NeurIPS 2019
Sun Dec 8th through Sat the 14th, 2019 at Vancouver Convention Center
Paper ID:4734
Title:Unlocking Fairness: a Trade-off Revisited

This paper received considerable discussion. Ultimately the reviewers reached consensus that this paper should be accepted -despite- finding the paper lacking in both the strength of the technical contribution and the experimental validation. Reviewers found that the ideas introduced in this work were more interesting and potentially significant than most papers. One reviewer noted that the paper, despite all of its shortcomings, has already changed the way s/he thinks about some elements of algorithmic fairness. *IMPORTANT* Again, I iterate that, while there is strong enthusiasm for the ideas, reviewers remain dissatisfied with the execution. The authors would be doing a disservice to their own work by not making every effort to improve the manuscript to strengthen the technical and experimental contributions prior to publication. This is a case where reviewers are willing to take a chance on a paper in hopes that the authors take initiative to improve their work in the camera-ready.