Automatic Posterior Transformation for Likelihood-free Inference on March 24

[swagnercarena]

Greenberg et al. 2019 uses normalizing flows to perform Bayesian inference on simulators with intractable likelihoods. This seems close in spirit to what we're trying to do here, and a good review of this paper would be helpful.

[makagan]

A few other related references:

A paper on doing sequential neural likelihood estimation https://arxiv.org/abs/1805.07226

And probably too long for us, but a recent review on normalizing flows for density estimation and inference: https://arxiv.org/abs/1912.02762