In late 2025, students and employees at GFI had the opportunity to join an evening course, spread over four sessions across five weeks, aimed at giving participants hands-on experience with building and tuning ML models.

Organised by Kjersti Birkeland Daae and Nilgun Kulan, and held by Robin Guillaume-Castel and Sigrid Passano Hellan, the course went through the basics of what an ML model really is and how a subset of models from the python packages SciKitLearn and PyTorch work. During the course, the participants solved exercises related to data pre-processing, hyperparameter selection and tuning.

For the last two sessions, the participants chose one of three half finished Neural Networks that consisted of different real world problems, with the goal of tuning them themselves. These were presented to the rest of the class at the end, with feedback and learning shared across the groups and projects.

The project’s goal is to use co-creation to promote active learning and communities of practice at the Geophysical Institute at the University of Bergen. We work towards this goal in four work packages (WP).

In many courses at GFI, the seeds of co-creation are in place and being cultivated already. Our WP1 is about supporting and strengthening those efforts by evaluating and iteratively improving them in specific courses to gain more experience at our institution and create pilot projects that can serve as proof of concept that we and others might learn from. In WP2, we help ground those efforts by creating supportive boundary conditions at GFI regarding how the organisation is structured, whether there are places where student voices could be elevated, and whether the administrative framework could better support co-creation at an institutional level. WP3 is then about engaging more and more teachers and students in other courses in co-creation and supporting this development by creating meeting places and conversations about the topic and supporting evaluation and discussion of results. Lastly, we are not doing this alone: WP4 brings together expert advice we are receiving as well as our efforts to share what we are learning. We have the support of the iEarth community, and specifically an advisory board with internationally renowned experts on co-creation and leading academic change processes to help us. As our efforts flower and bear fruit, we will produce a range of publications, infographics, “how-to guides,” and many other formats to share our learnings with the scientific community and interested practitioners.