In feedforward neural networks the sign passes in only one course.[111] Recurrent neural networks feed the output sign again to the enter, which permits limited-time period Recollections of prior enter events. lengthy temporary memory is the most successful network architecture for recurrent networks.[112] Tanya Berger-Wolf is a professor of C