Helsinki University of Technology → Department of Computer Science and Engineering →
Laboratory of Computer and Information Science → Teaching →
T-61.3050 Machine Learning: Basic Principles → 2007

News:

• The 2007 course has ended. You can still pass the 2007 course by passing the examination (if you have not already passed the examination) before the first lecturing period of Autumn 2008 (see the examination schelude) and by submitting the term project (if you have not already passed the term project) as per 2007 instructions on 1 September 2008, at latest (you can submit the term project earlier, but you not guaranteed to get the grading of the term project before 1 October 2008). There will be a new course in Autumn 2008 which may, for example, have a different term project and a different selection of topics (the course contents will likely remain mostly the same, though).

T-61.3050 Machine Learning: Basic Principles (5 cr)

Autumn 2007

26 + 26 (2 + 2)

Kai Puolamäki, PhD, lecturing researcher, lecturer
Antti Ukkonen, MSc, course assistant

Email: t613050@james.hut.fi
Newsgroup: opinnot.tik.t613050

Lectures:

In T1 on Tuesdays at 10-12 o'clock (from 11 September to 11 December 2007, no lecture on 30 October).

Problem sessions:

In T1 on Fridays at 10-12 o'clock (from 14 September to 7 December 2007, problem sessions are not every week).

Requirements:

Examination and term project.

Prerequisites:

Basic mathematics and probability courses (Mat-1.1010, Mat-1.1020, Mat-1.1031/1032 and Mat-1.2600/2620; or equivalent), basics of programming (T-106.1200/1203/1206/1207 or equivalent), data structures and algorithms (T-106.1220/1223 or equivalent).

Literature:

Alpaydin, 2004. Introduction to Machine Learning. The MIT Press. The course will be based mostly on topics covered in the Alpaydin's book, with some parts left out, and with minor parts based on material to be distributed from the course web site. Additionally, there may be some useful material, such as lecture notes or copies of the slides, that can be downloaded from the course web site.

Language:

English.

Additional information:

Replaces course T-61.3030 Principles of Neural Computing.

You must sign in to the course using WebTOPI.

The lectures and problem sessions take place in the lecture hall T1 of the Computer Science Building (Konemiehentie 2, Espoo) during the autumn 2007 semester.

The Alpaydin's book can be borrowed from the TKK library, or bought, for example, from vendors listed in Alpaydin's website or Yliopistokirjakauppa in Otaniemi.

Course poster: PDF, TXT.

This course as a part of the studies of methodological principles or postgraduate studies.

Material

• Lectures
• Problem sessions
• Examination
• Term project
• Mid-course survey 2007
• Grading principles
• Software

Contents

The topics include the background principles needed to understand and apply the models of machine learning. After the course, the student is able to apply the basic methods to data and understand new models based on these principles.

Machine learning (Wikipedia) is programming computers to learn or optimize a performance criterion using example data or past experience. Machine learning is closely related to data mining and statistics, and also to the theoretical computer science, especially analysis of algorithms.

Machine learning comes into play when computers have to deal with natural or otherwise "noisy" data. Applications include syntactic pattern recognition, information retrieval and search engines, user modeling, bioinformatics and cheminformatics, analysis of economic data, natural language processing, speech and handwriting recognition, object recognition in computer vision, game playing and robot locomotion. These applications are mostly not covered at this course but in subsequent courses, for which this course is prerequisite or recommend knowledge.

The course covers the principles of probabilistic modeling, as well as algorithmic considerations, in data analysis.

The course introduces some basic machine learning methods that can be used in classification, regression and unsupervised learning.

The principles are of a little use for the student if he or she cannot apply them to his or her own data, which is why the course includes hands-on applications of these methods and principles in real data analysis examples.

After this course, the student should...

1. be able to apply the basic methods to real world data;
2. understand the basic principles of the methods; and
3. have necessary prerequisites to understand and apply new concepts and methods that build on the topics covered in the course.

For further introduction to the contents of the course, please read the Preface and Introduction (Chapter 1) of the Alpaydin's book.