When one develops quantum information processors with larger numbers of qubits, correlated dephasing errors and crosstalk can have an important effect on the overall accuracy of the device. Detecting these correlations can require a large number of measurements. We consider the special case where the correlations are sparse, i.e., for a system of n qubits, there are k pairs of qubits that are correlated, where k << n(n-1)/2. We propose a method that uses compressed sensing and entangled measurements to find these correlated pairs of qubits more efficiently, using poly(k log n) measurement settings.
