Modern Optimization for Statistical Learning

Catalog description

Introduce several modern optimization algorithms useful in statistics and machine learning problems with applications in biostatistics and healthcare from a computational perspective. As most statistics and machine learning problems can be formulated as optimization problems, it is important for students to have a powerful toolbox of optimization algorithms. The course will also demonstrate the algorithms' applications in different large-scale biological and healthcare problems. After taking the course, students are expected to acquire reasonable working skills to apply different algorithms to solve optimization problems practically and model different estimation/inference problems as optimization problems. Students are expected to have a reasonable working knowledge of probability and linear algebra. Taking a programming/computing in the past is helpful but not required. Pre-requisites: Permission only for MB students. Non-program students must obtain permission from the director of graduate studies or the instructor. Instructor: TBA. Credits 3