Course Description
STAT 535 is a graduate-level statistical optimization course that discusses the topics in modern optimization and high-dimensional statistics, including convex optimization, duality and KKT conditions, debiased Lasso, stability selection, projected and proximal gradient descent methods, low-rank estimation, mixed integer programming, conformal prediction, generative models, etc. Besides maintaining the course logistics throughout the quarter, the TA will
- hold weekly hour office;
- propose homework problems and grade assignments;
- answer questions on the course discussion platform.