CSE 254: Seminar on Learning Algorithms

TuTh 3.30-5 in HSS 2305B

Sanjoy Dasgupta
Office hours Mon 1-3 in EBU3B 4138

CSE 254 is a graduate seminar devoted to recent research on AI learning methods and applications.  This quarter the theme is inference in graphical models.
Prerequisite: background on graphical models (for instance, my class on probabilistic AI).

In each class meeting, a student will give a talk lasting about 60 minutes presenting a recent technical paper in detail.  In questions during the talk, and in the final 20 minutes, all seminar participants will discuss the paper and the issues raised by it.

Date Presenter
September 22 organizational meeting    


September 27 no meeting    


September 29 Sanjoy Presentation guidelines; overview of inference  


October 4 Antoni An introduction to variational methods for inference Jordan, Ghahramani, Jaakkola, Saul here
October 6 Daniel Large deviation methods for approximate probabilistic inference Kearns, Saul here
October 11 Evan Fast approximate energy minimization via graph cuts Boykov, Veksler, Zabih here
October 13 Nan What energy functions can be minimized via graph cuts? Kolmogorov, Zabih here
October 18 Brian Loopy-belief propagation: an empirical study and Efficient belief propagation for early vision Murphy, Weiss, Jordan; Felzenszwalb, Huttenlocher here
October 20 Thomas Understanding belief propagation and its generalizations Yedidia, Freeman, Weiss here
October 25 Brian Expectation propagation for approximate Bayesian inference Minka here
October 27 Buhm Tree-based reparameterization framework for analysis of sum-product and related algorithms Wainwright, Jaakkola, Willsky here
November 1 Sanjoy A review of some information geometry  


November 3 and 8 Daniel A variational principle for graphical models Wainwright, Jordan here
November 10 Evan MAP estimation via agreement on trees Wainwright, Jordan, Willsky here
November 15 Thomas Correctness of local probability propagation in graphical models with loops Weiss here
November 17 Antoni Nonparametric belief propagation Sudderth, Ihler, Freeman, Willsky here
November 22 Nan Classification problems with pairwise relationships Kleinberg, Tardos here
November 24 Thanksgiving    


November 29 Buhm Pairwise clustering and graphical models Shental, Zomet, Hertz, Weiss


December 1 Project Presentations    


This is a four unit course in which the work consists of presentations and a project. Guidelines for the term project can be found here.  Project reports will be evaluated using these grading criteria.  There is a schedule for handing in a detailed project proposal, a draft project report, and then the final report.

The seminar will have no final exam.  Letter grades will be based on the presentations and the final project report, but participation in class and the intermediate project deliverables are important also.


The procedure for each student presentation is as follows: Please read, reflect upon, and follow these presentation guidelines, courtesy of Prof Charles Elkan.  Presentations will be evaluated, in a friendly way but with high standards, using this feedback form.

Each presentation should be prepared using LaTeX or Powerpoint, and should consist of about 40 slides.  You must copy all important equations, diagrams, charts, and tables from the paper into your slides.

The schedule of presentations will be determined as much as possible on Thursday September 22.  Here is a list of papers.

If you want to change your presentation date, please arrange a swap with another student and notify me at least two weeks in advance.