UN/Costa Rica Workshop on Machine Learning for Space Weather & GNSS

The UN/Costa Rica Workshop on Machine Learning Applied to Space Weather and GNSS brought together an international community of researchers, forecasters, and policy experts to explore the growing role of artificial intelligence in monitoring and predicting solar and geomagnetic activity and its effects on satellite navigation systems. The workshop, organized under the auspices of the United Nations, provided a unique setting at the intersection of fundamental solar physics, data-driven modelling, and operational space weather services.

Prof. Dario Del Moro represented the Solar and Space Physics group of the University of Rome Tor Vergata with a presentation titled “Probabilistic Solar Corona Evolution Forecasting with Denoising Diffusion Models”. The contribution introduced a deep learning approach based on Denoising Diffusion Probabilistic Models (DDPMs) applied to EUV imagery from the NASA Solar Dynamics Observatory, framing corona evolution forecasting as a probabilistic video prediction problem. A central theme of the talk was the complementarity between deterministic AI approaches — such as large-scale solar foundation models — and probabilistic ensemble methods that are capable of quantifying uncertainty and supporting risk-based decision-making. The work is carried out in the framework of the CORNERSTONE project, funded by the Italian Ministry of University and Research.