Vector representations of graphs and distinguishing quantum product states with one-way LOCC (Q2185850)

This article is devoted to communication measurement for the purpose of quantum information theory; it uses ideas from the theory of local quantum operations and classical communication (LOCC). The idea is briefly introduced in Section 1: Given a set of states in a tensor product of two (finite-dimensional, complex) Hilbert spaces \(\mathcal H_A\) and \(\mathcal H_B\), can Alice and Bob distinguish them by measurements in \(\mathcal H_A\) and \(\mathcal H_B\), respectively? The paper focuses on a specific sub-problem: namely, that of one-way LOCC, in which Alice and Bob are required to perform measurements in a given order. To tackle this problem, the authors translate the assignment in a graph theoretical language and study this reformulation in terms of vector representation of graphs: the details are explained in Section 2. Essentially, given a set of \(r\) product states \(\{|\psi^A_k\rangle\otimes |\psi^B_k\rangle\}_{1\le k\le r}\subset \mathcal H_A\otimes \mathcal H_B\), the authors introduce two unique graphs \(G_A, G_B\) (one for Alice, one for Bob) that can be associated with this system, provided the measurements are performed in a given order; and ask whether the states are distinguishable only based on the graphs. The main results in this paper are Theorems 1 and 2, presented in Section 3 and Section 4, respectively: Theorem 1 gives a sufficient and necessary condition for distinguishablity (via one-way LOCC with Alice measuring first) in terms of properties of an intermediate graph between \(G_A\) and the complement of \(G_B\). Remarkably, this condition is not symmetric in \(A,B\); Theorem 2 thus provides a sufficient and necessary condition for distinguishablity via one-way LOCC regardless of who starts measuring. Sections 5 and 6 delve into the analysis of special cases where graphs belonging to relevant classes (trees, chordal graphs, etc.) arise from the measurements. The article assumes a certain familiarity with the theory of LOCC; a few examples help the layperson to grasp the main ideas. In any case, the reader is gently introduced to virtually all graph theoretical notions and results that are necessary to understand the paper.