Topic models for text analysis are most commonly trained using either Gibbs sampling or variational Bayes. Recently, hybrid variational-Gibbs algorithms have been found to combine the best of both worlds. Variational algorithms are fast to converge and more efficient for inference on new documents. Gibbs sampling enables sparse updates since each token is only associated with one topic instead ...

Bayesian networks offer great potential for use in automating large scale diagnostic rea­ soning tasks. Gibbs sampling is the main technique used to perform diagnostic reason­ ing in large richly interconnected Bayesian networks. Unfortunately Gibbs sampling can take an excessive time to generate a represen­ tative sample. In this paper we describe and test a number of heuristic strategies for ...

Gibbs sampling is a Markov chain Monte Carlo technique commonly used for estimating marginal distributions. To speed up Gibbs sampling, there has recently been interest in parallelizing it by executing asynchronously. While empirical results suggest that many models can be efficiently sampled asynchronously, traditional Markov chain analysis does not apply to the asynchronous case, and thus asy...

The Gibbs sampler, Metropolis’ algorithm, and similar iterative simulation methods are related to rejection sampling and importance sampling, two methods which have been traditionally thought of as non-iterative. We explore connections between importance sampling, iterative simulation, and importance-weighted resampling (SIR), and present new algorithms that combine aspects of importance sampli...

This paper describes the use of stochastic simulation techniques to reconstruct biomedical signals not directly measurable. In particular, a deconvolution problem with an uncertain clearance parameter is considered. The problem is addressed using a Monte Carlo Markov Chain method, called the Gibbs Sampling, in which the joint posterior probability distribution of the stochastic parameters is es...

Particle Markov chain Monte Carlo (PMCMC) is a systematic way of combining the two main tools used for Monte Carlo statistical inference: sequential Monte Carlo (SMC) and Markov chain Monte Carlo (MCMC). We present a new PMCMC algorithm that we refer to as particle Gibbs with ancestor sampling (PGAS). PGAS provides the data analyst with an off-the-shelf class of Markov kernels that can be used ...