Data-Driven Pseudo-spectral Full Waveform Inversion via Deep Neural Networks
Christopher Zerafa, Pauline Galea, Cristiana Sebu · 2025
Read Data-Driven Pseudo-spectral Full Waveform Inversion via Deep Neural Networks by Christopher Zerafa, Pauline Galea, Cristiana Sebu (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2502-17608.txt.
Description
Full waveform inversion builds a detailed picture of underground rock from seismic waves, but it depends on a physical model of how waves travel. Christopher Zerafa, Pauline Galea, and Cristiana Sebu took a different route in 2025. They rebuilt the method, in its pseudo-spectral form, as a deep neural network that learns from data rather than from a fixed wave model.
The network was tested on synthetic data and on a standard two-dimensional test model called Marmousi. It beat classical inversion in deeper zones and in areas where rock layers thrust over each other. The authors credit the network's global fitting power, which frees it from the limits of ray-based forward models.
Catalog record
| CREATOR | Christopher Zerafa, Pauline Galea, Cristiana Sebu |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 23 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2502-17608