Time: Monday, Wednesday
4:005:20pm
Location: Sennott Square, Room 5313
Instructor: Milos
Hauskrecht
Computer Science Department
5329 Sennott Square
phone: x48845
email: milos@cs.pitt.edu
office hours: by appointment
Course description
Lectures
Readings
Paper presentations
Projects
The Advanced Topics in AI is a seminar course and this year it will focus on probabilistic graphical models. Probabilistic graphical models are used to model stochasticity (uncertainty) in the world and are very popular in AI and machine learning. The course will cover two classes of graphical models: Bayesian belief networks (also called directed graphical models) and Markov Random Fields (undirected models). After introducing the two frameworks the course will focus on recent advances in inferences and learning with graphical models, including topics such as loopy belief propagation, variational approximations, conditional Markov random fields and others.
Prerequisites
CS 2710 Foundations of AI, CS 2750 Machine Learning , or the permission of the instructor.
Tentative topics to be covered:
Other books related to the topic:
Lectures  Topic(s)  

August 29  Administrivia and Course Overview.
Readings:
 
August 31  Representing uncertainty with probabilities. BBNs
Readings:
 
September 7  Probabilistic
graphical models. Intro to BBN inference. Markov random fields.
Readings:
 
September 12  BBN
representation (David Essary)
Readings:
 
September 14  BBN
inference. Complexities. Variable elimination algorithm. (Amruta Parundare)
Readings:
 
September 19  Variable
elimination in undirected graphs. Treewidth. (Milos) Complexities. Conditioning. ( Collin Lynch) Readings:
 
September 21  Clique
tree. Message passing framework. (Tomas Singliar)
Readings:
 
September 26  Monte Carlo
approximation I. (presented by Hua Ai)
Readings:
 
September 28  Monte Carlo
approximation II. (presented by Chang Liu)
Readings:
 
October 3 
Readings:
 
October 5  Local
probability models (presented by Jiang Zheng)
Readings:
 
October 10 
Readings:
 
October 12 
Readings:
 
October 17 
Readings:
 
October 19 
Readings:
 
October 24 
Readings:
 
October 26 
Readings:
 
October 31 
Readings:
 
November 2 
Readings:
 
November 8 
Readings:  
November 14 
Readings:
 
November 16 
Readings:
 
November 21 
Readings:
 
November 28  No class  
November 30 
Readings:
 
December 5 
Readings:

Readings will be assigned before the class at which the discussion on the topic covered by the paper takes place. Most of the readings will be electronic, however, some readings will be in the paper form or from the books.
Course material:
A Brief Introduction to Graphical Models and Bayesian Networks by Kevin Murphy
Each student is expected to lead discussions on 1 or 2 topics
and related papers during the semester.
There are no homeworks and exams in this course. However, students will be asked to prepare, submit and present two projects. The first project will be due in the middle of the semester. The topic of the final project (due at the end of the semester) will be chosen by a student from the list of topics. The final project report should take a form and read like an AI conference submission.
All the work in this course should be done independently. Collaborations on projects are not permitted. Cheating and any other antiintellectual behavior, including giving your work to someone else, will be dealt with severely. If you feel you may have violated the rules speak to us as soon as possible.
Please make sure you read, understand and abide by the Academic Integrity Code for the Faculty and College of Arts and Sciences.
Students With Disabilities
If you have a disability for which you are or may be requesting an accommodation, you are encouraged to contact both your instructor and Disability Resources and Services, 216 William Pitt Union, (412) 6487890/(412) 3837355 (TTY), as early as possible in the term. DRS will verify your disability and determine reasonable accomodations for this course.