January Technical Meeting: April Moreno-Ward

Presentation Title: Machine Learning for Geoscientists: Where to Start for Improved Deepwater Channel Interpretations

As machine learning becomes more mainstream in geoscience research and workflows, we see time and again how the identification and interpretation of geologic environments can be greatly improved through systematic multi-attribute analysis. We should be mindful of the standards of seismic interpretation workflows though, beginning with the choice of seismic attributes and using appropriate advanced statistical and algorithmic techniques, and understanding that the goal of the improved geologic interpretation process cannot happen without a knowledgeable geoscientist. Our reminder is to be mindful of finding a balance between quantitative and qualitative evaluation methods, don’t fall into the trap of becoming over-reliant on machine learning output nor underappreciate the geologic expertise necessary for a defendable interpretation. Additionally, there is the added complexity of evaluating the issue of bias and how it 1) infiltrates the data through statistical methods and 2) through the interpreter’s knowledge/experience.

Speaker: April Moreno-Ward

April Moreno-Ward is a PhD candidate in Geophysics at the University of Oklahoma, Mewbourne College of Earth and Energy School of Geoscience, under the guidance of Dr. Heather Bedle. She received her MSc. in Geology from the University of Texas at Arlington with a focus on fluvial geomorphology. Combining her six years of active petroleum geology consulting with her academic background, she now focuses on applying machine-learning workflows to enhance geoscience research and provide guidance to make these advanced technologies more accessible within her field of study. Her research aims to develop more intuitive machine learning approaches to bridge industry requirements and needs with novel technologies.