arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2304.08166quant-ph
收藏

保流 ZX-演算重写规则用于优化与混淆

Flow-preserving ZX-calculus Rewrite Rules for Optimisation and Obfuscation

Tommy McElvanney, Miriam Backens

AI总结:

本文提出多个增加量子比特并保持泡利流的ZX-演算重写规则,可将任意测量模式转为XY平面内测量,并首次允许任意改变测量角度,从而可能加速双量子比特门优化。

摘要:

在基于测量的量子计算(MBQC)的单向模型中,计算通过对资源态进行测量来进行。所谓的流条件确保整体计算在适当意义上是确定性的,其中泡利流是这些条件中最一般的。表示为测量模式的计算可以被重写以优化资源使用并用于其他目的。此类重写需要保持流的存在,以确保新模式仍然可以确定性地实现。该领域现有的大多数工作集中于减少量子比特数量的重写,然而对于某些类型的优化以及混淆,增加量子比特数量可能是有益的。在本工作中,我们引入了几个增加量子比特数量并保持泡利流存在的ZX-演算重写规则。这些规则可用于将任何测量模式转换为仅包含XY平面内(一般或泡利)测量的模式。我们还给出了第一个允许任意改变测量角度的保流重写规则,并利用该规则证明了Staudacher等人的“邻域解融合”规则保持了泡利流的存在。这意味着通过消除定期运行昂贵的gflow查找算法的需要,可能可以减少其双量子比特门优化过程的运行时间。

英文摘要:

In the one-way model of measurement-based quantum computation (MBQC), computation proceeds via measurements on a resource state. So-called flow conditions ensure that the overall computation is deterministic in a suitable sense, with Pauli flow being the most general of these. Computations, represented as measurement patterns, may be rewritten to optimise resource use and for other purposes. Such rewrites need to preserve the existence of flow to ensure the new pattern can still be implemented deterministically. The majority of existing work in this area has focused on rewrites that reduce the number of qubits, yet it can be beneficial to increase the number of qubits for certain kinds of optimisation, as well as for obfuscation. In this work, we introduce several ZX-calculus rewrite rules that increase the number of qubits and preserve the existence of Pauli flow. These rules can be used to transform any measurement pattern into a pattern containing only (general or Pauli) measurements within the XY-plane. We also give the first flow-preserving rewrite rule that allows measurement angles to be changed arbitrarily, and use this to prove that the `neighbour unfusion' rule of Staudacher et al. preserves the existence of Pauli flow. This implies it may be possible to reduce the runtime of their two-qubit-gate optimisation procedure by removing the need to regularly run the costly gflow-finding algorithm.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑