Holographic dynamics simulations with a trapped-ion quantum computer
Holographic dynamics simulations with a trapped-ion quantum computer
复制标题
DOI：
10.1038/s41567-022-01689-7
复制
发表时间：
2022-08-04
期刊：
影响因子：
19.6
通讯作者：
Foss-Feig, Michael
中科院分区：
文献类型：
--
作者：
Chertkov, Eli;Bohnet, Justin;Foss-Feig, Michael
量子动力学的模拟是一个具有挑战性的任务，解决与经典资源。一项使用离子阱量子处理器的实验表明，量子计算机可以有效地模拟大规模多体量子系统的演化。量子计算机有望有效地模拟量子动力学，这是一项经典的棘手任务，对从化学到高能物理等领域至关重要。然而，量子计算的优势仅在随机电路采样等人工任务中得到证明，硬件限制和噪声限制了小规模系统的定性研究。能够高保真度测量和量子位重用的量子处理器使最近提出的全息技术成为可能，该技术采用量子张量网络状态（一类有效压缩量子数据的状态）来模拟无限长、纠缠初始状态的演化。使用少量的量子位。在这里，我们在捕获离子量子处理器中使用11个量子比特来模拟无限纠缠态的动力学，从而对这种全息技术进行基准测试。我们观察到量子混沌和光锥传播的相关性的标志，并找到了极好的定量协议与理论预测的无限大小的限制，最小的后处理或错误缓解的实施模型。这些结果表明，量子张量网络方法与最先进的量子处理器功能相结合，为近期科学和技术直接感兴趣的问题提供了一条可行的量子计算优势。
The simulation of quantum dynamics is a challenging task to solve with classical resources. An experiment with a trapped-ion quantum processor now shows the efficient simulation of the evolution of large-scale many-body quantum systems.Quantum computers promise to efficiently simulate quantum dynamics, a classically intractable task central to fields ranging from chemistry to high-energy physics. Yet, quantum computational advantage has only been demonstrated for artificial tasks such as random circuit sampling, and hardware limitations and noise have limited experiments to qualitative studies of small-scale systems. Quantum processors capable of high-fidelity measurements and qubit reuse enable a recently proposed holographic technique that employs quantum tensor-network states, a class of states that efficiently compress quantum data, to simulate the evolution of infinitely long, entangled initial states using a small number of qubits. Here we benchmark this holographic technique in a trapped-ion quantum processor using 11 qubits to simulate the dynamics of an infinite entangled state. We observe the hallmarks of quantum chaos and light-cone propagation of correlations, and find excellent quantitative agreement with theoretical predictions for the infinite-size limit of the implemented model with minimal post-processing or error mitigation. These results show that quantum tensor-network methods, paired with state-of-the-art quantum processor capabilities, offer a viable route to practical quantum computational advantage on problems of direct interest to science and technology in the near term.