SELECT-ML is a strategic project at the University of Skövde to strengthen research and education in artificial intelligence, with data science as a core research area of the Skövde Artificial Intelligence Lab (SAIL). Its full name says how: strategic recruitment of a senior lecturer in informatics with a specialisation in machine learning. It is funded by the Knowledge Foundation together with three industrial partners, AstraZeneca, Sandvik Materials Technology and SSAB.
The project was started within the university’s INFINIT initiative and has continued to build the university’s capacity in data science and AI since that programme ended.
Why it matters
Demand for machine learning expertise is growing fast, in academia and across industry. Meeting it takes research and teaching that build on each other: research that keeps courses up to date, and teaching that trains the next generation of researchers and engineers. The industrial partners bring perspectives from pharmaceuticals, materials and steel.
The approach
The project is framed within federated learning, with an emphasis on transparent and trustworthy decision support. Its main areas of interest are:
- advanced analytics and machine learning methods for real-world data;
- explainable AI, so that the people who use a model can understand and trust its output;
- the usefulness and usability of machine learning: methods that are more adaptive, interactive, contextual and automated.
Related research
My research at SAIL during the project period follows its focus on federated learning and trustworthy decision support: robust aggregation, clients that join late, federated fine-tuning of large language models and privacy-preserving medical imaging, together with the PhD students I co-supervise. The related publications are listed below. Related software includes mcdakit, a Python library for multi-criteria decision analysis, and Decentrai, an open-source framework for hierarchical federated learning. The courses I teach at Skövde are on the Teaching page.