Robust Characterization of Loss Rates

Understand

Many physical implementations of qubits---including ion traps, optical lattices and linear optics---suffer from loss.

• A nonzero probability of irretrievably losing a qubit can be a substantial obstacle to fault-tolerant methods of processing quantum information, requiring new techniques to safeguard against loss that introduce an additional overhead that depends upon the loss rate.
• Here we present a scalable and platform-independent protocol for estimating the average loss rate (averaged over all input states) resulting from an arbitrary Markovian noise process, as well as an independent estimate of detector efficiency.
• Moreover, we show that our protocol gives an additional constraint on estimated parameters from randomized benchmarking that improves the reliability of the estimated error rate and provides a new indicator for non-Markovian signatures in the experimental data.