Ion-Trap-Based Quantum Computers: From Benchmarking to Outperforming Classical Digital Computers

我们对量子物理的使用始于20世纪之交，当时科学家们开始探索原子和电子的超微小世界。虽然量子力学的大部分理论基础都是在1920年代的S时期建立起来的，但我们仍然没有使用重要的技术来考虑量子世界可能出现的奇怪行为(如隐形传态或量子干涉)。最近，一种名为量子计算的新计算范式取得了重大进展。由于它的运行基于其组成部分之间的量子力学效应，它可以开创一个新的计算领域--在这个领域，一台小型量子计算机可以容纳比任何曾经建造(或将要建造)的传统计算机都要多的内存，或者它可以迅速解决一些难题，比如寻找大数的质因数，从而打破我们目前的加密标准。虽然我们还没有“微软量子表面”设备可以在商店里买到，但研究实验室已经生产了所谓的量子模拟器，它模拟耦合量子系统的行为，并允许在计算结束时读取它们的性质。这项研究项目将致力于寻找方法来验证这些量子计算机在超过传统计算机的能力时的正确性。这是通过寻找特殊情况来完成的，在这些特殊情况下，传统计算机可以预测结果，并确定如何在这些特殊情况下运行量子计算机。除了拟议的研究工作外，该项目还有一个重要的推广部分，该小组将于2016年秋在EdX上推出一门大型开放式在线课程(MOOC)，题为《面向每个人的量子力学》，该课程将用最小的数学(不比开平方根更复杂)教授量子力学原理，并提供一系列交互式计算机演示、教程、这项工作将涉及三个主要主题：(1)通过考察自旋压缩效应(海森堡测不准原理可以部分击败)和在长程相互作用极限下的计算，考察晶格振动对基于离子陷阱的量子模拟器性能的影响；(2)确定将量子计算从10s KHz的时钟频率加速到可以推到MHz范围或更远的时钟频率的技术；以及(3)发现宏观量子隧道效应或脆弱量子态(如所谓的中午态)是否可以在这些量子模拟器中形成和使用。我们使用的量子模拟器类型是由捕获的离子建造的，这些离子由激光驱动，对离子施加自旋相关的力(其中自旋自由度通常是离子的超精细状态)。激光根据离子的内部量子状态推动离子，而与晶格振动的耦合推动了这些内部状态之间的后续耦合。通过适当地设计这些耦合，人们可以创造出一台量子计算机/模拟器。这项理论工作将与NIST Boulder(彭宁陷阱)、马里兰(线性Paul陷阱)和UCLA(平面Paul陷阱)的实验工作密切配合。

Our use of of quantum physics began around the turn of the 20th century, when scientists started exploring the ultrasmall world of atoms and electrons. While much of the theoretical basis for quantum mechanics was worked out in the 1920's, we still do not employ significant technologies that take into account the bizarre behaviors that are possible due to the quantum world (like teleportation or quantum interference). Recently, there has been significant advance in a new paradigm for computing, called quantum computing. With its operation based on the quantum-mechanical effects between its constituents, it can usher in a new realm of computation--where a small quantum computer can hold more memory than that of every conventional computer ever built (or to be built), or it can rapidly solve hard problems like finding the prime factors of large numbers and thereby breaking our current standards for encryption. While we don't yet have the "Microsoft quantum surface" devices available in stores, research labs have produced so-called quantum simulators, which emulate the behavior of coupled quantum systems and allow their properties to be read off at the end of the computation. This research project will work on finding ways to verify the correctness of these quantum computers when they surpass the abilities of conventional computers. This is done by finding special cases, where conventional computers can predict the results, and determining how to run the quantum computers for those special cases. In addition to the research work proposed, this project has a significant outreach component, where the group will be launching a Massive Open Online Course (MOOC) entitled "Quantum mechanics for everyone" on EdX in the fall of 2016, which will teach quantum mechanics principles with minimal math (no more complicated than taking square roots) and with a series of interactive computer demonstrations, tutorials, and illustrations.Three main themes will be pursued in this work: (1) Examining effects of the lattice vibrations on the performance of ion-trap-based quantum simulators by looking at both spin squeezing effects (where the Heisenberg uncertainty principal can be partially beaten) and on computation in the long-range interacting limit; (2) Determining techniques to speed up quantum computation from clock frequencies in the 10s of KHz to clock frequencies that can be pushed in the MHz range or further; and (3) Discovering whether macroscopic quantum tunneling effects, or fragile quantum states (like the so-called NOON states) can be formed and employed in these quantum simulators. The type of quantum simulator we work with is one that is built out of trapped ions that are driven by lasers that apply spin-dependent forces on the ions (where the spin degree of freedom is often a hyperfine state of the ion). The lasers push the ions according to their internal quantum states, while the coupling to the lattice vibrations drives a subsequent coupling between those internal states. By properly engineering these couplings, one creates a quantum computer/simulator. This theoretical work will pair closely with experimental work at NIST Boulder (Penning trap), Maryland (linear Paul trap), and UCLA (planar Paul trap).