The Gaussian Latent Machine: Efficient Prior and Posterior Sampling for Inverse Problems
Muhamed Kuric, Martin Zach, Andreas Habring, Michael Unser, Thomas Pock · 2025
Read The Gaussian Latent Machine: Efficient Prior and Posterior Sampling for Inverse Problems by Muhamed Kuric, Martin Zach, Andreas Habring, Michael Unser, Thomas Pock (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-2505-12836.txt.
Introduction
Kuric and co-authors present a sampling method for the probability models that sit behind Bayesian image reconstruction. Many standard prior and posterior distributions in imaging can be written as products of expert terms, which are often awkward to sample from. The paper shows that such a model can be lifted into a new latent variable form, which the authors call a Gaussian latent machine. That move turns a hard sampling job into a simpler one.
In the general case, the approach yields a two-block Gibbs sampler that the authors describe as fast and effective. In special cases it reduces to direct sampling methods. Numerical experiments across a range of imaging problems back the claims. For researchers who need samples, not just point estimates, from imaging posteriors, the paper offers one framework that covers many models at once.
Catalog record
| CREATOR | Muhamed Kuric, Martin Zach, Andreas Habring, Michael Unser, Thomas Pock |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 169 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2505-12836