CS179: Introduction to Graphical Models

Assignments and Exams:

HW2 10/13/15Soln 
HW3 10/20/15Soln 
HW4 10/27/15Soln 
HW7 12/09/15Soln 

Lecture: M/W/F 3pm-3:50pm, SSL 290

Weekly review session: W 6-6:50pm, DBH1600

Instructor: Prof. Alex Ihler (ihler@ics.uci.edu), Office Bren Hall 4066

  • Office Hours: Fri 10:30-12:00pm, Bren Hall 4066, or by appointment

Teaching Assistant: Qi Lou

  • Office Hours: Mon 10:30-12:00pm in Bren Hall 3013; or by appointment (Bren Hall 4051)

Reader: Shu Kong

Introduction to graphical models

Graphical models are a powerful framework for efficiently representing and reasoning about large systems. Common examples include solving constraint satisfaction and constraint optimization problems, learning and reasoning about high-dimensional, multivariate probability models, and structured prediction tasks. Graphical models are used in a broad spectrum of scientific fields, including computer vision, natural language processing, computational biology, communications and information theory, physics, and more. In this course we will study the common representations of graphical models, including Bayesian networks and Markov random fields; methods for and the complexity of exact reasoning in these models; computationally efficient algorithms for approximate reasoning; and techniques for learning models from data.


We will assume basic familiarity with the concepts of probability. Some programming will be required; we will primarily use Python.

Textbook and Reading

There is no required textbook for the class. However, some useful books on the subject for supplementary reading, on hold at the science library, include:

and available online,

Discussion on Piazza

I prefer that students ask questions within a discussion forum so that everyone (staff and other students) can see the questions, participate in answering them, and learn from each other. I also expect students to read the posts regularly to keep current on the class issues. Please avoid duplicate posts: before posting, first check whether another student has already posted on that topic. When you do post a question, choose a clear and descriptive title. I currently prefer to use Piazza to manage student discussions and questions. Our class link is: http://piazza.com/uci/fall2015/cs179/home.

Collaboration Policy

I encourage you to collaborate in the sense of discussing the course material and ideas with your fellow students. However, it is an important part of the learning process to do your work yourself. To this end, do not show your homework solutions or code to your fellow students, or examine another students solutions or code. Similarly, do not post solution code on Piazza (unless privately to myself and the TA). An excellent guide is the ICS-33 academic integrity handout.

Syllabus (subject to change)

W0: 09/25pdfIntroduction
W1: 09/28pdfCSPs, Local SearchR&N CSPs (2nd Ed Ch5), pyGM CSP Example, Dechter Ch1-2
W1: 09/30pdfLocal SearchpyGM Local Search Example
W1: 10/02pdfBacktracking Search
W2: 10/05pdfVariable Elimination
W2: 10/07pdfProbability
W2: 10/09  
W3: 10/12pdfBayesian Networks
W3: 10/14  
W3: 10/16pdfUndirected graphical models
W4: 10/19pdfMarkov chains
W4: 10/21  python HMM example
W4: 10/23 Hidden Markov modelsR&N Ch15, P&M 6.5
W5: 10/26 Variable elimination (2)python Bayes Net & variable elimination example
W5: 10/28pdfLearning from dataR&N 20.2
W5: 10/30  python maximum likelihood example
W6: 11/02 Learning in undirected models
W6: 11/04 Learning structure in Bayesian networks
W6: 11/06pdfMonte Carlo methods IR&N 14.5
W7: 11/09 Monte Carlo methods II
W7: 11/11 Veteran's Day
W7: 11/13 Mini-bucket elimination
W8: 11/16pdfVariational methods Ipython dual decomposition
W8: 11/18 Variational methods IIpython loopy BP
W8: 11/20 Variational methods III
W9: 11/23  
W9: 11/25 Thanksgiving
W9: 11/27 Thanksgiving
W10: 11/30  
W10: 12/02pdfLatent Dirichlet Allocation
W10: 12/04 Boltzmann machines; RBMs, DBMs

External resources of interest

Last modified February 03, 2016, at 01:25 PM
Bren School of Information and Computer Science
University of California, Irvine