EEPS Colloquium: Marine Denolle

Associate Professor Marine Denolle, University of Washington College of the Environment

AI as a new paradigm for Geosciences

Artificial Intelligence (AI) is transforming the geoscience. This presentation explores the integration of machine learning (ML) techniques with physical modeling to address pressing challenges in climate change, natural hazards, and resource management. Key topics include the development of digital twins for the Earth, leveraging deep learning for earthquake detection, and using knowledge-guided machine learning to incorporate physical laws into predictive models. Examples highlight ensemble learning for seismic event detection, sparse sensor reconstruction for data efficiency, and the fusion of multi-modal datasets to advance scientific discovery. I also identify and present potential opportunities for AI to enhance real-time monitoring, improve understanding of Earth systems, and promote sustainability and resilience in the face of global environmental changes. Finally, I will conclude by outlining grand challenges for AI in geoscience over the next decade, emphasizing collaboration and innovation in leveraging multi-dimensional and multi-modal data.

Hosts: Doug Wiens and Walid Ben Mansour

EEPS colloquia are made possible by the William C. Ferguson Fund