REALI3SE · Bridging Igneous Petrology and Machine Learning through Intersectoral, Interdisciplinary, and Intercontinental Staff Exchanges

The REALI3SE Staff Exchange project will advance research and discovery at the intersection of igneous petrology, geochemistry, and artificial intelligence, bringing together Earth scientists, data scientists, and industry partners from Europe, South America, Africa, and the USA.By combining multimodal data acquisition with advanced machine-learning (ML) methodologies, REALI3SE will transform how geological processes are observed, modeled, and interpreted, from magma generation and evolution to volcanic eruptions and ore formation. Through structured secondments, researchers from academia, research institutions, and industry will jointly develop standardized protocols for data harmonization, spatial co-registration, and fusion of multimodal datasets derived from different analytical techniques, including EPMA, LA-ICP-MS, Raman and IR spectroscopy, µ-XRF, and synchrotron-based imaging. Integrating these datasets through physics-informed ML models will enable the development of scalable workflows for extracting meaningful patterns from complex geochemical archives, improving our understanding of magmatic processes, predictive models for volcanic hazards, and strategies for the sustainable exploration of critical raw materials.Beyond its scientific objectives, REALI3SE will strengthen long-term collaboration between European and international partners, bridging expertise in volcanology, ore geology, data science, and AI. Its outreach, dissemination, and open-science activities will ensure that the resulting knowledge and digital tools are accessible to the wider scientific community, industry stakeholders, and policymakers. Ultimately, REALI3SE will prepare a new generation of Earth scientists, capable of integrating AI-driven methods into the responsible management of geological resources.