Discrete-time cellular neural networks for associative memories: a new design method via iterative learning and forgetting algorithms

In this paper a synthesis procedure of discrete-time Cellular Neural Networks (DTCNN's) for associative memories with iterative learning and forgetting algorithms is developed, by which each pattern to be stored is learnt one at a time and each pattern to be forgotten is deleted one at a time. The proposed approach exploits the properties of pseudo inverse matrices and preserves the local connection feature of DTCNN's.