Our Research Approach
Our research blends Earth sciences and data science, with a specialization in computer vision with deep learning, a branch of artificial intelligence focused on interpreting and analyzing visual data such as geological images, satellite data, and seismic volumes. We integrate deep learning and data sciences with the meticulous acquisition of quantitative field data as well as advanced isotopic measurements in the lab. This dual approach ensures we master the entire data pipeline from acquisition all the way to machine learning applications: this is the cornerstone of our success.
Our research centers on applying AI to achieve two interconnected objectives: (1) understanding both paleo and current climate and environmental changes and (2) advancing the transition to cleaner geo-energies to meet global energy demands. These efforts directly support the United Nations Sustainability Goals, SDG 13 (Climate Action) and SDG 7 (Affordable and Clean Energy). More details about these two broach research themes are available below.
AI For Climate Action
Climate change poses one of the greatest challenges to humanity, with rising temperatures, extreme weather events, and environmental shifts impacting ecosystems and societies worldwide. Understanding both past and present climate change is essential for predicting future impacts and guiding adaptation strategies. Our research leverages Artificial Intelligence (AI) to analyze complex climate datasets, from paleo-environmental records to modern Earth Observation data, enabling more accurate climate reconstructions and real-time monitoring of environmental changes.
AI For the Energy Transition
The other side of the ‘climate coin’ is energy: the world’s energy demand is increasing, and to usher solutions to greenhouse warming requires supporting the global shift towards ‘NetZero’ carbon emissions. We leverage AI to tackle challenges in automated subsurface characterization for carbon capture and storage, hydrogen storage, and geothermal energy. We also pioneered the use of advanced isotopic techniques coupled with numerical modelling and AI to better constrain the thermal histories of sedimentary basins.
