反射是量子算法中无处不在的工具。我们考虑通过对可实现的酉矩阵的特征空间进行反射的任务。此类反射通常通过相位估计或酉矩阵的线性组合来设计。这些方法在所需精度和酉矩阵的谱隙方面具有有利的规模和深度。然而，它们所需的辅助量子比特数量随着这两个参数的增长而增加。在这里，我们提出了一种简单的算法，该算法具有相同的规模和深度缩放特性，但对所有问题实例仅需4个辅助量子比特。因此，预计此算法将成为通过酉矩阵特征空间进行反射的标准方法。

Reflections are omnipresent tools in quantum algorithms. We consider the task of reflecting through the eigenspace of an implementable unitary. Such reflections are generally designed using phase estimation or linear combination of unitaries. These methods have size and depth that scale favorably with the desired precision and the spectral gap of the unitary. However, they require a number of ancilla qubits that grows with both parameters. Here, we present a simple algorithm with the same size and depth scaling but requiring only 4 ancilla qubits for all problem instances. As such, this algorithm is expected to become the reference method to reflect through eigenspaces of unitaries.