论文部分内容阅读
The approximation of the posterior distribution associated to a Bayesian inverse problem is decomposed into a hierarchy consisting of a telescoping sum of increments with decreasing variance and increasing cost,and then probed with a sequential Monte Carlo sampler.