Adaptive real-time operations in small spin qubit arrays

Gate-controlled spin qubits offer tunability and programmable operation by voltage signals. Maturing devices from simple single- and two-qubit geometries to arrays of interconnected quantum dots poses challenges that arise for example from cross coupling and gate-offset drifts, necessitating efficient and autonomous operations in the high-dimensional control parameter space associated with qubit arrays.

I will present selected examples from our work encompassing multi-dot devices implemented in silicon structures and gallium-arsenide heterostructures, including active-learning algorithms that autonomously explore the charge stability diagram of capacitively coupled quantum dot arrays and spin rotations via real-time estimation of uncontrolled qubit parameters. These techniques will be important not only for the maturing of high-fidelity spin qubit devices via automated high-throughput device characterization, but also for the stabilization of operating parameters of quantum processors against environmental fluctuations.