The lecturers are also keen to make more use of videos. “Today's students have grown up with images,” says Lazar. “I occasionally play a video during my lectures when I notice that my students' attention is starting to wander, but ideally these videos would be an integral part of the lecture series.” The lecturers are considering teaching the theory involved in an experiment and following this up with a live video connection to a laboratory showing what that experiment looks like in practice.

Moreover, they are keen to have students do more practice-based challenges.“You can have students solve a software problem with numbers, but you can also have them calculate how they can achieve the fastest lap time with a car,” says Lazar. As a first step, at the end of the fourth quartile of this academic year, he wants to add a challenge to the regular questions in order to test how this will work in practice.

All the changes to the two courses should be implemented, with the aid of student assistants, during the coming two years.

Students taking courses in the field of image analysis or machine learning spend a fair amount of time preparing datasets and setting up experiments. For their part, their lecturers have to spend time inspecting the set-ups - a minor aberration in the code can lead to a major aberration in the results. Time that these lecturers - and students - would prefer to spend on other things. Listening to twenty groups of students give their presentations is also pretty time-consuming, tells Mitko Veta, one of the project's proposers and assistant professor at the Medical Image Analysis research group. Together with Veronika Cheplygina (Assistant Professor Medical Image Analysis) and Joaquin Vanschoren (Assistant Professor Machine Learning), he wrote the project proposal. “What's more, students don't see each other's presentations, they don't have the opportunity to learn from each other nor to cooperate at an earlier stage and swap ideas,” says Veta.

A solution to these ‘problems’, caused in part by the increase in student numbers, could be provided, the project proposers believe, by an online platform. At the press of a button, data and examples could be ready for use, so that students are freed of these concerns to focus on their assignments and the analysis. Furthermore, data should be taken from real-life and should reflect reality, says Veta. The data and images used on these courses relate, among other things, to the detection of cancer and the analysis of blood vessels. The project will also introduce a shared work environment, fostering the exchange of ideas, solutions and models. One of the benefits this will bring is that students will receive interim feedback and will know sooner whether they are on the right track.

Platform

The project's proposers are keen to implement their proposal on eleven courses, and also expect it to be applicable to the bachelor's final project and various master's projects. To facilitate this, they want to use OpenML, a platform developed at TU/e that enables researchers to share and discuss experiments, data and algorithms in the field of machine learning. OpenML lends itself to this, says Veta, but various new functionalities will be needed. The search for a suitably qualified programmer is already underway. If all goes according to plan, the upgraded platform will be ready for use in a year's time.

No fewer than 34 proposers of digital and innovative education projects at TU/e now have the chance to go ahead under the banner of BOOST. Most of these projects are currently in the startup phase. With the abolition of the basic grant and the introduction of the government's loan system, Ministry of Education (OCW) funding is also available, studievoorschotmiddelen as it is called, to support projects financially. In the past Cursor has written about BOOST, what stands for Be the Owner of your Own STudy. Next week we will present another two BOOST projects.