Sampling with time-changed Markov processes
2025
Read Sampling with time-changed Markov processes (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-2501-15155.txt.
Description
Andrea Bertazzi and Giorgos Vasdekis speed up sampling algorithms in twenty twenty-five. Markov chain Monte Carlo methods sample from complex distributions by wandering randomly. Their convergence can be slow. The authors study time-changed processes, where a base process runs at a state-dependent speed.
A user-chosen function sets how fast time flows in each region. Applied to Langevin diffusions and piecewise deterministic processes, the trick yields new algorithms. It also rediscovers some known ones from a fresh angle. Theory connects the changed process's stationary distribution to the original. Geometric and uniform ergodicity are studied, along with a central limit theorem. A comparison with space transformations clarifies the two ideas. Readers get a dial, time itself, to tune samplers.
Catalog record
| YEAR | 2025 |
|---|---|
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 99 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2501-15155