论文标题
一种去极化噪声感知的转侧，以实现最佳振幅扩增
A Depolarizing Noise-aware Transpiler for Optimal Amplitude Amplification
论文作者
论文摘要
当在与环境完全隔离的量子机上运行时，振幅放大为一系列量子算法提供了二次加速。但是，由于NISQ时代的量子机缺乏提供误差校正所需的大量量子，因此优势大大降低了。计算中的噪声随电路中的门数数量而增长，每次遍历振幅扩增。经过一定数量的扩增，栅极噪声的准确性损失开始掩盖由于扩增而导致的准确性增长，从而形成了一个拐点。除此之外，精度继续恶化，直到机器达到结果均匀随机的最大混合状态为止。因此，量子转侧应考虑基础量子机的噪声参数，以便可以优化电路以实现该机器的最大精度。在这项工作中，我们提出了向转板器的扩展，该扩展通过将纯贝叶斯分析应用于单个门噪声速率来预测每个放大时结果的准确性。使用此信息，它找到了拐点并通过停止扩增来优化电路。进行预测无需在量子模拟器或实际量子机上执行电路。
Amplitude amplification provides a quadratic speed-up for an array of quantum algorithms when run on a quantum machine perfectly isolated from its environment. However, the advantage is substantially diminished as the NISQ-era quantum machines lack the large number of qubits necessary to provide error correction. Noise in the computation grows with the number of gate counts in the circuit with each iteration of amplitude amplification. After a certain number of amplifications, the loss in accuracy from the gate noise starts to overshadow the gain in accuracy due to amplification, forming an inflection point. Beyond this point, accuracy continues to deteriorate until the machine reaches a maximally mixed state where the result is uniformly random. Hence, quantum transpilers should take the noise parameters of the underlying quantum machine into consideration such that the circuit can be optimized to attain the maximal accuracy possible for that machine. In this work, we propose an extension to the transpiler that predicts the accuracy of the result at every amplification with high fidelity by applying pure Bayesian analysis to individual gate noise rates. Using this information, it finds the inflection point and optimizes the circuit by halting amplification at that point. The prediction is made without needing to execute the circuit either on a quantum simulator or an actual quantum machine.
