Optimal friction matrix for underdamped Langevin sampling
From MaRDI portal
Abstract: A systematic procedure for optimising the friction coefficient in underdamped Langevin dynamics as a sampling tool is given by taking the gradient of the associated asymptotic variance with respect to friction. We give an expression for this gradient in terms of the solution to an appropriate Poisson equation and show that it can be approximated by short simulations of the associated first variation/tangent process under concavity assumptions on the log density. Our algorithm is applied to the estimation of posterior means in Bayesian inference problems and reduced variance is demonstrated when compared to the original underdamped and overdamped Langevin dynamics in both full and stochastic gradient cases.
Recommendations
- Using perturbed underdamped Langevin dynamics to efficiently sample from probability distributions
- High-dimensional MCMC with a standard splitting scheme for the underdamped Langevin diffusion
- Appropriate state-dependent friction coefficient accelerates kinetic Langevin dynamics
- Is there an analog of Nesterov acceleration for gradient-based MCMC?
- Adaptive stepsize algorithms for Langevin dynamics
Cites work
- Almost sure contraction for diffusions on \(\mathbb{R}^d\). Application to generalized Langevin diffusions
- Analysis of SPDEs arising in path sampling. I: The Gaussian case
- Analysis of SPDEs arising in path sampling. II: The nonlinear case
- Applications of Malliavin calculus to Monte Carlo methods in finance
- Asymptotic analysis of the Green-Kubo formula
- Calculating the Greeks by cubature formulae
- Central limit theorems for additive functionals of ergodic Markov diffusions processes
- Control variates for stochastic gradient MCMC
- Degenerate Fokker-Planck equations: Bismut formula, gradient estimate and Harnack inequality
- Fluctuations in Markov processes. Time symmetry and martingale approximation.
- Free energy computations. A mathematical perspective
- High-dimensional MCMC with a standard splitting scheme for the underdamped Langevin diffusion
- High-order Langevin diffusion yields an accelerated MCMC algorithm
- scientific article; zbMATH DE number 3951715 (Why is no real title available?)
- scientific article; zbMATH DE number 1375552 (Why is no real title available?)
- scientific article; zbMATH DE number 3505982 (Why is no real title available?)
- scientific article; zbMATH DE number 2006037 (Why is no real title available?)
- scientific article; zbMATH DE number 7626757 (Why is no real title available?)
- Hypoelliptic estimates and spectral theory for Fokker-Planck operators and Witten Laplacians
- Hypoelliptic second order differential equations
- Infinite dimensional analysis. A hitchhiker's guide.
- Malliavin matrix of degenerate SDE and gradient estimate
- MCMC METHODS FOR DIFFUSION BRIDGES
- MCMC methods for sampling function space
- Non-equilibrium statistical mechanics of strongly anharmonic chains of oscillators.
- On sampling from a log-concave density using kinetic Langevin diffusions
- On the functional central limit theorem and the law of the iterated logarithm for Markov processes
- On the Generalized Langevin Equation for Simulated Annealing
- Optimal linear drift for the speed of convergence of an hypoelliptic diffusion
- Optimal non-reversible linear drift for the convergence to equilibrium of a diffusion
- Partial differential equations and stochastic methods in molecular dynamics
- Rational construction of stochastic numerical methods for molecular sampling
- Sampling conditioned hypoelliptic diffusions
- Simulation of conditioned diffusion and application to parameter estimation
- Some methods of speeding up the convergence of iteration methods
- Some Remarks on Preconditioning Molecular Dynamics
- Stochastic gradient Markov chain Monte Carlo
- Stochastic processes and applications. Diffusion processes, the Fokker-Planck and Langevin equations
- Stochastic stability of differential equations. With contributions by G. N. Milstein and M. B. Nevelson
- The computation of averages from equilibrium and nonequilibrium Langevin molecular dynamics
- The Smoluchowski-Kramers limit of stochastic differential equations with arbitrary state-dependent friction
- Trend to equilibrium and particle approximation for a weakly selfconsistent Vlasov-Fokker-Planck equation
- Uniform in time estimates for the weak error of the Euler method for SDEs and a pathwise approach to derivative estimates for diffusion semigroups
- Using perturbed underdamped Langevin dynamics to efficiently sample from probability distributions
- Variance reduction for Markov chains with application to MCMC
- Variance reduction using nonreversible Langevin samplers
Cited in
(6)- Regularity preservation in Kolmogorov equations for non-Lipschitz coefficients under Lyapunov conditions
- Appropriate state-dependent friction coefficient accelerates kinetic Langevin dynamics
- Optimizing the diffusion coefficient of overdamped Langevin dynamics
- Optimal importance sampling for overdamped Langevin dynamics
- Title not available (Why is no real title available?)
- Reflection coupling for unadjusted generalized Hamiltonian Monte Carlo in the nonconvex stochastic gradient case
This page was built for publication: Optimal friction matrix for underdamped Langevin sampling
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6181262)
