EXPERT-GUIDED

HOW OUR MACHINE LEARNING WORKS

Our Expert-Guided Machine Learning approach leverages experts’ knowledge to power advanced analytics on the oilfield. Experts mark problems and solutions across well history and we use this labeled data to train models to detect and solve the same problems. Expert-Guided Machine Learning helps you label data across hundreds of wells quickly and effectively and uncover operational issues through the power of automation.

EMPOWER YOUR OPERATORS AND OILFIELDS WITH EXPERT-GUIDED MACHINE LEARNING

Field operators, engineers, and subject matter experts bring years of experience in learning, diagnosing and fixing problems. Empower them to accurately identify problem conditions by participating in model training.

LABEL EVENTS QUICKLY

Our artificial lift SMEs begin by marking events and conditions on your historical data for a small group of wells. Our models then generate labels which feed the machine learning process. We then expand this process to run on additional groups of wells until all data has been labeled, giving you the power to label years with of events quickly.

MODEL COLLABORATION

Your operators will have a shared place within Production Analytics to validate model results and their reviews and comments all serve to make models smarter. Our Auto-Validation tools will also speed up the training process on certain types of events. Once they match what your operators want, we get them running on live data to deliver real-time value.

CONTROL THE MODELS

Tune the models that can differentiate between critical events of interest and seemingly similar false positives. Models adaptively adjust their sensitivity according to the experts’ ratings of past and current diagnostic indicators. This ensures that the model emulates expert judgment, unlike other methods that follow traditional data science procedures.

AUTOMATE DATA LABELING

OUR EXPERT-GUIDED MACHINE LEARNING TECHNOLOGY STARTS TO LABEL EVENTS FOR YOU OVER TIME

Expert-Guided Machine Learning includes auto-validation features that save SMEs significant time. Models can be validated efficiently by observing well performance. If a model detects an expected condition, they can auto-validate their own diagnostic indicators. As your models mature, your experts will need to review an increasingly smaller amount of events.

ONE PLACE TO COLLABORATE

GIVE YOUR TEAM A SHARED ENVIRONMENT FOR TRAINING AND DEPLOYMENT

Harnessing advanced analytics requires the alignment of your data scientists to work with operators and engineers to inform models, deploy them and take action. Expert-Guided Machine Learning provides the shared space for these specialists to train and deploy models. Multiple team members can readily review and comment on model results in a simple, web-based application.

EVERYTHING OPERATORS NEED TO UNDERSTAND RECOMMENDATIONS

To increase adoption and act on model recommendations, operators need to understand the trends and rationale behind them. They also need visualizations that highlight the timing of changes in key signals, performance trends and other underlying indicators so they gain confidence in models and make timely decisions.

GAIN BUY-IN BY EMPOWERING EVERY OPERATOR ON YOUR TEAM

EGML can help you bring your entire organization on board with the objectives of the digital oilfield. Your teams will gain active insights into assets and learn to use Advanced Analytics to improve operational success. As your organization starts seeing the benefits scale, motivation will rise and adoption will become more widespread.