Monte Carlo sampling for stochastic weight functions

@article{Frenkel2017MonteCS,
title={Monte Carlo sampling for stochastic weight functions},
author={D. Frenkel and K. J. Schrenk and Stefano Martiniani},
journal={Proceedings of the National Academy of Sciences},
year={2017},
volume={114},
pages={6924 - 6929}
}

Significance Markov chain Monte Carlo is the method of choice for sampling high-dimensional (parameter) spaces. The method requires knowledge of the weight function (or likelihood function) determining the probability that a state is observed. However, in many numerical applications the weight function itself is fluctuating. Here, we present an approach capable of tackling this class of problems by rigorously sampling states proportionally to the average value of their fluctuating likelihoodâ€¦Â Expand