English
The coordinate sampler: a non-reversible Gibbs-like MCMC sampler
Virtualconference
Auteurs : Robert, Christian P (Auteur de la conférence)
CIRM (Editeur )
60J25 - Continuous-time Markov processes on general state spaces
62F15 - Bayesian inference
CIRM (Editeur )
Loading the player...
|
Résumé :
In this talk, we derive a novel non-reversible, continuous-time Markov chain Monte Carlo (MCMC) sampler, called Coordinate Sampler, based on a piecewise deterministic Markov process (PDMP), which can be seen as a variant of the Zigzag sampler. In addition to proving a theoretical validation for this new sampling algorithm, we show that the Markov chain it induces exhibits geometrical ergodicity convergence, for distributions whose tails decay at least as fast as an exponential distribution and at most as fast as a Gaussian distribution. Several numerical examples highlight that our coordinate sampler is more efficient than the Zigzag sampler, in terms of effective sample size.
[This is joint work with Wu Changye, ref. arXiv:1809.03388]
Mots-Clés : MCMC; no-reversibility; Gibbs sampling
Codes MSC : [This is joint work with Wu Changye, ref. arXiv:1809.03388]
Mots-Clés : MCMC; no-reversibility; Gibbs sampling
60J25 - Continuous-time Markov processes on general state spaces
62F15 - Bayesian inference
|
Informations sur la Vidéo
Réalisateur : Hennenfent, GuillaumeLangue : Anglais Date de Publication : 02/11/2020 Date de Captation : 02/11/2020 Sous Collection : Research School Catégorie arXiv : Statistics Theory Domaine(s) : Probabilités & Statistiques Format : MP4 (.mp4) - HD Durée : 00:32:44 Audience : Chercheurs Download : https://videos.cirm-math.fr/2020-11-05_Robert.mp4 
|
Informations sur la Rencontre
Nom de la Rencontre : Jean-Morlet Chair 2020 - Research School: Quasi-Monte Carlo Methods and Applications / Chaire Jean-Morlet 2020 - Ecole: Méthode de quasi-Monte-Carlo et applicationsOrganisateurs de la Rencontre : Rivat, Joël ; Thonhauser, Stefan ; Tichy, Robert Dates : 02/11/2020 - 07/11/2020 Année de la rencontre : 2020 URL de la Rencontre : https://www.chairejeanmorlet.com/2255.html
Données de citation
DOI : 10.24350/CIRM.V.19664703Citer cette vidéo: Robert, Christian P (2020). The coordinate sampler: a non-reversible Gibbs-like MCMC sampler. CIRM. Audiovisual resource. doi:10.24350/CIRM.V.19664703 URI : http://dx.doi.org/10.24350/CIRM.V.19664703 |
Voir Aussi
- [Virtualconference] PDMPs in risk theory and QMC integration III / Auteur de la conférence Thonhauser, Stefan.
- [Virtualconference] On the estimation of conditional quantiles - lecture 3 / Auteur de la conférence Maume-Deschamps, Véronique.
- [Virtualconference] On the estimation of conditional quantiles - lecture 2 / Auteur de la conférence Maume-Deschamps, Véronique.
- [Virtualconference] Stochastic differential equations / Auteur de la conférence Leobacher, Gunther.
- [Virtualconference] Number sequences for simulation - lecture 2 / Auteur de la conférence Ökten, Giray.
- [Virtualconference] PDMPs and Integrals PDMPs in risk theory and QMC integration II / Auteur de la conférence Thonhauser, Stefan.
- [Virtualconference] Derivative pricing, simulation from non-uniform distributions - lecture 3 / Auteur de la conférence Ökten, Giray.
- [Multi angle] Quasi-Monte Carlo methods and applications: introduction / Auteur de la conférence Tichy, Robert.
- [Virtualconference] Risk theory: models and objectives PDMPs in risk theory and QMC integration I / Auteur de la conférence Thonhauser, Stefan.
- [Virtualconference] On the estimation of conditional quantiles - lecture 1 / Auteur de la conférence Maume-Deschamps, Véronique.
- [Virtualconference] Brief introduction of Quasi-Monte Carlo Methods and their Applications / Auteur de la conférence Leobacher, Gunther.
- [Virtualconference] Two concrete FinTech applications of QMC / Auteur de la conférence Larcher, Gerhard.
- [Virtualconference] Number sequences for simulation - lecture 1 / Auteur de la conférence Ökten, Giray.
- [Virtualconference] Optimizing dividends and limited capital injections via exponential approximations I / Auteur de la conférence Avram, Florin.
Bibliographie
- WU, Changye et ROBERT, Christian P. Coordinate sampler: a non-reversible Gibbs-like MCMC sampler. Statistics and Computing, 2020, vol. 30, no 3, p. 721-730. - https://doi.org/10.1007/s11222-019-09913-w
