The Stanford Centre for Earth Resource Forecasting (SCERF), led by Jef Caers and Tapan Mukerji, invited Deep Earth Imaging to send a delegation to their annual affiliates meeting. Luk Peeters, Guillaume Rongier, Roman Beloborodov and James Gunning travelled to Stanford to engage with their researchers and get updates on their ongoing research. Like Deep Earth Imaging, SCERF is active across various domains, including oil and gas, mineral resources and groundwater. Much of their ongoing research is organised according to the Bayesian Evidential Learning (BEL) protocol, which provides a systematic, pragmatic approach to uncertainty quantification for problems with computationally expensive forward models and spatially variable parameters. The BEL protocol combines a wide variety of dimensionality reduction, sensitivity analysis, machine learning and sampling techniques. SCERF’s ambitions are to develop this into an automated uncertainty quantification workflow.
The trip was a perfect opportunity to get to know the students and post-docs at SCERF and we look forward to continue our collaboration and exchange students and postdocs between DEI and SCERF.