Taking Bigger Metropolis Steps by Dragging Fast Variables

Abstract

I show how Markov chain sampling with the Metropolis-Hastings algorithm can be modified so as to take bigger steps when the distribution being sampled from has the characteristic that its density can be quickly recomputed for a new point if this point differs from a previous point only with respect to a subset of 'fast' variables. I show empirically that when using this method, the efficiency of sampling for the remaining 'slow' variables can approach what would be possible using Metropolis updates based on the marginal distribution for the slow variables.

Cited by in corpus (35)

• DESI 2024 VI: Cosmological Constraints from the Measurements of Baryon Acoustic Oscillations
• Early Dark Energy Does Not Restore Cosmological Concordance
• GetDist: a Python package for analysing Monte Carlo samples
• MCMC for doubly-intractable distributions
• Constraining Early Dark Energy with Large-Scale Structure
• Robust Preference for Dynamical Dark Energy in DESI BAO and SN Measurements
• The Atacama Cosmology Telescope: Constraints on Pre-Recombination Early Dark Energy
• Nonequilibrium candidate Monte Carlo: A new tool for efficient equilibrium simulation
• A double take on early and interacting dark energy from JWST
• New cosmological bounds on hot relics: Axions $\&$ Neutrinos
• Gravity in the infrared and effective nonlocal models
• Estimating Ratios of Normalizing Constants Using Linked Importance Sampling
• Bayesian evidence for the tensor-to-scalar ratio $r$ and neutrino masses $m_{\nu }$: Effects of uniform vs logarithmic priors
• Cosmology with 6 parameters in the Stage-IV era: efficient marginalisation over nuisance parameters
• MCMC Using Ensembles of States for Problems with Fast and Slow Variables such as Gaussian Process Regression
• Constraining the dynamical dark energy parameters: Planck-2013 vs WMAP9
• Optimal parameterizations for observational constraints on thawing dark energy
• The BAHAMAS project: Effects of a running scalar spectral index on large-scale structure
• Direct cosmological inference from three-dimensional correlations of the Lyman-$\alpha$ forest
• The Einstein Ring 0047-2808 Revisited: A Bayesian Inversion
• A generalized method of constraining Warm Inflation with CMB data
• A multitracer analysis for the eBOSS galaxy sample based on the effective field theory of large-scale structure
• On the utility of Metropolis-Hastings with asymmetric acceptance ratio
• Constraining Holographic Dark Energy and Analyzing Cosmological Tensions
• Testing $\alpha$-attractor quintessential inflation against CMB and low-redshift data
• The Hubble tension in the non-flat Super-$\Lambda \mathrm{CDM}$ model
• Interpreting Internal Consistency of DES Measurements
• Constraints on dark energy from TDCOSMO & SLACS lenses
• Efficient parameter sampling for Markov jump processes
• One Extension to Explain Them All, One Scale-Invariant Spectrum to Test Them All, and in One Model Bind Them
• Scalable Monte Carlo inference for state-space models
• Metropolis-Hastings with Averaged Acceptance Ratios
• Constraint on Neutrino Statistics from Cosmological Data
• A Python compressed low-$\ell$ Planck likelihood for temperature and polarization
• Resolution exchange simulation with incremental coarsening