On Stochastic Gradient Langevin Dynamics with Dependent Data Streams: The Fully Nonconvex Case

  • Ngoc Huy Chau
  • , Éric Moulines
  • , Miklós Rásonyi
  • , Sotirios Sabanis
  • , Ying Zhang

Research output: Contribution to journalArticlepeer-review

Abstract

We consider the problem of sampling from a target distribution, which is not necessarily log-concave, in the context of empirical risk minimization and stochastic optimization as presented in [M. Raginsky, A. Rakhlin, and M. Telgarsky, Proc. Mach. Learn. Res., 65 (2017), pp. 1674–1703]. Non-asymptotic results are established in the L1-Wasserstein distance for the behavior of stochastic gradient Langevin dynamics algorithms. We allow gradient estimates based on dependent data streams. Our convergence estimates are sharper and uniform in the number of iterations, in contrast to those in previous studies.

Original languageEnglish
Pages (from-to)959-986
Number of pages28
JournalSIAM Journal on Mathematics of Data Science
Volume3
Issue number3
DOIs
Publication statusPublished - 1 Jan 2021

Keywords

  • Langevin dynamics
  • contraction
  • convergence guarantees
  • nonconvex optimization
  • stochastic gradient

Fingerprint

Dive into the research topics of 'On Stochastic Gradient Langevin Dynamics with Dependent Data Streams: The Fully Nonconvex Case'. Together they form a unique fingerprint.

Cite this