Multilevel-Langevin pathwise average for Gibbs approximation - Archive ouverte HAL Accéder directement au contenu
Pré-Publication, Document De Travail Année : 2024

Multilevel-Langevin pathwise average for Gibbs approximation

Résumé

We propose and study a new multilevel method for the numerical approximation of a Gibbs distribution $\pi$ on $\mathbb{R}^d$, based on (overdamped) Langevin diffusions. This method inspired by [PP18] and [GMS20] relies on a multilevel occupation measure, i.e. on an appropriate combination of R occupation measures of (constant-step) Euler schemes with respective steps $\gamma_r=\gamma_0 2^{-r}$, r=0,...,R. We first state a quantitative result under general assumptions which guarantees an epsilon-approximation (in a $L^2$-sense) with a cost of the order $\epsilon^{-2}$ or $\epsilon^{-2}|\log \epsilon|^3$ under less contractive assumptions. We then apply it to overdamped Langevin diffusions with strongly convex potential $U:\mathbb{R}^d\rightarrow \mathbb{R}$ and obtain an $\epsilon$-complexity of the order $O(d \epsilon^{-2}\log^3(d \epsilon^{-2}))$ or $O(d \epsilon^{-2})$ under additional assumptions on $U$. More precisely, up to universal constants, an appropriate choice of the parameters leads to a cost controlled by ${\bar{\lambda}_U\vee 1)^2}{\underline{\lambda}_U^{-3}} d\varepsilon^{-2}$ (where $\bar{\lambda}_U$ and $\underline{\lambda}_U$ respectively denote the supremum and the infimum of the largest and lowest eigenvalue of $D^2U$).\smallskip We finally complete these theoretical results with some numerical illustrations including comparisons to other algorithms in Bayesian learning and opening to non strongly convex setting.
Fichier principal
Vignette du fichier
Multilevel_langevin_final_version_9_11_23.pdf (586.38 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-03345717 , version 1 (15-09-2021)
hal-03345717 , version 2 (08-02-2024)

Identifiants

Citer

Maxime Egéa, Fabien Panloup. Multilevel-Langevin pathwise average for Gibbs approximation. 2024. ⟨hal-03345717v2⟩
237 Consultations
98 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More