Spatial Quantum Optical Annealer for Spin Hamiltonians

光学模拟器在速度、可扩展性、功耗和量子优势方面是最有希望推动未来技术突破的候选者之一，它服务于广泛的有用的优化问题。然而，这类模拟器的运行目前仍然受到噪声、它们可以嵌入的算法问题的程度以及它们与超级计算机竞争的经典制度的限制。HeISINGBERG的目标是将我们最先进的空间光子自旋模拟器(通过耦合10,000个自旋的空间光调制器进行的全光处理的迭代周期)引入量子体系，方法是将相干驱动器升级为压缩光，通过矢量矩阵乘法方案使其完全可编程，使用全息照相、辅助自旋和有效磁场，以及设计专用的定制和专用算法。场的一次正交起伏的减少将使我们能够扩大和优化现有机器的性能，使其超越经典超级计算机和竞争对手的自旋模拟器的能力。HeISINGBERG设备将在室温下运行100,000个自旋，并在改进的XY架构上处理新的量子退火算法。此外，通过分束器进行调制、混合和相位控制的压缩态的非经典资源，如多光子态的纠缠或叠加，将有望利用量子优势，推动我们的机器进入其量子模拟机制。这一发展将通过具体阐述算法复杂性的问题来刺激量子信息处理社区，并澄清退火炉和模拟器中可用量子优势的性质。这些进展将使我们能够在云平台上演示可以有效地解决NP-Hard问题的退火法和绝热算法。

Optical simulators rank among the most promising candidates to power future technological breakthroughs in terms of speed, scalability, power-consumption and quantum advantage, serving a wide range of useful optimization problems. However, the operation of such simulators remains currently limited by noise, the extent of algorithmic problems they can embed and to the classical regime where they compete with supercomputers. HEISINGBERG aims to bring our state-of-the-art spatial photonic spin simulator (an iterated cycle of all-optical processing through a spatial light modulator that couples 10,000 spins) into the quantum regime by upgrading its coherent drive to squeezed light, making it fully programmable through vector-matrix multiplication schemes, use of holography, ancillary spins & effective magnetic fields, and designing dedicated custom-tailored and purpose-built algorithms. The reduced fluctuations in one quadrature of the fields will allow us to scale up and optimize the performances of the existing machine to bring it beyond the capabilities of both classical supercomputers and competing spin-simulators. HEISINGBERG devices will operate 100,000 spins at room temperature and process new quantum annealing algorithms on an improved XY architecture. Besides, the nonclassical resources of squeezed states when modulated, admixed and phase-controlled through beam splitters, such as entanglement or superpositions of multiphoton states will be prospected to harness a quantum advantage and boost our machine into its quantum simulation regime. This development will stimulate the quantum information processing community by concretely articulating problems of algorithmic complexity and clarify the nature of the quantum advantage available in annealers and simulators. These advances will allow us to demonstrate, on a cloud platform, annealing and adiabatic algorithms that can efficiently solve NP-hard problems.