Prerequisites: EAS 520/DSC 520, EAS 501, and EAS 502; or DSC graduate student; or permission of the instructor
Rigorous and systematic introduction to the theory and practice of deep learning. Topics include approximation, generalization, optimization, and the mathematical concepts behind the various kinds of learning such as Supervised (regression and classification), unsupervised (clustering, dimension estimation), semi-supervised, and reinforcement. The course addresses the computational aspects of deep learning such as efficient computation of gradients using backpropagation and batch normalization.