Bidirectional Associative Memory (BAM) introduced by Bart Kosko in 1988 is a type of recurrent neural network that enables bidirectional data flow for pattern recognition and retrieval. It uses a symmetrical weight matrix and Hebbian learning