This paper was reviewed by three experts in the field and received 667 recommendations. Based on the reviewers' feedback, the decision is to recommend the paper for acceptance to NeurIPS 2019. The reviewers did raise some valuable concerns among which are further interpretation of the dependency matrix, evidence of the mutual complement of the two networks, cross-dataset generalization, etc. These questions should be addressed in the final camera-ready version of the paper. The authors are encouraged to make the necessary changes to the best of their ability. We congratulate the authors on the acceptance of their paper!