Dr. Priyank Jaiswal - Associate Professor, Oklahoma State University

  • January 12, 2023
  • 11:30 AM - 1:00 PM
  • Baxter's InterUrban Grill and online (zoom) option

Registration

  • In-person attendance at Baxter's available for TGS members at GST member price.

Registration is closed

January 12th -  Baxter's Interurban Grill (in-person gathering)

Title:     XRD-XRF Integration using Machine Learning

Dr. Priyank Jaiswal - Oklahoma State University

Abstract:

In Earth Science, integrating non-invasive continuous data streams with discrete invasive measurements remains an open challenge. In this paper, we address such a problem, that of predicting whole-core mineralogy using discrete measurements with the help of machine learning (ML). Our targets are sparsely sampled mineralogy from X-Ray Diffraction (XRD) and features are continually sampled elemental oxides from X-Ray Fluorescence (XRF). Both datasets are acquired on a core cut from Mississippian age mixed siliciclastic-carbonate formation in the US mid-continent. The novelty of this paper is predicting multiple classes of output targets from input features in a small multi-dimensional data setting. Our workflow has three salient aspects. First, it shows how single output models are more effective in relating selective target-feature subsets than using a multi-output model for simultaneously relating the entire target-feature set. Specifically, we adopt a competitive ensemble strategy comprising three classes of regression algorithms - elastic-net (linear regression), XGBoost (tree-based), and feedforward neural networks (non-linear regression). Second, it shows that feature selection and engineering, when done using statistical relationships within the dataset and domain knowledge, can significantly improve target predictability. Thirdly, it incorporates k-fold cross-validation and grid-search-based parameter tuning to predict targets within 4-6% accuracy using 40% training data. Results open doors to generating a wealth of information in energy, environmental and climate sciences where remotely sensed data is cheap and abundant, while physical sampling may be limited due to analytical, logistic, or economic issue.


Bio: 

Dr. Priyank Jaiswal

   Associate Professor - Boone Pickens School of geology

   Director Professional Science Master in Geoscience


Education:

   Ph.D. 2008 Rice University

   B.S. 1999 Indian Institute of Technology Kharagpur


Specialization:

   Petroleum Systems Gas Hydrates

   Rock Physics Near Surface

   Poroelasticity


Professional Affiliations

   Society of Explorations Geophysicists (SEG)

   American Association of Petroleum Geologists (AAPG)

   Geological Society of America (GSA)

   American Geophysical Union (AGU)

   European Association of Geoscientists and Engineers (EAGE)



 



  



"Geophysical Society of Tulsa (GST)" is a 501(c)3 non-profit organization. Oklahoma, P.O. Box 2784, Tulsa, OK 174101

Powered by Wild Apricot Membership Software