Vasdekis, G;
Roberts, GO;
(2023)
Speeding up the Zig-Zag Process.
In:
Bayesian Statistics, New Generations New Approaches: BAYSM 2022, Montréal, Canada, June 22–23.
(pp. pp. 35-46).
Springer: Cham, Switzerland.
Preview |
Text
Speeding Up the Zig Zag process.pdf - Accepted Version Download (320kB) | Preview |
Abstract
The Zig-Zag process is a Piecewise Deterministic Markov Process (PDMP), efficiently used for simulation in an MCMC setting. A generalisation of this process, the Speed Up Zig-Zag (SUZZ) process, was later suggested in Vasdekis G. and Roberts G. O. (2023+) [28] as a way to explore the tails of the distribution faster, making it an ideal candidate for heavy tailed targets. In this article we will describe the SUZZ process, we will review the main theoretical results and we will present a numerical study on some more practical models than the ones discussed in Vasdekis G. and Roberts G. O. (2023+) [28], showing that the advantages of using SUZZ may also extend to lighter tailed targets.
| Type: | Proceedings paper |
|---|---|
| Title: | Speeding up the Zig-Zag Process |
| Event: | BAYSM 2022: International Conference on Bayesian Statistics in Action |
| ISBN-13: | 9783031424120 |
| Open access status: | An open access version is available from UCL Discovery |
| DOI: | 10.1007/978-3-031-42413-7_4 |
| Publisher version: | https://doi.org/10.1007/978-3-031-42413-7_4 |
| Language: | English |
| Additional information: | This version is the author accepted manuscript. For information on re-use, please refer to the publisher’s terms and conditions. |
| Keywords: | Piecewise Deterministic Markov Process, Markov Chain Monte Carlo |
| UCL classification: | UCL UCL > Provost and Vice Provost Offices > UCL BEAMS UCL > Provost and Vice Provost Offices > UCL BEAMS > Faculty of Maths and Physical Sciences UCL > Provost and Vice Provost Offices > UCL BEAMS > Faculty of Maths and Physical Sciences > Dept of Statistical Science |
| URI: | https://discovery-pp.ucl.ac.uk/id/eprint/10185381 |
Archive Staff Only
 |
View Item |
