Accurate control of quantum systems requires a precise measurement of the parameters that govern the dynamics, including control fields and interactions with the environment. Parameters will drift in time and experiments interleave protocols that perform parameter estimation, with protocols that measure the dynamics of interest. Here we specialize to a system made of qubits where the dynamics correspond to a quantum computation. We propose setting aside some qubits, which we call spectator qubits, to be measured periodically during the computation, to act as probes of the changing experimental and environmental parameters. By using control strategies that minimize the sensitivity of the qubits involved in the computation, we can acquire sufficient information from the spectator qubits to update our estimates of the parameters and improve our control. As a result, we can increase the length of experiments where the dynamics of the data qubits are highly reliable. In particular, we simulate how spectator qubits can keep the error level of operations on data qubits below a 10⁻⁴ threshold in two scenarios involving coherent errors: a classical magnetic field gradient dynamically decoupled with sequences of two or four π-pulses, and laser beam instability detected via crosstalk with neighboring atoms in an ion trap.

量子系统的精确控制要求对控制动力学的参数进行精确测量，包括控制场和与环境的相互作用。参数会随时间漂移，实验会交错执行执行参数估计的协议，而协议会测量感兴趣的动态。在这里，我们专门研究由量子位构成的系统，其中动力学对应于量子计算。我们建议预留一些量子比特，我们称为旁观者量子比特，在计算过程中定期对其进行测量，以作为不断变化的实验和环境参数的探针。通过使用控制策略来最小化计算中所涉及的量子位的敏感性，我们可以从观众的量子位中获得足够的信息，以更新我们对参数的估计并改善我们的控制。结果，我们可以增加数据量子位的动态高度可靠的实验时间。特别是，我们模拟观众量子位如何将数据量子位的操作错误级别保持在10以下在涉及相干误差的两种情况下的⁻⁴阈值：与两个或四个π脉冲序列动态解耦的经典磁场梯度，以及通过与离子阱中相邻原子的串扰检测到的激光束不稳定性。