The UiT Machine Learning Group at UiT The Arctic University of Norway conducts both foundational and applied research across machine learning and artificial intelligence. Our work spans core methodology and real-world applications in medicine, marine science, energy, and earth observation. The group leads Visual Intelligence, co-leads Integreat, and is involved in numerous other projects.
The selection below reflects recent work from the group, spanning top publication venues in both methodology and application domains.
| Theme | Focus |
|---|---|
| 🖼️ Computer vision | Visual recognition, segmentation, detection, and representation learning |
| 🕸️ Graph learning | Graph neural networks, structured data, and relational reasoning |
| 📐 Information theory | Theoretical foundations of learning, compression, and uncertainty |
| 🧠 Deep learning | Neural network architectures, training methods, and generalization |
| 📊 Machine learning | Statistical learning, optimization, and fundamental ML theory |
| Domain | Focus |
|---|---|
| 🩺 Medicine and health | Medical image analysis, digital pathology, and clinical decision support |
| 🌊 Marine science | Echosounder data, underwater exploration, and marine monitoring |
| ⚡ Energy | Power line inspection, microfossil analysis, and energy applications |
| 🛰️ Earth observation | Remote sensing, SAR, land cover mapping, and change detection |
We offer courses in machine learning and related fields at UiT The Arctic University of Norway.
| Course | Title | Material |
|---|---|---|
| FYS-2010 | Image Processing | |
| FYS-2021 | Machine Learning | 📁 Repository |
| FYS-3012/8012 | Pattern Recognition | 📓 Handbook |
| FYS-3033/8033 | Deep Learning | |
| FYS-3032/8032 | Health Data Analysis | |
| — | Generative AI | 📁 Repository |
We also co-teach STA-2002 (Theoretical Statistics), STA-2003 (Time Series Analysis), and TEK-3601 (Machine Vision). Master's thesis topics are available across all group research themes — see open positions for current proposals.
We organize the annual Northern Lights Deep Learning Conference (NLDL), an international venue that brings together researchers to exchange ideas and present cutting-edge work in deep learning and machine learning. The conference is held each January in Tromsø, Norway, and includes a winter school alongside the main program.
| Edition | Dates |
|---|---|
| NLDL 2027 | January 12–14, 2027 |
Tutorials and educational material from past editions can be found in our repositories, such as GraphMLTutorialNLDL22.
Group members have access to Springfield, our internal GPU-powered Kubernetes compute cluster. The cluster nodes are named after characters from The Simpsons, with the cluster itself named after the town of Springfield. Infrastructure specifications and deployment manifests are maintained in the springfield repository.