Statistical Learning Theory

Graduate / Professional

This course covers classical topics and recent advances in statistical learning theory. This includes topics such as PAC learning, VC theory, boosting, and online learning. We explore statistical and computational aspects, with an emphasis on developing a rigorous quantitative understanding of key machine learning concepts. A second aim is to introduce technical tools that help with designing learning algorithms and proving learning guarantees. Prerequisites: Mathematical maturity and comfort with proof-oriented courses. Background in probability (e.g., S&DS 241), machine learning (e.g., S&DS 265), and algorithms (e.g., CPSC 365). Familiarity with basic concepts in computational complexity (e.g., NP-hardness) is helpful but not required. T 4pm-5:55pm