2026: Uncertainty Quantification for Scientific Computing and Intelligent Systems

The 26th edition of the Geilo Winter School, in collaboration with the Norwegian AI Cloud (NAIC), took place in Geilo, Norway from Sunday January 18 to Friday January 23, 2026.

Program

Fundamentals of Modern HPC Systems (Hicham Agueny)

In this lecture, we will cover the fundamentals of high-performance computing systems, from CPU and GPU architectures to high-speed interconnects and parallel file systems, including optimization strategies for GPU-aware data movement for performance and scalability.

Uncertainty quantification in physical systems (Kjetil Olsen Lye)

This lecture series will introduce the motivation for and core concepts behind uncertainty quantification in physical systems. We will cover Monte Carlo methods and their variants, and explore alternative algorithms for computing uncertainty in specific settings. We will also use the uncertainty quantification framework to discuss the efficient construction and best-practice use of surrogate models. The presentation will be accompanied by small, self-contained coding exercises to help participants engage with and better understand the material.

Monte Carlo methods and Polynomial Chaos Expansion for uncertainty propagation (David Métivier)

This series of lectures introduces sampling-based methods, focusing on Monte Carlo and Quasi-Monte Carlo techniques for numerical integration and uncertainty propagation. The second part presents Polynomial Chaos Expansion as an alternative approach. Hands-on exercises will reinforce the concepts through computational applications.

Practical Bayesian inference using MCMC methods (Anne Reinarz)

This lecture series will focus on practical Bayesian inference using MCMC methods, including multilevel extensions, supported by hands-on use of UM-Bridge to run complex models efficiently.

European Environment for Scientific Software Installations (EESSI) - from zero to science in minutes (Thomas Röblitz)

Thomas will demonstrate how EESSI allows you to go “from zero to science in minutes” on virtually any Linux system in the world. Beyond the introduction, he will be available to guide you through the hands-on sessions, helping you get your code running smoothly on the available HPC systems.

Lecturers

Hicham Agueny

Hicham Agueny is a senior engineer in high-performance computing (HPC) at the University of Bergen. His work focuses on optimizing HPC and AI applications using GPU accelerators. Previously, he worked as a researcher for nearly a decade in theoretical physics, with an emphasis on mathematical modeling and large-scale computing applied to physics and chemistry. He holds a PhD in computational physics and chemistry from both Sorbonne University and Moulay Ismail University, and a master’s degree in theoretical nanophysics.

Kjetil Olsen Lye

Kjetil Olsen Lye is a Research Scientist at SINTEF Digital in Oslo, working at the intersection of uncertainty quantification, machine learning, and high-performance computing, with a strong focus on industrially relevant applications. He holds a PhD in applied mathematics from ETH Zurich and a master’s degree in pure mathematics from the University of Oslo.

David Métivier

David Métivier is a permanent researcher at INRAE Montpellier, working in Applied Mathematics and Physics. His current research interests focus on statistical models (e.g., Hidden Markov Models) and Deep Learning to tackle environmental and climate change problems. This includes developing stochastic weather generators to estimate risks in agronomy and energy, as well as modeling water quality in lakes and the Seine River. He also works on practical robust statistics. He completed a PhD and postdoctoral work in theoretical physics on mean-field dynamics of particle systems, studying bifurcations, synchronization, instabilities, and partial differential equations. He is a Julia enthusiast and contribute to several packages.

Anne Reinarz

Dr Anne Reinarz is an Associate Professor of Computer Science at Durham University working in the Scientific Computing group. Her research focuses on developing scalable algorithms and software for Bayesian inference, with particular emphasis on high-performance computing and multilevel methods. She is a core developer of UM-Bridge (Uncertainty Quantification and Modeling Bridge), a unified model interface designed to make complex scientific simulations accessible to a wide range of uncertainty-quantification tools.

Thomas Röblitz

Thomas Röblitz is a Senior Engineer at the University of Bergen, bringing years of hands-on experience in operating HPC infrastructure and supporting researchers. As a core contributor to the European Environment for Scientific Software Installations (EESSI), he focuses on ensuring that scientific software is fast, reliable, and accessible everywhere.