8 Deep Learning (1989—2016)

When we last discussed perceptrons, Minsky and Papert had just pretty much squelched all research on them with Perceptrons, their volume on the mathematics and, critically, the limitations of this class of models. However, a group in San Diego centered around the mathematical psychologist David Rumelhart retained their interest in perceptron-like mechanisms and, in 1987, together with James McClelland of Stanford, published a two-volume set [86] titled Parallel Distributed Processing: Explorations in the Microstructure of Cognition, which was responsible for a resurgence of interest in the topic.