Improving quantum hardware performance using inverse noise matrices

With the advent of fault-tolerant quantum computing still far on the horizon, highly-entangled state preparation algorithms are not realizable on large numbers of qubits. Due to their versatility and the demand they place on the hardware, these algorithms can be useful for studying the noise on the hardware. We investigate how we can improve the performance of a quantum circuit by exploiting knowledge about the particular noise that affects the qubits.

To this end, we use the noisy output of the state-preparation algorithm to identify the noise on IBM quantum hardware. By identifying the dominant source of error, we reconstruct a noise model of the given hardware. Partial error mitigation is achieved by approximately mapping the noisy channel to a noiseless one. We further discuss our work in the context of the Klco-Salvage state preparation algorithm, as a minimally-entangled approximation to a discretized Gaussian.