COMP 572 - Neural Networks

Units: 3
Three hours of lecture in the lab per week
Prerequisite(s): Admission to the Computer Science or Mathematics Graduate Program
Covers the basic ideas of distributed computation with many simple processing units, similar to the neurons of the brain. Topics include: Hopfield style networks applied to optimization problems, and the backpropagation method applied to pattern classification problems. Additional topics include associate memory, binary vs. analog networks, simulated annealing.