Search

When applying quantum computing to machine learning tasks, one of the first considerations is the design of the quantum machine learning model itself. Conventionally, the design of quantum machine learning algorithms relies on the ``quantisation" of classical learning algorithms, such as using quantum linear algebra to implement important subroutines of classical algorithms, if not the entire algorithm, seeking to achieve quantum advantage through possible run-time accelerations brought by quantum computing. However, recent research has started questioning whether quantum advantage via speedup is the right goal for quantum machine learning [1]. Research also has been undertaken to exploit properties that are unique to quantum systems, such as quantum contextuality, to better design quantum machine learning models [2]. In this paper, we take an alternative approach by incorporating the heuristics and empirical evidences from the design of classical deep learning algorithms to the design of quantum neural networks. We first construct a model based on the data reuploading circuit [3] with the quantum Hamiltonian data embedding unitary [4]. Through numerical experiments on images datasets, including the famous MNIST and FashionMNIST datasets, we demonstrate that our model outperforms the quantum convolutional neural network (QCNN)[5] by a large margin (up to over 40% on MNIST test set). Based on the model design process and numerical results, we then laid out six principles for designing quantum machine learning models, especially quantum neural networks., Comment: 11 figures, 31 pages. Code available on https://github.com/peiyong-addwater/HamEmbedding. Author affiliation updated for v2. Acknowledgements and funding information added for v2

Quantum teleportation is not merely a fascinating corollary of quantum entanglement, it also finds utility in quantum processing and circuit compilation. In this paper, we measure and track the entanglement and fidelity of two-qubit states prepared on a 127-qubit IBM Quantum device, as one of the qubits is teleported across 19 qubits. We design, evaluate and compare two distinct approaches to teleportation: post-selected measurement categorisation and dynamic circuit corrections based on mid-circuit measurements, and compare with direct state transportation using SWAP gates. By optimally choosing the teleportation path which exhibits the highest total negativity entanglement measure across nearest-neighbour pairs, we show the entanglement of a two-qubit graph state is sustained after at least 19 hops in teleportation using the post-selection approach and 17 hops using the dynamic circuit approach. We observe a higher level of teleported entanglement in paths determined from two-qubit negativities compared to those obtained from gate errors, demonstrating an advantage in using the negativity map over the gate error map for compiling quantum circuits.

As quantum technology advances and the size of quantum computers grow, it becomes increasingly important to understand the extent of quality in the devices. As large-scale entanglement is a quantum resource crucial for achieving quantum advantage, the challenge in its generation makes it a valuable benchmark for measuring the performance of universal quantum devices. In this work, we study entanglement in Greenberger-Horne-Zeilinger (GHZ) and graph states prepared on the range of IBM Quantum devices. We generate GHZ states and investigate their coherence times with respect to state size and dynamical decoupling techniques. A GHZ fidelity of $0.519 \pm 0.014$ is measured on a 32-qubit GHZ state, certifying its genuine multipartite entanglement (GME). We show a substantial improvement in GHZ decoherence rates for a 7-qubit GHZ state after implementing dynamical decoupling, and observe a linear trend in the decoherence rate of $\alpha=(7.13N+5.54)10^{-3} \mu s^{-1}$ for up to $N=15$ qubits, confirming the absence of superdecoherence. Additionally, we prepare and characterise fully bipartite entangled native graph states on 22 IBM Quantum devices with qubit counts as high as 414 qubits, all active qubits of the 433-qubit Osprey device. Analysis of the decay of 2-qubit entanglement within the prepared states shows suppression of coherent noise signals with the implementation of dynamical decoupling techniques. Additionally, we observe that the entanglement in some qubit pairs oscillates over time, which is likely caused by residual ZZ-interactions. Characterising entanglement in native graph states, along with detecting entanglement oscillations, can be an effective approach to low-level device benchmarking that encapsulates 2-qubit error rates along with additional sources of noise, with possible applications to quantum circuit compilation., Comment: 25 pages, 21 figures

The determination of ground state properties of quantum systems is a fundamental problem in physics and chemistry, and is considered a key application of quantum computers. A common approach is to prepare a trial ground state on the quantum computer and measure observables such as energy, but this is often limited by hardware constraints that prevent an accurate description of the target ground state. The quantum computed moments (QCM) method has proven to be remarkably useful in estimating the ground state energy of a system by computing Hamiltonian moments with respect to a suboptimal or noisy trial state. In this paper, we extend the QCM method to estimate arbitrary ground state observables of quantum systems. We present preliminary results of using QCM to determine the ground state magnetisation and spin-spin correlations of the Heisenberg model in its various forms. Our findings validate the well-established advantage of QCM over existing methods in handling suboptimal trial states and noise, extend its applicability to the estimation of more general ground state properties, and demonstrate its practical potential for solving a wide range of problems on near-term quantum hardware., Comment: 5 pages, 2 figures. Accepted paper at the 2023 IEEE International Conference on Quantum Computing and Engineering (QCE23)

The data encoding circuits used in quantum support vector machine (QSVM) kernels play a crucial role in their classification accuracy. However, manually designing these circuits poses significant challenges in terms of time and performance. To address this, we leverage the GASP (Genetic Algorithm for State Preparation) framework for gate sequence selection in QSVM kernel circuits. We explore supervised and unsupervised kernel loss functions' impact on encoding circuit optimisation and evaluate them on diverse datasets for binary and multiple-class scenarios. Benchmarking against classical and quantum kernels reveals GA-generated circuits matching or surpassing standard techniques. We analyse the relationship between test accuracy and quantum kernel entropy, with results indicating a positive correlation. Our automated framework reduces trial and error, and enables improved QSVM based machine learning performance for finance, healthcare, and materials science applications.

The accurate computation of properties of large molecular systems is classically infeasible and is one of the applications in which it is hoped that quantum computers will demonstrate an advantage over classical devices. However, due to the limitations of present-day quantum hardware, variational-hybrid algorithms introduced to tackle these problems struggle to meet the accuracy and precision requirements of chemical applications. Here, we apply the Quantum Computed Moments (QCM) approach combined with a variety of noise-mitigation techniques to an 8 qubit/spin-orbital representation of the water molecule (H$_2$O). A noise-stable improvement on the variational result for a 4-excitation trial-state (circuit depth 25, 22 CNOTs) was obtained, with the ground-state energy computed to be within $1.4\pm1.2$ mHa of exact diagonalisation in the 14 spin-orbital basis. Thus, the QCM approach, despite an increased number of measurements and noisy quantum hardware (CNOT error rates c.1% corresponding to expected error rates on the trial-state circuit of order 20%), is able to determine the ground-state energy of a non-trivial molecular system at the required accuracy (c.0.1%). To the best of our knowledge, these results are the largest calculations performed on a physical quantum computer to date in terms of encoding individual spin-orbitals producing chemically relevant accuracy, and a promising indicator of how such hybrid approaches might scale to problems of interest in the low-error/fault-tolerant regimes as quantum computers develop., Comment: 20 pages, 8 figures (incl. SI)

With quantum devices rapidly approaching qualities and scales needed for fault tolerance, the validity of simplified error models underpinning the study of quantum error correction needs to be experimentally evaluated. In this work, we have assessed the performance of IBM superconducting quantum computer devices implementing heavy-hexagon code syndrome measurements with increasing circuit sizes up to 23 qubits, against the error assumptions underpinning code threshold calculations. Circuit operator change rate statistics in the presence of depolarizing and biased noise were modelled using analytic functions of error model parameters. Data from 16 repeated syndrome measurement cycles was found to be inconsistent with a uniform depolarizing noise model, favouring instead biased and inhomogeneous noise models. Spatial-temporal correlations investigated via $Z$ stabilizer measurements revealed significant temporal correlation in detection events. These results highlight the non-trivial structure which may be present in the noise of quantum error correction circuits, revealed by operator measurement statistics, and support the development of noise-tailored codes and decoders to adapt.

One of the fundamental tasks in information theory is the compression of information. To achieve this in the quantum domain, quantum autoencoders that aim to compress quantum states to low-dimensional ones have been proposed. When taking a pure state as the reference state, there exists an upper bound for the encoding fidelity. This bound limits the compression rate for high-rank states that have high entropy. To overcome the entropy inconsistency between the initial states and the reconstructed states, we allow the reference state to be a mixed state. A new cost function that combines the encoding fidelity and the quantum mutual information is proposed for compressing general input states. In particular, we consider the reference states to be a mixture of maximally mixed states and pure states. To achieve efficient compression for different states, two strategies for setting the ratio of mixedness (in the mixture of maximally mixed states and pure states) are provided based on prior knowledge about quantum states or observations obtained from the training process. Numerical results on thermal states of the transverse-field Ising model, Werner states, and maximally mixed states blended with pure states illustrate the effectiveness of the proposed method. In addition, quantum autoencoders using mixed reference states are experimentally implemented on IBM Quantum devices to compress and reconstruct thermal states and Werner states.

Quantum machine learning (QML) is emerging as an application of quantum computing with the potential to deliver quantum advantage, but its realisation for practical applications remains impeded by challenges. Amongst those, a key barrier is the computationally expensive task of encoding classical data into a quantum state, which could erase any prospective speed-ups over classical algorithms. In this work, we implement methods for the efficient preparation of quantum states representing encoded image data using variational, genetic and matrix product state based algorithms. Our results show that these methods can approximately prepare states to a level suitable for QML using circuits two orders of magnitude shallower than a standard state preparation implementation, obtaining drastic savings in circuit depth and gate count without unduly sacrificing classification accuracy. Additionally, the QML models trained and evaluated on approximately encoded data display an increased robustness to adversarially generated input data perturbations. This partial alleviation of adversarial vulnerability, possible due to the "drowning out" of adversarial perturbations while retaining the meaningful large-scale features of the data, constitutes a considerable benefit for approximate state preparation in addition to lessening the requirements of the quantum hardware. Our results, based on simulations and experiments on IBM quantum devices, highlight a promising pathway for the future implementation of accurate and robust QML models on complex datasets relevant for practical applications, bringing the possibility of NISQ-era QML advantage closer to reality., Comment: 14 pages, 8 figures

Magnetic van der Waals materials are often proposed for use in future spintronic devices, aiming to leverage the combination of long-range magnetic order and near-atomic thinness to produce energy-efficient components. One class of material that has been discussed in this context are the iron halides FeCl$_2$ and FeBr$_2$, which are A-type antiferromagnets with strong uniaxial magnetocrystalline anisotropy. However, despite characterization of the bulk materials, the possibility for sustaining the magnetic behaviors that would underpin such applications in thin flakes has not been investigated. In this work, we use nitrogen-vacancy (NV) center microscopy to quantitatively image magnetism in individual exfoliated flakes of these iron halides, revealing the absence of magnetic remanence, a weak induced magnetization under bias field and variable behavior versus temperature. We show that our results are consistent with the antiferromagnetic behavior of the bulk material with a soft ferromagnetic uncompensated layer, indicating that extended ($>1~\mu$m) ferromagnetic domains are not sustained even at low temperatures (down to 4 K). Finally, we find that the magnetic order is strongly affected by the sample preparation, with a surprising diamagnetic order observed in a thin, hydrated sample., Comment: 15 pages, 13 figures

Machine learning algorithms are powerful tools for data driven tasks such as image classification and feature detection, however their vulnerability to adversarial examples - input samples manipulated to fool the algorithm - remains a serious challenge. The integration of machine learning with quantum computing has the potential to yield tools offering not only better accuracy and computational efficiency, but also superior robustness against adversarial attacks. Indeed, recent work has employed quantum mechanical phenomena to defend against adversarial attacks, spurring the rapid development of the field of quantum adversarial machine learning (QAML) and potentially yielding a new source of quantum advantage. Despite promising early results, there remain challenges towards building robust real-world QAML tools. In this review we discuss recent progress in QAML and identify key challenges. We also suggest future research directions which could determine the route to practicality for QAML approaches as quantum computing hardware scales up and noise levels are reduced., Comment: 10 Pages, 4 Figures

The efficient preparation of quantum states is an important step in the execution of many quantum algorithms. In the noisy intermediate-scale quantum (NISQ) computing era, this is a significant challenge given quantum resources are scarce and typically only low-depth quantum circuits can be implemented on physical devices. We present a genetic algorithm for state preparation (GASP) which generates relatively low-depth quantum circuits for initialising a quantum computer in a specified quantum state. The method uses a basis set of R_x, R_y, R_z, and CNOT gates and a genetic algorithm to systematically generate circuits to synthesize the target state to the required fidelity. GASP can produce more efficient circuits of a given accuracy with lower depth and gate counts than other methods. This variability of the required accuracy facilitates overall higher accuracy on implementation, as error accumulation in high-depth circuits can be avoided. We directly compare the method to the state initialisation technique based on an exact synthesis technique by implemented in IBM Qiskit simulated with noise and implemented on physical IBM Quantum devices. Results achieved by GASP outperform Qiskit's exact general circuit synthesis method on a variety of states such as Gaussian states and W-states, and consistently show the method reduces the number of gates required for the quantum circuits to generate these quantum states to the required accuracy.

Machine learning (ML) methods such as artificial neural networks are rapidly becoming ubiquitous in modern science, technology and industry. Despite their accuracy and sophistication, neural networks can be easily fooled by carefully designed malicious inputs known as adversarial attacks. While such vulnerabilities remain a serious challenge for classical neural networks, the extent of their existence is not fully understood in the quantum ML setting. In this work, we benchmark the robustness of quantum ML networks, such as quantum variational classifiers (QVC), at scale by performing rigorous training for both simple and complex image datasets and through a variety of high-end adversarial attacks. Our results show that QVCs offer a notably enhanced robustness against classical adversarial attacks by learning features which are not detected by the classical neural networks, indicating a possible quantum advantage for ML tasks. Contrarily, and remarkably, the converse is not true, with attacks on quantum networks also capable of deceiving classical neural networks. By combining quantum and classical network outcomes, we propose a novel adversarial attack detection technology. Traditionally quantum advantage in ML systems has been sought through increased accuracy or algorithmic speed-up, but our work has revealed the potential for a new kind of quantum advantage through superior robustness of ML models, whose practical realisation will address serious security concerns and reliability issues of ML algorithms employed in a myriad of applications including autonomous vehicles, cybersecurity, and surveillance robotic systems., Comment: 10 pages, 5 Figures

In the lead up to fault tolerance, the utility of quantum computing will be determined by how adequately the effects of noise can be circumvented in quantum algorithms. Hybrid quantum-classical algorithms such as the variational quantum eigensolver (VQE) have been designed for the short-term regime. However, as problems scale, VQE results are generally scrambled by noise on present-day hardware. While error mitigation techniques alleviate these issues to some extent, there is a pressing need to develop algorithmic approaches with higher robustness to noise. Here, we explore the robustness properties of the recently introduced quantum computed moments (QCM) approach to ground state energy problems, and show through an analytic example how the underlying energy estimate explicitly filters out incoherent noise. Motivated by this observation, we implement QCM for a model of quantum magnetism on IBM Quantum hardware to examine the noise-filtering effect with increasing circuit depth. We find that QCM maintains a remarkably high degree of error robustness where VQE completely fails. On instances of the quantum magnetism model up to 20 qubits for ultra-deep trial state circuits of up to ~500 CNOTs, QCM is still able to extract reasonable energy estimates. The observation is bolstered by an extensive set of experimental results. To match these results, VQE would need hardware improvement by some 2 orders of magnitude on error rates., Comment: 14 pages, 4 figures

The emergence of Quantum Machine Learning (QML) to enhance traditional classical learning methods has seen various limitations to its realisation. There is therefore an imperative to develop quantum models with unique model hypotheses to attain expressional and computational advantage. In this work we extend the linear quantum support vector machine (QSVM) with kernel function computed through quantum kernel estimation (QKE), to form a decision tree classifier constructed from a decision directed acyclic graph of QSVM nodes - the ensemble of which we term the quantum random forest (QRF). To limit overfitting, we further extend the model to employ a low-rank Nystr\"{o}m approximation to the kernel matrix. We provide generalisation error bounds on the model and theoretical guarantees to limit errors due to finite sampling on the Nystr\"{o}m-QKE strategy. In doing so, we show that we can achieve lower sampling complexity when compared to QKE. We numerically illustrate the effect of varying model hyperparameters and finally demonstrate that the QRF is able obtain superior performance over QSVMs, while also requiring fewer kernel estimations., Comment: 33 pages, 11 figures

Quantum algorithms based on variational approaches are one of the most promising methods to construct quantum solutions and have found a myriad of applications in the last few years. Despite the adaptability and simplicity, their scalability and the selection of suitable ans\"atzs remain key challenges. In this work, we report an algorithmic framework based on nested Monte-Carlo Tree Search (MCTS) coupled with the combinatorial multi-armed bandit (CMAB) model for the automated design of quantum circuits. Through numerical experiments, we demonstrated our algorithm applied to various kinds of problems, including the ground energy problem in quantum chemistry, quantum optimisation on a graph, solving systems of linear equations, and finding encoding circuit for quantum error detection codes. Compared to the existing approaches, the results indicate that our circuit design algorithm can explore larger search spaces and optimise quantum circuits for larger systems, showing both versatility and scalability., Comment: 32 pages, 24 figures

Quantum process tomography (QPT), where a quantum channel is reconstructed through the analysis of repeated quantum measurements, is an important tool for validating the operation of a quantum processor. We detail the combined use of an existing QPT approach based on tensor networks (TNs) and unsupervised learning with TN contraction path finding algorithms in order to use TNs of arbitrary topologies for reconstruction. Experiments were conducted on the 7-qubit IBM Quantum Falcon Processor ibmq_casablanca, where we demonstrate this technique by matching the topology of the tensor networks used for reconstruction with the topology of the processor, allowing us to extend past the characterisation of linear nearest neighbour circuits. Furthermore, we conduct single-qubit gate set tomography (GST) on each individual qubit to correct for separable errors during the state preparation and measurement phases of QPT, which are separate from the channel under consideration but may negatively impact the quality of its reconstruction. We are able to report a fidelity of 0.89 against the ideal unitary channel of a single-cycle random quantum circuit performed on ibmq_casablanca, after obtaining just $1.1 \times 10^5$ measurements for the reconstruction of this 7-qubit process. This represents more than five orders of magnitude fewer total measurements than the number needed to conduct full, traditional QPT on a 7-qubit process., Comment: 8 pages, 7 figures

Imaging individual conformational instances of generic, inhomogeneous, transient or intrinsically disordered protein systems at the single molecule level in situ is one of the notable challenges in structural biology. Present techniques access averaged structural information by measuring over large ensembles of proteins in nearly uniform conformational states in synthetic environments. This poses significant implications for diagnostics and drug design which require a detailed understanding of subtle conformational changes, small molecule interactions and ligand dynamics. Here we tackle the problem by designing a single molecule imaging platform technology embracing the advantages silicon-based spin qubits offer in terms of quantum coherence and established fabrication pathways. We demonstrate through detailed simulation, that this platform enables scalable atomic-level structure-determination of individual molecular systems in native environments. The approach is particularly well suited to the high-value lipid-membrane context, and as such its experimental implementation could have far-reaching consequences for the understanding of membrane biology and drug development.

Quantum computers hold promise to circumvent the limitations of conventional computing for difficult molecular problems. However, the accumulation of quantum logic errors on real devices represents a major challenge, particularly in the pursuit of chemical accuracy requiring the inclusion of dynamical effects. In this work we implement the quantum computed moments (QCM) approach for hydrogen chain molecular systems up to H$_6$. On a superconducting quantum processor, Hamiltonian moments, $\langle \mathcal{H}^p\rangle$ are computed with respect to the Hartree-Fock state, which are then employed in Lanczos expansion theory to determine an estimate for the ground-state energy which incorporates electronic correlations and manifestly improves on the variational result. Post-processing purification of the raw QCM data takes the estimate through the Hartree-Fock variational limit to within 99.9% of the exact electronic ground-state energy for the largest system studied, H$_6$. Calculated dissociation curves indicate precision at about 10mH for this system and as low as 0.1mH for molecular hydrogen, H$_2$, over a range of bond lengths. In the context of stringent precision requirements for chemical problems, these results provide strong evidence for the error suppression capability of the QCM method, particularly when coupled with post-processing error mitigation. Greater emphasis on more efficient representations of the Hamiltonian and classical preprocessing steps may enable the solution of larger systems on near-term quantum processors., Comment: 17 pages, 6 figures

Surface code error correction offers a highly promising pathway to achieve scalable fault-tolerant quantum computing. When operated as stabilizer codes, surface code computations consist of a syndrome decoding step where measured stabilizer operators are used to determine appropriate corrections for errors in physical qubits. Decoding algorithms have undergone substantial development, with recent work incorporating machine learning (ML) techniques. Despite promising initial results, the ML-based syndrome decoders are still limited to small scale demonstrations with low latency and are incapable of handling surface codes with boundary conditions and various shapes needed for lattice surgery and braiding. Here, we report the development of an artificial neural network (ANN) based scalable and fast syndrome decoder capable of decoding surface codes of arbitrary shape and size with data qubits suffering from the depolarizing error model. Based on rigorous training over 50 million random quantum error instances, our ANN decoder is shown to work with code distances exceeding 1000 (more than 4 million physical qubits), which is the largest ML-based decoder demonstration to-date. The established ANN decoder demonstrates an execution time in principle independent of code distance, implying that its implementation on dedicated hardware could potentially offer surface code decoding times of O($\mu$sec), commensurate with the experimentally realisable qubit coherence times. With the anticipated scale-up of quantum processors within the next decade, their augmentation with a fast and scalable syndrome decoder such as developed in our work is expected to play a decisive role towards experimental implementation of fault-tolerant quantum information processing., Comment: 11 pages, 6 figures

Multi-time quantum processes are endowed with the same richness as many-body physics, including temporal entanglement and well-defined causal structures. We dub this many-time physics, and show how surprisingly accessible, yet under-explored, these phenomena are in nascent quantum processors. Here, we develop a family of tools that allow us access to many-time physics on quantum information processors, which are then demonstrated. First, we access short-range microscopic properties, such as genuine multi-time entanglement and estimators for non-Markovian memory. Then, adapting classical shadow tomography to many-time scenarios, we access macroscopic features like long-range correlations and compact representations of large processes. We showcase this for a 20-step process (42-qubit state) by accurately capturing numerous facets of the dynamics, including multi-time correlations -- for example, in the prediction of mid-circuit measurement distributions. Our techniques are pertinent to generic quantum stochastic dynamical processes, with a scope ranging across condensed matter physics, quantum biology, and in-depth diagnostics of NISQ era quantum devices., Comment: 12 pages, 6 figures + SI

Quantum machine learning (QML) is a rapidly growing area of research at the intersection of classical machine learning and quantum information theory. One area of considerable interest is the use of QML to learn information contained within quantum states themselves. In this work, we propose a novel approach in which the extraction of information from quantum states is undertaken in a classical representational-space, obtained through the training of a hybrid quantum autoencoder (HQA). Hence, given a set of pure states, this variational QML algorithm learns to identify, and classically represent, their essential distinguishing characteristics, subsequently giving rise to a new paradigm for clustering and semi-supervised classification. The analysis and employment of the HQA model are presented in the context of amplitude encoded states - which in principle can be extended to arbitrary states for the analysis of structure in non-trivial quantum data sets., Comment: 15 pages, 14 figures

Phosphorus donor spins in silicon offer a number of promising characteristics for the implementation of robust qubits. Amongst various concepts for scale-up, the shared-control concept takes advantage of 3D scanning tunnelling microscope (STM) fabrication techniques to minimise the number of control lines, allowing the donors to be placed at the pitch limit of $\geq$30 nm, enabling dipole interactions. A fundamental challenge is to exploit the faster exchange interaction, however, the donor spacings required are typically 15 nm or less, and the exchange interaction is notoriously sensitive to lattice site variations in donor placement. This work presents a proposal for a fast exchange-based surface-code quantum computer architecture which explicitly addresses both donor placement imprecision commensurate with the atomic-precision fabrication techniques and the stringent qubit pitch requirements. The effective pitch is extended by incorporation of an intermediate donor acting as an exchange-interaction switch. We consider both global control schemes and a scheduled series of operations by designing GRAPE pulses for individual CNOTs based on coupling scenarios predicted by atomistic tight-binding simulations. The architecture is compatible with the existing fabrication capabilities and may serve as a blueprint for the experimental implementation of a full-scale fault-tolerant quantum computer based on donor impurities in silicon., Comment: 7 pages, 5 figures

Characterisation protocols have so far played a central role in the development of noisy intermediate-scale quantum (NISQ) computers capable of impressive quantum feats. This trajectory is expected to continue in building the next generation of devices: ones that can surpass classical computers for particular tasks -- but progress in characterisation must keep up with the complexities of intricate device noise. A missing piece in the zoo of characterisation procedures is tomography which can completely describe non-Markovian dynamics over a given time frame. Here, we formally introduce a generalisation of quantum process tomography, which we call process tensor tomography. We detail the experimental requirements, construct the necessary post-processing algorithms for maximum-likelihood estimation, outline the best-practice aspects for accurate results, and make the procedure efficient for low-memory processes. The characterisation is a pathway to diagnostics and informed control of correlated noise. As an example application of the hardware-agnostic technique, we show how its predictive control can be used to substantially improve multi-time circuit fidelities on superconducting quantum devices. Our methods could form the core for carefully developed software that may help hardware consistently pass the fault-tolerant noise threshold., Comment: 21 pages, 10 figures

The interaction between a central qubit spin and a surrounding bath of spins is critical to spin-based solid state quantum sensing and quantum information processing. Spin-bath interactions are typically strongly anisotropic, and rapid physical rotation has long been used in solid-state nuclear magnetic resonance to simulate motional averaging of anisotropic interactions, such as dipolar coupling between nuclear spins. Here, we show that the interaction between electron spins of nitrogen-vacancy centers and a bath of $^{13}$C nuclear spins in a diamond rotated at up to 300,000rpm introduces decoherence into the system via frequency-modulation of the nuclear spin Larmor precession. The presence of an off-axis magnetic field necessary for averaging of the dipolar coupling leads to a rotational dependence of the electron-nuclear hyperfine interaction, which cannot be averaged out with experimentally achievable rotation speeds. Our findings offer new insights into the use of physical rotation for quantum control with implications for quantum systems having motional and rotational degrees of freedom that are not fixed., Comment: 9 pages, four figures, comments welcome

The ability to generate large-scale entanglement is an important progenitor of quantum information processing capability in noisy intermediate-scale quantum (NISQ) devices. In this paper, we investigate the extent to which entangled quantum states over large numbers of qubits can be prepared on current superconducting quantum devices. We prepared native-graph states on the IBM Quantum 65-qubit $\textit{ibmq_manhattan}$ device and the 53-qubit $\textit{ibmq_rochester}$ device and applied quantum readout-error mitigation (QREM). Connected entanglement graphs spanning each of the full devices were detected, indicating bipartite entanglement over the whole of each device. The application of QREM was shown to increase the observed entanglement within all measurements, in particular, the detected number of entangled pairs of qubits found within $\textit{ibmq_rochester}$ increased from 31 to 56 of the total 58 connected pairs. The results of this work indicate full bipartite entanglement in two of the largest superconducting devices to date., Comment: 15 pages, 8 figures

Generating and detecting genuine multipartite entanglement (GME) of sizeable quantum states prepared on physical devices is an important benchmark for highlighting the progress of near-term quantum computers. A common approach to certify GME is to prepare a Greenberger-Horne-Zeilinger (GHZ) state and measure a GHZ fidelity of at least 0.5. We measure the fidelities using multiple quantum coherences of GHZ states on 11 to 27 qubits prepared on the IBM Quantum ibmq_montreal device. Combinations of quantum readout error mitigation (QREM) and parity verification error detection are applied to the states. A fidelity of $0.546 \pm 0.017$ was recorded for a 27-qubit GHZ state when QREM was used, demonstrating GME across the full device with a confidence level of 98.6%. We benchmarked the effect of parity verification on GHZ fidelity for two GHZ state preparation embeddings on the heavy-hexagon architecture. The results show that the effect of parity verification, while relatively modest, led to a detectable improvement of GHZ fidelity., Comment: 16 pages, 7 figures. V3 includes a revised analysis of the data in Figures 5 and 6

Metalated phthalocyanines (Pc's) are robust and versatile molecular complexes, whose properties can be tuned by changing their functional groups and central metal atom. The electronic structure of magnesium Pc (MgPc) - structurally and electronically similar to chlorophyll - adsorbed on the Ag(100) surface is investigated by low-temperature scanning tunneling microscopy (STM) and spectroscopy (STS), non-contact atomic force microscopy (ncAFM) and density functional theory (DFT). Single, isolated MgPc's exhibit a flat, four-fold rotationally symmetric morphology, with doubly degenerate, partially populated (due to surface-to-molecule electron transfer) lowest unoccupied molecular orbitals (LUMOs). In contrast, MgPc's with neighbouring molecules in proximity undergo a lift of LUMOs degeneracy, with a near-Fermi local density of states with reduced two-fold rotational symmetry, indicative of a long-range attractive intermolecular interaction. The latter is assigned to a surface-mediated two-step electronic hybridization process. First, LUMOs interact with Ag(100) conduction electrons, forming hybrid molecule-surface orbitals with enhanced spatial extension. Then, these delocalized molecule-surface states further hybridize with those of neighbouring molecules. This work highlights how the electronic structure of molecular adsorbates - including orbital degeneracies and symmetries - can be significantly altered via surface-mediated intermolecular hybridization, over extended distances (beyond 3 nm), having important implications for prospects of molecule-based solid-state technologies., Comment: 28 pages, 4 figures; supplementary 35 pages, 19 figures

The variational principle of quantum mechanics is the backbone of hybrid quantum computing for a range of applications. However, as the problem size grows, quantum logic errors and the effect of barren plateaus overwhelm the quality of the results. There is now a clear focus on strategies that require fewer quantum circuit steps and are robust to device errors. Here we present an approach in which problem complexity is transferred to dynamic quantities computed on the quantum processor - Hamiltonian moments, $\langle H^n\rangle$. From these quantum computed moments, estimates of the ground-state energy are obtained using the "infinum" theorem from Lanczos cumulant expansions which manifestly correct the associated variational calculation. With system dynamics encoded in the moments the burden on the trial-state quantum circuit depth is eased. The method is introduced and demonstrated on 2D quantum magnetism models on lattices up to 5 $\times$ 5 (25 qubits) implemented on IBM Quantum superconducting qubit devices. Moments were quantum computed to fourth order with respect to a parameterised antiferromagnetic trial-state. A comprehensive comparison with benchmark variational calculations was performed, including over an ensemble of random coupling instances. The results showed that the infinum estimate consistently outperformed the benchmark variational approach for the same trial-state. These initial investigations suggest that the quantum computed moments approach has a high degree of stability against trial-state variation, quantum gate errors and shot noise, all of which bodes well for further investigation and applications of the approach., Comment: 10 pages, 5 figures

For universal quantum computation, a major challenge to overcome for practical implementation is the large amount of resources required for fault-tolerant quantum information processing. An important aspect is implementing arbitrary unitary operators built from logical gates within the quantum error correction code. A synthesis algorithm can be used to approximate any unitary gate up to arbitrary precision by assembling sequences of logical gates chosen from a small set of universal gates that are fault-tolerantly performable while encoded in a quantum error-correction code. However, current procedures do not yet support individual assignment of base gate costs and many do not support extended sets of universal base gates. We analysed cost-optimal sequences using an exhaustive search based on Dijkstra's pathfinding algorithm for the canonical Clifford+$T$ set of base gates and compared them to when additionally including $Z$-rotations from higher orders of the Clifford hierarchy. Two approaches of assigning base gate costs were used. First, costs were reduced to $T$-counts by recursively applying a $Z$-rotation catalyst circuit. Second, costs were assigned as the average numbers of raw (i.e. physical level) magic states required to directly distil and implement the gates fault-tolerantly. We found that the average sequence cost decreases by up to $54\pm 3\%$ when using the $Z$-rotation catalyst circuit approach and by up to $33\pm 2 \%$ when using the magic state distillation approach. In addition, we investigated observed limitations of certain assignments of base gate costs by developing an analytic model to estimate the proportion of sets of $Z$-rotation gates from higher orders of the Clifford hierarchy that are found within sequences approximating random target gates., Comment: 19 pages, 7 figures

In the scale-up of quantum computers, the framework underpinning fault-tolerance generally relies on the strong assumption that environmental noise affecting qubit logic is uncorrelated (Markovian). However, as physical devices progress well into the complex multi-qubit regime, attention is turning to understanding the appearance and mitigation of correlated -- or non-Markovian -- noise, which poses a serious challenge to the progression of quantum technology. This error type has previously remained elusive to characterisation techniques. Here, we develop a framework for characterising non-Markovian dynamics in quantum systems and experimentally test it on multi-qubit superconducting quantum devices. Where noisy processes cannot be accounted for using standard Markovian techniques, our reconstruction predicts the behaviour of the devices with an infidelity of $10^{-3}$. Our results show this characterisation technique leads to superior quantum control and extension of coherence time by effective decoupling from the non-Markovian environment. This framework, validated by our results, is applicable to any controlled quantum device and offers a significant step towards optimal device operation and noise reduction.

The recent isolation of two-dimensional van der Waals magnetic materials has uncovered rich physics that often differs from the magnetic behaviour of their bulk counterparts. However, the microscopic details of fundamental processes such as the initial magnetization or domain reversal, which govern the magnetic hysteresis, remain largely unknown in the ultrathin limit. Here we employ a widefield nitrogen-vacancy (NV) microscope to directly image these processes in few-layer flakes of magnetic semiconductor vanadium triiodide (VI$_3$). We observe complete and abrupt switching of most flakes at fields $H_c\approx0.5-1$ T (at 5 K) independent of thickness down to two atomic layers, with no intermediate partially-reversed state. The coercive field decreases as the temperature approaches the Curie temperature ($T_c\approx50$ K), however, the switching remains abrupt. We then image the initial magnetization process, which reveals thickness-dependent domain wall depinning fields well below $H_c$. These results point to ultrathin VI$_3$ being a nucleation-type hard ferromagnet, where the coercive field is set by the anisotropy-limited domain wall nucleation field. This work illustrates the power of widefield NV microscopy to investigate magnetization processes in van der Waals ferromagnets, which could be used to elucidate the origin of the hard ferromagnetic properties of other materials and explore field- and current-driven domain wall dynamics., Comment: includes SI

Microscopic imaging based on nitrogen-vacancy (NV) centres in diamond, a tool increasingly used for room-temperature studies of condensed matter systems, has recently been extended to cryogenic conditions. However, it remains unclear whether the technique is viable for imaging temperature-sensitive phenomena below 10 K given the inherent laser illumination requirements, especially in a widefield configuration. Here we realise a widefield NV microscope with a field of view of 100 $\mu$m and a base temperature of 4 K, and use it to image Abrikosov vortices and transport currents in a superconducting Nb film. We observe the disappearance of vortices upon increase of laser power and their clustering about hot spots upon decrease, indicating that laser powers as low as 1 mW (4 orders of magnitude below the NV saturation) are sufficient to locally quench the superconductivity of the film ($T_c = 9$ K). This significant local heating is confirmed by resistance measurements, which reveal the presence of large temperature gradients (several K) across the film. We then investigate the effect of such gradients on transport currents, where the current path is seen to correlate with the temperature profile even in the fully superconducting phase. In addition to highlighting the role of temperature inhomogeneities in superconductivity phenomena, this work establishes that, under sufficiently low laser power conditions, widefield NV microscopy enables imaging over mesoscopic scales down to 4 K with a submicrometer spatial resolution, providing a new platform for real-space investigations of a range of systems from topological insulators to van der Waals ferromagnets.

Quantum system characterization techniques represent the front line in the identification and mitigation of noise in quantum computing, but can be expensive in terms of quantum resources and time to repeatedly employ. Another challenging aspect is that parameters governing the performance of various operations tend to drift over time, and monitoring these is hence a difficult task. One of the most promising characterization techniques, gate set tomography (GST), provides a self-consistent estimate of the completely positive, trace-preserving (CPTP) maps for a complete set of gates, as well as preparation and measurement operators. We develop a method for performance optimization seeded by tomography (POST), which couples the power of GST with a classical optimization routine to achieve a consistent gate improvement in just a short number of steps within a given calibration cycle. By construction, the POST procedure finds the best available gate operation given the hardware, and is therefore robust to the effects of drift. Further, in comparison to other quantum error mitigation techniques, it builds upon a one-time application of GST. To demonstrate the performance of this method on a real quantum computer, we map out the operations of six qubit pairs on the superconducting \emph{ibmq\_poughkeepsie} quantum device. Under the restriction of logical-only control, we monitor the performance of the POST approach on a chosen CNOT gate over a period of six weeks. In this time, we achieve a consistent improvement in gate fidelity, averaging a fidelity increase of 21.1\% as measured by randomized benchmarking. The POST approach should find wide applicability as it is hardware agnostic, and can be applied at the upper logical level or at a deeper pulse control level., Comment: 11 pages, 4 figures

We introduce a framework for mapping NP-Hard problems to adiabatic quantum computing (AQC) architectures that are heavily restricted in both connectivity and dynamic range of couplings, for which minor-embedding -- the standard problem mapping method -- cannot be directly applied. Separating the mapping into two distinct stages, we introduce problem-specific reductions for both quadratic unconstrained binary optimisation (QUBO) and satisfiability (SAT) and develop the subdivision-embedding method that is suitable for directly embedding onto these heavily restricted architectures. The theory underpinning this framework provides tools to aid in the manipulation of Ising Hamiltonians for the purposes of Ising energy minimisation, and could be used to assist in developing and optimising further problem mapping techniques. For each of the problem mapping methods presented, we examine how the physical qubit count scales with problem size on architectures of varying connectivity., Comment: 35 pages, 28 figures

The theory of angular momentum connects physical rotations and quantum spins together at a fundamental level. Physical rotation of a quantum system will therefore affect fundamental quantum operations, such as spin rotations in projective Hilbert space, but these effects are subtle and experimentally challenging to observe due to the fragility of quantum coherence. Here we report a measurement of a single-electron-spin phase shift arising directly from physical rotation, without transduction through magnetic fields or ancillary spins. This phase shift is observed by measuring the phase difference between a microwave driving field and a rotating two-level electron spin system, and can accumulate nonlinearly in time. We detect the nonlinear phase using spin-echo interferometry of a single nitrogen-vacancy qubit in a diamond rotating at 200,000rpm. Our measurements demonstrate the fundamental connections between spin, physical rotation and quantum phase, and will be applicable in schemes where the rotational degree of freedom of a quantum system is not fixed, such as spin-based rotation sensors and trapped nanoparticles containing spins., Comment: Five pages, three figures, Supplementary Info

The application of imaging techniques based on ensembles of nitrogen-vacancy (NV) sensors in diamond to characterise electrical devices has been proposed, but the compatibility of NV sensing with operational gated devices remains largely unexplored. Here we fabricate graphene field-effect transistors (GFETs) directly on the diamond surface and characterise them via NV microscopy. The current density within the gated graphene is reconstructed from NV magnetometry under both mostly p- and n-type doping, but the exact doping level is found to be affected by the measurements. Additionally, we observe a surprisingly large modulation of the electric field at the diamond surface under an applied gate potential, seen in NV photoluminescence and NV electrometry measurements, suggesting a complex electrostatic response of the oxide-graphene-diamond structure. Possible solutions to mitigate these effects are discussed., Comment: 13 pages, 9 figures

Atomic level qubits in silicon are attractive candidates for large-scale quantum computing, however, their quantum properties and controllability are sensitive to details such as the number of donor atoms comprising a qubit and their precise location. This work combines machine learning techniques with million-atom simulations of scanning-tunnelling-microscope (STM) images of dopants to formulate a theoretical framework capable of determining the number of dopants at a particular qubit location and their positions with exact lattice-site precision. A convolutional neural network was trained on 100,000 simulated STM images, acquiring a characterisation fidelity (number and absolute donor positions) of above 98\% over a set of 17,600 test images including planar and blurring noise. The method established here will enable a high-precision post-fabrication characterisation of dopant qubits in silicon, with high-throughput potentially alleviating the requirements on the level of resource required for quantum-based characterisation, which may be otherwise a challenge in the context of large qubit arrays for universal quantum computing., Comment: 8 pages, 4 figures and a supplementary information document

Towards realising larger scale quantum algorithms, the ability to prepare sizeable multi-qubit entangled states with full qubit control is used as a benchmark for quantum technologies. We investigate the extent to which entanglement is found within a prepared graph state on the 20-qubit superconducting quantum computer, IBM Q Poughkeepsie. We prepared a graph state along a path consisting of all twenty qubits within Poughkeepsie and performed full quantum state tomography on all groups of four connected qubits along this path. We determined that each pair of connected qubits was inseparable and hence the prepared state was entangled. Additionally, a genuine multipartite entanglement witness was measured on all qubit subpaths of the graph state and we found genuine multipartite entanglement on chains of up to three qubits., Comment: 10 pages, 7 figures

Iron biomineralisation is critical for life. Nature capitalises on the physical attributes of iron biominerals for a variety of functional, structural and sensory applications. Although magnetism is an integral property of iron biominerals, the role it plays in their nano-assembly remains a fundamental, unanswered question. This is well exemplified by the magnetite-bearing radula of chitons. Chitons, a class of marine mollusc, create the hardest biomineral of any animal in their abrasion-resistant, self-sharpening teeth4. Despite this system being subjected to a range of high resolution imaging studies, the mechanisms that drive mineral assembly remain unresolved. However, the advent of quantum imaging technology provides a new avenue to probe magnetic structures directly. Here we use quantum magnetic microscopy, based on nitrogen-vacancy centres in diamond, to attain the first subcellular magnetic profiling of a eukaryotic system. Using complementary magnetic imaging protocols, we spatially map the principal mineral phases (ferrihydrite and magnetite) in the developing teeth of Acanthopleura hirtosa with submicron resolution. The images reveal previously undiscovered long-range magnetic order, established at the onset of magnetite mineralisation. This is in contrast to electron microscopy studies that show no strong common crystallographic orientation. The quantum-based magnetic profiling techniques presented in this work have broad application in biology, earth science, chemistry and materials engineering and can be applied across the range of systems for which iron is vital.

Diamond materials are central to an increasing range of advanced technological demonstrations, from high power electronics, to nano-scale quantum bio-imaging with unprecedented sensitivity. However, the full exploitation of diamond for these applications is often limited by the uncontrolled nature of the diamond material surface, which suffers from Fermi-level pinning and hosts a significant density of electro-magnetic noise sources. These issues occur despite the oxide-free and air-stable nature of the diamond crystal surface, which should be an ideal candidate for functionalization and chemical-engineering. In this work we reveal a family of previously unidentified and near-ubiquitous primal surface defects which we assign to differently reconstructed surface vacancies. The density of these defects is quantified with X-ray absorption spectroscopy, their energy structures are elucidated by ab initio calculations, and their effect on near-surface quantum probes is measured directly. Subsequent ab-initio calculation of band-bending from these defects suggest they are the source of Fermi-level pinning at most diamond surfaces. Finally, an investigation is conducted on a broad range of post-growth surface treatments and concludes that none of them can reproducibly reduce this defect density below the Fermi-pinning threshold, making this defect a prime candidate as the source for decoherence-limiting noise in near-surface quantum probes., Comment: 12 pages, 3 figures; SI: 10 pages, 10 figures

Optical bio-markers have been used extensively for intracellular imaging with high spatial and temporal resolution. Extending the modality of these probes is a key driver in cell biology. In recent years, the NV centre in nanodiamond has emerged as a promising candidate for bio-imaging and bio-sensing with low cytotoxicity and stable photoluminescence. Here we study the electrophysiological effects of this quantum probe in primary cortical neurons. Multi-electrode array (MEA) recordings across five replicate studies showed no statistically significant difference in 25 network parameters when nanodiamonds are added at varying concentrations over various time periods 12-36 hr. The physiological validation motivates the second part of the study which demonstrates how the quantum properties of these biomarkers can be used to report intracellular information beyond their location and movement. Using the optically detected magnetic resonance from the nitrogen vacancy defects within the nanodiamonds we demonstrate enhanced signal-to-noise imaging and temperature mapping from thousands of nanodiamond probes simultaneously. This work establishes nanodiamonds as viable multi-functional intraneuronal sensors with nanoscale resolution, that may ultimately be used to detect magnetic and electrical activity at the membrane level in excitable cellular systems., Comment: 27 pages, 5 Figures

A controlled qubit in a rotating frame opens new opportunities to probe fundamental quantum physics, such as geometric phases in physically rotating frames, and can potentially enhance detection of magnetic fields. Realising a single qubit that can be measured and controlled during physical rotation is experimentally challenging. In this work, we demonstrate quantum control of a single nitrogen-vacancy (NV) centre within a diamond rotated at 200,000rpm, a rotational period comparable to the NV spin coherence time $T_2$. We stroboscopically image individual NV centres that execute rapid circular motion in addition to rotation, and demonstrate preparation, control and readout of the qubit quantum state with lasers and microwaves. Using spin-echo interferometry of the rotating qubit, we are able to detect modulation of the NV Zeeman shift arising from the rotating NV axis and an external DC magnetic field. Our work establishes single NV qubits in diamond as quantum sensors in the physically rotating frame, and paves the way for the realisation of single-qubit diamond-based rotation sensors., Comment: 19 pages, 5 figures, accompanying Supplementary material

Precise control of the resonant frequency of a spin qubit is of fundamental importance to quantum sensing protocols. We demonstrate a control technique on a single nitrogen-vacancy (NV) centre in diamond where the applied magnetic field is modified by fine-tuning a permanent magnet's magnetisation via temperature control. Through this control mechanism, nanoscale cross-relaxation spectroscopy of both electron and nuclear spins in the vicinity of the NV centre are performed. We then show that through maintaining the magnet at a constant temperature an order of magnitude improvement in the stability of the NV qubit frequency can be achieved. This improved stability is tested in the polarisation of a small ensemble of nearby $^{13}$C spins via resonant cross-relaxation and the lifetime of this polarisation explored. The effectiveness and relative simplicity of this technique may find use in the realisation of portable spectroscopy and/or hyperpolarisation systems., Comment: 8 pages, 6 figures including Supporting Information

We detail techniques to optimise high-level classical simulations of Shor's quantum factoring algorithm. Chief among these is to examine the entangling properties of the circuit and to effectively map it across the one-dimensional structure of a matrix product state. Compared to previous approaches whose space requirements depend on $r$, the solution to the underlying order-finding problem of Shor's algorithm, our approach depends on its factors. We performed a matrix product state simulation of a 60-qubit instance of Shor's algorithm that would otherwise be infeasible to complete without an optimised entanglement mapping., Comment: 8 pages, 2 figures, 2 tables. v2 using PDFLaTeX compiler. v3 to include extra references. v4 for publication in Quantum

The hyperpolarisation of nuclear spins within target molecules is a critical and complex challenge in magnetic resonance imaging (MRI) and nuclear magnetic resonance (NMR) spectroscopy. Hyperpolarisation offers enormous gains in signal and spatial resolution which may ultimately lead to the development of molecular MRI and NMR. At present, techniques used to polarise nuclear spins generally require low temperatures and/or high magnetic fields, radio-frequency control fields, or the introduction of catalysts or free-radical mediators. The emergence of room temperature solid-state spin qubits has opened exciting new pathways to circumvent these requirements to achieve direct nuclear spin hyperpolarisation using quantum control. Employing a novel cross-relaxation induced polarisation (CRIP) protocol using a single nitrogen-vacancy (NV) centre in diamond, we demonstrate the first external nuclear spin hyperpolarisation achieved by a quantum probe, in this case of $^1$H molecular spins in poly(methyl methacrylate). In doing so, we show that a single qubit is capable of increasing the thermal polarisation of $\sim 10^6$ nuclear spins by six orders of magnitude, equivalent to an applied magnetic field of $10^5$\,T. The technique can also be tuned to multiple spin species, which we demonstrate using both \C{13} and $^1$H nuclear spin ensembles. Our results are analysed and interpreted via a detailed theoretical treatment, which is also used to describe how the system can be scaled up to a universal quantum hyperpolarisation platform for the production of macroscopic quantities of contrast agents at high polarisation levels for clinical applications. These results represent a new paradigm for nuclear spin hyperpolarisation for molecular imaging and spectroscopy, and beyond into areas such as materials science and quantum information processing., Comment: 6 pages, 4 figures