The reviewers all felt positively about the paper, though the scores were somewhat borderline, meaning this paper was right on the threshold for inclusion in the conference. A discussion was initiated to discuss the authors' rebuttal, which resulted in some positive movement of reviewer scores. Overall reviewers felt that the rebuttal seemed reasonable: most of the issues are clarifications, that the authors do a good job of clarifying. A few are hard to address in the rebuttal, such as the inclusion of additional baselines. Overall the reviewers are positive regarding the release of new datasets, and the significance of the results, though the reviewers broadly agreed that additional experimental comparisons would make this paper more convincing.