Skip to main content
QUICK REVIEW

[论文解读] Reducing the Complexity of Matrix Multiplication to $O(N^2log_2N)$ by an Asymptotically Optimal Quantum Algorithm

Jiaqi Yao, Ding Liu|arXiv (Cornell University)|Feb 5, 2026
Quantum Computing Algorithms and Architecture被引用 0
一句话总结

论文提出基于量子核的矩阵乘法(QKMM)算法,在渐近意义上实现最优的量子门复杂性 O(N^2 log N),超越经典最佳 O(N^2.371552),通过无噪声和有噪声仿真验证其理论与实际优势。

ABSTRACT

Matrix multiplication is a fundamental classical computing operation whose efficiency becomes a major challenge at scale, especially for machine learning applications. Quantum computing, with its inherent parallelism and exponential storage capacity, offers a potential solution to these limitations. This work presents a quantum kernel-based matrix multiplication algorithm (QKMM) that achieves an asymptotically optimal computational complexity of $ O(N^2 \log_2 N) $, outperforming the classical optimal complexity of $ O(N^{2.371552}) $, where $N$ denotes the matrix dimension. Through noiseless and noisy quantum simulation experiments, we demonstrate that the proposed algorithm not only exhibits superior theoretical efficiency but also shows practical advantages in runtime performance and stability.

研究动机与目标

  • 推动对计算密集型任务(如 AI 与科学计算)中快速矩阵乘法的需求
  • 提出一个量子核框架(QKMM)以加速矩阵乘法
  • 证明 QKMM 达到渐近最优的 O(N^2 log N) 量子门复杂度
  • 将 QKMM 与已知的经典上界及早先的量子方法进行比较
  • 通过无噪声和有噪声量子仿真与电路级分析验证该方法

提出的方法

  • 提出带振幅编码数据与基于核的内积的量子核基矩阵乘法(QKMM)
  • 发展四种电路族:V2V(向量到向量)、V2M(向量到矩阵)、M2M(矩阵到矩阵)和 MMM(多矩阵),以在单一电路中实现并行计算
  • 使用多重受控振幅编码与索引寄存器将行/列映射到单位矩阵
  • 分析资源计数,给出 N×N 矩阵的门级复杂度与量子比特需求
  • 提供电路设计(含图示)与每个子程序的详细演化步骤
  • 在无噪声和现实噪声模型下进行仿真以评估保真度与鲁棒性
Figure 1: Comparison of time complexity between QKMM algorithm and classical algorithms.
Figure 1: Comparison of time complexity between QKMM algorithm and classical algorithms.

实验结果

研究问题

  • RQ1量子核基框架是否能够为矩阵乘法实现渐近最优的量子门复杂度?
  • RQ2就资源计数和可扩展性而言,QKMM 与经典最佳上界 O(N^2.371552) 及先前量子方法相比如何?
  • RQ3噪声对 QKMM 性能有何影响,并且并行化如何影响鲁棒性与保真度?
  • RQ4V2V、V2M、M2M 与 MMM 各组成部分对整体加速与实际可行性有何贡献?

主要发现

  • QKMM 实现渐近最优的 O(N^2 log N) 量子门复杂度,且在大 N 时接近 O(N^2) 的下界
  • QKMM 的表现超越已知的经典矩阵乘法上界 O(N^2.371552)
  • 无噪声仿真显示在 V2V、V2M、M2M、MMM 变体中,矩阵规模增大时存在显著的并行加速
  • 在有噪声仿真中,保真度随维度和噪声耦合下降,门噪声在更深电路中是主要因素
  • QKMM 在真实噪声下的内积任务中,相较 Hadamard 测试和 Swap 测试基准,保真度更高且误差增长更慢
  • 该工作提供完整的电路实现和基于门数的性能指标,便于与经典及先前的量子方法直接比较
Figure 2: Time consumption analysis of matrix multiplication based on different quantum circuits; (a) Comparison of V 2 V computational efficiency; (b) Comparison of V 2 M computational efficiency; (c) Comparison of M 2 M computational efficiency; (d) Comparison of M-MM computational efficiency.
Figure 2: Time consumption analysis of matrix multiplication based on different quantum circuits; (a) Comparison of V 2 V computational efficiency; (b) Comparison of V 2 M computational efficiency; (c) Comparison of M 2 M computational efficiency; (d) Comparison of M-MM computational efficiency.

更好的研究,从现在开始

从论文设计到论文写作,大幅缩短您的研究时间。

无需绑定信用卡

本解读由 AI 生成,并经人工编辑审核。