URTeC 2020 was a unique occasion to learn the latest on unconventional technologies and beyond. The need to streamline and simplify processes is a common denominator in most discussions regarding artificial intelligence and machine learning. DeepCast, through his CEO Hector Klie, was excited to contribute in two distinct venues: participating in the panel discussion "Data Analytics and Physics-Infomed Models: The Next Generation Takes Shape" and presenting the paper "Transfer Learning for Scalable Optimization of Unconventional Field Operations".
The panel count on an audience of more than 800 people who enjoyed a vivid discussion of experts on how physics may be reproduced or discovered from data. Dr. Klie had the opportunity to exchange thoughts with Professor Rami Younis (University of Tulsa), Cedric Fraces (Tenokonda), and Ruben Rodriguez (OriGen.AI). The session was moderated by Alejandro Lerza (Chevron).
The URTeC paper N. 2719, coauthored by Dr. H. Klie and Dr. B. Yan, describes a successful application of transfer learning for scaling the optimization on multiple pads or fields. The idea is to leverage previous learnings from the solution of each reservoir project and start ahead in the next project. This is the first account of transfer learning in Oil and Gas and will surely trigger derived works along this direction.
This seminar took place via Zoom on April 14, 2020