A BIDIRECTIONAL ASSOCIATIVE MEMORY-BASED ON OPTIMAL LINEAR ASSOCIATIVE MEMORY

A new bidirectional associative memory is presented. Unlike many exist ing BAM algorithms, the presented BAM uses an optimal associative memo ry matrix in place of the standard Hebbian or quasi-correlation matrix . The optimal associative memory matrix is determined by using only si mple correlation learning, requiring no pseudoinverse calculation. Gua ranteed recall of all training pairs is ensured by the present BAM. Th e designs of a linear DAM (LBAM) and a nonlinear BAM (NBAM) are given, and the stability and other performances of the BAMs are analyzed, Th e introduction of a nonlinear characteristic enhances considerably the ability of the BAM to suppress the noises occurring in the output pat tern, and reduces largely the spurious memories, and therefore improve s greatly the recall performance of the DAM. Due to the nonsymmetry of the connection matrix of the network, the capacities of the present B AMs are far higher than that of the existing BAMs. Excellent performan ces of the present BAMs are shown by simulation results.