论文标题
使用过滤时间序列的混合量子间隙估计算法
Hybrid quantum gap estimation algorithm using a filtered time series
论文作者
论文摘要
量子模拟优于经典记忆限制将使紧凑型量子电路能够深入了解棘手的量子多体问题，但是在量子时间演变中，大电路深度的相互关联障碍物和噪声似乎排除了近期无偏的量子模拟。我们证明，经典的后处理，即对离线时间序列的长期过滤，可以呈指数级改善量子时间演变所需的电路深度。我们将过滤方法应用于构建杂种量子古典算法以估计能量差距，这是不受变异定理控制的重要观察到的重要观测。我们在过滤的工作范围内证明了算法在概念验证模拟中的成功，用于最小旋转模型的有限尺寸缩放。我们的发现为公正的量子模拟奠定了基础，以在短期内提供内存优势。
Quantum simulation advantage over classical memory limitations would allow compact quantum circuits to yield insight into intractable quantum many-body problems, but the interrelated obstacles of large circuit depth in quantum time evolution and noise seem to rule out unbiased quantum simulation in the near term. We prove that classical post-processing, i.e., long-time filtering of an offline time series, exponentially improves the circuit depth needed for quantum time evolution. We apply the filtering method to the construction of a hybrid quantum-classical algorithm to estimate energy gap, an important observable not governed by the variational theorem. We demonstrate, within an operating range of filtering, the success of the algorithm in proof-of-concept simulation for finite-size scaling of a minimal spin model. Our findings set the stage for unbiased quantum simulation to offer memory advantage in the near term.
