Efficient scheduling of noise characterization protocols in quantum computing architectures

Spectator qubits embedded in quantum computing architectures enable in situ detection of noise processes affecting quantum hardware. Common spatial correlations between processing and spectator qubits present a new resource which may be exploited in efficient scheduling of measurements. We present an algorithmic framework for 2D field mapping in quantum computing architectures using sparse measurements. We adapt classical simultaneous localisation and mapping (SLAM) techniques to enable spatial field characterization using idle qubits, where idle qubits are a static or a dynamically available resource. A Quantum SLAM (QSLAM) framework is implemented via a particle filter that shares information between neighboring qubits while discovering neighborhood sizes relevant to the physical system; an adaptive controller then schedules future measurements based on this algorithm. We use experimental measurements on a linear array of trapped ions subject to an observed but uncontrolled magnetic field gradient. Numeric simulations demonstrate that QSLAM outperforms a brute force approach for estimating the magnetic field gradient by over an order of magnitude across a range of operating parameter regimes. Extensions to incorporating time dynamics are discussed.