Quantum Information Protocols with limited resources

经过多年的深入研究，现在可以高精度地控制和操纵数十个量子位。这已经通过捕获离子、冷原子、超导体和光子实现了，其他技术很可能很快就会赶上来。尽管成熟的量子计算机在短期内仍然遥不可及，但预计在未来几年内，由多达100个量子比特组成的量子处理器将可用，并且可以可靠地执行超过1000个量子门而无需诉诸纠错方案。在这些条件下，我们能利用这些小系统吗？我们能学到什么呢？该项目的长期目标和愿景是：(i)开发可以用小型量子处理器执行的应用程序和协议，并且优于现有和计划中的经典设备；（ii）重新审视受量子信息处理启发的经典算法和方法，将其应用于量子设备和多体系统；（iii）弥合抽象结果与具体实验装置之间的差距。为了实现上述目标，前四年将致力于开发50-100个量子比特和1000-10000个量子门的量子算法：这些算法将应用于优化问题，量子多体态分析，因为它们出现在原子和凝聚态物理中，以及量子增强机器学习中。•经典机器学习在量子系统中的应用：将张量网络方法与弦键态相结合，旨在通过机器学习辅助进行与实验相关的多体系统的表征和量子多体特性的认证。•研究如何用不同的技术实现这些想法和其他想法：这将包括捕获离子、冷中性原子、光子、超导量子比特，以及新的场景。从长远来看，随着实验的进行，我们将研究包括更多量子比特、更复杂的通信通道以及混合量子/经典算法和协议的协议。

After many years of intensive research, it is now possible to control and manipulate tens of qubits with high precision. This has been achieved with trapped ions, cold atoms, superconductors, and photons, and it is very likely that other technologies will soon catch up. Even though a full-fledged quantum computer is still out of reach in the near term, it is expected that in the next few years quantum processors composed of up to hundred qubits will be available, and that one will be able to reliably perform more than a thousand quantum gates without having to resort to error correction schemes. Under these conditions, can we take advantage of those small systems? And, what can we learn?The long-term goals and visions of this project are to: (i) develop applications and protocols which can be carried out with small quantum processors, and that outperform existing and planned classical devices; (ii) revisit classical algorithms and methods inspired by quantum information processing to apply them to quantum devices and many-body systems; (iii) bridge the gap between abstract results and specific experimental setups.For the goals outlined above, the first four years will be dedicated to• Developing quantum algorithms for 50-100 qubits and 1000-10000 quantum gates: The algorithms will have applications in optimization problems, the analysis of quantum many-body states as they appear in atomic and condensed matter physics, as well as in quantum-enhanced machine learning.• The application of classical machine learning for quantum systems: This will combine the tensor network approach with string bond states, and aim at the characterization of many-body systems relevant for experiments and the certification of quantum many-body properties assisted by machine learning.• The investigation of how to implement those and other ideas with different technologies: This will include trapped ions, cold neutral atoms, photons, superconducting qubits, as well as novel scenarios.In the long run, as experiments take over, we will investigate protocols that include more qubits, more sophisticated communication channels and also hybrid quantum/classical algorithms and protocols.