While the encoding of the problem is clever and likely of interest to the NeurIPS community, the reviewers highlighted important concerns for this applied paper. First, clear discussion of related work should be in the main paper (in my opinion) in a single location. Second, and most importantly, an applied paper that uses noiseless data is a hard sell. The authors need to show the results on data that has appropriate, real-world noise. Third, while the authors say it will be self-contained, it is hard to imagine with the page limits how this will happen.