Analyzes three main traditions in models of learning: unsupervised (self-organized) learning, supervised learning (learning with a teacher), and reinforcement learning. Architectures studied include adaptive filters, back propagation, competitive learning, self-organizing feature maps, gradient descent procedures, Boltzmann machines, simulated annealing, neocognitron, and gated dipoles. CAS CN 510 and 520 may be taken concurrently.