Integrated Geophysics and Machine Learning for Risk Mitigation in Exploration Geosciences

Abstract

Summary We discuss how Machine Learning (ML) can support the integration workflow of heterogeneous geophysical data sets in the process of exploration risk evaluation and/or in the process of field appraisal. Data set includes seismic, electromagnetic, gravity and borehole measurements. We combine sequential geophysical modelling and inversion with statistical and automatic classification approaches commonly used in the field of Machine Learning. We applied this “hybrid approach” to two multidisciplinary geophysical data sets recorded in different geological settings, obtaining encouraging results in both cases.