Examples:

• variational inference via (reverse-mode AD on a subset of Metaprob) applied to (Bayesian linear regression, with a naive mean field variational approximation)
• importance resampling, single-site "lightweight" MH, and rejection sampling, on the toy Gaussian example
• hmm forward filtering w/ backward sampling, maybe on a real dataset from Marco's incremental inference paper
• sequential Monte Carlo filtering, given a sequence of addresses, applied to the same HMM example *** VKM can help define and formulate
• using MH inference to infer the posterior distribution on accepted particles and rejected particles in the toy Gaussian example *** VKM can help formulate