On the partially symmetric rank of tensor products of \(W\)-states and other symmetric tensors (Q2418836)

In this paper, the authors focus on the submultiplicativity of the partially symmetric rank, in particular on the tensor \(W_d\in S^d\mathbb{C}^2\subseteq (\mathbb{C}^2)^{\otimes d}\), which is called \(W\)-state in the physics literature, and is defined as \[ W_d=y\otimes \otimes \dots \otimes x+x\otimes y\otimes x\otimes \dots x+\dots+x\otimes \dots\otimes x\otimes y\] \noindent where \(\{x,y\} \) is a basis of \(\mathbb{C}^2\). Tensor decomposition for structured tensors is a classical topic in algebraic geometry that finds numerous applications such as Entanglement, Communication Complexity, Algebraic Complexity Theory, \(W\)-states For instance, \(W\)-states,namely monomials of the form \(x^{d1}y\), are of key importance both in algebraic geometry and quantum information theory. Tensors of type \(W\) are the simplest examples showing that tensor rank fails to be upper semicontinuous. The argument is very interesting and the applications, the tecniques and the tools used in the paper are well described by the authors. In Section 2, the authors provide preliminary results that will be useful in the rest of the paper. In Section 3, they prove several results on the upper bounds of partially symmetric rank. In Section 4, they give lower bounds on the partially symmetric rank. Section 5 is dedicated to results on the set of rank one tensors, and more generally on the zero-dimensional scheme supported at a set of rank 1 tensors, that spans given partially symmetric tensor. Section 7 contains a brief discussion on the classical Sylvester's Theorem for binary forms which inspires most of the techniques used in the rest of the paper, some results about flattening techniques, which are useful tools for lower bounds on several notions of rank and an example giving some insight on the subtleties of zero-dimensional schemes minimally spanning a point.