mala_kernel: Get adaptive coupled Metropolis adjusted Langevin algorithm...

In pierrejacob/unbiasedpathsampling: Unbiased Path Sampling with couplings

Description

This function takes descriptions of the target and tuning parameters, and returns a list containing the keys single_kernel, coupled_kernel corresponding to marginal and coupled MALA kernels. These kernels can then be used in the function coupled_chains.

Usage

get_mala_kernel(logtarget, gradlogtarget, stepsize, dimension,
  adaptive_tol)

Arguments

logtarget	function to compute target log-density, e.g. see get_mvnormal
gradlogtarget	function to compute gradient of target log-density, e.g. see get_mvnormal
stepsize	standard deviation of Gaussian increment
dimension	dimension of the target distribution
adaptive_tol	threshold below which maximal coupling is triggered

Value

A list containing the keys single_kernel, coupled_kernel.

pierrejacob/unbiasedpathsampling documentation built on May 17, 2019, 12:03 p.m.