The CORNERSTONE project was represented at the United Nations/Costa Rica Workshop on Machine Learning Applied to Space Weather and Global Navigation Satellite Systems, held in San José, Costa Rica, from 16 to 20 February 2026. The workshop, organised by the United Nations Office for Outer Space Affairs (UNOOSA) in cooperation with the Space Research Centre (CINESPA) of the University of Costa Rica, brought together scientists, engineers, and early-career researchers from across the Latin American region and beyond to advance the application of machine learning to space weather forecasting and GNSS.
Prof. Dario Del Moro (Università degli Studi di Roma Tor Vergata) delivered a presentation on the probabilistic forecasting of solar corona evolution, a research thread that complements CORNERSTONE’s core flare prediction pipeline. The work leverages Denoising Diffusion Probabilistic Models (DDPMs) combined with a Light Video UNet architecture applied to EUV solar imagery from NASA’s SDO/AIA instrument. Unlike deterministic approaches, this framework generates an ensemble of physically plausible future solar states, offering a richer uncertainty quantification that is crucial for operational space weather services.
The participation reflects CORNERSTONE’s commitment to disseminating its results in international venues dedicated to capacity building and knowledge transfer — particularly in communities where the societal impact of space weather on infrastructure, navigation, and communications is most acute.
The workshop, co-sponsored by COSPAR, INGV, SCOSTEP, and the International Committee on GNSS (ICG), convened approximately 80 participants and served as a platform for forging new collaborations between research institutions and space weather service providers worldwide.