Information on course projects: EECS 281A / STAT 241A
Description
The course project provides you with an opportunity to explore more
deeply a particular aspect of the course that interests you. Of
course, the topic of the project should be related to the material
(concepts, models, algorithms, applications etc.) discussed in the
course, but otherwise you have a fair bit of freedom in selecting a
topic.
Possible types of course projects include:

 a survey project: an overview of several research papers around a
coherent theme.
 an applications research project: demonstrating the application
of some techniques from the course in an application of interest
(e.g., vision, natural language processing, signal processing, coding,
bioinformatics, artificial intelligence, communication, control,
neuroscience etc.)
 a theoretical or methodological research project: examples studying different classes of models; proving convergence guarantees
for a known algorithm; developing a new method.

Naturally, the expectations are somewhat lower for a research project
than a survey project (since research is a capricious thing, and not
always easy to predict!) However, in planning your project, you
should ensure that it has several stages, and includes at least some
goals that you expect to be able to make within the semester.
You are free to work either individually, or in groups of 2 to 3
people total.
Evaluation
The course project is worth roughly 40% of your grade; evaluation
will be based upon:

 A technical writeup describing your project area and results.
This writeup should be aimed at a technically knowledgeable audience,
but should not assume expertise in your particular area. (It should
be readable, for instance, by one of your classmates.) It will be
useful to include sections giving the background of your problem, past
work (if relevant), and a conclusion in which you state what future
directions you might take this work in. You should aim for a writeup
that is roughly 1015 pages. Note that the length per se is not the
main concern; rather, it should be the clarity of the presentation.
Sample projects from previous semesters
Here are a few examples of projects from previous years. They
are not supposed to be the best ones, they're just
some pseudorandom sample to give you an idea of what people
can do.
 Graphical Models for Game Theory: A survey of some recent papers, by
Ambuj Tewari  report
[ps] 
poster
[ppt]
[jpeg]
 A Naive Bayes Spam Filter, by Kai Wei
 report 
poster [jpeg]
 Localization of Robots Using Particle Filters, by Phoebus Chen 
poster [jpeg]
 Nonlinear Dimensionality Reduction on Human Facial Expressions, by Ryan White  report
[pdf] 
poster [jpeg]
 Combining SVM with graphical models for supervised classification: an introduction to
MaxMargin Markov Networks, by Simon LacosteJulien 
report 
poster [jpeg]
List of projects for 20022003
Background material
This section contains a number of suggestions for background reading
that could be useful in generating project ideas.

 Many of these papers are semitutorial in nature, and are useful
to read to gain a general sense of the field. If you are interested
in a more specific topic, then it is worthwhile to follow up on the
references in a given paper.
 This section is a (rapidly) evolving beast. Free free to
contribute suggestions or other
relevant papers that you found interesting, or think other classmates
might find useful.
 Any of the papers listed here that appear in IEEE journals can be
downloaded from the IEEExplore website, accessible for free from any
computer on a Berkeley domain. See the proxy instructions at
http://proxy.lib.berkeley.edu/ if you want to connect to this service
from a computer at home.
 Any of the papers listed here that appear in most statistics
journals (e.g., JASA, Annals of Statistics etc.) can be download from
JSTOR at www.jstor.org from a computer on the Berkeley domain.
General (graphical models; messagepassing)
Graphical models: Probabilistic inference. M. I. Jordan and
Y. Weiss. In The Handbook of Brain Theory and Neural
Networks , 2002.
Graphical Models M. I. Jordan. General survey paper on
graphical models, their applications and algorithms. Appeared in
Statistical Science , 2004.
Graphical
models, exponential families, and variational methods.
M. J. Wainwright and M. I. Jordan. Semitutorial paper on graphical
models, exponential families, and variational methods. Appeared as
book chapter in New Directions in Statistical Signal Processing
. 2003, 2005.
J.S. Yedidia, W. T. Freeman and Y. Weiss. Constructing free energy
approximations and generalized belief propagation algorithms.
Appeared in IEEE Transactions on Information Theory , Vol. 51,
pp. 22822312.
Signal processing
A. S. Willsky (2002). Multiresolution Markov models for signal
and image processing. Appeared in Proceedings of the IEEE
Vol. 90, pp. 13961458.
Loeliger (2004). An Introduction to Factor Graphs. Appeared in
IEEE Signal Processing Magazine , Vol. 21, pp. 2841.
Communication and coding
F.R. Kschischang et al. (2001). Factor graphs and the sumproduct
algorithm. Appeared in IEEE Transactions on Information Theory
. Vol. 47, pp. 498519.
S.M. Aji and R.J. McEliece (2000), The Generalized Distributive Law.
Appeared in IEEE Transactions on Information Theory .
Vol. 46, pp. 325343.
Markov chain Monte Carlo
S. Geman and D. Geman (1984). Stochastic Relaxation, {Gibbs}
Distributions, and the {Bayesian} Restoration of Images. Appeared in
IEEE Transactions on Pattern Analysis and Machine Intelligence
Vol. 6, pp. 721741.
J. Besag, P. Green, D. Higdon and K. Mengersen (1995). Bayesian
computation and stochastic systems, Appeared in Statistical
Science . Vol. 10, pp. 341.
Natural language processing
S. Vogel, H. Ney, and C. Tillmann.
HMMbased word alignment in statistical translation. In Proceedings of
the 16th conference on Computational linguistics, pp. 836841,
Morristown, NJ, USA, 1996. Association for Computational Linguistics.
Bioinformatics
M.I. Jordan. Chapter 23 of book in preparation. (See the course reading list.)
Computer vision; image processing
S. Geman and D. Geman (1984). Stochastic Relaxation, {Gibbs}
Distributions, and the {Bayesian} Restoration of Images. Appeared in
IEEE Transactions on Pattern Analysis and Machine Intelligence
Vol. 6, pp. 721741.
W. T. Freeman, E. C. Pasztor and O. T. Carmichael (2000).
Learning LowLevel Vision. Appeared in International Journal
of Computer Vision , Vol. 40, pp. 2547.
K. Murphy, A. Torralba, and W. T. Freeman, Using the forest to see the trees: a graphical model
relating features, objects, and scenes,
in Advances in Neural Information Processing Systems 16 (NIPS),
Vancouver, BC, MIT Press, 2004
Neuroscience
home
Last modified: 09/01/2008.