By Dr Chris Mansell, Senior Scientific Writer on Terra Quantum
Below is a summary of some of the interesting research papers in quantum technology we’ve looked at over the past month.
Title: Measurement and rearrangement of mid-circuit qubits in the Yb-171 atomic array
Organization: Atom Computing, Inc.
Measuring ancilla qubits repeatedly during quantum computation could enable error-correction procedures to remove entropy from the quantum processor. Several different approaches have been tried for processors based on ultracold atoms, so the challenge is now to identify or develop the most reliable and scalable path forward. In this experiment, ancilla atoms are imaged using atomic transitions which have narrow line widths that can be controlled in and out of resonance by Zeeman shifts or light shifts. This means imaging can be performed in both country-selective and site-selective ways. Impressively, the other atoms in the experiment maintained their coherence even when the magneto-optical trap was loaded and the ancilla atoms were optically pumped and imaged repeatedly.
Title: Suppression of quantum immersion in superconducting circuits by immersion cooling
Organization: Royal Holloway University London; Chalmers University of Technology; Google; National Physics Laboratory, UK
It may be possible to increase the coherence of superconducting quantum circuits by shielding them from sources of ionizing radiation, such as cosmic rays, and by filtering out unwanted photons. In this paper, the researchers immersed the superconductor in helium-3 to reduce the noise it experiences from two-level defects in its dielectric environment. Helium-3 acts as an efficient heat sink for this environment, increasing the rate at which defects relax to their ground state by a factor of 1000. Helium-3 also cools the spin of surface electrons that uncontrollably pairs with the superconductor. Overall, this research demonstrates a very effective way of increasing the coherence of superconducting qubits by managing their temperature-dependent degrees of freedom.
Title: Eagle: QuEra’s 256-qubit neutral atomic quantum computer
Organization: QuEra Computing Inc.
This white paper, aimed at educators and aspiring users, provides an overview of QuEra’s 256-qubit neutral atomic quantum computer, “Aquila.” It recommends various best practices to make the most effective use of its capabilities. Helpful links are provided for sample notebooks on the company’s GitHub page. The document transparently describes the strengths and weaknesses of the system. It starts with the basics of an atom-neutral platform and progresses to protocols such as combinatorial optimization.
Title: Forty thousand kilometers under quantum protection
Organization: Terra Quantum
Existing approaches to quantum key distribution (QKD) all have a secret key rate that decreases exponentially with its transmission distance. Even with hardware upgrades, they can’t go beyond certain fundamental boundaries that characterize this relationship. The authors of this paper present a new way to overcome this limitation. They examined the traditional assumption that eavesdroppers could collect and use every photon that was lost from a remote communication channel. By considering the physics of light scattering in modern optical fibers and devising ways to monitor these channels, they showed how to ensure that no effective, yet undetectable, eavesdropping could occur. Ultimately, this allows their protocol to overcome the constraints mentioned above and send keys at a higher speed over much longer distances.
Title: Separating phonons: Building a platform for linear mechanical quantum computing
Organization: University of Chicago; Argonne National Laboratory
Phonons are quanta of sound waves, just as photons are for electromagnetic waves. This similarity means that a linear optical quantum computing scheme can be rearranged and applied to phonons in dense systems. In a remarkable experiment, the two-phonon Bell state was created to an accuracy of 0.816, which demonstrated that the phonons, each representing a collective oscillation of about one quadrillion atoms, could be made to interact in a coherent and controlled manner. The arrangement consists of two superconducting Xmon qubits coupled together. However, for phonons to become another string in the wake of superconducting architectures or to become self-scalable computing platforms, phonon losses need to be reduced.
Title: Pipelined quantum processor architecture for silicon spin qubits
Organization: Quantum Motion; University College London; University of Oxford
When you read a quantum circuit diagram, you start from the left with the qubits in their initial state. The gate layers are then applied in such a way that each subsequent layer is displayed to the right of the previous layer. Usually, the physical qubits are stationary and the direction to the right in the diagram indicates a progression in time. The idea expressed in this paper is to physically move the qubit column to the right each time a new gate layer is to be applied. Specifically, qubits are electrons in a silicon processor that can be collectively cascaded to the right by applying a global pulse. For a given type of algorithm or application, the same gate types need to be implemented in the same order. The main advantage is that the processor can be pre-designed with this in mind, so that the built-in components are in their appropriate places on the chip, ready to tune whenever electrons arrive.
Title: Efficient tensor network simulation of IBM’s Ising experiment
Organization: Flatiron Institute, New York University
This paper reports an accurate, memory-saving, and time-saving classical simulation of a 127-qubit Ising quantum system kicked on a heavy hexagonal lattice. Simulation of this system on a quantum processor was recently performed using noise mitigation techniques to improve accuracy (Nature volume 618, p. 500-505 (2023)). Now, this paper demonstrates that, by adopting a tensor network approach that reflects the device’s qubit connectivity, classical simulations can be performed that are significantly more accurate than the results obtained from quantum devices in a verifiable regime and can be compared with quantum simulation results for greater depth. big.
Title: Interpretable Quantum Gain in Neural Sequence Learning
Organizations: Massachusetts Institute of Technology; University of California San Diego; Harvard University
Minimally extending classical machine learning models is a very practical way to create new quantum machine learning models. Empirically, this means that tests can be performed directly with similar model sizes, allowing fair comparisons. Theoretically, this means that if quantum superiority is found, then it can be attributed to the introduced features of quantum mechanics. In this paper, the linear recurrent neural network is extended so that no definite classical values can be assigned to the measurement operators. This is known as quantum contextuality. The researchers applied this new model to a Spanish to English translation task and found it useful in practice. They also show, without making any complexity theoretical assumptions, that quantum contextuality gives it a memory advantage over the original model.
Title: Variational quantum non-orthogonal optimization
Organization: Multiverse Computing; Donostia International Physics Center; Ikerbasque Science Foundation
We usually use two orthogonal qubit states to represent the classical variables 0 and 1, but what if there was a better way? By writing the dodecahedron in the Bloch sphere of the qubits, the authors of this paper suggest that the 20 non-orthogonal states located at the vertices of this shape can be used to represent 20 different classical variables. This proposal allows current NISQ processors to make better use of their somewhat limited number of qubits and thus tackle larger problems. They classically simulated this idea, including the required single-qubit tomography, for optimization tasks and found good performance. The main unanswered question is whether these encoding benefits will hold up over noisy quantum processors.
Title: How to calculate a 256-bit elliptic curve private key with only 50 million Toffoli gates
Elliptic curve cryptography (ECC) is widely adopted because of its resistance to attacks from classical computers. However, like the RSA cryptographic scheme, it is vulnerable to the quantum computers running Shor’s algorithm. Last year, a new approach to fault-tolerant quantum computation was devised for photonic processors. In this paper, Daniel Litinski introduces three refinements to this approach and then estimates the amount of resources needed to solve ECC. He concluded that, compared to breaking the RSA, about ten times less resources were needed.
Title: Quantum dropout: Continuous quantum approximation optimization algorithm
Organization: Peking University; Shanghai Jiao Tong University; Quantum Matter Collaborative Innovation Center, Beijing
In satisfaction problems, clauses specify what requirements the variable must satisfy, and more difficult examples of problems tend to have more clauses. In this paper, numerical simulation of a quantum approximate optimization (QAOA) algorithm is used to maximize the number of fulfilled clauses. To make the QAOA parameters easier to update, a clever trick was introduced: some clauses are not included in the QAOA circuit. The rest of the protocol is preceded as usual: the qubits are measured at the end of the chain and the results are checked on a classical computer (against all clauses). This results in a higher probability of success compared to standard QAOA.
June 28, 2023