All reviewers are uninamously for accept. The work has clear and interesting leveraging hierarchical priors as an extension of GECO and with importance weighted losses. The literature review and related work is well-written (which the reviewers also agree).