Sun Dec 8th through Sat the 14th, 2019 at Vancouver Convention Center
This paper introduces an algorithm for performing Metropolis-Hastings sampling with implicit distributions. All three reviewers found the work to be sufficiently interesting and novel, and felt that the empirical analysis suggests the method is promising, viewing the paper overall as borderline but positive. The author response addressed some of the review concerns, as well. NeurIPS is quite competitive but hopefully there will be space to accept this one.