Deep learning tools are helping decision-making in virtually every industry, but traditionally, when applied to geoscience, they rely on a network trained on unknown data. Processing can take months, and you can only supply feedback to the network once training is complete. A network trained on data by geographically separated interpreters with variable expertise, will not work where you need it to. Bluware is pioneering the concept of ‘interactive deep learning’. You remain in the ‘driver’s seat’ and interact with the network via real-time adjustments to the training set as it learns and adapts.
‘Black Box’ vs ‘Interactive’ Deep Learning
