Uncertainty quantification in imaging and automatic horizon tracking: a Bayesian deep-prior based approach

  title={Uncertainty quantification in imaging and automatic horizon tracking: a Bayesian deep-prior based approach},
  author={Ali Siahkoohi and Gabrio Rizzuti and F. Herrmann},
In inverse problems, uncertainty quantification (UQ) deals with a probabilistic description of the solution nonuniqueness and data noise sensitivity. Setting seismic imaging into a Bayesian framework allows for a principled way of studying uncertainty by solving for the model posterior distribution. Imaging, however, typically constitutes only the first stage of a sequential workflow, and UQ becomes even more relevant when applied to subsequent tasks that are highly sensitive to the inversion… 

Figures from this paper

Deep Bayesian inference for seismic imaging with tasks
This method is designed to handle large scale Bayesian inference problems with computationally expensive forward operators as in seismic imaging and introduces a flexible inductive bias that is a surprisingly good fit for many diverse domains in imaging.
Parameterizing uncertainty by deep invertible networks, an application to reservoir characterization
This work proposes an approach characterized by training a deep network that "pushes forward" Gaussian random inputs into the model space (representing, for example, density or velocity) as if they were sampled from the actual posterior distribution, designed to solve a variational optimization problem based on the Kullback-Leibler divergence between the posterior and the network output distributions.
Learning by example: fast reliability-aware seismic imaging with normalizing flows
This is the first attempt to train a conditional network on what the authors know from neighboring images to improve the current image and assess its reliability, and is a data-driven variational inference approach where a normalizing flow is trained, capable of cheaply sampling the posterior distribution given previously unseen seismic data from neighboring surveys.
An encoder-decoder deep surrogate for reverse time migration in seismic imaging under uncertainty
An encoder-decoder deep learning surrogate model for RTM under uncertainty is proposed that can reproduce the seismic images accurately, and, more importantly, the uncertainty propagation from the input velocity fields to the image ensemble.
Velocity continuation with Fourier neural operators for accelerated uncertainty quantification
Seismic imaging is an ill-posed inverse problem that is challenged by noisy data and modeling inaccuracies—due to errors in the background squared-slowness model. Uncertainty quantification is
Faster Uncertainty Quantification for Inverse Problems with Conditional Normalizing Flows
This work proposes a two-step scheme that makes use of normalizing flows and joint data to train a conditional generator to approximate the target posterior density of a invertible generator, and presents some synthetic results that demonstrate considerable training speedup when reusing the pretrained network as a warm start or preconditioning for approximating a prior.
Surface wave dispersion inversion using an energy likelihood function
Seismic surface wave dispersion inversion is used widely to study the subsurface structure of the Earth. The dispersion property is usually measured by using frequency-phase velocity (f-c) analysis
Seismic characterization of deeply buried paleocaves based on Bayesian deep learning


A deep-learning based Bayesian approach to seismic imaging and uncertainty quantification
This work proposes to use the functional form of a randomly initialized convolutional neural network as an implicit structured prior, which is shown to promote natural images and excludes images with unnatural noise.
A Bayesian Perspective on the Deep Image Prior
It is shown that the deep image prior is asymptotically equivalent to a stationary Gaussian process prior in the limit as the number of channels in each layer of the network goes to infinity, and derive the corresponding kernel, which informs a Bayesian approach to inference.
Assessing uncertainties in velocity models and images with a fast nonlinear uncertainty quantification method
Seismic imaging is conventionally performed using noisy data and a presumably inexact velocity model. Uncertainties in the input parameters propagate directly into the final image and therefore into
Uncertainty quantification for inverse problems with weak partial-differential-equation constraints
This work considers inverse problems with partial-differential-equation (PDE) constraints, which are applicable to many seismic problems, and develops a bilinear posterior distribution with weak PDE constraints that is more amenable to uncertainty quantification because of its special structure.
The importance of transfer learning in seismic modeling and imaging
Accurate forward modeling is essential for solving inverse problems in exploration seismology. Unfortunately, it is often not possible to afford being physically or numerically accurate. To overcome
Learned imaging with constraints and uncertainty quantification
The aim, and main contribution of this work, is the combination of handcrafted constraints with deep convolutional neural networks, as a way to harness their remarkable ease of generating natural images.
Two ways to quantify uncertainty in geophysical inverse problems
We present two approaches to invert geophysical measurements and estimate subsurface properties and their uncertainties when little is known a priori about the size of the errors associated with the
Task adapted reconstruction for inverse problems
The suggested framework is generic, yet adaptable, with a plug-and-play structure for adjusting both the inverse problem and the task at hand, and any task that is encodable as a trainable neural network can be used.
Regularized elastic full-waveform inversion using deep learning
Obtaining high-resolution models of the earth, especially around the reservoir, is crucial to properly image and interpret the subsurface. We have developed a regularized elastic full-waveform
Multi-resolution neural networks for tracking seismic horizons from few training images
A projected loss-function for training convolutional networks with a multi-resolution structure, including variants of the U-net is proposed and labels are proposed as the convolution of a Gaussian kernel and the known horizon locations that indicate uncertainty in the labels.