The inaugural meet of the Society of Petroleum Engineers PD2A technical section was held in San Antonio’s Iron Cactus Grill this month. PD2A promotes the application of data-driven modeling, data mining and predictive analytics R&D in the upstream.
Keynote speaker Matt Denesuk (IBM) set the scene, describing the ‘three big things’ in data today as 1) a meeting of the physical and digital worlds, 2) data-driven analytics and 3) heterogeneity and integration. The data driven approach, a.k.a. machine learning means ‘doing stuff with data without understanding the underlying processes.’ ‘Big data’ is essentially a marketing term. What is more interesting is the relationship between data complexity and size. This separates the data boys from the men—with Google fine for low complexity, high volume data and IBM’s Watson coming into its own at the other end of the spectrum.
IBM’s offering to industry actually combines the data-driven approach with domain knowledge. This is helping overcome resistance to previous attempts at machine learning which generated too many false positives.
Adco’s Fareed Abdulla then presented several case histories of Adco’s use of machine learning which we will report on in next month’s Oil IT Journal. The SPE PD2A meet was sponsored by BP, ExxonMobil and Hess. More on the embryonic section here.
© Oil IT Journal - all rights reserved.