Stochastic Seismic Waveform Inversion Using Generative Adversarial Networks as a Geological Prior

We present an application of deep generative models in the context of partial differential equation constrained inverse problems. We combine a generative adversarial network representing an a priori model that generates geological heterogeneities and their petrophysical properties, with the numerical solution of the partial-differential equation governing the propagation of acoustic waves within the earth’s interior. We perform Bayesian inversion using an approximate Metropolis-adjusted Langevin algorithm to sample from the posterior distribution of earth models given seismic observations. Gradients with respect to the model parameters governing the forward problem are obtained by solving the adjoint of the acoustic wave equation. Gradients of the mismatch with respect to the latent variables are obtained by leveraging the differentiable nature of the deep neural network used to represent the generative model. We show that approximate Metropolis-adjusted Langevin sampling allows an efficient Bayesian inversion of model parameters obtained from a prior represented by a deep generative model, obtaining a diverse set of realizations that reflect the observed seismic response.

[1]  Ajinkya Kadu,et al.  A parametric level-set approach for seismic full-waveform inversion , 2016 .

[2]  Martin J. Blunt,et al.  Conditioning of Generative Adversarial Networks for Pore and Reservoir Scale Models , 2018, 80th EAGE Conference and Exhibition 2018.

[3]  拓海 杉山,et al.  “Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks”の学習報告 , 2017 .

[4]  S. Kabanikhin Definitions and examples of inverse and ill-posed problems , 2008 .

[5]  D. Grana,et al.  Bayesian Gaussian Mixture Linear Inversion for Geophysical Inverse Problems , 2017, Mathematical Geosciences.

[6]  Alan Richardson,et al.  Generative Adversarial Networks for Model Order Reduction in Seismic Full-Waveform Inversion , 2018, 1806.00828.

[7]  Martin J. Blunt,et al.  Reconstruction of three-dimensional porous media using generative adversarial neural networks , 2017, Physical review. E.

[8]  Kurt Hornik,et al.  Multilayer feedforward networks are universal approximators , 1989, Neural Networks.

[9]  George Biros,et al.  Parallel Multiscale Gauss-Newton-Krylov Methods for Inverse Wave Propagation , 2002, ACM/IEEE SC 2002 Conference (SC'02).

[10]  Yann LeCun,et al.  A theoretical framework for back-propagation , 1988 .

[11]  Denis Lukovnikov,et al.  On the regularization of Wasserstein GANs , 2017, ICLR.

[12]  Amir Adler,et al.  Deep-learning tomography , 2018 .

[13]  Martin J. Blunt,et al.  Stochastic Reconstruction of an Oolitic Limestone by Generative Adversarial Networks , 2017, Transport in Porous Media.

[14]  Olivier Dubrule,et al.  Geostatistics for Seismic Data Integration in Earth Models , 2003 .

[15]  R. Tweedie,et al.  Exponential convergence of Langevin distributions and their discrete approximations , 1996 .

[16]  Ahmed H. Elsheikh,et al.  Parametric generation of conditional geological realizations using generative neural networks , 2018, Computational Geosciences.

[17]  A. Buland,et al.  Bayesian linearized AVO inversion , 2003 .

[18]  Sepp Hochreiter,et al.  GANs Trained by a Two Time-Scale Update Rule Converge to a Local Nash Equilibrium , 2017, NIPS.

[19]  Diego Klabjan,et al.  Generative Adversarial Nets for Multiple Text Corpora , 2017, 2021 International Joint Conference on Neural Networks (IJCNN).

[20]  Chun-Liang Li,et al.  One Network to Solve Them All — Solving Linear Inverse Problems Using Deep Projection Models , 2017, 2017 IEEE International Conference on Computer Vision (ICCV).

[21]  J. Rosenthal,et al.  Optimal scaling of discrete approximations to Langevin diffusions , 1998 .

[22]  Lukas Mosser,et al.  Rapid seismic domain transfer: Seismic velocity inversion and modeling using deep generative neural networks , 2018, 80th EAGE Conference and Exhibition 2018.

[23]  Ernesto Della Rossa,et al.  Probabilistic petrophysical-properties estimation integrating statistical rock physics with seismic inversion , 2010 .

[24]  Prafulla Dhariwal,et al.  Glow: Generative Flow with Invertible 1x1 Convolutions , 2018, NeurIPS.

[25]  Alexandros G. Dimakis,et al.  Compressed Sensing using Generative Models , 2017, ICML.

[26]  Samy Bengio,et al.  Density estimation using Real NVP , 2016, ICLR.

[27]  Wojciech Zaremba,et al.  Improved Techniques for Training GANs , 2016, NIPS.

[28]  Anil A. Bharath,et al.  Inverting the Generator of a Generative Adversarial Network , 2016, IEEE Transactions on Neural Networks and Learning Systems.

[29]  T. Mukerji,et al.  Seismic inversion for reservoir properties combining statistical rock physics and geostatistics: A review , 2010 .

[30]  R. Plessix A review of the adjoint-state method for computing the gradient of a functional with geophysical applications , 2006 .

[31]  Eero P. Simoncelli,et al.  Image quality assessment: from error visibility to structural similarity , 2004, IEEE Transactions on Image Processing.

[32]  Alex Graves,et al.  Conditional Image Generation with PixelCNN Decoders , 2016, NIPS.

[33]  Tapan Mukerji,et al.  Seismic inversion combining rock physics and multiple-point geostatistics , 2008 .

[34]  Albert Tarantola,et al.  Neural networks and inversion of seismic data , 1994 .

[35]  Albert Tarantola,et al.  Inverse problem theory - and methods for model parameter estimation , 2004 .

[36]  Philippe Doyen,et al.  ebook - Seismic Reservoir Characterization: An Earth Modelling Perspective (EET 2) , 2007 .

[37]  Felix J. Herrmann,et al.  Full-waveform inversion, Part 1: Forward modeling , 2017 .

[38]  Jasper A. Vrugt,et al.  High‐dimensional posterior exploration of hydrologic models using multiple‐try DREAM(ZS) and high‐performance computing , 2012 .

[39]  Carla Carvajal,et al.  Petrophysical seismic inversion conditioned to well-log data: Methods and application to a gas reservoir , 2009 .

[40]  Mrinal K. Sen,et al.  Bayesian inference, Gibbs' sampler and uncertainty estimation in geophysical inversion , 1996 .

[41]  Yoshua Bengio,et al.  Plug & Play Generative Networks: Conditional Iterative Generation of Images in Latent Space , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[42]  Daniel Rueckert,et al.  Real-Time Single Image and Video Super-Resolution Using an Efficient Sub-Pixel Convolutional Neural Network , 2016, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[43]  T. Hansen,et al.  Efficient Monte Carlo sampling of inverse problems using a neural network-based forward—applied to GPR crosshole traveltime inversion , 2017 .

[44]  Alan C. Bovik,et al.  Mean squared error: Love it or leave it? A new look at Signal Fidelity Measures , 2009, IEEE Signal Processing Magazine.

[45]  Klaus Mosegaard,et al.  Resolution analysis of general inverse problems through inverse Monte Carlo sampling , 1998 .

[46]  Yann Le Cun,et al.  A Theoretical Framework for Back-Propagation , 1988 .

[47]  R. M. Srivastava,et al.  Multivariate Geostatistics: Beyond Bivariate Moments , 1993 .

[48]  David R. O'Hallaron,et al.  High Resolution Forward And Inverse Earthquake Modeling on Terascale Computers , 2003, SC.

[49]  Eric Laloy,et al.  Training‐Image Based Geostatistical Inversion Using a Spatial Generative Adversarial Neural Network , 2017, ArXiv.

[50]  Bruce S. Hart,et al.  A visual data-mining methodology for seismic facies analysis: Part 1 — Testing and comparison with other unsupervised clustering methods , 2009 .

[51]  Jeroen Tromp,et al.  Computational Efficiency of Full Waveform Inversion Algorithms , 2015 .

[52]  Niels Bohr,et al.  Monte Carlo sampling of solutions to inverse problems , 2004 .

[53]  Matthias Bethge,et al.  A note on the evaluation of generative models , 2015, ICLR.

[54]  Andre G. Journel,et al.  Constraining Stochastic Images to Seismic Data , 1993 .

[55]  Dario Grana,et al.  Geostatistical rock physics AVA inversion , 2018, Geophysical Journal International.

[56]  R. Swennen,et al.  Advances in Oil and Gas Exploration & Production , 2017 .

[57]  Rishi Sharma,et al.  A Note on the Inception Score , 2018, ArXiv.

[58]  Max Welling,et al.  Auto-Encoding Variational Bayes , 2013, ICLR.

[59]  O. Dubrule,et al.  Geostatistical inversion - a sequential method of stochastic reservoir modelling constrained by seismic data , 1994 .

[60]  Clayton V. Deutsch,et al.  Hierarchical object-based stochastic modeling of fluvial reservoirs , 1996 .