Supervising Students who use AI in their UG Research Projects (CPD Case Study)
Jochen Kursawe (School of Maths and Statistics)
[email protected]
1. What motivated you to use AI in this module, and what goals or challenges were you aiming to address?
I chose machine learning and AI-related topics for undergraduate dissertation projects. My research involves analysing time series data from cell biology, for example readouts of gene expression over time, or microscopy movies of moving and deforming cells. Therefore, my work naturally involves machine learning and deep learning methods, and incorporating these methods into undergraduate supervision is a good way to expose students to state-of-the-art research methodology. Each project was associated with real-life research questions, such as designing new algorithms, training neural networks on labelled data, or generating manually curated datasets for training.
2. How did you design or adapt the assessment, and how did you prepare students for using AI appropriately?
We typically start out with reading research papers, then co-design a strategy for the project, and then the student will work on coding and data-analysis/ neural network training etc. under my supervision. The students get exposed to the theory through the papers, and if computational skills are required that the students don’t already have I guide them to online tutorials and also spend a few meetings jointly working on the first steps of the code until they are able to continue independently.
3. What challenges did you encounter, and how did you address them?
Sometimes a project is too ambitious, and we dial back the aims. The idea is that the student learns about how research is conducted in principle, and how to critically assess their results. For this it’s not necessary to fully solve each problem, but rather to learn how to conceptualise and discuss challenges that arise.
4. What benefits did you see for students and for your own teaching practice?
Because all projects are jointly designed with the students, they feel invested, and I like to think they enjoy their excursion into research rather than seeing it as a chore. Sometimes our discussions will venture into areas that I had not worked on before either, and this exposure is rewarding.
5. How did you evaluate the usefulness of this assessment to ensure that it reflected the desirable learning outcomes?
Undergraduate dissertations are marked using established procedures in the School, and the topic does not influence marking criteria.
6. What would you do differently next time, and what advice would you give to colleagues?
I think an important aspect of undergraduate projects is to design the project to meet the requirements of the degree (e.g. 15/30/60 credits and associated marking criteria), and to fall within the capabilities and expectations of the individual student.