Summer internship: Machine learning meets programming education

We are offering a summer internship position at Department of Computer Science at University of Helsinki. For details on how to apply, see the internship call. The deadline is February 20th.

Topic

University of Helsinki programming courses, both local and those offered at mooc.fi, produce data from thousands of students annually. The data includes all students' actions within the programming environment -- keystrokes, save points, compilation attempts and submissions. At each point we know if the (partial) solution compiles and which portion of the tests passes, and which constructs the students have used. Every event, including even the individual keystrokes, naturally comes with accurate time information.

We are looking for a practical machine learning oriented student to work with the data, in a project that combines development of new statistical machine learning models and addressing practical questions on programming education based on this unique data. Some examples of practical questions include "how do individual time management skills and behavior affect students' learning outcomes?", "what sorts of learning paths do students take?", and "which factors explain behaviors such as excessive use of copy-pasting or no copy-pasting at all?".

The project is suitable for a student who has some background in machine learning, data analysis, or statistics, and who is interested in analyzing complex log data. Interest in (programming) education is a strong assett.

The work will be supervised by Arto Klami and Arto Vihavainen.