On the synthesis of nonlinear continuous neural networks
S.R. Das · IEEE Transactions on Systems Man and Cybernetics · 1991
A synthesis technique for a class of nonlinear neural networks that subsumes J.J. Hopfield's network with graded response (1984) is presented. The network is synthesized by designing its energy function E such that the function has local minima at the prescribed attractor points. Implementation of the synthesis technique as a nonlinear programming problem is also suggested.>