An entropic approach for Hamiltonian Monte Carlo: the idealized case

Abstract

Quantitative long-time entropic convergence and short-time regularization are established for an idealized Hamiltonian Monte Carlo chain which alternatively follows an Hamiltonian dynamics for a fixed time and then partially or totally refreshes its velocity with an auto-regressive Gaussian step. These results, in discrete time, are the analogous of similar results for the continuous-time kinetic Langevin diffusion, and the latter can be obtained from our bounds in a suitable limit regime. The dependency in the log-Sobolev constant of the target measure is sharp and is illustrated on a mean-field case and on a low-temperature regime, with an application to the simulated annealing algorithm. The practical unadjusted algorithm is briefly discussed.

References in corpus (22)

• Hypocoercivity and exponential time decay for the linear inhomogeneous relaxation Boltzmann equation
• Log-concave sampling: Metropolis-Hastings algorithms are fast
• Randomized Hamiltonian Monte Carlo
• Poincaré and logarithmic Sobolev inequalities by decomposition of the energy landscape
• Coupling and Convergence for Hamiltonian Monte Carlo
• Uniform long-time and propagation of chaos estimates for mean field kinetic particles in non-convex landscapes
• On explicit $L^2$-convergence rate estimate for underdamped Langevin dynamics
• Mixing Time Guarantees for Unadjusted Hamiltonian Monte Carlo
• Convergence of unadjusted Hamiltonian Monte Carlo for mean-field models
• Fast mixing of Metropolized Hamiltonian Monte Carlo: Benefits of multi-step gradients
• Optimal Convergence Rate of Hamiltonian Monte Carlo for Strongly Logconcave Distributions
• On explicit $L^2$-convergence rate estimate for piecewise deterministic Markov processes in MCMC algorithms
• Logarithmic bounds for isoperimetry and slices of convex sets
• Bounding the error of discretized Langevin algorithms for non-strongly log-concave targets
• Small eigenvalues of the low temperature linear relaxation Boltzmann equation with a confining potential
• Second order quantitative bounds for unadjusted generalized Hamiltonian Monte Carlo
• HMC and underdamped Langevin united in the unadjusted convex smooth case
• Uniform minorization condition and convergence bounds for discretizations of kinetic Langevin dynamics
• Score-Based Diffusion meets Annealed Importance Sampling
• Speeding up Langevin Dynamics by Mixing
• Nonlinear recombinations and generalized random transpositions
• Hamiltonian Monte Carlo for efficient Gaussian sampling: long and random steps