Separation between Entanglement Criteria and Entanglement Detection Protocols (2403.01664v1)

Abstract: Entanglement detection is one of the most fundamental tasks in quantum information science, playing vital roles in theoretical studies and quantum system benchmarking. Researchers have proposed many powerful entanglement criteria with high detection capabilities and small observable numbers. Nonetheless, entanglement criteria only represent mathematical rules deciding the existence of entanglement. The relationship between a good entanglement criterion and an effective experimental entanglement detection protocol (EDP) is poorly understood. In this study, we introduce postulates for EDPs about their detection capabilities and robustness and use them to show the difference between entanglement criteria and EDPs. Specifically, we design an entanglement detection task for unknown pure bipartite states and demonstrate that the sample complexity of any EDP and the number of observables for a good entanglement criterion can have exponential separation. Furthermore, we discover that the optimal EDP with the lowest sample complexity does not necessarily correspond to the optimal entanglement criterion with the fewest observables. Our results can be used to prove the exponential speedups achieved through quantum memory and be generalized to multipartite entanglement detection. By highlighting the significance and independence of EDP design, our work holds practical implications for entanglement detection experiments.