ABSTRACT
One of the most extensively analyzed classes of artificial neural networks is the class of associative networks or associative neural memories. These memory models can be classified in various ways depending on their architecture (static versus recurrent), their retrieval mode (synchronous versus asynchronous), the nature of the stored associations (autoassociative versus heteroassociative), the complexity and capability of the memory storage/recording algorithm, and so on. This section discusses various architectures and recording algorithms for the storage and retrieval of information in neural memories with emphasis on dynamic (recurrent) associative memory (DAM) architectures. The Hopfield model and the bidirectional associative memory are discussed in detail, and criteria for high-performance dynamic memories are outlined for the purpose of comparing the various models.
