Enhancing qubit readout with Bayesian learning

We introduce an efficient and accurate readout measurement scheme for single and multiqubit states. Our method uses Bayesian inference to build an assignment probability distribution for each qubit state based on a reference characterization of the detector response functions. This allows us to account for system imperfections and thermal noise within the assignment of the computational basis. We benchmark our protocol on a quantum device with five superconducting qubits, testing initial state preparation for single- and two-qubit states and an application of the Bernstein-Vazirani algorithm executed on five qubits. Our method shows a substantial reduction of the readout error and promises advantages for near -term and future quantum devices.