Evaluating machine learning-predicted subsurface properties via seismic data reconstruction

In recent years, machine-learning (ML) approaches have gained significant attention in seismic-based subsurface property estimation problems. However, because of the data-driven nature of these methods, it is challenging to evaluate the quality of the estimated properties in regions without ground-truth data. In this paper, we discuss evaluating the quality of ML-predicted subsurface properties through ML-based seismic data reconstruction. We use a deep-learning workflow to reconstruct the poststack seismic data, then use the misfit between the measured data and the reconstructed data as a proxy for the quality of ML-predicted subsurface properties. We also use self-supervised learning to improve the model generalization when training the deep-learning model for reconstruction. Our method is particularly valuable for subsurface properties without direct physical relation to seismic data. We provide synthetic and field data examples to demonstrate the consistency of our method.