Moonen is proud to be able to present the list on TU/e’s sustainability website now. “As you can see, it’s been a major undertaking to create that list.”

Helpful tool

It is important that we now have an overview of exactly which SDGs are addressed in each course, making it easier for students to choose which courses to take if they want to learn more about certain SDGs. “It has definitely become easier to find courses at other departments now, or if you’re interested in a particular SDG, for example.”

Moonen calculated that about five percent of all courses (both master’s and bachelor’s) at TU/e are related to sustainability. That may not seem much, but Moonen is quick to say that there is more to the story. “Don’t forget that there are also courses that are indirectly related to it. I studied Building Environment myself (which was also the focus of her EngD, Ed.) where, for example, in materials science you learn how the materials you use are extracted or recycled, but the course does not have a sustainability term in its name. Tip for lecturers: Tell students why you teach them something. For example, that you can apply that particular way of programming to work more sustainably.”

Read about two concrete examples below the photo.

Concrete

The list created by the SustainableTU/e team includes as many as 44 courses related to sustainability at the Department of Industrial Engineering & Innovation Sciences. At Applied Physics, there are 6 and at Built Environment 28. Cursor spoke to lecturer Hilde Weerts, artificial intelligence engineer at the department of Mathematics & Computer Science, who gives the course Responsible Data Science, the only course that Data Science got on the list, and to Dirk Fahland, who shared with us how he takes the SDGs into account as much as possible in CBL.

The course that Hilde Weerts gives is related to SDGs 8, 10 and especially 5, which are about Fair Work, Reducing inequality within and between countries and Gender Equality, respectively. This is the third year that this elective is being taught and about a hundred students are taking the course. Most are Data Science and Computer Science students, but she also sees Biomedical Engineering students and Psychology &Technology students in the lecture halls.

“What they learn there is how to use Machine Learning to make the world a little better rather than worse”, says Weerts, and she clarifies: “They have to ensure that their models work equally well for all groups of people. Because that’s where things go wrong sometimes. Just think of the problem with Proctorio where dark skin colors weren’t recognized properly. Also, when you create a model that automatically screens job applicants’ resumes, you have to prevent there being – or being introduced – a bias based on gender, nationality or age.”

In the group project, students build their own model, which introduces them to the practical challenges of more responsible design of machine learning systems. “How do you test whether a model is working properly for a very small subgroup in the data? For example, when only a few respondents say they’re nonbinary while there are two large groups of men and women.” And at an even higher level, pure ethics comes into play: “Is machine learning even the appropriate method here? I want to train students to be critical engineers who don’t just do what’s possible because it’s possible.”

Reducing inequality

Weerts says Responsible design plays a bigger role in preventing or aggravating inequality within countries rather than between countries. “One of the things that will be addressed is algorithmic accountability. We will discuss the increasing influence of large (tech) companies, mostly in the Western world and also in China, and how we can hold these companies accountable for their impact. Large tech companies sometimes use data from poorer countries without properly compensating them for it.” In all honesty, she adds that the latter will only be reflected to a lesser extent in the course. This is because of time constraints.

Step by step

"Each year, the students came up with new ideas and deeper insights into the problem, helping the organization to get a better grip on the problem and to adopt new ways of problem solving. But, as courses run only 8 weeks, the students usually cannot fully solve the problem. Learning how to handle this is for me a very important element of educating responsible Data Scientists. We are teaching our students the skills and mindset to take on a really large, societally relevant problem that they cannot solve themselves by identifying an important element in this large problem that they investigate thoroughly to make a small but valid contribution. One small piece at a time."

In some years, the challenge could coninue and build on the insights from the previous year and work on the next step of the challenge to help the organization get closer to its aim. In some years, they had to find a new challenge because together with the students the point had been reached where now the organization could take next steps themselves.

"When looking for new challenges, we so far did not explicitly look for ones related to SDG. But the focus on working with public and non-profit organizations always led us there.To be honest, we also have to be a bit opportunistic when finding these challenges: not every organization has a several experts and data available to support us in running such a challenge for over 120 students."

Future

"Now that SDGs are coming more into focus, I think we should make an effort to more consciously look for challenges that do relate to the SDGs. What I do see is that students greatly enjoy working on truly relevant problems. Our students really "take ownership" of the problem and give everything to solve it. We try to give the students as much freedom as possible in their work and so far it always has resulted in many new ideas."