Convergence of Langevin Monte Carlo in Chi-Squared and Renyi Divergence

arXiv2007.11612MaRDI QIDQ6345649FDOQ6345649

Authors: Murat A. Erdogdu, Rasa Hosseinzadeh, Matthew S. Zhang 

Publication date: 22 July 2020

Abstract: We study sampling from a target distribution $u_{\ast }=e^{-f}$ using the unadjusted Langevin Monte Carlo (LMC) algorithm when the potential $f$ satisfies a strong dissipativity condition and it is first-order smooth with a Lipschitz gradient. We prove that, initialized with a Gaussian random vector that has sufficiently small variance, iterating the LMC algorithm for $widetildemathcalO(lambd{a}^{2}depsilo{n}^{-1})$ steps is sufficient to reach $epsilon$-neighborhood of the target in both Chi-squared and Renyi divergence, where $lambda$ is the logarithmic Sobolev constant of $u_{\ast }$. Our results do not require warm-start to deal with the exponential dimension dependency in Chi-squared divergence at initialization. In particular, for strongly convex and first-order smooth potentials, we show that the LMC algorithm achieves the rate estimate $widetildemathcalO(depsilo{n}^{-1})$ which improves the previously known rates in both of these metrics, under the same assumptions. Translating this rate to other metrics, our results also recover the state-of-the-art rate estimates in KL divergence, total variation and $2$-Wasserstein distance in the same setup. Finally, as we rely on the logarithmic Sobolev inequality, our framework covers a range of non-convex potentials that are first-order smooth and exhibit strong convexity outside of a compact region.

This page was built for publication: Convergence of Langevin Monte Carlo in Chi-Squared and Renyi Divergence

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6345649)