Compressible generalized hybrid Monte Carlo.

@article{Fang2014CompressibleGH,
  title={Compressible generalized hybrid Monte Carlo.},
  author={Youhan Fang and Jes{\'u}s Mar{\'i}a Sanz-Serna and Robert D. Skeel},
  journal={The Journal of chemical physics},
  year={2014},
  volume={140 17},
  pages={
          174108
        }
}
One of the most demanding calculations is to generate random samples from a specified probability distribution (usually with an unknown normalizing prefactor) in a high-dimensional configuration space. One often has to resort to using a Markov chain Monte Carlo method, which converges only in the limit to the prescribed distribution. Such methods typically inch through configuration space step by step, with acceptance of a step based on a Metropolis(-Hastings) criterion. An acceptance rate of… 
View on PubMed
arxiv.org
54 Citations
Highly Influential Citations
3
Background Citations
21
Methods Citations
15

54 Citations

Citation Type

Citation Type

On Irreversible Metropolis Sampling Related to Langevin Dynamics
TLDR
This work shows that as the step size tends to 0, the HAMS proposal satisfies a class of stochastic differential equations including Langevin dynamics as a special case, and provides theoretical results for HAMS under the univariate Gaussian setting, including the stationary variance, the expected acceptance rate, and the spectral radius.
Variable length trajectory compressible hybrid Monte Carlo
TLDR
A framework to further extend HMC to a situation in which the dynamics is reversible but not necessarily Hamiltonian and enables an effective application of variable step size integrators to HMC-type sampling algorithms based on reversible dynamics.
Hamiltonian Assisted Metropolis Sampling
  • Zexi Song, Z. Tan
  • Computer Science
    Journal of the American Statistical Association
  • 2021
TLDR
A broad class of irreversible sampling algorithms, called Hamiltonian assisted Metropolis sampling (HAMS), is proposed, and two specific algorithms with appropriate tuning and preconditioning strategies are developed, which facilitate possible further development of irreversible MCMC algorithms.
Geometry & Dynamics for Markov Chain Monte Carlo
TLDR
The aim of this review is to provide a comprehensive introduction to the geometric tools used in Hamiltonian Monte Carlo at a level accessible to statisticians, machine learners and other users of the methodology with only a basic understanding of Monte Carlo methods.
Randomized Hamiltonian Monte Carlo
Tuning the durations of the Hamiltonian flow in Hamiltonian Monte Carlo (also called Hybrid Monte Carlo) (HMC) involves a tradeoff between computational cost and sampling quality, which is typically…
Modified Hamiltonian Monte Carlo for Bayesian inference
TLDR
It is shown that performance of HMC can be significantly improved by incorporating importance sampling and an irreversible part of the dynamics into a chain, and is called Mix & Match Hamiltonian Monte Carlo (MMHMC).
Geometrically Tempered Hamiltonian Monte Carlo
TLDR
A new class of metrics aimed at improving HMC's performance on multi-modal target distributions is proposed, referred to as geometrically tempered HMC (GTHMC) due to its connection to other tempering methods and a novel variable step size integrator is developed.
Mix & Match Hamiltonian Monte Carlo
TLDR
It is shown that performance of HMC can be dramatically improved by incorporating importance sampling and an irreversible part of the dynamics into the chain, and the resulting generalized HMC importance sampler is called Mix & Match Hamiltonian Monte Carlo.
Emulation of higher-order tensors in manifold Monte Carlo methods for Bayesian Inverse Problems
Geometry and Dynamics for Markov Chain Monte Carlo
TLDR
The aim of this review is to provide a comprehensive introduction to the geometric tools used in Hamiltonian Monte Carlo at a level accessible to statisticians, machine learners, and other users of the methodology with only a basic understanding of Monte Carlo methods.
...
1
2
3
4
5
...

References

SHOWING 1-10 OF 47 REFERENCES
Optimal tuning of the hybrid Monte Carlo algorithm
TLDR
It is proved that, to obtain an O(1) acceptance probability as the dimension d of the state space tends to, the leapfrog step-size h should be scaled as h=l ×d−1/ 4, which means that in high dimensions, HMC requires O(d1/ 4 ) steps to traverse the statespace.
Hybrid Monte Carlo on Hilbert spaces
Monte Carlo Sampling Methods Using Markov Chains and Their Applications
SUMMARY A generalization of the sampling method introduced by Metropolis et al. (1953) is presented along with an exposition of the relevant theory, techniques of application and methods and…
Riemann manifold Langevin and Hamiltonian Monte Carlo methods
The paper proposes Metropolis adjusted Langevin and Hamiltonian Monte Carlo sampling methods defined on the Riemann manifold to resolve the shortcomings of existing Monte Carlo algorithms when…
A separable shadow Hamiltonian hybrid Monte Carlo method.
TLDR
The separable shadow Hamiltonian hybrid Monte Carlo (S2HMC) method is introduced, which gives the acceptance rate of a fourth order integrator at the cost of a second-order integrator and has the additional advantage of not requiring any user parameters beyond those of HMC.
Shadow hybrid Monte Carlo: an efficient propagator in phase space of macromolecules
Lagrangian Dynamical Monte Carlo
TLDR
This work proposes an explicit geometric integrator that replaces the momentum variable in RMHMC by velocity and shows that the resulting transformation is equivalent to transforming Riemannian Hamilton dynamics to Lagrangian dynamics.
Numerical Integrators for the Hybrid Monte Carlo Method
TLDR
Limited, proof-of-concept numerical experiments suggest that the new integrators constructed may provide an improvement on the efficiency of the standard Verlet method, especially in problems with high dimensionality.
A generalized guided Monte Carlo algorithm
Hamiltonian Monte Carlo Without Detailed Balance
TLDR
A method for performing Hamiltonian Monte Carlo that largely eliminates sample rejection for typical hyperparameters and significantly suppresses the random walk behavior and wasted function evaluations that are typically the consequence of update rejection is presented.
...
1
2
3
4
5
...