Geometric integrators and the Hamiltonian Monte Carlo method

Abstract

This paper surveys in detail the relations between numerical integration and the Hamiltonian (or hybrid) Monte Carlo method (HMC). Since the computational cost of HMC mainly lies in the numerical integrations, these should be performed as efficiently as possible. However, HMC requires methods that have the geometric properties of being volume-preserving and reversible, and this limits the number of integrators that may be used. On the other hand, these geometric properties have important quantitative implications on the integration error, which in turn have an impact on the acceptance rate of the proposal. While at present the velocity Verlet algorithm is the method of choice for good reasons, we argue that Verlet can be improved upon. We also discuss in detail the behavior of HMC as the dimensionality of the target distribution increases.

Final version will appear in Acta Numerica 2018

References in corpus (3)

• Palindromic 3-stage splitting integrators, a roadmap
• Couplings and quantitative contraction rates for Langevin dynamics
• Cayley Splitting for Second-Order Langevin Stochastic Partial Differential Equations

Cited by in corpus (40)

• Computing quantum dynamics in the semiclassical regime
• Coupling and Convergence for Hamiltonian Monte Carlo
• Cayley modification for strongly stable path-integral and ring-polymer molecular dynamics
• Dimension-free path-integral molecular dynamics without preconditioning
• On the convergence of Hamiltonian Monte Carlo
• A generalized class of strongly stable and dimension-free T-RPMD integrators
• Unbiased Markov chain Monte Carlo with couplings
• Efficient Learning of the Parameters of Non-Linear Models using Differentiable Resampling in Particle Filters
• Mixing Time Guarantees for Unadjusted Hamiltonian Monte Carlo
• Convergence of unadjusted Hamiltonian Monte Carlo for mean-field models
• On explicit $L^2$-convergence rate estimate for piecewise deterministic Markov processes in MCMC algorithms
• Multi-stage splitting integrators for sampling with modified Hamiltonian Monte Carlo methods
• Simulating Coulomb gases and log-gases with hybrid Monte Carlo algorithms
• Quantum Dynamical Hamiltonian Monte Carlo
• Two-scale coupling for preconditioned Hamiltonian Monte Carlo in infinite dimensions
• Higher Order Hamiltonian Monte Carlo Sampling for Cosmological Large-Scale Structure Analysis
• Geometric Methods for Sampling, Optimisation, Inference and Adaptive Agents
• Constraint-Based Regularization of Neural Networks
• Cores for Piecewise-Deterministic Markov Processes used in Markov Chain Monte Carlo
• Efficient Sampling of Thermal Averages of Interacting Quantum Particle Systems with Random Batches
• MetFlow: A New Efficient Method for Bridging the Gap between Markov Chain Monte Carlo and Variational Inference
• Mixing of Metropolis-Adjusted Markov Chains via Couplings: The High Acceptance Regime
• Sampling algorithms in statistical physics: a guide for statistics and machine learning
• Adaptive multi-stage integration schemes for Hamiltonian Monte Carlo
• GPU-Accelerated Particle Methods for Evaluation of Sparse Observations for PDE-Constrained Inverse Problems
• Estimating Normalizing Constants for Log-Concave Distributions: Algorithms and Lower Bounds
• HMC: avoiding rejections by not using leapfrog and some results on the acceptance rate
• A Unifying and Canonical Description of Measure-Preserving Diffusions
• Incorporating Local Step-Size Adaptivity into the No-U-Turn Sampler using Gibbs Self Tuning
• Markov-chain sampling for long-range systems without evaluating the energy
• Unadjusted Hamiltonian MCMC with Stratified Monte Carlo Time Integration
• Zigzag path connects two Monte Carlo samplers: Hamiltonian counterpart to a piecewise deterministic Markov process
• Is the NUTS algorithm correct?
• Wasserstein distance estimates for the distributions of numerical approximations to ergodic stochastic differential equations
• Improving sampling efficacy on high dimensional distributions with thin high density regions using Conservative Hamiltonian Monte Carlo
• Motion equations in a Kerr-Newman-de Sitter spacetime: some methods of integration and application to black holes shadowing in Scilab
• Randomized Runge-Kutta-Nyström Methods for Unadjusted Hamiltonian and Kinetic Langevin Monte Carlo
• Simple third order operator-splitting schemes for stochastic mechanics and field theory
• Entropy-based adaptive Hamiltonian Monte Carlo
• Symmetrically processed splitting integrators for enhanced Hamiltonian Monte Carlo sampling