ON NEURAL NETWORKS THAT DESIGN NEURAL ASSOCIATIVE MEMORIES

The design problem of generalized brain-state-in-a-box (GBSB) type ass ociative memories is formulated as a constrained optimization program, and ''designer'' neural net works for solving the program in real tim e are proposed, The stability of the designer networks is analyzed usi ng Barbalat's lemma, The analyzed and synthesized neural associative m emories do not require symmetric weight matrices, Two types of the GBS B-based associative memories are analyzed, one when the network trajec tories are constrained to reside in the hypercube [-1,1](n) and the ot her type when the network trajectories are confined to stay in the hyp ercube [0,1](n). Numerical examples and simulations are presented to i llustrate the results obtained.