Arvind Sharma
Dr. Arvind Sharma earned both his BSc in applied geology and MSc in exploration geophysics at the Indian Institute of Technology and his Ph.D. in Geophysics at Virginia Tech.
Leadership: Sharma was promoted to the position of chief geophysicist for data and analytics in 2018 and took on a leadership role in machine learning capabilities. As the person responsible for setting strategy for the company in this sector, he oversees IT, software development, high-performance computing and machine learning.
In taking on the chief geoscientist role, Sharma assumed responsibility for creating a new division within TGS that allows the company to enhance subsurface knowledge through better analysis of subsurface geodata.
Sharma spearheaded the first crowd sourcing project for oil and gas, the Salt Identification Challenge pursued in partnership with Kaggle, an online community of data scientists and machine learners, owned by Google LLC. This was a hugely successful project, with more than 3,800 participants devoting their data science skills to solving one the industry’s biggest challenges to date – how to use 3-D seismic to identify salt bodies in deep water. Using the 80,000 collective responses gained from this competition, Sharma and his team incorporated the results into a new machine learning aid called SaltNet.
Sharma’s vision has resulted in a product that is making a significant contribution to advancing well data technology that will help operators find hydrocarbons faster than ever before by quickly detecting salt in preliminary seismic processing through an image classification algorithm.
Accomplishments: Sharma also led the team that developed Analytics Ready LAS (ARLAS) log prediction algorithms that are part of the TGS AI initiative. ARLAS uses a series of high-density data and blind testing, using more than 75 machine learning models in each basin to produce exceptional details and accuracies greater than 90% for most curves
ARLAS incorporates nearly 2 million wellbores digitized throughout North America, constituting the largest commercial digital log library in the world. Sharma and his team have developed machine learning algorithms that predict missing curve responses in digital well log data. The algorithms calculate missing log curves and fill in gaps to provide complete reservoir coverage from top to bottom, generating results that are standardized and ready for use. Thanks to the leading efforts of Sharma’s team, ARLAS can work from a single Gamma Ray, Bulk Density, Neutron Porosity, Sonic Pwave, or Deep Resistivity curve and produce a complete Quad-Combo suite.
This industry-leading accomplishment is only one of the initiatives being pursued by Sharma and his team. They also are dedicating efforts to use interpretation AI to eliminate cumbersome desktop applications. Sharma is overseeing development of a web-based, AI platform directly from the cloud that allows companies to train the AI to learn and interpret their data for individualized seismic interpretations.
Insight: “As a technology leader, I am committed to the digital transformation of the oil and energy industry and am passionate about introducing new ways to accelerate and improve decision-making so E&P companies can quickly and more efficiently further their developments, production and divestment opportunities.”