From Church Wiki

To return to the top level: Language Models Tutorial.

enumeration isn't possible, unless it is and recognizing when it is in fact possible is interesting and useful. sampling is always applicable, unless it isn't, because the samplers we build come with basically zero guarantees about their performance

Contents 1 Exact Sampling 2 Rejection Sampling 3 Importance Sampling 3.1 Particle Filtering: Sequential Importance Sampling 3.1.1 Adaptive Importance Sampling 3.2 Annealed Importance Sampling 3.3 MCMC 3.3.1 Gibbs Sampling 3.3.2 Metropolis Hastings 3.4 Psychological Plausibility of Sampling as Inference

Exact Sampling

The simplest kind

Rejection Sampling

Importance Sampling

Particle Filtering: Sequential Importance Sampling

Adaptive Importance Sampling

Annealed Importance Sampling

MCMC

Gibbs Sampling

Metropolis Hastings

Psychological Plausibility of Sampling as Inference