NeurIPS 2020

Gaussian Process Bandit Optimization of the Thermodynamic Variational Objective

Meta Review

Reviewers are positive about the contributions of this paper and feel that the authors addressed their primary concerns. Please update the paper with the revisions proposed in the rebuttal. Further, R4 suggests that a fair comparison would include the search time for the GP hyperparameters or show insensitivity. I agree and think this would strengthen the paper.