All reviewers and myself think that the proposed representation is nontrivial and useful. One reviewer still has some concern about the cubic computational complexity, but it's not a major concern. Overall this is a solid paper and deserves publication.