International Network on Quantum Annealing (INQA)

量子退火（QA）是通用门基量子计算（GBQC）的一种特定于应用的替代范式。QA最初提出的应用是优化，尽管最近也提出了量子模拟和机器学习方面的应用。与GBQC相比，QA对底层量子比特的相干性要求不那么严格。这使得实验人员能够使用QA演示许多不同的优化用例原理证明，利用D-Wave系统开发的超导流体量子比特退火系统。这些应用包括材料模拟、金融组合优化、流量路由、欺诈检测、电路故障诊断和网站推荐引擎。然而，目前很少有（如果有的话）实验证据表明任何量子加速的任何现实世界的应用，至少有争议的是，仅仅增加通量量子比特的数量在现有的硬件没有任何新的定性设计特征将永远不会导致缩放加速。因此，在美国、英国、欧盟和日本建立了几个全球合作，以确定哪些定性特征可以添加到量子退火设备的未来实现中，通过与经典基准相比，这些特性将实现转型的缩放加速（至少对于某些类型的难题）。与现有的D-Wave实现相比，这些协作的一个关键区别在于前者专注于相干量子位。例如，在美国的QAFS项目中，使用麻省理工学院林肯实验室开发的铝电容分流通量量子位，使得相干寿命比D-Wave目前使用的铌量子位长三个数量级以上，并且超过了典型的一年持续时间约为1微秒。完全相干退火可以提供许多好处，最值得注意的是能够通过最小间隙过渡，这在困难（即小间隙）问题中作为计算加速的瓶颈。完全相干退火在精神上也更接近于封闭系统绝热量子计算（AQC），其中有一个与GBQC等价的证明（因此可证明量子加速）。国际量子退火网络（INQA）将首次建立一个机制，通过该机制，四个全球合作组织将聚集在一起，分享技术和知识知识，并批判性地分析量子退火理论和实验研究的发展。该网络将由伦敦大学学院的Paul Warburton教授领导，他是英国量子计算与模拟（QCS）中心和最近宣布的QEVEC项目的联合研究员。他还曾在美国领导的QEO和QAFS合作项目中担任联合研究员。其他被命名为INQA网络参与者的英国研究人员包括格拉斯哥大学的Martin Weides教授（QCS Hub和EU-AVAQUS），斯特拉斯克莱德大学的Viv Kendon教授（QCS Hub和QEVEC），达勒姆大学的Nick Chancellor博士（QCS Hub和QEVEC）和伦敦大学学院的Andrew Green教授。该网络将主办每周的在线技术研讨会，为合作大学之间的国际交流访问提供资金，并举办年度会议。这些会议将与现有的AQC年度会议（将由INQA部分赞助）穿插举行，以便每年举行两次专门讨论退火问题的科学会议。

Quantum annealing (QA) is an application-specific alternative paradigm to universal gate-based quantum computation (GBQC). The application for which QA was originally proposed is optimisation, though more recently applications in quantum simulation and machine learning have also been proposed. QA makes less stringent demands on the coherence of the underlying qubits than does GBQC. This has enabled experimentalists to demonstrate many diverse optimisation use-case proofs of principle using QA, exploiting the superconducting flux-qubit-based annealing system developed by D-Wave Systems. Such applications include materials simulation, financial portfolio optimisation, traffic routing, fraud detection, circuit fault diagnosis and website recommendation engines. Nevertheless at present there is scant (if any) experimental evidence of any quantum speedup for any real world application, and it is at least arguable that merely increasing the number of flux qubits in the existing hardware without any new qualitative design features will never lead to a scaling speedup. Several global collaborations have therefore been set up in the US, UK, EU and Japan to identify what qualitative features could be added to future implementations of quantum annealing devices which would enable a transformational scaling speedup (at least for some class of hard problems) by comparison with classical benchmarks. A key discriminator for these collaborations by comparison with the existing D-Wave implementation is the focus of the former on coherent qubits. In the QAFS programme in the US, for example, use of the aluminium capacitively-shunted flux qubits developed at MIT Lincoln Lab has enabled coherence lifetimes which are both more than three orders of magnitude longer than the niobium qubits currently used by D-Wave and which exceed the typical anneal durations of around 1 microsecond. Fully coherent annealing may offer a number of benefits, most notable being the ability to transition through the minimum gap which acts as a bottleneck for computational speedup in hard (i.e. small gap) problems. Fully coherent annealing is also closer in spirit to closed-system adiabatic quantum computation (AQC) for which there is a proven equivalence to GBQC (and therefore provable quantum speedup). The International Network in Quantum Annealing (INQA) will for the first time establish a mechanism by which four global collaborations come together to share technical and intellectual know-how and critically analyse developments in theoretical and experimental research in quantum annealing. The network will be led by Prof. Paul Warburton of UCL, who is a co-investigator in the UK's Quantum Computation and Simulation (QCS) Hub and in the recently-announced QEVEC project. He was also previously a co-investigator in the US-led QEO and QAFS collaborations. Other UK researchers who are named network participants in INQA are Prof. Martin Weides of the University of Glasgow (QCS Hub and EU-AVAQUS), Prof. Viv Kendon of Strathclyde University (QCS Hub and QEVEC), Dr Nick Chancellor of Durham University (QCS Hub and QEVEC) and Prof. Andrew Green of UCL. The network will host weekly on-line technical seminars, offer funding to enable international exchange visits between collaborating universities, and run annual conferences. These conferences will be interlaced with the existing annual AQC conferences (which will be part-sponsored by INQA) so as to provide two annealing-specific scientific meetings per year.