Over the past decade, the application of machine learning has become one of the most studied topics in the oil and gas industry. The recent success of some small-scale pilot projects has resulted in more attention being paid to deploying machine learning models at scale for operations. While cloud-based deployment (data backhauled for analysis in the cloud) and edge-based deployment (data being analyzed locally at a remote site without persistent Internet connection) are the two major approaches, each strategy has specific advantages and drawbacks. In this paper, we proposed "mesh learning", a new approach that optimally balances cloud computing and edge computing for deploying machine learning solutions in an industrial internet of things (IIoT) environment. We present how the system enables some new strategies to manage machine learning model lifecycles and how it helps with relevant use cases for the oil and gas industry.

Keywords:
artificial intelligence, cloud computing, use case, upstream oil & gas, neural network, mesh, hierarchical learning, detection, technology conference, application
Subjects:
Information Management and Systems, Artificial intelligence, Neural networks
Copyright 2020, Offshore Technology Conference
You can access this article if you purchase or spend a download.