Laboratory of Computer and Information Science >
Teaching > T-122.102
Please sign in using Topi.
T-122.102 Informaatiotekniikan erikoiskurssi VI 3-4 ov L V
T-122.102 Special Course in Information Technology VI 3-4 cr P V
Yhteisesiintymämenetelmät diskreetin datan analysoinnissa
Co-occurence methods in analysis of discrete data
Lecturer: PhD Kai
Puolamäki, Prof. Jaakko Hollmén
Assistant: MSc Arto Klami
Time: In T4 on Tuesdays at 14-16 o'clock (from 20
January to 27 April 2004)
Requirements: To pass the course with 3 credits you have to
- participate actively (read the articles before presentations, be
present in the sessions, participate to the discussion),
- give one lecture on a given topic (instructions at the
bottom of the page) and
- submit a summary article of your presentation, to be published in
the course web site.
To pass the course with 4 credits you have to additionally
- carry out a small research project on a given topic and write a
brief report on the results.
Prerequisites: first two years' mathematics courses; the target
audience of the course are graduate students and pre-graduate students
interested in reasearch topics .
Contents of the course
Topic of this research seminar is the recent advances in the
analysis of discrete data sets, the common theme being co-occurences
and how they are being dealt with. The emphasis of the course is on
information theoretically (e.g. mutual information) or
probabilistically motivated approaches (e.g. generative models). The
related question is the definition distance or similarity measures
in discrete data sets. The purpose of this course is to review
various topics of the analysis of discrete data in this framework.
A preliminary sketch of topics:
- Information Bottleneck
- Discriminative Clustering
- Kernel methods
- ICA motivated
- Discretizing data
A whole session is reserved for each presentation. It means that the
presentations should be roughly 45 minutes long and the
remaining time is used for discussion and possible organizational
issues. It is recommended that the slides of the presentation are
given as handouts at the beginning of the session. Please also send
the slides to the course assistant in PostScript or PDF format so that
they can be published in the web.
In addition to the presentation, you have to prepare a
summary article of the topic. The summaries are published
on the web page, and they are supposed to be quite short
(a few pages) and easy to read. The purpose is that other people
can understand the basic idea and the most important aspects
of the presented topic without having to take the effort of
reading the original article(s) thoroughly. Clever illustrations
and examples are highly preferable.
The summaries (in postscript or PDF format) are to be submitted
to the course assistant (Arto Klami)
within 12 days after the presentation. Because the deadline is after
the actual presentation, you can (and should) take the
comments given at the presentation into account.
You are requested (but not required) to release your summary
article under a Creative
Commons License of your choosing. You can do this for example
by including the following text to your article:
This work is licensed under the Creative Commons
Attribution-NonCommercial License. To view a copy of this license,
send a letter to Creative Commons, 559 Nathan Abbott Way, Stanford,
California 94305, USA.
Friday, 17-Dec-2004 14:46:17 EET