Beyond it’s application to design a new MCMC algorithm, we can use our coupling to construct a perfect sampling algorithm for the q-flow representation of the Potts model. This uses the idea of coupling of MCMC chains, see here our work here: https://t.co/
I look forward to reading this paper, not only because I published about a surprising connections between couplings, extreme value theory and MCMC relaxation / mixing times of some (monotone) heat bath chains. https://t.co/msiDh1kWvl
Below is some of the work I mention (somehow the references to not show): https://t.co/msiDh1kWvl
A similar version of the paper lives on arXiv: https://t.co/ise2hPBF1x
How to estimate the relaxation time of families of monotone heatbath chains via simple forward coupling: check out https://t.co/8eJz5KpNMR (I plan to write a non-physics intro to this)
Collevecchio, Elci, ... : On the coupling time of the heat-bath process for the Fortuin-Kasteleyn random-cluste... https://t.co/qD0eQ5WTiV
On the coupling time of the heat-bath process for the Fortuin-Kasteleyn random-cluster model. https://t.co/p3eVCZY3T8
https://t.co/IuwVQGn2cy A Collevecchio et. al. On the coupling time of the heat-bath process for the Fortuin-Kasteleyn random-cluster mode https://t.co/CyUX4pD3i7