Superconducting circuits are being used for large-scale quantum devices, and a major challenge is to perform accurate numerical simulations of device parameters. One of the most advanced methods for analyzing superconducting circuit designs is the energy-participation-ratio (EPR) method, which constructs quantum Hamiltonians based on the energy distribution extracted from classical electromagnetic simulations. In the EPR approach, we extract linear terms from finite-element simulations and add nonlinear terms using the energy participation ratio extracted from the classical simulations. However, the EPR method relies on a low-order expansion of nonlinear terms, which is prohibitive for accurately describing highly anharmonic circuits. An example of such a circuit is the fluxonium qubit, which has recently attracted increasing attention due to its high lifetimes and low error rates. In this work, we extend the EPR approach to effectively address highly nonlinear superconducting circuits, and, as a proof of concept, we apply our approach to a fluxonium qubit. Specifically, we design, fabricate, and experimentally measure a fluxonium qubit coupled to a readout resonator. We compare the measured frequencies of both the qubit and the resonator to those extracted from the EPR analysis, and we find an excellent agreement. Furthermore, we compare the dispersive shift as a function of external flux obtained from experiments with our EPR analysis and a simpler lumped-element model. Our findings reveal that the EPR results closely align with the experimental data, providing more accurate estimations compared to the simplified lumped-element simulations.

Hybrid quantum systems integrating spin qubits and superconducting qubits have emerged as promising candidates for scalable quantum information processing. In hybrid architectures, the development of high-fidelity quantum buses is a crucial but challenging element. Here we design a quantum bus that utilizes engineered virtual photons to couple spin and superconducting modules. We tailor the driving pulse to accelerate the evolution of the hybrid system in a noncyclic geometric phase way. We demonstrate the enhanced functionality of the quantum bus in three ways. First, we propose a scheme for realizing universal gate operation between different modules in a short time of 40 ns with 99.05% fidelity, exceeding the fault-tolerance threshold. Second, we theoretically demonstrate the preparation of remote entanglement between different modules with a fidelity of 99.21%, a valuable resource for hybrid architectures. Finally, we show the robustness of this approach to control parameter imperfections compared to usual dynamical pulses. These results provide a toolbox for fast, high-fidelity, and robust quantum information processing on hybrid platforms.

Quantum processors require rapid and high-fidelity simultaneous measurements of many qubits. While superconducting qubits are among the leading modalities toward a useful quantum processor, their readout remains a bottleneck. Traditional approaches to processing measurement data often struggle to account for crosstalk present in frequency-multiplexed readout, the preferred method to reduce the resource overhead. Recent approaches to address this challenge use neural networks to improve the state-discrimination fidelity. However, they are computationally expensive to train and evaluate, resulting in increased latency and poor scalability as the number of qubits increases. We present an alternative machine-learning approach based on next-generation reservoir computing that constructs polynomial features from the measurement signals and maps them to the corresponding qubit states. This method is highly parallelizable, avoids the costly nonlinear activation functions common in neural networks, and supports real-time training, enabling fast evaluation, adaptability, and scalability. Despite its lower computational complexity, our reservoir approach is able to maintain high qubit-state-discrimination fidelity. Relative to traditional methods, our approach achieves error reductions of up to 50% and 11% on single- and five-qubit datasets, respectively, and delivers up to $2.5\times$ crosstalk reduction on the five-qubit dataset. Compared with recent machine-learning methods, evaluating our model requires $100\times$ fewer multiplications for single-qubit and $2.5\times$ fewer for five-qubit models. This work demonstrates that reservoir computing can enhance qubit-state discrimination while maintaining scalability for future quantum processors.

The increasing number of qubits in quantum processors necessitates a corresponding increase in the number of control lines between the processor, which is typically operated at cryogenic temperatures, and external electronics. Scaling poses significant challenges in terms of the thermal loads, forming a major bottleneck in the realization of large-scale quantum computers. In this study, we analyze simultaneous gate operations on multiple qubits using microwaves transmitted via a single cable in a frequency-division multiplexing (FDM) scheme. By employing rectangular control microwave pulses, we reveal the contribution of drive frequency spacing to gate fidelity. Through theoretical and numerical analyses, we demonstrate that orthogonal and quasiorthogonal microwave signals suppress interference in simultaneously driven qubits, thereby ensuring high gate fidelity. Additionally, we provide design guidelines for key parameters, including pulse length, number of multiplexed microwave signals, and rotation angle, to achieve precise qubit operations. Our findings enable a scalable FDM-based microwave control scheme suitable for quantum processors with a large number of qubits.

Efficient qubit reset and leakage reduction are essential for scalable superconducting quantum computing, particularly in the context of quantum error correction. However, such operations often require additional on-chip components. Here, we propose and experimentally demonstrate a hardware-efficient approach to qubit reset and readout using a multimode Purcell filter in a superconducting quantum circuit. We exploit the inherent multimode structure of a coplanar waveguide resonator, using its fundamental and second-order modes for qubit reset and readout, respectively, thereby avoiding additional components. Implemented in a flip-chip architecture, our device achieves unconditional reset with residual excitation below 1% in 220 ns, and a leakage-reduction unit that selectively resets the second excited state within 62 ns, with a residual $|f⟩$ population of 6.1%, after accounting for the readout error. Despite the qubits being directly coupled to the filter in our configuration, the measured relaxation times are not degraded, owing to intrinsic Purcell protection provided by an auxiliary mode. To our knowledge, this work constitutes the first experimental demonstration exploiting different-order modes of a microwave resonator for distinct qubit operations, representing a new direction toward scalable, hardware-efficient quantum processor design.

Unlike other quantum hardware, photonic quantum architectures can produce millions of qubits from a single device. However, controlling photonic qubits remains challenging, even at small scales, due to their weak interactions, making nondeterministic gates in linear optics unavoidable. Nevertheless, a single photon can readily spread over multiple modes and create entanglement within the multiple modes deterministically. Rudolph’s concept of third quantization leverages this feature by evolving multiple single photons into multiple modes, distributing them uniformly and randomly to different parties, and creating multipartite entanglement without interactions between photons or nondeterministic gates. This method requires only classical communication and deterministic entanglement within multimode single-photon states and enables universal quantum computing. The multipartite entanglement generated within the third-quantization framework is nearly deterministic, where “deterministic” is achieved in the asymptotic limit of a large system size. In this work, we propose a near-term experiment using antimony donor in a silicon chip to realize third quantization. Utilizing the eight energy levels of antimony, one can generate two eight-mode single-photon states independently and distribute them to parties. This enables a random multipartite Bell-state experiment, achieving a Bell state with an upper-bound efficiency of 87.5% among 56 random pairs without nondeterministic entangling gates. This approach opens alternative pathways for silicon-based photonic quantum computing.

We present a practical design framework for high-fidelity quantum control in coupled Kerr-nonlinear oscillators, directly addressing the challenge of spectral crowding. We show that systematic spectral degeneracies, which hinder selective addressing, are a direct consequence of rational Kerr-nonlinearity ratios ($K_1/K_2$). Our solution is a universal architectural principle: engineer this ratio to be a complex rational value, approximating an incommensurate number to systematically eliminate parasitic resonances. Using a Magnus expansion, we derive a complete effective Hamiltonian, including all Stark-shift corrections, to accurately target transitions. We numerically validate this framework by demonstrating protocols for the deterministic synthesis of NOON states, and high-photon-number Fock states, achieving ideal fidelities exceeding $\mathcal{F}>99.7\mathrm{\%}$. The protocols are shown to be robust against environmental decay and thermal effects. This work provides an architectural blueprint for bosonic processors in circuit quantum electrodynamics and establishes foundational principles that could inform future designs of multimode quantum systems.

High-fidelity coherent population transfer plays a vital role in the realization of quantum memories. However, population transfer with high performance across a broad frequency range is still challenging due to the finite Rabi coupling strength limited by laser powers. Here we propose a population-transfer scheme by suturing adiabatic control pulses where each pulse covers a certain frequency interval and that are connected in a way that neighboring adiabatic pulses have opposite chirping directions. Taking the widely used hyperbolic-square-hyperbolic pulse as an example, we demonstrate that rapid and robust population transfer can be achieved. The transfer bandwidth scales linearly with the number of suture pulses while maintaining high fidelity, even at the suture points where adiabaticity breaks down. Crucially, these pulses can be realized by a single laser by means of temporal multiplexing. For a given bandwidth, this strategy substantially reduces the operational time, which is necessary for on-demand read out and suppression of decoherence effects. Our scheme enables a dramatic increase in multimode storage capacity, and paves the way for realizing practical quantum networks.

The active stabilization of polarization channels is a task of growing importance as quantum networks move to deployed demonstrations over existing fiber infrastructure. However, the uniquely strict requirements for high-fidelity qubit transmission complicate the extent to which classical solutions may apply to future quantum networks, particularly in terms of recognizing noise sources present in low-flux, nonunitary channels. Here we introduce an in situ benchmarking approach that augments a classical polarization tracking system, limited to unitary correction, with simultaneously transmitted quantum light for ancilla-assisted process tomography of the full quantum map. Implemented in a quantum local-area network, our method uses the reconstructed map both to validate the classical compensation and to expose noise sources it fails to capture. A sliding measurement window that continuously updates the estimated quantum process further increases sensitivity to rapid channel fluctuations. Our results should unlock new opportunities for in situ channel characterization in quantum-classical coexistence networks.

The maximum-independent-set (MIS) problem is a fundamental combinatorial optimization task that can be naturally mapped onto the Ising Hamiltonian of neutral-atom quantum processors. Given its connection to NP-hard problems and real-world applications, there has been significant interest in exploring quantum advantage for the MIS problem. Pioneering experiments on King’s lattice graphs have suggested a quadratic speedup over simulated annealing; however, recent benchmarks using state-of-the-art methods found no clear advantage, likely due to the structured nature of the tested instances. In this work, we generate hard instances of unit-disk graphs by leveraging complexity-theory results and varying key hardness parameters such as density and treewidth. For a fixed graph size, we show that increasing these parameters reveals a large number of hard instances that can lead to prohibitive classical run-time increases of several orders of magnitude. We then compare classical and quantum approaches on small instances and find that, at this scale, quantum solutions are slower than classical ones for finding exact solutions. Based on extended classical benchmarks at larger problem sizes, we estimate that scaling up to 1000 atoms with a 1-kHz repetition rate is a necessary step toward demonstrating a computational advantage with quantum methods.

Kernel methods serve as powerful tools to capture nonlinear patterns behind data in machine learning. Quantum kernels, by mapping data into exponentially large Hilbert spaces, offer the potential to solve problems intractable for classical models. However, existing studies encounter performance bottlenecks when processing complex data with localized structural patterns, stemming from the limitation in single-view feature representation and the exclusive reliance on global data structure. In this paper, we propose quantum multiview kernel learning with local information, called L-QMVKL. Specifically, based on the multiple kernel learning, a representative method for multiview data processing, we construct the quantum multiple kernel that combines view-specific quantum kernels to effectively fuse cross-view information. Further leveraging local information to capture intrinsic structural information, we design a sequential training strategy for the quantum circuit parameters and combination weights with the use of the hybrid global-local kernel alignment. We evaluate the effectiveness of L-QMVKL through comprehensive numerical simulations on the Mfeat dataset, demonstrating significant accuracy improvements achieved through leveraging multiview methodology and local information. Meanwhile, the results show that L-QMVKL exhibits competitive accuracy compared to its classical counterparts. Our work holds promise for advancing the theoretical and practical understanding of quantum kernel methods.

Quantum optimization solvers typically rely on one-variable-to-one-qubit mapping. However, the low qubit count of current quantum computers is a major obstacle to competing against classical methods. Here, we develop a qubit-efficient algorithm that overcomes this limitation by mapping a candidate bit-string solution to an entangled wave function of fewer qubits. We propose a variational quantum circuit generalizing the quantum approximate optimization ansatz. Extremizing the ansatz for Sherrington-Kirkpatrick spin-glass problems, we show valuable properties such as the concentration of ansatz parameters and derive performance guarantees. This approach could benefit near-term intermediate-scale and future fault-tolerant small-scale quantum devices.

The Mølmer-Sørensen (MS) gate is a two-qubit controlled-phase gate in ion traps that is highly valued due to its ability to preserve the motional state of the ions. However, its fidelity is obstructed by errors affecting the motion of the ions as well as the rotation of the qubits. In this work, we propose an amplitude-modulated composite MS gate, which features high fidelity, robust to gate timing, detuning, and coupling errors and is also tolerant of ac Stark shifts and drifting detuning errors. With the proposed amplitude-modulated composite sequence, we maintain gate infidelities below ${10}^{-3}$ even under simultaneous detuning, timing, and coupling errors as large as 10–15%, outperforming both standard and multitone MS gate implementations. The physical origin of this improvement is due to a separation of the error channels: the amplitude modulation suppresses residual spin-motion displacement (hence reducing sensitivity to detuning and timing imperfections), while the composite phase schedule cancels the dominant over- and under-rotation of the entangling phase produced by coupling miscalibration.

With the rapid advancement of generative artificial intelligence, efficiently training complex neural networks has emerged as a central challenge in deep learning. Ising machines, as specialized quantum-inspired hardware, show considerable promise for solving combinatorial optimization problems and training neural networks. However, current existing approaches are often limited to shallow architectures due to inefficient sampling strategies. In this study, we introduce a deep belief network (DBN) architecture that utilizes the sampling mechanism of a coherent Ising machine integrated with a spiking neural network (SNN-CIM). By stacking two restricted Boltzmann machines (RBMs), the model supports hierarchical feature learning while significantly enhancing training efficiency. The SNN-CIM mitigates amplitude heterogeneity through dissipative pulse injection and employs Gaussian noise to drive the system toward thermal equilibrium, enabling effective deep network training. Experimental results demonstrate that a single-layer RBM achieves $96.4\mathrm{\%}$ accuracy on the MNIST dataset, while the full DBN reaches $98.9\mathrm{\%}$ on the test set. Compared to Metropolis-Hastings-based MCMC sampling, the SNN-CIM exhibits broader sampling coverage and higher efficiency, achieving orders-of-magnitude performance improvements on large-scale problems. These findings underscore the potential of SNN-CIMs for scalable deep generative modeling and offer alternative pathways for enhancing performance in quantum-inspired computing and deep learning.

Determining the ground and low-lying excited states is critical in numerous scenarios. Recent work has proposed the ancilla-entangled variational quantum eigensolver (AEVQE) that utilizes entanglement between ancilla and physical qubits to simultaneously target multiple low-lying energy levels. In this work, we report the implementation of the AEVQE on a superconducting quantum cloud platform, demonstrating the full procedure of solving the low-lying energy levels of the $\mathrm{H}$${}_2$ molecule and the transverse-field Ising models (TFIMs). We obtain the potential energy curves of $\mathrm{H}$${}_2$ and show an indication of the ferromagnetic to paramagnetic phase transition in the TFIMs from the average absolute magnetization. Moreover, we investigate multiple factors that affect the algorithmic performance and provide a comparison with ancilla-free VQE algorithms. Our work demonstrates the feasibility of the AEVQE algorithm and offers guidance for the VQE approach in solving realistic problems on publicly accessible quantum platforms.

Phase correlations are an underexplored vulnerability in quantum key distribution. Here, we present an experimental and simulated characterization of correlations arising from electro-optic phase encoding, over repetition rates up to the GHz level. To mitigate this vulnerability (and all side channels arising from active phase modulators), we propose a “path-selection modulation” source that eliminates the need for active phase modulation altogether. Encoding is achieved by randomly selecting between multiple paths, each path corresponding to one of the desired encoded states. Phase randomization is achieved using gain switching. We characterize this source at a clock rate of 1 GHz.

Dispersive readout is widely used to perform high-fidelity measurement of superconducting qubits. Much work has been focused on the qubit readout fidelity, which depends on the achievable signal-to-noise ratio and the qubit relaxation time. As groups have pushed to increase readout fidelity by increasing readout photon number, dispersive readout has been shown to strongly affect the postmeasurement qubit state. Such effects hinder the effectiveness of quantum error correction, which requires measurements that both have high readout fidelity and are quantum nondemolition (QND). Here, we experimentally investigate non-QND effects in the fluxonium. We map out the state evolution of fluxonium qubits in the presence of resonator photons and observe that these photons induce transitions in the fluxonium both within and outside the qubit subspace. We numerically model our system and find that transitions to higher-excited states and coupling to an external spurious mode are necessary to explain observed non-QND effects.