A Neural Network Architecture Based on Attention Gate Mechanism for 3D Magnetotelluric Forward Modeling

TL;DR

This work tackles the computational intensity of 3D magnetotelluric forward modeling by introducing MTAGU-Net, a 3D neural network that integrates a dual-path attention gate within encoder–decoder skip connections to better capture anomalous region features. It combines FEM-based forward modeling with a deep learning surrogate, supported by a synthetic 3D Gaussian random field dataset that realisticly represents subsurface resistivity structures. Empirical results show MTAGU-Net achieves SSIM values consistently above 0.98 and lower RMSE than a conventional 3D U-Net, while offering predicted forward responses thousands of times faster than FEM. The method demonstrates strong generalization to unseen complex models, highlighting its potential for large-scale, real-time MT forward simulations in geophysical exploration.

Abstract

Traditional three-dimensional magnetotelluric (MT) numerical forward modeling methods, such as the finite element method (FEM) and finite volume method (FVM), suffer from high computational costs and low efficiency due to limitations in mesh refinement and computational resources. We propose a novel neural network architecture named MTAGU-Net, which integrates an attention gating mechanism for 3D MT forward modeling. Specifically, a dual-path attention gating module is designed based on forward response data images and embedded in the skip connections between the encoder and decoder. This module enables the fusion of critical anomaly information from shallow feature maps during the decoding of deep feature maps, significantly enhancing the network's capability to extract features from anomalous regions. Furthermore, we introduce a synthetic model generation method utilizing 3D Gaussian random field (GRF), which accurately replicates the electrical structures of real-world geological scenarios with high fidelity. Numerical experiments demonstrate that MTAGU-Net outperforms conventional 3D U-Net in terms of convergence stability and prediction accuracy, with the structural similarity index (SSIM) of the forward response data consistently exceeding 0.98. Moreover, the network can accurately predict forward response data on previously unseen datasets models, demonstrating its strong generalization ability and validating the feasibility and effectiveness of this method in practical applications.

Figures (16)

• Figure 1: Schematic diagram of 3D computational domain grid and hexahedral element. (a) is 3D computational domain grid, and (b) is the hexahedral element
• Figure 2: MTAGU-Net Attention gate network architecture diagram.
• Figure 3: Diagram of the Attention gate mechanism module.
• Figure 4: Examples of different conductivity models generated using Gaussian random fields (when $\alpha$ takes different values).
• Figure 5: Forward modeling response data of the example resistivity model (at the Fig. \ref{['fig3']}(a1)). (a1) and (a2) represent the apparent resistivity in the XY and YX directions, respectively, while (b1) and (b2) represent the phase in the XY and YX directions, respectively.