Feature-based Inversion of 2.5D Controlled Source Electromagnetic Data using Generative Priors

2026-01-05

IEEE Transactions on Geoscience and Remote Sensing

AI Summary

This study addresses the challenges of effectively integrating prior information and achieving high reconstruction accuracy in marine controlled-source electromagnetic (mCSEM) inversion. To this end, a feature-driven 2.5D inversion method is proposed, which innovatively embeds a conductivity-generating prior learned by a variational autoencoder (VAE) into a Gauss–Newton iterative framework in a plug-and-play manner. Explicit synergy between the prior and data misfit is achieved through model-space projection constraints. Coupled with finite-difference forward modeling, the method demonstrates significantly improved accuracy and generalization capability in reconstructing conductivity models, as validated by both synthetic and field data experiments, thereby enabling deep integration of geological priors into the inversion process.

Technology Category

• Machine Learning: Deep Generative Models & Autoencoders
• Constraint Satisfaction and Optimization: Constraint Learning and Acquisition
• Search and Optimization: Mixed Discrete/Continuous Search

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

In this study, we investigate feature-based 2.5D controlled source marine electromagnetic (mCSEM) data inversion using generative priors. Two-and-half dimensional modeling using finite difference method (FDM) is adopted to compute the response of horizontal electric dipole (HED) excitation. Rather than using a neural network to approximate the entire inverse mapping in a black-box manner, we adopt a plug-andplay strategy in which a variational autoencoder (VAE) is used solely to learn prior information on conductivity distributions. During the inversion process, the conductivity model is iteratively updated using the Gauss Newton method, while the model space is constrained by projections onto the learned VAE decoder. This framework preserves explicit control over data misfit and enables flexible adaptation to different survey configurations. Numerical and field experiments demonstrate that the proposed approach effectively incorporates prior information, improves reconstruction accuracy, and exhibits good generalization performance.

Problem

• controlled source electromagnetic
• 2.5D inversion
• generative priors
• conductivity reconstruction
• marine CSEM

Innovation

• generative priors
• variational autoencoder
• 2.5D CSEM inversion
• plug-and-play inversion
• Gauss-Newton method