ffe10334251de1dc98339d99ae4743ba-AuthorFeedback.pdf

To3 our knowledge, techniques up to this point have been mostly focused on special case models. As Reviewer 1 pointed out, it is not always clear that the increase in running time is worth the savings in7 memory. Ifthatmodel8 doubled in size, it could take weeks of engineering work to get it to run. We provide a formalization of the problem with rigorous guarantees. As several reviewers noted, this is a13 useful contribution.