Nice work. The group testing methods and connections to VI for neural connectivity inference will be of great interest to the NeurIPS community. In addition to the incorporating the reviewers' feedback into the final paper, please also address these minor issues: - You simply assert the marginal polytope constraints in eq (7). Please explain how these are derived from the exponential family form of Q. - Please also cite this relevant past work at NeurIPS: Shababo B, Paige B, Pakman A, Paninski L. Bayesian inference and online experimental design for mapping neural microcircuits. In Advances in Neural Information Processing Systems 2013 (pp. 1304-1312). - Shababo and Paninski (together with Shizhe Chen and Hillel Adesnik) have been extending and applying these methods, but I don't know if it has been published.