Export Reviews, Discussions, Author Feedback and Meta-Reviews

The paper was interesting to read and contained many nice ideas, intuitions, and theoretical connections. The technical contributions in Sections 3-5 are sound and clever, although in some cases building on previous works [2,4]. Moreover, the problem formulation as a linear optimization and the least squared solution make the approach computationally favorable. However, there are points that need more clarification/elaboration in the paper too, in order to better judge the practical application and impact of the work. For instance, it's never discussed where the set of low order marginal come from.