Abstracts

The 2018  Oil & Gas Machine Learning Symposium features a full day of dynamic presentations, engaging breakout sessions, and effective networking opportunities.

27 September 2018
07:30 am – 5:30 pm Houston, Texas

John Adamick

Senior VP Data & Analytics | TGS

Leveraging Cloud and Machine Learning Technologies to Transform Seismic and Geoscience Data Use

Abstract: Machine learning has been used by the oil and gas industry for some time but has only recently begun to truly move into exploration workflows.  However, things are now moving fast and the question has quickly become “How can I find more oil and gas using machine learning?”  This presentation will describe one company’s experience in making its seismic and geoscience data library readily available to E&P companies via a Data Lake structure. The importance of creating “analytics-ready” datasets will also be discussed.  The latter part of the presentation will focus on some of the company’s own analytics research to solve a variety of geological and geophysical problems.

Dr. Sumit Gupta

VP, AI, Machine Learning, & HPC | IBM Cognitive Systems

AI & Machine Learning: Opportunities for Oil & Gas E&P

Abstract: We will talk about the recent advances in AI, machine learning and deep learning, and the opportunities for the oil and gas industry to take advantage of these technology advances to improve exploration and production.  The talk will also include an overview of work done by IBM Research on using AI for seismic interpretation, well log analytics, visual analytics for reservoirs, and enhanced oil recovery.

Dr. Mauricio Araya

Sr. Researcher, Computer Science | Shell

Digitalization of Exploration workflows with Machine Learning

Abstract: How machine learning fits into our digitization strategy and a review our experiences when applying deep learning techniques to a set of challenging problems at Shell, from fault detection to velocity modelling. Also, another ground-breaking application will be showcased, this one related to Digital Rock. All introduced applications are implemented on our proprietary platform ShellNet, which is based on open source components.

Satyen Yadav

General Manager, Internet of Things, Machine Learning, Edge Computing Services | Amazon Web Services 

Life on the Edge

Abstract: Today, an increasing number of objects are being connected to the internet at an unprecedented rate. As a result, customers are collecting vast amounts of valuable IoT data that was previously unavailable. We are already seeing the benefits of applying machine learning models to process data at the source where it is being generated- farmers predict crop yield, power companies predict energy demand, vehicles can identify distracted drivers, and doctors deliver improved care with real-time insight from medical devices … the possibilities for applying intelligence at the edge are countless. However, processing and analyzing this vast amount of IoT data is not possible with traditional business intelligence tools. In this session, we will showcase how customers can use AWS’s IoT, Artificial Intelligence, and machine learning services to gain predictive insights and take intelligent, real-time actions on their IoT data, from the cloud to the edge.

Dr. Tom Smith

President & CEO | Geophysical Insights

Machine Learning: New Tools and Thinking for Geoscientists

Abstract: The topic is machine learning (ML) in seismic interpretation. Seismic reflection data, the indisputable champion in reducing risks in drilling for oil and gas, is well suited for ML. In this talk we survey new tools for this technology and include examples of convolutional neural networks and self-organizing maps. We present three kinds of seismic samples for ML – simple, multi-attribute and self-defining – to discuss how ML classification results in geobodies as the goal of both supervised and unsupervised learning. We illustrate geobody examples of geobody filtering and geobody shape classification. But these are just tools. Robust answers are more important to avoid over-fitting which is generally minimized with multi-model cross-validation techniques. Geoscientists must reinvent themselves to embrace these new tools. We conclude with a discussion of places at the new workbench for geophysicists, geologists and data scientists in this new era.

Dr. Christian Noll

Geoscience Manager, Advanced Analytics & Emerging Technology | Anadarko

Embracing a future of advanced analytics in Geoscience: a view from Anadarko

Abstract: At Anadarko, we are actively integrating advanced analytical techniques such as machine learning into an array of next-generation toolsets that are targeted to directly enable our corporate strategy, which spans the DJ Basin in Colorado, the Delaware Basin in West Texas, and the Deepwater Gulf of Mexico. By integrating Data Scientists into defined teams with traditional Geoscientists, we have been able to define a systematic maturation process where solutions evolve from ideas to proof of concept, to local prototypes and eventual enterprise products for deployment into our asset teams. Our journey has been one of trial and error, learning from both our successes and failures in the past 18 months while we explore new and creative digital solutions to enable this corporate strategy. It is clear, however, that the process is providing exciting solutions of unprecedented scale, accuracy and efficiency for our asset and technology teams across our portfolio, illustrating that the future of machine learning in Geoscience is really already upon us.

Dania Kodeih

Industry Technical Analyst | Microsoft

Applying AI in Oil and Gas Scenarios, What, Where and How

Abstract: In recent years, AI has become an established class of technologies encompassing machine learning, visual analytics and deep neural networks. It includes an expanding catalog of cognitive capabilities such as natural language processing, artificial and custom vision and digital assistants. These technologies are being introduced to various areas of the oil and gas business, with everyday tools incorporating one or more of these capabilities. In this session, we will explore examples of where AI is delivering business value, unpack the components necessary to enable it and describe the role that edge and cloud play in delivering intelligent solutions.

Ruben Rodriguez Torrado

Artificial Intelligence AdvisorRepsol

Artificial Intelligence for Field Development Plan Optimization

Abstract: This presentation will focus on a novel approach for planning the development of hydrocarbon fields, considering the sequential nature of well drilling decisions, and the possibility to react to future information. In a dynamic fashion, we want to optimally decide where to drill each well conditional on every possible piece of information that could be obtained from previous wells. In addition, We will introduce the main limitations to apply AI in the upstream industry. We show that our new approach leads to better results compared to the current standard in the oil and gas (O&G) industry.

Preston Cody

Head of Analytics LabWood Mackenzie

Signal vs Noise: the (dis)connection between Predictive Analytics and Value

Abstract: There is great potential for Machine Learning to unlock new value in the development of unconventional resources, but the potential is limited by the underlying data those methods rely on. While statistical approaches show promise, there is a steep learning curve still ahead, with questions remaining as to how to make “Analytics” actionable by decisions makers. Examples from a case study of predictive analytics on Bakken well performance demonstrate that while ML methods enable complex pattern detection across diverse datasets, limitations of the underlying data make it difficult to discern signal from noise and the risk of misinterpreting results is high.

Fabian Rada

Technical Advisor to PEMEX | Petroleum Oil & Gas Services

Net Reservoir Discrimination in Southern Mexico using Machine Learning

Abstract: A new approach has been applied in order to discriminate net reservoir using seismic multi-attribute analysis at single sample resolution. Also, a big effort was done in the calibration stage by means of standard visualization techniques and particularly with bivariate statistical analysis from petrophysical evaluations from well logs and classification volume related to lithological contrast detection, to ensure a fair representation of the reservoir static properties and reduce the uncertainty related to reservoir distribution and storage capacity. Some uses of the results might be in the adjustment or refinement of the sedimentary model to more accurately identify the lateral and vertical distribution of the facies of economic interest and capture them within maps for each stratigraphic unit, also, in making decisions when proposing new locations and reducing the uncertainty associated with field exploitation, among others.

Dr. Iván Marroquín

Sr. Research Geophysicist | Geophysical Insights

Rise of Machine Learning – Enabled Seismic Interpretation

Abstract: The following quote from Calvin and Hobbs – a comic strip – describes how well automation efforts can make processes more efficient: “Day by day nothing seems to change. But pretty soon … everything’s different”. In the last couple of years, Machine Learning has rapidly become an important pillar of many industries’ digital transformation. Companies have increasingly been looking at new technologies related to data analytics with the objective to reduce costs, improve efficiency, and minimize downtime. The Oil and Gas industry is also not immune to this transformation. Machine Learning, a subfield of the big data revolution, uses algorithms that learn from the data and independently adapt to produce reliable and repeatable results. The ultimate goal is to derive performant predictive models. Therefore, the excitement about using Machine Learning is the ability to discover and identify interesting patterns in geological, geophysical, engineering data and use them for a more guided exploration and production of hydrocarbons.

At Geophysical Insights, we are committed at developing technological solutions to support geoscientists in their daily interpretive tasks. In this presentation, I share with you our vision and research advances in Machine Learning to bring forward a software platform oriented to solve specific interpretation challenges.

Paul Holzhauer

Director of Oil and Gas Nvidia

Certainty About Uncertain Things

Abstract: We are continually looking for ways to deal with uncertainty and minimize risk. In fact, within the oil & gas industry, risk often takes on more weight than reward. So what do you do? Do you use simulation, analytic models, subsurface models, prototype construction, or analysis of field data for an actual product, which is nothing new? However, knowing the important variables in near real-time can offer benefits. Knowing which variables to control, measure and ignore gives you the insight into new information and knowledge.

Artificial intelligence is the result of the information era. Advances in machine learning and deep learning have combined with more powerful and an ever-expanding pool of data that allows for AI to reach companies across all industries. Technology advancements continue to address optimization issues. NVIDIA believes that the ability to leverage AI will become a defining attribute in gaining a competitive advantage for companies in the near-term and will deliver a new levels in productivity in the future.

Randall Hunt

Sr. Staff Geoscientist Range Resources

Machine-learning analysis of 3D seismic for detection of potential deep karsting in the Pt Pleasant

Abstract: Rocks of the Utica/Pt Pleasant Shale play of Appalachia are carbonate-rich, ranging from calcareous shales (Utica), to marls (Pt Pleasant). The play overlies the legacy Trenton/Black River play, which is known to be affected by extensive vertical zones of hydrothermal alteration and deep karsting. Drilling results, correlated with 3D seismic attributes, suggest these processes could have affected the Utica/Pt Pleasant. Paradise machine-learning techniques are used to locate these zones of deep karsting, aiding the detection of major geohazards and exploratory drilling sweet spots.

Bill Abriel

Past President (SEG), CEOOrinda Geophysical

The Role of Cooperative Strategies for Advancing Machine Learning in Applied Geophysics

Abstract: Machine Learning (ML) has the potential to generate significant breakthroughs in applied geophysics. As new concepts emerge and evolve at a rapid pace, it will challenge both the business model and the time-to-application differently for international operators, service companies, national oil companies and independent operators. These conditions pose an opportunity to employ cooperative strategies in both the development and application of ML. Value through cooperation can be gained via publications, conferences and online forums. Advancements can also be significantly accelerated by pooling professional and financial capital investments for significant leverage in cooperative research while still retaining proprietary advantage. This talk covers the anticipated opportunities for such industrial cooperation based on past success in petroleum geophysics, and predictions of the near-future.

Become a Sponsor

Gain valuable brand awareness and recognition as a leading company in data analytics and machine learning in the upstream. Sponsorships available at several levels.  Contact us for more information.

Contact Us