Google TensorFlow, Scikit-learn extend Ikon’s RokDoc

Machine learning used in facies classification and shale porosity evaluation.

Ikon Science has embedded Google’s TensorFlow open source toolbox for machine learning into its RokDoc quantitative seismic interpretation flagship. TensorFlow is exposed via a Python API that allows for ‘rapid identification of novel predictive relationships across large, multi-disciplinary, multi-scale datasets in both shale and conventional petroleum systems.’ The new functionality was introduced with the RokDoc 6.5.2 release. Along with TensorFlow integration the Python interface also includes Scikit-learn0702, another ML toolbox that is used, inter alia, by Spotify, and, more a propos, by France’s Inria for energy-related predictive modeling (see above).

Ikon’s Ehsan Zabihi Naeini speaking at the recent London Geological Society’s big data event presented two use cases for the new tool in facies classification and shale volume and porosity prediction. Naeini observed that while there is no novelty in such applications, what is new is how easy implement has become with modern machine learning tools and the publicly-available libraries.

Click here to comment on this article

Click here to view this article in context on a desktop

© Oil IT Journal - all rights reserved.