Skip to content


Track: Data Mining

Customizing Natural Language Processing Models for Domain-specific Entity Recognition for Geoscience

Monday, April 12, 4:30-5:10pm EDT

Geospatial data is used extensively by exploration geologists to reduce uncertainty related to finding and producing oil and gas resources. Much of the data needed by geologist is trapped in unstructured text documents. Geologists need a mechanism to identify and extract key terms from text in order to efficiently filter documents and find the data needed to improve decision making. This presentation demonstrates how a generic natural language processing (NLP) model was extended and training to identify geologic-specific terms. The trained model was deployed as an API endpoint and the presentation will show how the endpoint can be invoked to provide real time classification of unstructured text. The audience should leave with a better understanding of creating and deploying domain-specific natural language processing models in order to start applying the same principles to address their business needs.

Kyle Jones image

Kyle Jones

Kyle Jones

Principal Solutions Architect at Amazon Web Services

Kyle Jones is a Principal Solutions Architect at Amazon Web Services (AWS) specializing in the Energy industry. Prior to joining AWS, he worked across the energy value chain ExxonMobil. Before ExxonMobil, he was a diplomat and special assistant to the U.S. Ambassador to Canada. He is a project management professional (PMP), certified analytics professional (CAP), and AWS Certified Machine Learning Specialist and holds a doctorate in systems engineering. He is a Lecturer of business analytics at a Sam Houston State University.