Adiabatic operations are powerful tools for robust quantum control in numerous fields of physics, chemistry, and quantum information science. The inherent robustness due to adiabaticity can, however, be impaired in applications requiring short evolution times. We present a single versatile gradient-based optimization protocol that combines adiabatic control with effective Hamiltonian engineering in order to design adiabatic operations tailored to the specific imperfections and resources of an experimental setup. The practicality of the protocol is demonstrated by engineering a fast, 2.3 Rabi cycle-long adiabatic inversion pulse for magnetic resonance with built-in robustness to Rabi field inhomogeneities and resonance offsets. The performance and robustness of the pulse is validated in a nanoscale force-detected magnetic resonance experiment on a solid-state sample, indicating an ensemble-averaged inversion accuracy of 99.997%. We further showcase the utility of our protocol by providing examples of adiabatic pulses robust to spin-spin interactions, parameter-selective operations, and operations connecting arbitrary states, each motivated by experiments.

中文翻译：

---

鲁棒绝热操作的数值工程

绝热操作是在物理，化学和量子信息科学的许多领域中实现强大量子控制的强大工具。但是，由于绝热性导致的固有鲁棒性在需要较短进化时间的应用中会受到损害。我们提出了一个单一的，基于梯度的通用优化协议，该协议将绝热控制与有效的汉密尔顿工程相结合，以设计针对特定缺陷和实验装置资源量身定制的绝热操作。该协议的实用性通过设计一个快速的2.3 Rabi周期长的绝热反转脉冲来进行磁共振，并具有对Rabi磁场非均匀性和共振偏移的内置鲁棒性。脉冲的性能和鲁棒性在固态样品的纳米力检测磁共振实验中得到验证，表明整体平均反演精度为99.997％。我们通过提供对自旋-自旋相互作用，参数选择操作以及连接任意状态的操作均具有鲁棒性的绝热脉冲的示例，进一步展示了本协议的实用性，每种操作均受实验激发。