Examination arrangement

Course content

The course gives an introduction to the principles and methods for automatic learning in computer systems. Classical syntax-based learning methods as well as more knowledge-intensive methods are described. Main empahsis is on symbolic methods, where explicit concepts and relationships are learned. Statistical generalizations, ensemble methods, and deep learning are also included. The strengths and weaknesses of various methods are discussed.
Learning methods in case-based reasoning is integrated with problem solving within the CBR cycle. Numerical and cognitive models for similarity asessment will be discussed, together with different learning system architectures. Methods that combine case-based and generalisation-based inferences will be discussed as well.

Learning outcome

The aim of the course is to introduce principles of machine learning methods in general, to give an understanding of basic mechanisms underlying various specific methods. In case-based reasoning the integration of learning and problem solving is focused.

Learning methods and activities

Lectures, colloquia, self study, exercises.

Further on evaluation

Portfolio assessment is the basis for the grade in the course. The portfolio includes a final written exam (80%) and exercises (20%). The results for the parts are given in %-scores, while the entire portfolio is assigned a letter grade.
If there is a re-sit examination, the examination form may be changed from written to oral.
In the case that the student receives an F/Fail as a final grade after both ordinary and re-sit exam, then the student must retake the course in its entirety. Submitted work that counts towards the final grade will also have to be retaken.

Recommended previous knowledge

TDT4136 Introduction to Artificial Intelligence, TDT4171 Artificial Intelligence Methods or similar.

Course materials

Text book: Tom Mitchell: Machine learning, McGraw Hill, 1997.
Michael M. Richer and Rosina Weber: Case-Based Reasoning, Springer, 2013.
Selected papers.

Credit reductions