Gazprom Neft, Saudi Aramco, Tatweer Petroleum and Schlumberger hold international hackathon on machine learning in upstream operations

6–7 December will see digital divisions from oil and gas companies across Russia, Europe and the Middle East, including teams from Saudi Aramco, Tatweer Petroleum and Schlumberger, competing in the GeoHack 2019 hackathon in Manama, Bahrain. Microsoft is supporting the project as a technology partner for the event, providing participants with the necessary computing capacity.

The GeoHack will create an environment for software developers, engineers and geoscientists to spend intensive hours hacking, testing and experimenting with the latest advancements in machine learning (ML) algorithms, using these in conjunction with subsurface seismic logs, cores and other data to solve data production, geological interpretation and quantitative reservoir characterisation problems. The competition’s challenges will be set by the European Association of Geosciences and Engineers (EAGE), together with Russia’s leading company in AI in the oil and gas industry, Gazprom Neft.

The competition will give hackathon participants 48 hours to outstrip basic Gazprom Neft ML models, which are built based on three sets of anonymised data from real oilfields. Teams will need to solve several tasks in the course of training their models, including detecting formation horizons, working with seismic programmes, and predicting reservoir porosity based on petrophysical research data. The event is a great opportunity for companies to benchmark the performance of their ML algorithms, expose the most promising technologies, and understand the advantages and drawbacks of machine learning.

GeoHack will bring together ML enthusiasts, students, geoscientists and industry specialists to exchange ideas and develop solutions for the many complex earth-imaging issues the industry is facing today. The goal of this project is to develop open source tools that will contribute towards technological progress in the oil and gas industry. Machine learning has proved its efficiency in many upstream operations and is currently at its peak potential for use in oil extraction projects.

Find out more about GeoHack 2019 here.