Skip to main content

Brainchild: A Fault Tolerant Neural Network

19 June 1989

New Image

Brainchild is a neural network designed to bridge the gap between current neural models and the brain. It models the physical organization of neurons by using both feed forward and lateral connections. It also has a high depree of fault tolerance in keeping with neural connections. A series of tests were run on both Brainchild and a Hopfield model network to compare fault tolerance. Both hard and soft faults were used, as well as combinations of the two. Brainchild proved to be the more fault tolerant of the two.