1887

Abstract

Summary

Several studies on deep generative models for use in geomodeling show encouraging results with binary training data. An important question is what type of training data to use, since realistic 3D geology with natural variability is difficult to create. The advent of multiple types of remote sensing data of subaerial and subaqueous sedimentary patterns provides new possibilities in this context. Here, we train a Wasserstein GAN using 20,000 multispectral satellite images of subsections of 40 modern river deltas. The generated output has three facies and a background facies, and all quantitative evaluation methods of the unconditional output show a close overlap between the model and training data distributions. Standard MCMC sampling conditional on soft and hard data works well as long as the likelihood model is balanced against the prior model. Transfer learning, i.e. fine-training a small subset of the network parameters on smaller dataset of interest, such as highly non-stationary images of river deltas with similar characteristics, also shows promising results.

Loading

Article metrics loading...

/content/papers/10.3997/2214-4609.201902196
2019-09-02
2024-03-28
Loading full text...

Full text loading...

References

  1. Abed-Elmdoust, Armaghan, SinghArvind, and YangZong Liang. 2017. “Emergent Spectral Properties of River Network Topology: An Optimal Channel Network Approach.” Scientific Reports.
    [Google Scholar]
  2. Arjovsky, Martin, ChintalaSoumith, and BottouLéon. 2017. “Wasserstein Generative Adversarial Networks.” Proceedings of The 34th International Conference on Machine Learning:1–32.
    [Google Scholar]
  3. Deutsch, Clayton V., and PyrczMichael J.. 2012. Geostatistical Reservoir Modeling.2nd ed. Oxford University Press.
    [Google Scholar]
  4. Dubuisson, Marie-Pierre, and JainAnil K.. 1994. “A Modified Hausdorff Distance for Object Matching.” In ICPR,, 566–68.
    [Google Scholar]
  5. Dupont, Emilien2018. “Generating Realistic Geology Conditioned on Physical Measurements With Generative Adversarial Networks.” arXiv (8 Feb 2018).
    [Google Scholar]
  6. Gomez-Hernandez, Jaime J., and WenXian Huan. 1998. “To Be or Not to Be Multi-Gaussian? A Reflection on Stochastic Hydrogeology.” Advances in Water Resources.
    [Google Scholar]
  7. Goodfellow, Ian2014. “Generative Adversarial Nets.” Advances in Neural Information Processing Systems27: 2672–80.
    [Google Scholar]
  8. Gulrajani, Ishaan2017. “Improved Training of Wasserstein GANs Montreal Institute for Learning Algorithms.” In NIPS,.
    [Google Scholar]
  9. Laloy, Eric, HéraultRomain, JacquesDiederik, and LindeNiklas. 2018. “Training-Image Based Geostatistical Inversion Using a Spatial Generative Adversarial Neural Network.” Water Resources Research.
    [Google Scholar]
  10. Orton, G J, and ReadingH G. 1993. “Variability of Deltaic Processes in Terms of Sediment Supply, with Particular Emphasis on Grain Size.” Sedimentology40(3): 475.
    [Google Scholar]
  11. Salakhutdinov, Ruslan. 2015. SSRNLearning Deep Generative Models.
    [Google Scholar]
  12. Salimans, Tim2016. “Improved Techniques for Training GANs.” arXiv Machine Learning.
    [Google Scholar]
  13. Scheidt, Céline, FernandesAnjali, PaolaChris, and CaersJef. 2015. “Can Geostatistical Models Represent Nature's Variability? An Analysis Using Flume Experiments.” In Petroleum Geostatistics,, 1–16.
    [Google Scholar]
  14. Strebelle, Sebastien. 2002. “Conditional Simulation of Complex Geological Structures Using Multiple-Point Statistics.” Mathematical Geology34(1): 1–21.
    [Google Scholar]
  15. Torquato, S., BeasleyJ. D., and ChiewY. C.. 1988. “Two-Point Cluster Function for Continuum Percolation.” The Journal of Chemical Physics.
    [Google Scholar]
  16. Welling, Max. 2013. “Auto-Encoding Variational Bayes ArXiv: 1312. 6114v10 [ Stat. ML ] 1 May 2014.” ICLR 2014.
    [Google Scholar]
  17. BengioYoshua. 2012. “Deep Learning of Representations for Unsupervised and Transfer Learning.” JMLR: Workshop and Conference Proceedings.
    [Google Scholar]
http://instance.metastore.ingenta.com/content/papers/10.3997/2214-4609.201902196
Loading
/content/papers/10.3997/2214-4609.201902196
Loading

Data & Media loading...

This is a required field
Please enter a valid email address
Approval was a Success
Invalid data
An Error Occurred
Approval was partially successful, following selected items could not be processed due to error