In this paper, two conjectures which were proposed in Luo et al (2010 Phys. Rev. A 82 052122 ) on the correlations in a bipartite state ρ AB are studied. If the mutual information I (ρ AB ) between two quantum systems A and B before any measurement is considered as the total amount of correlations in the state ρ AB , then it can be separated into two parts: classical correlations and quantum correlations. The so-called classical correlations C (ρ AB ) in the state ρ AB are defined by the maximizing mutual information between two quantum systems A and B after von Neumann measurements on system B . We show that it is upper bounded by the von Neumann entropies of both subsystems A and B , which answered the conjecture on the classical correlation. If the quantum correlations Q (ρ AB ) in the state ρ AB are defined by Q (ρ AB ) = I (ρ AB ) − C (ρ AB ), we also show that it is upper bounded by the von Neumann entropy of subsystem B . It is also found that Q (ρ AB ) is upper bounded by the von Neumann entropy of subsystem A for a class of states.
In this paper, two conjectures which were proposed in [Phys. Rev. A \textbf{82}, 052122(2010)] on the correlations in a bipartite state
ρAB are studied. If the mutual information $I\Pa{\rho^{AB}}$ between two quantum systems
A and
B before any measurement is considered as the total amount of correlations in the state
ρAB, then it can be separated into two parts: classical correlations and quantum correlations. The so-called classical correlations $C\Pa{\rho^{AB}}$ in the state
ρAB, defined by the maximizing mutual information between two quantum systems
A and
B after von Neumann measurements on system
B, we show that it is upper bounded by the von Neumann entropies of both subsystems
A and
B, this answered the conjecture on the classical correlation. If the quantum correlations $Q\Pa{\rho^{AB}}$ in the state
ρAB is defined by $Q\Pa{\rho^{AB}}= I\Pa{\rho^{AB}} - C\Pa{\rho^{AB}}$, we show also that it is upper bounded by the von Neumann entropy of subsystem
B. It is also obtained that $Q\Pa{\rho^{AB}}$ is upper bounded by the von Neumann entropy of subsystem
A for a class of states.
The presence of correlations shared among the subparts of a given physical system, and which do not have a classical counterpart, is a clear signature of non-classicality for a quantum state. As we saw in Chap. 2, such correlations can be captured by entanglement, one of the most intriguing features predicted by quantum theory, which points directly to a violation of the Bell’s inequalities. In this chapter we provide a rigorous definition of entanglement, and review its basic properties together with the general framework that is needed to quantify it for bipartite states. We present the common criteria that are used to detect whether a state is entangled or not, and discuss the most common entanglement quantifiers in the bipartite context. The latter follow an introduction of the cornerstone concepts of classical Shannon entropy and quantum von Neumann entropy. Subsequently, the multipartite scenario is briefly addressed. Although entanglement is the most remarkable among non-classical correlations, even some separable (not entangled) mixed states can exhibit a non-classical behaviour. In this wider context we discuss the quantum discord, which is expressed as the difference between two entropic quantities and is extractable by performing a local measure on one of the subsystems. This, and other related quantifiers, enlighten the impossibility of recovering the information contained in a composite quantum system after the measurement process. A special-topic section on the different definitions of entropy used in physics closes the chapter.
On the premises that total correlations in a bipartite quantum state are measured by the quantum mutual information, and that separation of total correlations into quantum and classical parts satisfies an intuitive dominance relation, we examine to what extent various entropic entanglement measures, such as the distillable entanglement, the relative entropy entanglement, the squashed entanglement, the entanglement cost, and the entanglement of formation, can be regarded as consistent measures of quantum correlations. We illustrate that the entanglement of formation often overestimates quantum correlations and thus is too big to be a genuine measure of quantum correlations. This indicates that the entanglement of formation does not quantify the quantum correlations intrinsic to a quantum state, but rather characterizes the pure entanglement needed to build the quantum state via local operations and classical communication. Furthermore, it has the consequence that, if the additive conjecture for the entanglement of formation is true (as is widely believed), then the entanglement cost, which is an operationally defined measure of entanglement with significant physical meaning, cannot be a consistent measure of quantum correlations in the sense that it may exceed total correlations. Alternatively, if the entanglement cost is dominated by total correlations, as our intuition suggests, then we can immediately disprove the additive conjecture. Both scenarios have their counterintuitive and appealing aspects, and a natural challenge arising in this context is to prove or disprove that the entanglement cost is dominated by the quantum mutual information.