Statistical analysis of low rank tomography with compressive random measurements

We consider the statistical problem of `compressive' estimation of low rank states with random basis measurements, where the estimation error is expressed terms of two metrics - the Frobenius norm and quantum infidelity. It is known that unlike the case of general full state tomography, low rank states can be identified from a reduced number of observables' expectations. Here we investigate whether for a fixed sample size $N$, the estimation error associated to a `compressive' measurement setup is `close' to that of the setting where a large number of bases are measured. In terms of the Frobenius norm, we demonstrate that for all states the error attains the optimal rate $rd/N$ with only $O(r \log{d})$ random basis measurements. We provide an illustrative example of a single qubit and demonstrate a concentration in the Frobenius error about its optimal for all qubit states. In terms of the quantum infidelity, we show that such a concentration does not exist uniformly over all states. Specifically, we show that for states that are nearly pure and close to the surface of the Bloch sphere, the mean infidelity scales as $1/\sqrt{N}$ but the constant converges to zero as the number of settings is increased. This demonstrates a lack of `compressive' recovery for nearly pure states in this metric.