Remote generalized measurements (POVMs) require non-maximal entanglement

Abstract

We show that non-maximal entangled states can be used for implementing, with unit probability, remote generalized measurements (POVMs). We show how any n-qubit POVM can be applied remotely and derive its entanglement cost. The later turns out to be equal to the entanglement capability for a class of POVMs. This suggests a one-to-one relation between sub-sets of POVM operations and entanglement.