We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

stat.CO

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo

Physics > Geophysics

Title: Probabilistic model-error assessment of deep learning proxies: an application to real-time inversion of borehole electromagnetic measurements

Abstract: The advent of fast sensing technologies allows for real-time model updates in many applications where the model parameters are uncertain. Bayesian algorithms, such as ensemble smoothers, offer a real-time probabilistic inversion accounting for uncertainties. However, they rely on the repeated evaluation of the computational models, and deep neural network (DNN) based proxies can be useful to address this computational bottleneck. This paper studies the effects of the approximate nature of the deep learned models and associated model errors during the inversion of extra-deep borehole electromagnetic (EM) measurements, which are critical for geosteering. Using a deep neural network (DNN) as a forward model allows us to perform thousands of model evaluations within seconds, which is very useful for quantifying uncertainties and non-uniqueness in real-time. While significant efforts are usually made to ensure the accuracy of the DNN models, it is known that they contain unknown model errors in the regions not covered by the training data. When DNNs are utilized during inversion of EM measurements, the effects of the model errors could manifest themselves as a bias in the estimated input parameters and, consequently, might result in a low-quality geosteering decision. We present numerical results highlighting the challenges associated with the inversion of EM measurements while neglecting model error. We further demonstrate the utility of a recently proposed flexible iterative ensemble smoother in reducing the effect of model bias by capturing the unknown model errors, thus improving the quality of the estimated subsurface properties for geosteering operation. Moreover, we describe a procedure for identifying inversion multimodality and propose possible solutions to alleviate it in real-time.
Subjects: Geophysics (physics.geo-ph); Applications (stat.AP); Computation (stat.CO); Machine Learning (stat.ML)
Journal reference: Geophysical Journal International, Volume 230, Issue 3, September 2022, Pages 1800-1817
DOI: 10.1093/gji/ggac147
Cite as: arXiv:2205.12684 [physics.geo-ph]
  (or arXiv:2205.12684v1 [physics.geo-ph] for this version)

Submission history

From: Sergey Alyaev [view email]
[v1] Wed, 25 May 2022 11:44:48 GMT (1328kb,D)

Link back to: arXiv, form interface, contact.