Greedy Stein Variational Gradient Descent: An algorithmic approach for wave prospection problems
José L. Varona-Santana, Marcos A. Capistrán · 2025
Read Greedy Stein Variational Gradient Descent: An algorithmic approach for wave prospection problems by José L. Varona-Santana, Marcos A. Capistrán (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-2501-19370.txt.
Description
José Varona-Santana and Marcos Capistrán present a methods paper, published in 2025, on a faster way to estimate unknown values from wave data. Their setting is wave prospection, where waves sent through the ground help reveal what lies below the surface. The authors work in Bayesian inference, which weighs how well each guess fits the data against how plausible it was at the start.
Their method, called gradient-steered Stein variational gradient descent, adds a new loss that blends gradient information with a standard bound on fit. Tests on low contrast and high contrast ground models showed faster convergence than two established methods. The gain matters most when each gradient step is costly to compute, a common problem in this field.
Catalog record
| CREATOR | José L. Varona-Santana, Marcos A. Capistrán |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 79 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2501-19370