I just saw this paper recently which mentioned that the optimization on a Grassmanian Manifold can be used to get an achieve an best approximation of a multilinear rank of a tensor (in the sense of a multidimensional matrix, also called a hypermatrix). Does anyone happen to know what Grassmanians have to do with tensors and their analysis? And where could I learn more about Grassmanian manifolds (enough to understand the use here, more is of course welcome =)

The paper is here for those interested Paper

(I hope this isn't too specific)