Deep Preconditioners and their application to seismic wavefield processing

TL;DR

This paper introduces a deep learning-based preconditioning method for seismic inverse problems, using autoencoders to improve convergence and solution quality over traditional fixed-basis transforms.

Contribution

It proposes a novel nonlinear preconditioning approach using autoencoders to enhance seismic data inversion, outperforming traditional methods in convergence speed and accuracy.

Findings

Deep learned preconditioners outperform fixed-basis transforms.

Faster convergence to seismic inverse solutions.

Effective on both synthetic and field data.

Abstract

Seismic data processing heavily relies on the solution of physics-driven inverse problems. In the presence of unfavourable data acquisition conditions (e.g., regular or irregular coarse sampling of sources and/or receivers), the underlying inverse problem becomes very ill-posed and prior information is required to obtain a satisfactory solution. Sparsity-promoting inversion, coupled with fixed-basis sparsifying transforms, represent the go-to approach for many processing tasks due to its simplicity of implementation and proven successful application in a variety of acquisition scenarios. Leveraging the ability of deep neural networks to find compact representations of complex, multi-dimensional vector spaces, we propose to train an AutoEncoder network to learn a direct mapping between the input seismic data and a representative latent manifold. The trained decoder is subsequently used as a nonlinear preconditioner for the physics-driven inverse problem at hand. Synthetic and field data are presented for a variety of seismic processing tasks and the proposed nonlinear, learned transformations are shown to outperform fixed-basis transforms and convergence faster to the sought solution.