Learning Distributed Quantum State Discrimination with Noisy Classical Communications

Consider a distributed quantum sensing system in which Alice and Bob are tasked with detecting the state of a quantum system that is observed partly at Alice and partly at Bob via local operations and classical communication (LOCC). Prior work introduced LOCCNet, a distributed protocol that optimizes the local operations via parameterized quantum circuits (PQCs) at Alice and Bob. This paper presents Noise Aware-LOCCNet (NA-LOCCNet) for distributed quantum state discrimination in the presence of noisy classical communication. We propose specific ansatzes for the case of two observed qubit pairs, and we describe a noise-aware training design criterion. Through experiments, we observe that quantum, entanglement-breaking, noise on the observed quantum system can be useful in improving the detection capacity of the system when classical communication is noisy.