Error-Robust Quantum Logic Optimization Using a Cloud Quantum Computer Interface

We describe an experimental effort designing and deploying error-robust single-qubit operations using a cloud-based quantum computer analog-layer programming access. design numerically-optimized pulses that implement target exhibit robustness to various error processes including dephasing noise, instabilities in control amplitudes, crosstalk. Pulse optimization is performed flexible package incorporating device model physically-relevant constraints (e.g. bandwidth limits on the transmission lines of dilution refrigerator housing IBM Quantum hardware). present techniques for conversion calibration physical Hamiltonian definitions pulse waveforms programmed via Qiskit compare performance against hardware default DRAG five-qubit device. Experimental measurements reveal coherent errors order magnitude larger than tabulated randomized-benchmarking measurements; solutions designed be robust these outperform hardware-default all qubits across multiple metrics. demonstrate enhancements up to: $\sim10\times$ gate coherent-error reduction; $\sim5\times$ average reduction five qubit system; increase window one week valid calibration; $\sim12\times$ gate-error variability over time; $\sim9\times$ (including crosstalk) presence fully parallelized operations. Randomized benchmarking reveals rates Clifford gates constructed from optimized consistent with $T_{1}$ limits, demonstrates narrowing distribution outcomes randomizations associated suppression coherent-errors.