Deep Learning for Seismic Inverse Problems: Toward the Acceleration of Geophysical Analysis Workflows

Abstract

Seismic inversion is a fundamental tool in geophysical analysis, providing a window into the Earth. In particular, it enables the reconstruction of large scale subsurface earth models for hydrocarbon exploration, mining, earthquakes analysis, shallow hazard assessment and other geophysical tasks. This article provides a comprehensive and timely overview of emerging datadriven Deep Learning (DL) solutions to seismic inverse problems, including velocity, impedance, reflectivity model building and seismic bandwidth extension. The article reviews seismic wave propagation and signal acquisition principles using large scale sensor arrays in offshore and inland exploration. In addition, the relations between the seismic forward and inverse problems are established, and prominent model-based solutions including seismic tomography, Full-Waveform Inversion (FWI) and multiscale FWI are explained. A primer to DL principles is presented, including empirical risk minimization, stochastic gradient-based learning, Convolutional Neural Networks (CNN), Recurrent Neural Networks (RNN) and regularization. The article then presents DL-based solutions addressing full seismic inversion as well as DL-assisted FWI solutions including Encoder-Decoder architectures, CNN-based networks and RNN-based networks with Long Short Term Memory and Gated Recurrent Units. Seismic signals features extraction methods such as semblance velocity analysis and seismic spectrograms are discussed, as well as batch vs. sequential data processing architectures. The article further presents advanced solutions based on Generative Adversarial Networks , and Physics-guided DL architectures that incorporate numerical geophysical modeling into the DL training loop, enabling unsupervised and semi-supervised learning.

A Deep Learning-Based Electromagnetic Signal for Earthquake Magnitude Prediction

“…In contrast, DL can assemble simple features through a variety of nonlinear transformations containing billions of weight parameters and further abstract more complex features. The DL net gradually optimizes these parameters during the training process so as to establish an incomprehensible mapping from input to output [ 34 ]. Therefore, it has been widely used in earthquake monitoring with a great performance in recent years [ 35 , 36 ].…”

DuRIN: A Deep-unfolded Sparse Seismic Reflectivity Inversion Network

“…This is done through a sequential algorithm-based approach (SBL-SA) to update the sparsity-controlling hyperparameters [26], or through the expectation-maximization algorithm (SBL-EM) [8], [27]. The application of neural networks to inversion problems in geophysics [28], particularly reflectivity inversion, is a recent development [3], [29]. Kim and Nakata [29] used an elementary feedforward neural network to recover the sparse reflectivity.…”

“…To the best of our knowledge, model-based architectures have not been explored for solving the seismic reflectivity inversion problem. In fact, deep-unrolling has not been employed for solving seismic inverse problems [28]. The minimax-concave penalty corresponding to γ = 2, γ = 3 and γ = 100.…”

A Regularization Homotopy Strategy for the Constrained Parameter Inversion of Partial Differential Equations

“…Inverse problems of partial differential equations arise in a variety of practical problems in science and engineering. These range from biomedical and geophysical imaging to groundwater flow modeling [1][2][3][4][5][6][7][8][9]. It is very significant to conduct research into the theory of inverse problems and their applications.…”