research

This is an example of the use of auxiliary variables in improving the performance of an MCMC algorithm. The model is the Ising model: on the left, the Gibbs sampler updates lead to slow mixing, on the right, the Swendsen-Wang algorithm leads to dramatically improved performance.