Scholar
Joel Oskarsson
Google Scholar ID: YQaxGpkAAAAJ
Linköping University
Citations & Impact
Citations
178
H-index
8
i10-index
5
Publications
14
Co-authors
6
list available
Publications
Resume (English only)
Academic Achievements
- Successfully defended PhD thesis titled 'Modeling Spatio-Temporal Systems with Graph-based Machine Learning' in April 2025
- Paper 'Probabilistic Weather Forecasting with Hierarchical Graph Neural Networks' accepted as a Spotlight at NeurIPS 2024
- Paper 'Continuous Ensemble Weather Forecasting with Diffusion models' accepted to ICLR 2025
- Paper 'Diffusion-LAM: Probabilistic Limited Area Weather Forecasting with Diffusion' accepted to the Climate Change AI workshop
- Co-authored preprint 'Uncertainty Quantification of Pre-Trained and Fine-Tuned Surrogate Models using Conformal Prediction' with collaborators from UCL and UKAEA
- Presented research at multiple international venues including NeurIPS 2024, Royal Swedish Academy of Sciences 'AI for Science' symposium, and workshops on machine learning for Earth system modeling
Background
- PhD student at the Division of Statistics and Machine Learning, Department of Computer and Information Science, Linköping University, Sweden
- Research focuses on probabilistic machine learning methods for modeling data with spatial and temporal dependencies
- Motivated by applications of machine learning to earth system modeling, such as weather forecasting and climate modeling
- Addresses unique challenges from earth science: irregular observations, high-dimensional data, complex temporal dynamics, and the need for accurate uncertainty quantification
- Develops tailored machine learning methods for earth science applications by integrating domain knowledge
