Joint Learning for Seismic Inversion: An Acoustic Impedance Estimation Case Study

TL;DR

This paper introduces a joint learning approach for seismic inversion that leverages multiple datasets and shared neural network weights to improve generalization, achieving high accuracy with minimal training data.

Contribution

It proposes a novel joint training scheme with weight similarity constraints to enhance seismic inversion performance across different datasets.

Findings

Achieved an $r^{2}$ coefficient of 0.8399 using less than 3% of training data.

Joint learning improved generalization across datasets.

Method reduces data requirements for accurate seismic inversion.

Abstract

Seismic inversion helps geophysicists build accurate reservoir models for exploration and production purposes. Deep learning-based seismic inversion works by training a neural network to learn a mapping from seismic data to rock properties using well log data as the labels. However, well logs are often very limited in number due to the high cost of drilling wells. Machine learning models can suffer overfitting and poor generalization if trained on limited data. In such cases, well log data from other surveys can provide much needed useful information for better generalization. We propose a learning scheme where we simultaneously train two network architectures, each on a different dataset. By placing a soft constraint on the weight similarity between the two networks, we make them learn from each other where useful for better generalization performance on their respective datasets. Using less than 3$\%$ of the available training data, we were able to achieve an average $r^{2}$ coefficient of 0.8399 on the acoustic impedance pseudologs of the SEAM dataset via joint learning with the Marmousi dataset.