Objectives

1) To improve and test the efficiency of Backpropagation- and Hebbian- based learning algorithms after designing evolving neural learning algorithms with use of a supervising controller.

2) To design circuits and layout to allow for future TSMC 0.35µm chip- implementation of a dense and reconfigurable design, using analog and mixed-signal cell circuits to perform the necessary propagation and learning tasks.

3) To create a prototype demonstration of the routing design using inexpensive and readily available discrete components.