The best Scandinavian Ph. D. was awarded at the SCIA 2025 conference on Wednesday, June 25. The candidates were selected by each country (Denmark, Sweden, Norway, Finland and Iceland) could nominate up to two candidates. In total, we received nine highly qualified candidates. The candidates had all defended their thesis within two years of the SCIA 2025 conference. A panel consisting of:
- Professor Rasmus R. Paulsen, DTU Compute, Denmark
- Associate Professor Juho Kannala, Aalto University & University of Oulu, Finland
- Assistant Professor Valgerður Guðrún Halldórsdóttir, University of Iceland
- Professor Fredrik Kahl, Chalmers University of Technology, Sweden
- Associate Professor Kristoffer Wickstrøm, Arctic University of Norway.
Assessed and ranked the candidates based on scientific and technological impact, and relevance for SCIA.
Based on this the winner of the award is:
Frederik Warburg (Technical University of Denmark)
Supervisors: Søren Hauberg (Technical University of Denmark) and Serge Belongie (University of Copenhagen)
Thesis Title: Probabilistic 3D Reconstruction
Comments: Frederik has made substantial contributions to both fundamental and applied AI. His research on uncertainty in neural networks pushes the frontier of neural network’s ability to “know what they don’t know,” a critical step in improving the safety of AI systems. During his PhD, Frederik co-authored 16 papers at top venues such as ECCV (2023), NeurIPS (2023x2, 2022), CVPR (2023, 2020), 3DV (2023), MICCAI (2023), HCOMP (2022), WACV (2022), IROS (2022), ICRA (2021), ICCV (2021), and UAI (2020).
The two runners up for the best thesis are:
- Shuzhe Wang. Deep Learning Methods for Point Matching, Visual Localization and 3D Reconstruction. Finland
- Srishti Gautam. Towards Interpretable, Trustworthy and Reliable AI. Norway