Non-asymptotic convergence bounds for modified tamed unadjusted Langevin algorithm in non-convex setting

DOI10.1016/J.JMAA.2024.128892MaRDI QIDQ6640899zbMATH OpenFDO

Authors Ariel Neufeld, Matthew (Cheng En) Ng, Ying Zhang

Publication date 20 November 2024

Published in Journal of Mathematical Analysis and Applications (Search for Journal in Brave)

Full work available at URL https://doi.org/10.1016/J.JMAA.2024.128892

zbMATH Keywords

high-dimensional samplingLangevin SDEmodified tamed unadjusted Langevin algorithmnon-asymptotic convergence boundssuper-linearly growing diffusion coefficients

Mathematics Subject Classification ID

Stochastic ordinary differential equations (aspects of stochastic analysis) (60H10) Monte Carlo methods (65C05)

Recommendations

• The tamed unadjusted Langevin algorithm
• Nonasymptotic convergence analysis for the unadjusted Langevin algorithm
• Higher order Langevin Monte Carlo algorithm
• Kinetic Langevin MCMC sampling without gradient Lipschitz continuity -- the strongly convex case

Cites work

• A note on tamed Euler approximations
• An introduction to MCMC for machine learning
• Ergodicity for SDEs and approximations: locally Lipschitz vector fields and degenerate noise.
• High-dimensional Bayesian inference via the unadjusted Langevin algorithm
• scientific article; zbMATH DE number 54145 (Why is no real title available?)
• Kinetic Langevin MCMC sampling without gradient Lipschitz continuity -- the strongly convex case
• Laplace's method revisited: Weak convergence of probability measures
• MCMC methods for functions: modifying old algorithms to make them faster
• Nonasymptotic bounds for sampling algorithms without log-concavity
• Nonasymptotic convergence analysis for the unadjusted Langevin algorithm
• On Stochastic Gradient Langevin Dynamics with Dependent Data Streams: The Fully Nonconvex Case
• Partial differential equations and stochastic methods in molecular dynamics
• Quantitative Harris-type theorems for diffusions and McKean-Vlasov processes
• Rapid convergence of the unadjusted Langevin algorithm: isoperimetry suffices
• Strong and weak divergence in finite time of Euler's method for stochastic differential equations with non-globally Lipschitz continuous coefficients
• Strong convergence of an explicit numerical method for SDEs with nonglobally Lipschitz continuous coefficients
• Taming Neural Networks with TUSLA: Nonconvex Learning via Adaptive Stochastic Gradient Langevin Algorithms
• The tamed unadjusted Langevin algorithm
• Theoretical Guarantees for Approximate Sampling from Smooth and Log-Concave Densities
• User-friendly guarantees for the Langevin Monte Carlo with inaccurate gradient

Cited in

(7)

• The tamed unadjusted Langevin algorithm
• Linear implicit approximations of invariant measures of semi-linear SDEs with non-globally Lipschitz coefficients
• Non-asymptotic estimates for TUSLA algorithm for non-convex learning with applications to neural networks with ReLU activation function
• On modified Euler methods for McKean-Vlasov stochastic differential equations with super-linear coefficients
• Projected Langevin Monte Carlo algorithms in non-convex and super-linear setting
• Geometric ergodicity and strong error estimates for tamed schemes of super-linear SODEs
• Numerical ergodicity and optimal strong error estimates for a class of novel tamed schemes to superlinear SPDEs

This page was built for publication: Non-asymptotic convergence bounds for modified tamed unadjusted Langevin algorithm in non-convex setting

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6640899)