7d12b66d3df6af8d429c1a357d8b9e1a-AuthorFeedback.pdf

We thank the reviewers for their invested time and constructive criticism. Inference isdone through optimising27 asandwich bound of the ELBO, which isspecific to the choice of implicit distributions. That is an interesting question and we have run some selective experiments, where we combined REWO with the36 VampPrior andthestandard normal prior.