English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Deep learning techniques have been shown to be extremely effective forvarious classification and regression problems, but quantifying the uncertaintyof their predictions and separating them into the epistemic and aleatoricfractions is still considered challenging. In oil and gas exploration projects,tools consisting of seismic, sonic, magnetic resonance, resistivity, dielectricand/or nuclear sensors are sent downhole through boreholes to probe the earth'srock and fluid properties. The measurements from these tools are used to buildreservoir models that are subsequently used for estimation and optimization ofhydrocarbon production. Machine learning algorithms are often used to estimatethe rock and fluid properties from the measured downhole data. Quantifyinguncertainties of these properties is crucial for rock and fluid evaluation andsubsequent reservoir optimization and production decisions. These machinelearning algorithms are often trained on a "ground-truth" or core database.During the inference phase which involves application of these algorithms tofield data, it is critical that the machine learning algorithm flag data as outof distribution from new geologies that the model was not trained upon. It isalso highly important to be sensitive to heteroscedastic aleatoric noise in thefeature space arising from the combination of tool and geological conditions.Understanding the source of the uncertainty and reducing them is key todesigning intelligent tools and applications such as automated loginterpretation answer products for exploration and field development. In thispaper we describe a methodology consisting of a system of dual networkscomprising of the combination of a Bayesian Neural Network (BNN) and anArtificial Neural Network (ANN) addressing this challenge for geophysicalapplications.

Dual Neural Network Architecture for Determining Epistemic and Aleatoric Uncertainties