SmartTensors: Unsupervised and physics-informed machine learning framework for the geoscience applications
B. Ahmmed, V.V. Vesselinov, M.K. Mudunuru
Second International Meeting for Applied Geoscience & Energy2022DOI 10.1190/image2022-w8-01.1
Summary
SmartTensors is presented as an unsupervised and physics-informed machine learning framework for geoscience applications built around constrained matrix and tensor factorization methods. The platform emphasizes robustness and interpretability and is demonstrated on geothermal prospectivity analysis in the Great Basin, where regional geochemical data are used to infer subsurface reservoir characteristics.