Machine learning seminar [ LINGI2379 ]

> https://icampus.uclouvain.be/claroline/course/index.php?cid=lingi2379

having passed at least one of the following courses:

• INGI2262 Machine Learning
• ELEC2870 Artificial neural networks
• SINF2275 Data mining and decision making

Themes are chosen in the domain of machine learning

• To study in groups current issues in machine learning, pattern recognition or data analysis
• To summarize a technical or scientific paper of the domain, convey it to colleagues, and discuss it with a critical viewpoint

The evaluation focuses on the quality of the presentations made ''by each student in front of to other participants in the seminar.
The overall grade consists of:
- 50% on the educational quality of the presentation
- 50% on the accuracy of the scientific content of the presentation

In the second session, the evaluation is 100% on a written report to the teacher the first day of the examination session.

The course is organised as a seminar where student meet regularly to present and discuss recent scientific papers.

Les séminaires pourront être présentés en anglais ou en français par les étudiants.

Illustrative examples:

• Semi-supervised learning methods
• Structured data mining (graphs, trees, sequences, etc.)
• Kernel methods for classification and regression
• Variable selection methods
• Hidden Markov models and their applications
• Boosting and bagging algorithms
• Automata induction techniques

Scientific articles in Machine Learning, supplemented by one or the other textbooks depending on the choice of students's topics.

Examples:

• Statistics for High-Dimensional Data: Methods, Theory and Applications, Bühlmann and van Geer, Springer, 2011.
• Nonlinear Dimensionality Reduction, Lee and Verleysen, Springer, 2007.
• Computational Methods of Feture Selection, Liu and Motoda, Chapman & Hall / CRC, 2008.