Summary of Machine Learning-enabled Velocity Model Building with Uncertainty Quantification, by Rafael Orozco et al.
Machine learning-enabled velocity model building with uncertainty quantification
by Rafael Orozco, Huseyin Tuna Erdinc, Yunlin Zeng, Mathias Louboutin, Felix J. Herrmann
First submitted to arxiv on: 11 Nov 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Computer Vision and Pattern Recognition (cs.CV)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This paper proposes a scalable methodology for building migration velocity models, which is crucial for various geophysical applications such as hydrocarbon exploration and CO2 sequestration monitoring. Traditional methods like Full-Waveform Inversion (FWI) struggle with complexities like noise, limited bandwidth, receiver aperture, and computational constraints. The authors integrate generative modeling using Diffusion networks with physics-informed summary statistics to address these challenges. They define summary statistics in terms of subsurface-offset image volumes for poor initial velocity models, allowing for efficient generation of Bayesian posterior samples that provide a useful assessment of uncertainty. To validate their approach, the authors introduce tests measuring the quality of inferred velocity models and uncertainties. Synthetic datasets are used to reconfirm gains from using subsurface-image gathers as conditioning observables. The method is also demonstrated on complex velocity model building involving salt and shows how uncertainty in the velocity model can be propagated to final product reverse-time migrated images. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper helps us better understand how to build accurate migration velocity models for important geophysical projects. Traditional methods are good but have limitations, like noise and computational constraints. The authors combined two techniques: generative modeling and physics-informed summary statistics. This makes it easier to generate many possible solutions (called Bayesian posterior samples) that show how unsure we are about the result. They tested their method on fake data and real-world datasets to make sure it works well. |
Keywords
» Artificial intelligence » Diffusion
