International Catalyst on Computer-Aided Design Algorithms for Quantum Annealing

这笔Catalyst赠款的目的是在我们在UBC的团队和伦敦帝国理工学院(IC)的Wayne Luk领导的团队之间建立合作，并通过这一合作，启动量子计算机的基础计算机辅助设计(CAD)和应用研究。我们将专注于量子退火处理器(QAP)，如加拿大公司D-Wave生产的处理器。合作将结合IC在识别、构建、优化和评估适合在QAP上解决的应用程序(包括金融和机器学习)方面的专业知识，以及UBC在QAP和量子辅助机器学习(QAML)CAD方面的专业知识。具体的技术目标将包括(A)使量子比特映射算法针对金融领域的应用，如投资组合优化和因子投资，(B)设计和实现一种新的量子比特映射算法，该算法使用现有的机器学习模型(如强化学习)来优化量子比特全局布局的初始阶段，以及(C)通过使用基于模型的故障分析来开发用于量子退火优化的性能调试基础设施。培训中增加的重要价值将来自UBC和IC团队的交互。这种合作的好处有两个：(1)QAP的应用领域仍处于新兴阶段，这意味着从事嵌入CAD工作的研究人员需要能够对适用于这些设备的问题类型进行最先进的思考，并深入了解与这些领域相关的领域的应用专家的要求。这样的专业水平在世界上是罕见的，然而，IC团队处于有利地位，将在这里产生重大影响。与此同时，今天，领域专家可能对QAP和量子计算的性质几乎没有接触，所以将专家与那些拥有QAP知识的人配对将确保他们能够理解量子计算的局限性和前景，并将其用于设计新的应用程序。

The objective of this Catalyst grant is to build a collaboration between our group at UBC and a group led by Wayne Luk at Imperial College (IC) London, and through this, to initiate fundamental computer-aided design (CAD) and application investigation for quantum computers. We will focus on Quantum Annealing Processors (QAPs) such as those produced by Canadian company D-Wave. The collaboration will combine IC's expertise in identifying, building, optimizing, and evaluating applications (including finance, and machine learning) which are suitable for solving on a QAP, with UBC's expertise in CAD for QAPs and Quantum-assisted Machine Learning (QAML). Specific technical goals will include (a) adapting qubit-mapping algorithms to target applications in finance, such as portfolio optimization, and factor investing, (b) to design and implement a new qubit-mapping algorithm that uses established machine-learning models such as reinforcement-learning to optimize the initial stage of qubit global placement, and (c) to develop a performance debugging infrastructure for quantum annealing optimization through the use of model-based fault analysis.Significant value added in training will come from the interaction of the UBC and IC team. The benefits of this collaboration are two-fold: (1) Application domains of QAPs are still emerging, meaning researchers doing embedding CAD work need to have access to state-of-the-art thinking on the types of problems suitable for these devices, and access to a deep understanding of the requirements of application experts in areas related to these domains. This level of expertise is rare in the world, however, the IC team is well-positioned to have significant impact here. Meanwhile, today, domain experts may have little exposure to QAPs and the nature of quantum computation, so pairing experts with those with QAP knowledge will ensure that they are able to understand the limitations and promises of quantum computing and use that in designing new applications.