Course: Seminar in Artificial Intelligence

ICS: 279

Instructor: Max Welling

Time: Tu-Th: 3.30-4.50 PM

Location: CS 253.

Prerequisites: None

Goals: The main objective of this course is for the students to be introduced to some of the research
that is being conducted at ICS here in UCI. I have asked a number of faculty who do research in
areas related to artificial intelligence to either talk about their most representative work or to send me
an article about their most representative research, which we will treat in class.

If time allows, some timely topics in AI and statistics will be studied, in particular approximate inference methods
on graphs such as belief propagation, mean field and exact sampling.

Students will be asked study a paper from the literature and prepare a presentation on a topic related to the material treated in class.

Homework : To be announced


Sep.28 - M. Welling: Introduction to graphical models
Sep.30 - M. Welling: Introduction to approximate inference

Oct.05 - M. Welling: Introduction to Markov chain Monte Carlo sampling
Oct.07 - Vibhav Gogate:
Understanding Belief Propagation and its Generalizations - J. Yedida

Oct.12 - Vibhav Gogate: More on GBP.
Oct.14 - M. Welling: More on MCMC.

Oct.19 - P. Gehler: Perfect sampling: coupling from the past (David MacKay’s book, ch. 32 p.413)
Oct.21 - P. Gehler: Perfect sampling: CFTP (Propp & Wilson paper)

Oct.26 – Ian: Population MCMC
Oct.28 – Hal Stern

Nov.02 – Matthias Seeger – Discussion on Gaussian Processes
Nov.04 -

Nov.09 - Prof. P. Baldi
Nov.11 - holiday

Nov.16 - Prof. E. Mjolsness
Nov.18 - Elroy

Nov.23 - Prof. R. Dechter
Nov.25 - holiday

Nov.30 - Prof. D. VanDyk
Dec.02 - Anna

Presenters to be scheduled: Ian, Elroy, Anna, Earth.

Some papers:
Population MCMC, Laskey & Myers
LPCD Codes: An Introduction, Shokrollahi
Efficient Graphical Models for Processing Images, Tappen, Russel & Freeman
Loopy Belief Propagation in Image-Based Rendering, Sharon
Approximate Inference and Protein Folding, Yanover & Weiss

Grading Criteria: Satisfactory/Unsatisfactory