7cac11e2f46ed46c339ec3d569853759-AuthorFeedback.pdf

However, we do not view this as a fundamental4 weakness of the paper. Different problems require different trade-offs between, for example, precision and speed.5 Our evaluation metrics (ˆk, distance between the moments) reflect these specific goals. A more general study of6 the suite of tools we introduce applied to other problems would certainly be interesting, but is beyond the scope of7 our paper. Weonly became aware of[1] after our submission, sowe will add asubstantial discussion ofthis paper9 in the camera-ready version.