Figure 1

Scenarios for entanglement certification. Red denotes trusted states/devices. (a) Standard Bell scenario for device-independent entanglement certification. The estimated probabilities $p(ab|xy)$ are tested for the violation of a Bell inequality in order to certify the entanglement of the state ${\varrho }^{\mathrm{A}\mathrm{B}}$. (b) Scenario for MDI entanglement certification. Here, the inputs are given by trusted quantum states ${\psi }_x$ and ${\varphi }_y$. (c) Equivalent MDI scenario in which the inputting of the states ${\psi }_x$ and ${\varphi }_y$ in scenario (b) is replaced by giving Alice and Bob each one half of a maximally entangled state and performing local measurements on them. (d) Our proposal for DI entanglement certification. The entangled state ${\varrho }^{\mathrm{A}\mathrm{B}}$ to be detected is placed in a network containing additional auxiliary entangled states. Using self-testing techniques, these entangled states are certified to be maximally entangled and perform the expected measurements as required in (c).