MARKOVIA

CSE 590mv - Markovia

NEW meeting time: Tuesdays 9:00AM - 10:30AM

Allen Center 503

Organizers:

Tanzeem Choudhury (Intel Research), Sumit Basu (Microsoft Research), Dieter Fox (UW), Henry Kautz (UW)

This seminar will cover a variety of advanced topics in machine learning focusing on graphical models. The format of the course will be weekly presentations (i.e., powerpoint slides) to be given by teams of 2-3 students. Students should ideally have a good background in probability and at least a passing knowledge of graphical models, though we will review the basic mechanisms for exact inference and learning. The course can be taken for 1, 2, or 3 credits: 1 credit for attending and reading the papers, 2 credits for doing one presentation, 3 credits for doing two presentations or a presentation and a project.

Date & Discussion Leader	Topic	Reading Materials
September 30, 2004	Introduction
October 5, 2004 Discussion Leaders: Benson Limketkai & Pradeep Shenoy Slides from class	Basics of Probability Theory	Max Welling, General Statistical Modeling Tom Minka. Nuances of Probability Theory Iain Murray. Basic Math Needed for Machine Learning
October 12, 2004 Discussion Leaders: Maisie Wang & Mike Caferella Slides from class	Probabilistic Inference in Graphical Models	Jordan and Weiss. Probabilistic Inference in Graphical Models
October 19, 2004 Discussion Leaders: Karthik Gopalratnam & Julie Letchner	Graphical Models: Parameter Learning	Zoubin Ghahramani. Graphical Models: Parameter Learning Optional: Ross Shachter. Bayes Ball Optional: Frank Kschischang et. al. Factor Graphs and the Sum-Product Algorithm
October 26, 2004 Discussion Leaders: Lin Liao & Ana-Maria Popescu	Graphical Models: Structure Learning	David Heckerman. Graphical Models: Structure Learning
November 2, 2004 Discussion Leaders: Krzysztof Gajos & Don Patterson	Bayesian Model Selection	Robert Kass and Adrian Raftery. Bayes Factor and Model Uncertainty
November 9, 2004 Discussion Leaders: Michele Banko & Kevin Dun	Causal Inference	Richard Scheines. An Introduction to Causal Inference
November 16, 2004 Discussion Leaders: Danny Wyatt & Colin Zheng	Clustering - Mixture of Gaussians and k-means	David Mackay. Information Theory, Inference and Learning Algorithms. Chapter 20 Max Welling. Clustering
November 23, 2004	TBA	TBA
November 30, 2004 Discussion Leaders: Krzysztof Gajos & Douglas Downey	Time Series Models	Max Welling. Hidden Markov Models Optional: Max Welling. Kalman Filter
December 7, 2004 Discussion Leaders: Sumit Basu and Tanzeem Choudhury	Time Series Models	Tom Minka. From Hidden Markov Models to Linear Dynamical Systems Optional: Sam Roweis and Zoubin Ghahramani. A Unifying Review of Linear Gaussian Models