Quantum digital twins based on hardware-tailored tensor networks for computing quantum dynamics

基于硬件定制张量网络的量子数字孪生，用于计算量子动力学

• 批准号: EP/Y005007/1
• 负责人: Oleksandr Kyriienko
• 金额: $ 38.78万
• 依托单位: UNIVERSITY OF EXETER
• 依托单位国家: 英国
• 项目类别: Research Grant
• 财政年份: 2023
• 资助国家: 英国
• 起止时间: 2023 至 无数据
• 项目状态: 未结题
• 来源: https://gtr.ukri.org/projects?ref=EP%2FY005007%2F1
• 关键词: Quantum; digital; twins; based; hardware

Quantum computing offers a promise to solve problems that cannot be addressed by classical devices, and an early access to quantum computers is vital for UK national security. The ability to solve complex problems with quantum computers relies on optimising both the hardware (quantum devices) and the software (quantum algorithms). In this project, we will design the tools for improving the quantum software, which can largely save required resources for state preparation and simulation of dynamics. We call these tools the quantum digital twins - specific programmable models that help representing quantum devices in the most efficient way, and thus enabling their optimisation.Quantum computing (QC) offers a distinct paradigm for performing calculations. Unlike classical computers that operate with bits (taking binary values 0 or 1), quantum devices rely on two-level quantum systems - qubits - that are described by states |0> and |1> that can be put in a superposition. The collection of N qubits can be efficiently evolved on specialised hardware - quantum computers - thus processing information encoded in a quantum form. Classical processing of the same amount of information will require manipulating 2^N complex numbers, a task that becomes impossible already at the size of fifty qubits. We know that in the future quantum computing can exponentially speed up factoring (having a huge impact on cryptography) and help with areas such as simulating materials and chemicals at a scale impossible before (promoting substantial steps towards green energy and sustainability).Quantum hardware is developed by various industrial and academic institutions worldwide. Qubit counts grow every year, and this makes community hopeful for achieving a practical quantum advantage in the near term. Yet, the current level of noise does not allow for running circuits of sufficient depth. Specifically designed quantum software may help to alleviate this problem if tailored algorithms are developed. This challenge calls for imaginative approaches that account for hardware capabilities and limitations.To harness benefits from near-term quantum computing, UK needs to channel an effort on developing software tools that enable the scalable prototyping of quantum algorithms and allow for benchmarking quantum devices at the increased scale. We propose to do this by designing quantum digital twins as efficient tensor network emulators of quantum devices.This project is the collaboration between quantum researchers at the University of Exeter and the National Physical Laboratory. It is built on the three pillars, each representing an open challenge for advancing quantum software and applications: 1) developing efficient tools for quantum state preparation and quantum circuit emulation; 2) developing quantum digital twins of the dynamics; 3) benchmarking quantum algorithms for solving computationally-hard problems in material science.To tackle these challenges we will address three objectives:1. We will develop compact tensor network representations for low energy states of relevant quantum Hamiltonians, and translate these tensor networks into low-depth quantum circuits for efficient initial quantum state preparation.2. We will develop scalable tensor network-based emulators of quantum dynamics as quantum digital twins, taking advantage of the knowledge of the hardware-specific Hamiltonians.3. We will benchmark the scalability of quantum digital twins for emulating quantum devices in materials simulations, and determine the threshold for potential quantum advantage.As a result of the project, we will have the efficient tools that enable the scalable prototyping and improving of quantum simulation, thus maximizing the performance of quantum computers at increasing scale.

量子计算有望解决传统设备无法解决的问题，尽早使用量子计算机对于英国国家安全至关重要。使用量子计算机解决复杂问题的能力依赖于硬件（量子设备）和软件（量子算法）的优化。在这个项目中，我们将设计改进量子软件的工具，这可以很大程度上节省状态准备和动力学模拟所需的资源。我们将这些工具称为量子数字孪生——特定的可编程模型，有助于以最有效的方式表示量子设备，从而实现其优化。量子计算 (QC) 提供了执行计算的独特范例。与使用位操作的经典计算机（采用二进制值 0 或 1）不同，量子设备依赖于两级量子系统 - 量子位 - 由可以叠加的状态 |0> 和 |1> 描述。 N 个量子位的集合可以在专用硬件（量子计算机）上有效地演化，从而处理以量子形式编码的信息。同等信息量的经典处理将需要处理 2^N 个复数，而在 50 个量子位的大小下，这项任务就变得不可能了。我们知道，未来量子计算可以成倍地加速因式分解（对密码学产生巨大影响），并以以前不可能的规模模拟材料和化学品等领域（推动向绿色能源和可持续发展迈出实质性步伐）。量子硬件是由全球各个工业和学术机构开发。量子比特数量每年都在增长，这使得社区对在短期内实现实际的量子优势充满希望。然而，当前的噪声水平不允许运行足够深度的电路。如果开发出定制的算法，专门设计的量子软件可能有助于缓解这个问题。这一挑战需要考虑到硬件功能和限制的富有想象力的方法。为了利用近期量子计算的优势，英国需要努力开发软件工具，以实现量子算法的可扩展原型设计，并允许在量子设备上进行基准测试。规模扩大。我们建议通过设计量子数字孪生作为量子设备的高效张量网络模拟器来实现这一目标。该项目是埃克塞特大学和国家物理实验室的量子研究人员之间的合作。它建立在三大支柱之上，每一个支柱都代表着推进量子软件和应用的公开挑战：1）开发用于量子态准备和量子电路仿真的有效工具； 2）开发动力学的量子数字孪生体； 3）对量子算法进行基准测试，以解决材料科学中的计算难题。为了应对这些挑战，我们将实现三个目标：1。我们将为相关量子哈密顿量的低能态开发紧凑的张量网络表示，并将这些张量网络转化为低深度的量子电路，以实现有效的初始量子态准备。 2．我们将利用特定于硬件的哈密顿量的知识，开发可扩展的基于张量网络的量子动力学模拟器作为量子数字孪生。3。我们将对量子数字孪生的可扩展性进行基准测试，以在材料模拟中模拟量子器件，并确定潜在量子优势的阈值。该项目的结果是，我们将拥有有效的工具，能够实现可扩展的原型设计和量子模拟的改进，从而在规模不断扩大的情况下最大限度地提高量子计算机的性能。