounds on the required redundancy are analytically derived for some canonical problems. Results indicate that ANNs have good partial fault tolerance and degrade gracefully. A single extra replication is seen to considerably improve fault tolerance.<<ETX>>