By Dr. Chris Mansell

Shown below are summaries of a few interesting research papers in quantum computing and communications that have been published over the past month.

Hardware

Title: Single-Atom Trapping in a Metasurface-Lens Optical Tweezer
Organizations: National Institute of Standards and Technology; University of Colorado
Lenses are usually made of a spatially uniform material, like glass, that has been carefully manufactured into a precise shape. What if we flipped this so that the shape was very simple but the optical properties of the material varied spatially? In theory, these lenses, called metasurface lenses, could be extremely practical and have superb performance. Cold atom quantum computers require top-of-the-range optical components, such as high numerical aperture aspheric lenses, to control the atomic qubits. In the past couple of years, metasurface lenses have made their debut in cold atom experiments by manipulating the beams that make a magneto-optical trap. This new work shows that a metasurface lens can be used to create optical tweezers that can trap and image single atoms. Even though it did not outperform more conventional lens choices, it lends credence to the idea that the ongoing developments in metasurface design could be harnessed to dramatically improve the control we have over neutral-atom arrays. 
Link: https://journals.aps.org/prxquantum/abstract/10.1103/PRXQuantum.3.030316

Title: Cavity-Enhanced Optical Lattices for Scaling Neutral Atom Quantum Technologies to Higher Qubit Numbers
Organizations: Max-Planck-Institut für Quantenoptik; Munich Center for Quantum Science and Technology; Institute of Physics, Croatia; Ludwig-Maximilians-Universität München
Interfering laser beams can create a regularly spaced array of neutral-atom traps called an optical lattice. This system can be used for quantum simulation by making the traps shallow and for atomic clocks and quantum computers by making them deep. The above applications all benefit by increasing the number of traps and their similarity to one another. Unfortunately, due to limits on the power of the interfering lasers, there is a trade-off between the number of traps and their depth. A pair of mirrors known as a cavity can enhance the available laser power when there is only one line of atoms. For a two- or three-dimensional grid of atoms, the experimental challenges are foreboding.  In this article, the authors overcome issues related to mechanical and thermal stability by using a monolithic assembly and an ultra-low-expansion glass spacer to create a 2D cavity-enhanced optical lattice. With some feasible refinements, it seems likely that their system could have tens of thousands of atomic qubits.
Link: https://journals.aps.org/prxquantum/abstract/10.1103/PRXQuantum.3.030314

Title: Small-world complex network generation on a digital quantum processor
Organizations: National Renewable Energy Laboratory, USA; ColdQuanta; University of Texas; Colorado School of Mines; NVIDIA; Google
What are the most complex phenomena that can emerge from the simplest components acting according to the simplest rules? It is harder to get something simpler than a bit that can only be in one of two states, on or off. Likewise, it is difficult to think of behaviours more complex than life-like processes, such as self-replication. Nevertheless, a straightforward rule like “only on bits with two or three on neighbours stay on; only off bits with exactly three on neighbours get turned on” can allow a 2D grid of bits to look alive.  Quantum generalisations of these remarkable systems have been investigated for over three decades. Until recently, researchers had to use classical devices to simulate how these so-called quantum cellular automata (QCA) would behave. Now, a 1D QCA has been realised on a superconducting quantum computer related to Google’s famous Sycamore processor.  A key finding was that experimental errors did not stop some of the predicted structures from forming. For example, the mutual information between pairs of qubits was shown to have a small-world character. Further investigations could lean towards simulations of strongly-correlated matter or demonstrations of quantum advantage.
Link: https://www.nature.com/articles/s41467-022-32056-y

Software

Title: Quantum Computational Quantitative Trading: High-Frequency Statistical Arbitrage Algorithm
Organizations: Origin Quantum Computing; University of Science and technology of China; Hefei Comprehensive National Science Center 
In the world of high frequency trading, the quickest can take it all, so every fraction of a second counts. Typically, historical pricing data is incrementally updated with the latest tick data. This serves as the input to sophisticated algorithms whose output is automatically used to make bets on the market. Most of the time, the stock prices of, say, Coca-Cola and Pepsi are highly correlated – they go up together and they go down together. The task of statistical arbitrage is to notice if this briefly stops being the case and then bet on the underpriced company while also betting against the overpriced one. Profit is made when the transiently mispriced assets return to the longer-term trend. In this paper, two quantum algorithms are devised that work in tandem to quickly identify worthwhile arbitrage opportunities. They are analysed in terms of how well they would perform in a realistic scenario and how their runtimes scale asymptotically.
Link: https://iopscience.iop.org/article/10.1088/1367-2630/ac7f26

Title: Optimized SWAP networks with equivalent circuit averaging for QAOA
Organizations: University of California at Berkeley; Lawrence Berkeley National Lab;  ColdQuanta; University of Chicago
Sometimes, a quantum algorithm requires a logic gate to be performed between two qubits that cannot directly interact with one another in the physical quantum computing architecture. While SWAP gates can be used to overcome this issue, implementing too many of them could ruin the algorithm’s chances of ever demonstrating a substantial advantage over classical methods. In 2018, the SWAP network was introduced, whereby the total number of SWAP gates that would ever be needed could be made to scale linearly with the number of qubits. In this latest work, the SWAP network was cleverly optimised for executing the Quantum Approximate Optimization Algorithm. The authors introduced a technique called equivalent circuit averaging, where they randomly chose between equally good gate decompositions so that coherent errors would not build up. Testing on the Advanced Quantum Testbed at Lawrence Berkeley National Laboratory revealed that the optimisation lowered the gate count and also reduced errors by around 60%.
Link: https://journals.aps.org/prresearch/abstract/10.1103/PhysRevResearch.4.033028

Title: Classically verifiable quantum advantage from a computational Bell test
Organization: University of California at Berkeley
When quantum advantages have been observed in sampling experiments, the main way that we have verified that the quantum computer is operating correctly is by extrapolating. That is, we assume that because it produces the right output for small cases, it will continue to do so for the large instances that can’t be classically checked. In this paper, the authors explain how to produce a directly verifiable quantum advantage. In short, a quantum computer is given access to less information than the classical computer, so when it nevertheless performs a difficult calculation, the result can be assessed efficiently by classical means.  In more detail, a function is said to have a trapdoor if it is hard to figure out what inputs lead to a given output unless some extra information, called the trapdoor, is provided. Prior work made use of these functions but they had to meet a very stringent condition: knowing one input-output pair couldn’t make it easy to deduce the parity of the other inputs that also get mapped to that output. By relaxing this condition so that trapdoor claw-free functions (TCFs) can be used, the authors have made their scheme considerably more practical. A loose description of their protocol is as follows. The classical computer can generate a TCF and its trapdoor. When the quantum computer is told about the TCF but not about the trapdoor, it can identify two inputs that give the same output. The classical computer then checks this using the trapdoor. The paper has several other key points, including a proposal for an experimental implementation that uses ultracold Rydberg atoms.
Link: https://www.nature.com/articles/s41567-022-01643-7

August 27, 2022