Course Information (Statistical Methods in AI and ML, CS 6347)

Basic Information

• Semester: Spring 2021

• Where: Remote Learning/Online

• When: Tuesday and Thursday, 7 to 8:15 p.m.

• Instructor Office hours: Online: Tuesday: 5:45 p.m. to 6:45 p.m.

• TA: TBD

• TA Office: TBD

• TA Office hours: TBD

• Class Discussions: We will have a discussion board for the class on Piazza. This will be the main on-line forum for discussing assignments and course material, and interacting with other students, TA and me. We will also post course-wide announcements on Piazza.

Course Overview

Probabilistic Graphical models (PGMs) such as Bayesian and Markov networks have revolutionized the fields of AI and machine learning. They are used in a large number of application domains such as natural language processing, computer vision, sensor networks, computational biology, robotics, transportation science, medical diagnosis, data mining, and music parsing. Their main virtue is their ability to compactly represent highly uncertain, complex, structured, real-world domains and the availability of efficient manipulation algorithms.

This graduate level course will provide you with an in depth knowledge of probabilistic graphical models, enabling you to apply it to complex real-world problems. It will also provide you with the necessary background for pursuing research in them. In particular, the course will cover the following three aspects:

• Representation: How to encode real-world knowledge into a PGM?

• Inference (both exact and approximate): given a PGM and a query, how to compute an answer to the query?

• Learning: how to learn a PGM from data?

Textbooks

Required

• Reasoning with Probabilistic and Deterministic Graphical Models, 2nd Edition Exact Algorithms, by Rina Dechter.

• Modeling and Reasoning with Bayesian networks, by Adnan Darwiche.

• Probabilistic Graphical Models: Principles and Techniques, by Daphne Koller and Nir Friedman.

Optional

• Machine Learning: a Probabilistic Perspective by Kevin Murphy.

Grading

• We will have 4-5 mini-projects/problem sets (about 60% of the grade)

• We will have 1 Project (about 20% of the grade)

• A take-home final exam (20% of the grade)

• A (92 or above), A- (87 to 92), B+ (83 to 87), B (79 to 83), B- (75 to 79), Fail (75 and below)

• This class involves some work but we will have fun!