Geometry-informed irreversible perturbations for accelerated convergence of Langevin dynamics

DOI10.1007/s11222-022-10147-6zbMath1496.62025arXiv2108.08247OpenAlexW3194724043MaRDI QIDQ2080357

Benjamin J. Zhang, Konstantinos V. Spiliopoulos, Youssef M. Marzouk

Publication date: 7 October 2022

Published in: Statistics and Computing (Search for Journal in Brave)

Full work available at URL: https://arxiv.org/abs/2108.08247

zbMATH Keywords

Monte Carlo samplingBayesian computationgeometry-informed irreversibilityRiemannian manifold Langevin dynamicsstochastic gradient Langevin dynamics

Mathematics Subject Classification ID

Computational methods for problems pertaining to statistics (62-08) Monte Carlo methods (65C05) Diffusion processes (60J60)

Cites Work

• Unnamed Item
• Information-geometric Markov chain Monte Carlo methods using diffusions
• Improving the convergence of reversible samplers
• Non-reversible Metropolis-Hastings
• Optimal non-reversible linear drift for the convergence to equilibrium of a diffusion
• Accelerating diffusions
• Accelerating Gaussian diffusions
• Exponential convergence of Langevin distributions and their discrete approximations
• Normalizing constants of log-concave densities
• Analysis of a nonreversible Markov chain sampler.
• Inhomogeneous functionals and approximations of invariant distributions of ergodic diffusions: central limit theorem and moderate deviation asymptotics
• Optimal scaling of the MALA algorithm with irreversible proposals for Gaussian targets
• Measuring sample quality with diffusions
• High-dimensional Bayesian inference via the unadjusted Langevin algorithm
• Langevin diffusions and the Metropolis-adjusted Langevin algorithm
• Variance reduction for irreversible Langevin samplers and diffusion on graphs
• Stochastic simulation: Algorithms and analysis
• (Non-) asymptotic properties of Stochastic Gradient Langevin Dynamics
• Irreversible Langevin samplers and variance reduction: a large deviations approach
• The behavior of the spectral gap under growing drift
• Analysis of Multiscale Integrators for Multiple Attractors and Irreversible Langevin Samplers
• Riemann Manifold Langevin and Hamiltonian Monte Carlo Methods
• Stochastic Processes and Applications

This page was built for publication: Geometry-informed irreversible perturbations for accelerated convergence of Langevin dynamics