Prerequisite: STAT 652 or STAT 664, or permission of instructor.
Course Description:Introduces theory and methods for building computationally efficient software agents that reason, act, and learn environments characterized by noisy and uncertain information. Covers methods based on graphical probability and decision models. Students study approaches to representing knowledge about uncertain phenomena, and planning and acting under uncertainty. Topics include knowledge engineering, exact and approximate inference in graphical models, learning in graphical models, temporal reasoning, planning, and decision- making. Practical model building experience is provided. Students apply what they learn to semester-long project of their own choosing.