•    Understanding the challenges that complex structural deposits pose for autonomous block modelling
•    An unsupervised machine learning method is developed that can automatically quantify 3D, local trends (anisotropy) in multi-attribute drilling data
•    Demonstrate pairing of the local trend information with anisotropic block estimations and locally adaptive agglomerative clustering, to generate single attribute and multi-attribute domain proxies. 
•    Explore the workflow as a reliable, autonomous geologic modelling system that can adapt to the complexities of varied geologic scenarios.