Optimization for Machine Learning
Formerly: ELEC 503
The steepest-descent, Newton, conjugate, and quasi-Newton algorithms for unconstrained optimization. Inexact line-search techniques. Application of optimization methods to classification, logistic regression, and support vector machines for signal processing and machine intelligence involving audio, image, video and other types of data. Introduction to constrained optimization. Students are required to complete a project.
Graduate course in the Electrical and Computer Engineering program offered by the Faculty of Graduate Studies.