How Reusing Trained Machine Learning Models Accelerates and Improves the Work of Operational Geoscientists

We show examples of trained machine learning models that are applied ‘AS IS’ to solve similar problems on unseen datasets. A huge benefit of this methodology is the speed of application. Operational geoscientists simply select a model from a library of trained models and apply this to their data. Within minutes they have a result that they can either use, or discard. In many cases, the alternative solution involves expensive and time-consuming reprocessing efforts. We show examples of detecting features, enhancing data, and predicting missing data.