Course: STAT 37798
Title: Topics in Machine Learning: Scientific Computing Tools for High-Dimensional Problems
Instructor(s): Yuehaw Khoo
Teaching Assistant(s): TBA
Class Schedule: Sec 1: TR 11:00 AM-12:20 PM in Eckhart 207A
Textbook(s): None
Description: This course considers computational techniques in representing and optimizing a high-dimensional probability distribution in the context of computational physics and data science applications. Computationally tractable approximations to the probability distribution based on convex relaxations, tensor-network, neural-network, mean-field, Bethe approximation, and normalizing flow will be discussed. The course aims to present a practical view on these methods and illustrate when they can be successfully deployed in specific situations.
