Leiden University logo.

nl en

Interview - ‘I learn a lot from my students’

'I am very honored to be nominated for the Teacher of the Year Award,’ says dr. Jan van Rijn. Every autumn he teaches the course Automated Machine Learning, which he set up himself at Leiden University.

Jan van Rijn

The course is about training a data science model as efficiently as possible to analyze large amounts of data, explains Van Rijn. ‘To train a data science model, you first need good data, but it is also important to set the buttons and dials of such a model properly. In machine learning, there is someone who turns those buttons. But with automated machine learning, that process is automated. Not because we want to get people out of the loop. I think that artificial intelligence and humans should always work together, but humans can then focus on things that are more interesting and challenging.”

Important insights for the teacher as well

Van Rijn believes that a course should get a little better every year. ‘It has to evolve. What I teach coincides exactly with my field of research. That is why I can keep the teaching material up-to-date quite easily. If new scientific research appears that I think belongs in the course, I will add it. One of the nicest aspects of teaching is that students present a paper from the field. Some of those I don't know very well yet, so I learn a lot from the students myself. They sometimes come up with important insights that I hadn't thought of, and neither did the authors of the paper.’

‘A course should get a little better every year’

Open source book about metalearning

During the corona pandemic, Van Rijn missed the interaction with students. ‘They still gave presentations, but there was less feedback from the class.’ He tried to keep challenging students in the lectures he recorded. ‘Then I said for example, pause the video for two minutes and think about this question. After that I gave the answer.’ Together with colleagues, Van Rijn wrote the book Metalearning, which is available to everyone. Now that there is an open source book in which the most important material is explained, we also want to hold an exam this year.’ Until now, students were graded on the programming assignments they do.

Prior knowledge

Thanks to a team of 4 teaching assistants, the course Automated Machine Learning is a success, says Van Rijn. Students from almost all specializations follow the course. ‘In recent years we started with about 80 students each year, 60 of which completed the course. We recommend students who do not have the correct prior knowledge from their study program to first take the course Machine Learning.’

About Jan van Rijn

Thanks to a team of 4 teaching assistants, the course Automated Machine Learning is a success, says Van Rijn. Students from almost all specializations follow the course. ‘In recent years we started with about 80 students each year, 60 of which completed the course. We recommend students who do not have the correct prior knowledge from their study program to first take the course Machine Learning.’

This website uses cookies.