TU Delft
Year
print this page print this page     
NEDERLANDSENGLISH
Organization
2017/2018 Electrical Engineering, Mathematics and Computer Science Master Computer Science
IN4320
Machine learning
ECTS: 5
Responsible Instructor
Name E-mail
M. Loog    M.Loog@tudelft.nl
Instructor
Name E-mail
Dr. J.C. van Gemert    J.C.vanGemert@tudelft.nl
Dr.ing. J. Kober    J.Kober@tudelft.nl
Dr. D.M.J. Tax    D.M.J.Tax@tudelft.nl
Contact Hours / Week x/x/x/x
0/0/2/2
Education Period
3
4
Start Education
3
Exam Period
4
Course Language
English
Expected prior knowledge
IN4085
Course Contents
The course will treat a number of machine learning topics, approaches, and techniques in detail and on an advanced level. Possible topics are

- learning theory
- complexity
- semi-supervised learning
- multiple instance learning
- kernel methods
- reinforcement learning
- Gaussian processes
- sparsity
- conditional random fields
Study Goals
After the course, the student is able to recognize the (limits to the) practical applicability of the presented theory. Moreover, s/he is able to see the relationships of a novel technique to those discussed in the course, and has insight in what type of problem requires application of which type of machine learning technique.
Education Method
We follow a scheme in which every topic is treated in a two-week block. In the first week, one of lecturer will present a technique based on a tutorial paper or other reading material. In the second week, the student will work on an exercise that extends and deepens their knowledge and understanding of the technique under consideration in that particular block. A large part of the exercises involves programming. The final output to every exercise is a report covering the necessary derivations, snippets of code, figures, and general text.
Literature and Study Materials
For every block of two weeks in which a single topic is treated, specific literature will be provided through Blackboard.
Assessment
Assessment grades are based on the reports handed in (60-80%) and the final assignment (about 40-20%), the latter of which is based on a somewhat larger and more advanced machine learning challenge that the students will write a report on as well. There is no resit; not overall, nor for any of the elements.
maximum aantal deelnemers
50