AI 中文总结
该研究提出基于单次测量序贯蒙特卡洛估计的实时反馈框架，实现了单量子比特频率及双量子比特耦合的稳定，缩短了参数估计时间，提升了量子比特相干时间，对量子技术发展具重要意义。
AI 中文摘要
快速测量、信号处理及哈密顿参数的精确估计是量子-经典接口电路反馈控制的关键。然而，现有频率派和贝叶斯推理方法通常需要大量测量才能达到缓解量子比特退相干所需的精度，因此半导体量子比特的反馈控制在很大程度上仅限于单量子比特频率稳定，而双量子比特参数稳定在实验上仍未被探索。本文展示了一种基于序贯蒙特卡洛估计的实时反馈框架，该框架利用单次测量的1比特数据。使用四量子比特半导体量子点器件，我们快速估计单个量子比特频率，与传统贝叶斯策略相比，相干时间提高约两倍。此外，使用2比特数据进行序贯双量子比特参数估计可实现量子比特间耦合的稳定，能够估计并抑制准静态频率漂移和交换相互作用噪声。通过缩短精确参数估计所需的时间，这些结果表明经典与量子电子学的协同发展对在波动环境中构建鲁棒且可扩展的量子技术具有重要意义。
英文摘要
Fast measurement, signal processing, and accurate estimation of Hamiltonian parameters are essential for feedback control in quantum-classical interface circuitry. However, existing frequentist and Bayesian inference methods typically require a large number of measurements to achieve the accuracy needed to mitigate qubit decoherence. Consequently, feedback control of semiconductor qubits has largely been limited to single-qubit frequency stabilization, whereas two-qubit parameter stabilization remains experimentally unexplored. Here, we demonstrate a real-time feedback framework based on sequential Monte Carlo estimation using one bit of data from a single-shot measurement. Using a four-qubit semiconductor quantum dot device, we rapidly estimate individual qubit frequencies, yielding an approximately twofold increase in coherence time compared with a conventional Bayesian strategy. Moreover, sequential two-qubit parameter estimation using two bits of data enables stabilization of qubit-qubit coupling, allowing both quasi-static frequency drift and exchange-interaction noise to be estimated and suppressed. By shortening the time required for precise parameter estimation, these results demonstrate the importance of the synergistic development of classical and quantum electronics for building robust and scalable quantum technologies in fluctuating environments.
