“When we were working on the Capstone projects, I wish I was a student again,” says Tom Viering. He is the coordinator of the AI subject and also teaches the machine learning course. He is, as he himself describes: ‘a real AI guy’. He has been a teacher for a year, he took a Ph.D. in machine learning theory and now he is finishing his dissertation. He laughs a lot when he talks about the minor. It was a success, they could not even say the number of sign-ups; there were only a hundred seats this year. It is therefore the intention to increase the minor every year in the future: attract more teachers so that more students can participate. Next year there is already room for 125 students, in a few years this will be expanded to 500 students. Ultimately, all students at TU Delft must get in touch with AI education, which is one of the university’s goals. Despite the fact that there were still corona lockdowns in 2021 and many work consultations had to take place via the computer, this first step towards it was a great success. In addition to the ‘Introduction to Machine Learning’ course, students will receive the ‘Introduction to Python Programming’, the ‘Algorithms and Data Structures’ and the ‘Introduction to Responsible AI’ course, which deals with AI ethics. Finally, the Capstone project follows: students in small groups design an algorithm for a complex, existing, social problem.
Junzi Sun is a lecturer in Air Traffic Management at the Faculty of Aerospace Engineering and supervised students who participated in the minor. The Capstone question that Sun presented to his students was about flight delays. Can AI better predict delays? Sun: “Delays are an important and major problem in air traffic control. And it’s complex: because there are many airports in Europe, and many more planes and planes. I wanted to see if they could make accurate predictions through a machine learning algorithm. “To his surprise, the students passed brilliantly.” They started with simple models, but I was impressed with their progression. Eventually, they were able to develop very serious AI models, they really surprised me. “Sun was in contact with his students twice a week during the project. Finally, he also helped them write a conference paper for ICRAT 2022 (International Conference on Research in Air Transportation).
The topics for the Capstone projects varied widely. A chatbot was created for students with depressed emotions. Celebration says: “This project was about well-being on campus. The chatbot was able to communicate with the students. We adjusted the project along the way: the chatbot could ultimately answer general questions but not give specific psychological advice; it was too uncertain. But it “It’s nice to see that you can really combine machine learning with all the fields. All the studies have something to do with it.” The latter is confirmed by student and participant Huib Baetsen. “When I started on this minor, I did not know how broad its application is. I always thought physics and artificial intelligence were far apart, but they are not. All sorts of new possibilities arise when combining these areas. So it’s super interesting at least to learn what the possibilities are with AI in your field or in your studio, so you can always consider later whether you want to use it or not. AI offers extra opportunities in all courses at TU Delft. You can really get new information and new results using AI; it was the big eye opener for me. ”
He collaborated with nanobiology students on his own Capstone project. The topic was the use of AI to map a specific protein structure. This can usually be done with an electron microscope or with a fluorescence microscope, the student explains. But it is practically impossible to depict the same exact structure in both ways; because in the first method the protein structure is broken down, because one has to cool it down, in the second one it changes, because one has to add things to the one that attaches to the proteins. Baetsen says: “So the idea was to convert an image created through one method to another using AI.” Whether it was ‘successful’ is a difficult question to answer: “It was interesting to get a sense of how sensitive these models are and how many factors are involved. There was literally an endless list of things we could do. “Knowing the versatility, but also the limitations; we learned a lot from that. We now know: In principle, it is possible to do what we want, but it would still require a lot of work.” Teacher Junzi Sun adds: “In the final projects, we distinguished between a ‘must do’, a ‘must do’, a ‘could do’ and a ‘will not do’. Depending on their own interests and skills, students can do so simple or complicated as they wish. ”
Sun is one of 26 AI Capstone mentors. The supervisors of the projects, lecturers from all different faculties, are given the freedom to set up their projects together with the students according to their own wishes. Fifteen of these teachers are also involved in TU Delft AI Labs. The 24 cross-faculty TU Delft AI Labs are committed to education, research and innovation in AI, data and digitization. In these laboratories, the bridge between research ‘in’ and ‘with’ AI is decided. Celebration: “We quickly noticed that the supervisors had a lot of fun in these projects. And we also got positive feedback from the students in the evaluation. ” The subject that Viering teaches, Introduction to Machine Learning, was the favorite subject for student Tatjana Hertel, a third-year undergraduate student in mechanical engineering. She says: “The best thing was that we got the theory in the morning and could apply it immediately in the afternoon.” Hertel has always wanted to take an AI minor, even before it was available in Delft. “I would almost call AI my hobby. I do not want to go all the way in that direction, but I want to stay informed on the topic and what can be done with AI. I believe that this minor in Engineering with AI – in addition to minor in computer science – will help you the most during your studies and later. ” She sees a future for herself in the medical field. In her Capstone project, AI was used to develop a program with which the computer can see if someone is walking normally or unusually, with which, for example, a physiotherapist can detect abnormalities. “It was very hard to make, haha, but also a lot of fun!”
Ibrahim Hassan also participated in minors this year. He is an electrical engineering student and believes that AI is the future: “You hear the concept of AI everywhere these days. It made me curious. In addition, this minor is a good supplement to my studies. My Capstone project was about energy: how much energy is produced by solar panels and wind turbines? As we become more dependent on these energy sources in the future, it is important that we can also better predict how much energy they will supply. We have made models that determine exactly how much energy is produced. I truly believe that this sub-subject enriches your knowledge. AI provides insight and opens doors in your research. ”
Text: Pauline Bijster
Photos: Adam Klugkist