Graphical Models and Machine Learning

EE 639 Advanced Topics in Signal Processing and Communication

Spring 2008

 

.

[Home] [Lectures] [Homework]  [Related Links]

News

1/9

Visit course website at http://www.vis.uky.edu/~cheung/courses/ee639/index.html. Send email to me (cheung at engr.uky.edu) to obtain the password for the course material.  

 

Professor

Dr. Sen-ching Cheung (cheung at engr.uky.edu)
Office: Room 831 VizCenter at Kentucky Utility Building (7-1257 ext. 80299)
Office hours: TTh, 1:00pm-3:30pm or by appointment

Teaching Assistant

Vijay Venkatesh M (mvijay at engr.uky.edu)

Office: TBD

Office hours: TBD

Schedule

Regular class: TTh 3:45pm-5:00pm (VizCenter conference room)

Final Examination: 4/30/08 (VizCenter conference room)

 

Course Description

Probabilistic graphical models are probabilistic models that encode local information of conditional independency among a large number of random variables. By choosing an appropriate (sparse) graph to describe the data, powerful and rigorous techniques exist to perform prediction, estimation, data fusion and to handle uncertainty and missing data. While the probabilistic graphical model acts as a powerful probabilistic database to answer queries, we need machine learning techniques to learn an appropriate model from the data. There have been significant advances in both areas in the last two decades. The goals of this course are to first provide a solid foundation on the fundamental concepts in probabilistic graphical model and machine learning so that graduate students can apply them in their own research, and to discuss recent results in this area so as to stimulate new research ideas.

Tentative Topics

  1. Basic Probability and Statistics
  2. Introduction to Graphical Models
  3. Simple Graphical Models: Linear Classification and Regression
  4. Parameter Estimation for Completely Observed Graphical Models
  5. Expectation Maximization: Parameter Estimation for Incomplete Graphical Models
  6. Exact Inference on Graphical Models
  7. Factor Analysis, Hidden Markov Model and Kalman Filtering
  8. Approximate Inference on Graphical Models
  9. Neural Networks
  10. Kernel Methods and Sparse Kernel Machines
  11. Combining Models

 

Grading

Your grade will be based on:

 

Homework

30%

Midterm

15%

Final

15%

Final Project Report

30%

Final Project Poster

10%

              

  • The letter grade assignment is based on the following scale: from 100 to 90 pts => A, from 89 to 80 pts. => B, from 79 to 70 pts => C, from 60 to 69 pts. => D, from 59 to 0 pts. => E.       
  • Homework will be assigned biweekly. A random question in each homework will be graded. Late homework will not be accepted.
  • One midterm and one final will be given. They are open-book.
  • Each student must also complete a final project applying graphical models to research problems or surveying latest research areas in graphical models. All findings will be presented in a poster session and summarized in a final report.

 

Required Text

  • MIJ: Preprints of chapters from “An Introduction to Probabilistic Graphical Models" by M. I. Jordan will be available in the password-protected portion of this website.
  • CMB: “Pattern Recognition and Machine Learning” by Christopher M. Bishop (2006)

 

 

Prerequisites:

  • Required: Undergraduate level linear algebra (matrix operations), multivariate calculus, probability (random variables, discrete and continuous, especially Gaussian, distributions) and statistics (confidence level and hypothesis testing).
  • Desired: Basic understanding of graph theory, stochastic processes, information theory and optimization.
  • Students will need to be familiar with Matlab, S+, R or a related matrix-oriented programming language.

 

 

Sen-ching Samson Cheung

Last modified: 01/12/2006