These are the archived web pages of the spring 2007 course. The contents of this course vary yearly. The next course will probably take place in spring 2008. Please see http://www.cis.hut.fi/Opinnot/T-61.6020/ for current information.

Courses in previous years: [ 2000 | 2001 | 2002 | 2003 | 2004 | 2005 | 2006 ]

T-61.6020 Special Course in Computer and Information Science II (V, P):
Machine Learning: Basic Principles

Course poster: TXT, PDF

Instructor	Kai Puolamäki, PhD
Course assistant	Mikko Korpela, MSc
Course format	Seminar course
Credits (ECTS)	5
Semester	Spring 2007 (during periods III and IV)
Seminar sessions	On Mondays at 14–16 in lecture hall T4 in the computer science building Konemiehentie 2, Otaniemi, Espoo
Language	English
Web	http://www.cis.hut.fi/Opinnot/T-61.6020/
Email	t616020@cis.hut.fi

Course description

The seminar course will provide a comprehensive introduction to the fields of pattern recognition and machine learning. The course is especially suitable for advanced undergraduates and PhD students who want to have (or refresh) a comprehensive and solid overview of these fields.

Requirements for passing the course

To pass the course, you must:

• Have a seminar presentation on a given topic.
• Get at least 50% of the points from the exercise problems.
• Participate actively (over 70% presence at the lectures, read the chapter before the lecture).

To pass with distinction, you should additionally:

• Have a good presentation.
• Solve the exercise problems well (i.e., well over 50% of the points).

Estimate of time usage (5 cr equals approximately 133 hours of working time for a typical student):

• Lectures 14*2h=28h
• Preparing own presentation 26h
• Reading a chapter per lecture and solving exercise problems 13*6h=78h

Course material

The course is based on the following book:

• Christopher M. Bishop. Pattern Recognition and Machine Learning. Springer, 2006. ISBN 0387310738.

Information on how to get the book.

Schedule

The schedule follows the structure of the course book. At each seminar session, one chapter of the book is presented. Below, you can find the PDF slides of the seminar lectures by the students. (Notice that the presentations below have been prepared and given by the students, who also may or may not have the source files (LaTeX, PowerPoint etc.) for the PDF slides. The lecturer does not have the source files.)

Time	Subject	Lecturer	Material	Misc
Jan 15	Organization of the course	Kai Puolamäki	Slides (PDF)
Jan 22	Probability Distributions	Mikko Korpela	Slides (PDF)	A condensed introduction to the exponential family notation (including Bernoulli, Multinomial, Gaussian etc. distributions) is given by subsection 3.1 of Buntine ECML'02. Rest of the section 3, "Background Theory", covers the EM and variational approaches.
Jan 29	Linear Models for Regression	Nicolau Gonçalves	Slides (PDF)
Feb 5	Linear Models for Classification	Teemu Mutanen Antti Sorjamaa	Slides 1 (PDF) Slides 1 with comments (PDF) Slides 2 (PDF)
Feb 12	Neural Networks	Tino Ojala Marcus Dobrinkat	Slides (PDF)
Feb 19	Kernel Methods	Jayaprakash Rajasekharan	Slides (PDF)
Feb 26	Sparse Kernel Machines	Janne Argillander	Slides (PDF)
Mar 5	No lecture (exam period)
Mar 12	Graphical Models	Dmitrij Lagutin	Slides (PDF)
Mar 19	Mixture Models and EM	Yoan Miche	Slides (PDF)
Mar 26	Approximate Inference	Janne Toivola Ville Turunen	Slides 1 (PDF) Slides 2 (PDF)
Apr 2	Sampling Methods	Matti Pöllä	Slides (PDF)
Apr 9	No lecture (easter break)
Apr 16	Continuous Latent Variables	Mikaela Klami	Slides (PDF)
Apr 23	Sequential Data	Mari-Sanna Paukkeri	Slides (PDF)
Apr 30	Combining Models	Jukka Parviainen	Slides (PDF)

Please sign in using WebTOPI.