arXiv subjects

Xi Cao

Publications and source records attributed to Xi Cao.

At least 19 recordsLinked to original sources

arXiv2026

Parametrically induced strong coupling between a superconducting quantum circuit and a solid-state spin ensemble

Efficient quantum state transfer between superconducting circuits and solid-state spins would unlock high-coherence quantum memories for superconducting quantum processors. We demonstrate dynamically controlled strong coupling between a Josephson circuit and a rare-earth spin ensemble. Using a parametric pump, we realize on-demand coupling of several MHz, which will enable faithful state transfer between quantum circuits and spins. Our architecture enables quantum control of spin ensembles, and paves the way for hybrid memories with coherence far beyond those of superconducting circuits alone.

quant-ph↗

arXiv2025

Autonomous stabilization of remote entanglement in a cascaded quantum network

Remote entanglement between widely separated qubits is a fundamental quantum phenomenon and a critical resource for quantum information applications. Generating entanglement between independent qubits separated by arbitrary, potentially large distances requires propagating quantum states, and is typically achieved using pulsed protocols combining distinct steps of local entanglement generation followed by distribution. This necessity raises an intriguing question: Can remote entanglement be stabilized indefinitely, instead of only periodically regenerated and redistributed after decay? Here, we demonstrate that this is indeed possible, reporting autonomous stabilization of entanglement between two separate superconducting-qubit devices. Combining nonreciprocal waveguide coupling and local driving, we experimentally realize a symmetry-based coherent quantum-absorber scheme in a cascaded network. We quantify the degree of entanglement through quantum state tomography, finding that the protocol's entangling power is severely limited by imperfections that break the required symmetry. We show, however, that a modified protocol based on an alternate symmetry is far more robust, enabling us to achieve a concurrence approaching 0.5, a limit set only by local loss in the network. Our results enable on-demand delivery of high-fidelity entanglement in modular quantum processors and networks and pave the way for autonomously protecting distributed quantum information.

quant-ph↗

arXiv2025

The waves-in-space Purcell effect for superconducting qubits

Quantum information processing, especially with quantum error correction, requires both long-lived qubits and fast, quantum non-demolition readout. In superconducting circuits this leads to the requirement to both strongly couple qubits, such as transmons, to readout modes while also protecting them from associated Purcell decay through the readout port. So-called Purcell filters can provide this protection, at the cost of significant increases in circuit components and complexity. However, as we demonstrate in this work, visualizing the qubit fields in space reveals locations where the qubit fields are strong and cavity fields weak; simply placing ports at these locations provides intrinsic Purcell protection. For a $\lambda/2$ readout mode in the `chip-in-tube' geometry, we show both millisecond level Purcell protection and, conversely, greatly enhanced Purcell decay (qubit lifetime of 1~$\mu$s) simply by relocating the readout port. This method of integrating the Purcell protection into the qubit-cavity geometry can be generalized to other 3D implementations, such as post-cavities, as well as planar geometries. For qubit frequencies below the readout mode this effect is quite distinct from the multi-mode Purcell effect, which we demonstrate in a 3D-post geometry where we show both Purcell protection of the qubit while spoiling the quality factor of higher cavity harmonics to protect against dephasing due to stray photons in these modes.

quant-ph↗

arXiv2024

Human-in-the-Loop Generation of Adversarial Texts: A Case Study on Tibetan Script

DNN-based language models excel across various NLP tasks but remain highly vulnerable to textual adversarial attacks. While adversarial text generation is crucial for NLP security, explainability, evaluation, and data augmentation, related work remains overwhelmingly English-centric, leaving the problem of constructing high-quality and sustainable adversarial robustness benchmarks for lower-resourced languages both difficult and understudied. First, method customization for lower-resourced languages is complicated due to linguistic differences and limited resources. Second, automated attacks are prone to generating invalid or ambiguous adversarial texts. Last but not least, language models continuously evolve and may be immune to parts of previously generated adversarial texts. To address these challenges, we introduce HITL-GAT, an interactive system based on a general approach to human-in-the-loop generation of adversarial texts. Additionally, we demonstrate the utility of HITL-GAT through a case study on Tibetan script, employing three customized adversarial text generation methods and establishing its first adversarial robustness benchmark, providing a valuable reference for other lower-resourced languages.

cs.CL↗

arXiv2024

High dynamic-range quantum sensing of magnons and their dynamics using a superconducting qubit

Magnons can endow quantum devices with new functionalities. Assessing their potential requires precise characterization of magnon properties. Here, we use a superconducting qubit to probe magnons in a ferrimagnet over a range of about 2000 excitations. Using qubit control and parametrically induced qubit-magnon interactions we demonstrate few-excitation sensitive detection of magnons and are able to accurately resolve their decay. These results introduce quantum circuits as high-dynamic range probes for magnons and provide an avenue toward sensitive detection of nontrivial magnon dynamics.

quant-ph↗

arXiv2024

TSCheater: Generating High-Quality Tibetan Adversarial Texts via Visual Similarity

Language models based on deep neural networks are vulnerable to textual adversarial attacks. While rich-resource languages like English are receiving focused attention, Tibetan, a cross-border language, is gradually being studied due to its abundant ancient literature and critical language strategy. Currently, there are several Tibetan adversarial text generation methods, but they do not fully consider the textual features of Tibetan script and overestimate the quality of generated adversarial texts. To address this issue, we propose a novel Tibetan adversarial text generation method called TSCheater, which considers the characteristic of Tibetan encoding and the feature that visually similar syllables have similar semantics. This method can also be transferred to other abugidas, such as Devanagari script. We utilize a self-constructed Tibetan syllable visual similarity database called TSVSDB to generate substitution candidates and adopt a greedy algorithm-based scoring mechanism to determine substitution order. After that, we conduct the method on eight victim language models. Experimentally, TSCheater outperforms existing methods in attack effectiveness, perturbation magnitude, semantic similarity, visual similarity, and human acceptance. Finally, we construct the first Tibetan adversarial robustness evaluation benchmark called AdvTS, which is generated by existing methods and proofread by humans.

cs.CL↗

arXiv2024

Multi-Granularity Tibetan Textual Adversarial Attack Method Based on Masked Language Model

In social media, neural network models have been applied to hate speech detection, sentiment analysis, etc., but neural network models are susceptible to adversarial attacks. For instance, in a text classification task, the attacker elaborately introduces perturbations to the original texts that hardly alter the original semantics in order to trick the model into making different predictions. By studying textual adversarial attack methods, the robustness of language models can be evaluated and then improved. Currently, most of the research in this field focuses on English, and there is also a certain amount of research on Chinese. However, there is little research targeting Chinese minority languages. With the rapid development of artificial intelligence technology and the emergence of Chinese minority language models, textual adversarial attacks become a new challenge for the information processing of Chinese minority languages. In response to this situation, we propose a multi-granularity Tibetan textual adversarial attack method based on masked language models called TSTricker. We utilize the masked language models to generate candidate substitution syllables or words, adopt the scoring mechanism to determine the substitution order, and then conduct the attack method on several fine-tuned victim models. The experimental results show that TSTricker reduces the accuracy of the classification models by more than 28.70% and makes the classification models change the predictions of more than 90.60% of the samples, which has an evidently higher attack effect than the baseline method.

cs.CL↗

arXiv2024

Pay Attention to the Robustness of Chinese Minority Language Models! Syllable-level Textual Adversarial Attack on Tibetan Script

The textual adversarial attack refers to an attack method in which the attacker adds imperceptible perturbations to the original texts by elaborate design so that the NLP (natural language processing) model produces false judgments. This method is also used to evaluate the robustness of NLP models. Currently, most of the research in this field focuses on English, and there is also a certain amount of research on Chinese. However, to the best of our knowledge, there is little research targeting Chinese minority languages. Textual adversarial attacks are a new challenge for the information processing of Chinese minority languages. In response to this situation, we propose a Tibetan syllable-level black-box textual adversarial attack called TSAttacker based on syllable cosine distance and scoring mechanism. And then, we conduct TSAttacker on six models generated by fine-tuning two PLMs (pre-trained language models) for three downstream tasks. The experiment results show that TSAttacker is effective and generates high-quality adversarial samples. In addition, the robustness of the involved models still has much room for improvement.

cs.CL↗

arXiv2024

Engineering a multi-level bath for transmons with three-wave mixing and parametric drives

A quantum system with a tunable bath temperature provides an additional degree of freedom for quantum simulators. Such a system can be realized by parametrically modulating the coupling between the system and the bath. Here, by coupling a transmon qubit to a lossy Superconducting Nonlinear Asymmetric Inductive eLement (SNAIL) mode, we experimentally create a tunable bath for the qubit mode. The effective temperature of this bath can be precisely controlled, ranging from negative to positive values. We show that the qubit can be thermalized to equilibrium with different population distributions under different parametric pumping conditions. We further extend our method to the third level of the transmon, demonstrating its potential utility beyond the two-level case. Our results provide a useful tool that can be readily integrated with quantum simulators that would benefit from a nontrivial photon population distribution.

quant-ph↗

arXiv2024

A high-efficiency plug-and-play superconducting qubit network

Modular architectures are a promising approach to scale quantum devices to the point of fault tolerance and utility. Modularity is particularly appealing for superconducting qubits, as monolithically manufactured devices are limited in both system size and quality. Constructing complex quantum systems as networks of interchangeable modules can overcome this challenge through `Lego-like' assembly, reconfiguration, and expansion, in a spirit similar to modern classical computers. First prototypical superconducting quantum device networks have been demonstrated. Interfaces that simultaneously permit interchangeability and high-fidelity operations remain a crucial challenge, however. Here, we demonstrate a high-efficiency interconnect based on a detachable cable between superconducting qubit devices. We overcome the inevitable loss in a detachable connection through a fast pump scheme, enabling inter-module SWAP efficiencies at the 99%-level in less than 100 ns. We use this scheme to generate high-fidelity entanglement and operate a distributed logical dual-rail qubit. At the observed ~1% error rate, operations through the interconnect are at the threshold for fault-tolerance. These results introduce a modular architecture for scaling quantum processors with reconfigurable and expandable networks.

quant-ph↗

arXiv2024

MindSpore Quantum: A User-Friendly, High-Performance, and AI-Compatible Quantum Computing Framework

We introduce MindSpore Quantum, a pioneering hybrid quantum-classical framework with a primary focus on the design and implementation of noisy intermediate-scale quantum (NISQ) algorithms. Leveraging the robust support of MindSpore, an advanced open-source deep learning training/inference framework, MindSpore Quantum exhibits exceptional efficiency in the design and training of variational quantum algorithms on both CPU and GPU platforms, delivering remarkable performance. Furthermore, this framework places a strong emphasis on enhancing the operational efficiency of quantum algorithms when executed on real quantum hardware. This encompasses the development of algorithms for quantum circuit compilation and qubit mapping, crucial components for achieving optimal performance on quantum processors. In addition to the core framework, we introduce QuPack, a meticulously crafted quantum computing acceleration engine. QuPack significantly accelerates the simulation speed of MindSpore Quantum, particularly in variational quantum eigensolver (VQE), quantum approximate optimization algorithm (QAOA), and tensor network simulations, providing astonishing speed. This combination of cutting-edge technologies empowers researchers and practitioners to explore the frontiers of quantum computing with unprecedented efficiency and performance.

quant-ph↗

arXiv2024

Parametrically controlled chiral interface for superconducting quantum devices

Nonreciprocal microwave routing plays a crucial role for measuring quantum circuits, and allows for realizing cascaded quantum systems for generating and stabilizing entanglement between non-interacting qubits. The most commonly used tools for implementing directionality are ferrite-based circulators. These devices are versatile, but suffer from excess loss, a large footprint, and fixed directionality. For utilizing nonreciprocity in scalable quantum circuits it is desirable to develop efficient integration of low-loss and in-situ controllable directional elements. Here, we report the design and experimental realization of a minimal controllable directional interface that can be directly coupled to superconducting qubits. In the presented device, nonreciprocity is realized through a combination of interference and phase-controlled parametric pumping. We have achieved a maximum directionality of around 30\,dB, and the performance of the device is predicted quantitatively from independent calibration measurements. Using the excellent agreement of model and experiment, we predict that the circuit will be useable as a chiral qubit interface with inefficiencies at the one-percent level or below. Our work offers a promising route for realizing high-fidelity signal routing and entanglement generation in all-to-all connected microwave quantum networks, and provides a path for isolator-free qubit readout schemes.

quant-ph↗

arXiv2024

Simulating a Chern Insulator with C = $\pm$2 on Synthetic Floquet Lattice

The synthetic Floquet lattice, generated by multiple strong drives with mutually incommensurate frequencies, provides a powerful platform for the quantum simulation of topological phenomena. In this study, we propose a 4-band tight-binding model of the Chern insulator with a Chern number C = $\pm$2 by coupling two layers of the half-BHZ lattice and subsequently mapping it onto the Floquet lattice to simulate its topological properties. To determine the Chern number of our Floquet-version model, we extend the energy pumping method proposed by Martin et al. [Phys. Rev. X 7, 041008 (2017)] and the topological oscillation method introduced by Boyers et al. [Phys. Rev. Lett. 125, 160505 (2020)], followed by numerical simulations for both methodologies. The simulation results demonstrate the successful extraction of the Chern number using either of these methods, providing an excellent prediction of the phase diagram that closely aligns with the theoretical one derived from the original bilayer half-BHZ model. Finally, we briefly discuss a potential experimental implementation for our model. Our work demonstrates significant potential for simulating complex topological matter using quantum computing platforms, thereby paving the way for constructing a more universal simulator for non-interacting topological quantum states and advancing our understanding of these intriguing phenomena.

quant-ph↗

arXiv2024

Quantum Optimal Control Theory for the Shaping of Flying Qubits

The control of flying qubits carried by itinerant photons is ubiquitous in quantum networks. Beside their logical states, the shape of flying qubits must also be tailored for high-efficiency information transmission. In this paper, we introduce quantum optimal control theory to the shaping of flying qubits. Building on the flying-qubit control model established in our previous work, we design objective functionals for the generation of shaped flying qubits under practical constraints on the emitters and couplers. Numerical simulations employing gradient-descent algorithms demonstrate that the optimized control can effectively mitigate unwanted level and photon leakage caused by these non-idealities. Notably, while coherent control offers limited shaping capacity with a fixed coupler, it can significantly enhance the shaping performance when combined with a tunable coupler that has restricted tunability. The proposed optimal control framework provides a systematic approach to achieving high-quality control of flying qubits using realistic quantum devices.

quant-ph↗

arXiv2024

Loss resilience of driven-dissipative remote entanglement in chiral waveguide quantum electrodynamics

Establishing limits of entanglement in open quantum systems is a problem of fundamental interest, with strong implications for applications in quantum information science. Here, we study limits of entanglement stabilization between remote qubits. We theoretically investigate the loss resilience of driven-dissipative entanglement between remote qubits coupled to a chiral waveguide. We find that by coupling a pair of storage qubits to the two driven qubits, the steady state can be tailored such that the storage qubits show a degree of entanglement that is higher than what can be achieved with only two driven qubits coupled to the waveguide. By reducing the degree of entanglement of the driven qubits, we show that the entanglement between the storage qubits becomes more resilient to waveguide loss. Our analytical and numerical results offer insights into how waveguide loss limits the degree of entanglement in this driven-dissipative system, and offers important guidance for remote entanglement stabilization in the laboratory, for example using superconducting circuits.

quant-ph↗

arXiv2023

Robust Control of Single-Qubit Gates at the Quantum Speed Limit

Fastness and robustness are both critical in the implementation of high-fidelity gates for quantum computation, but in practice, a trade-off has to be made between them. In this paper, we investigate the underlying robust time-optimal control problem so as to make the best balance. Based on the Taylor expansion of the system's unitary propagator, we formulate the design problem as the optimal control of an augmented finite-dimensional system at its quantum speed limit (QSL), where the robustness is graded by the degree of series truncation. The gradient-descent algorithm is then introduced to sequentially seek QSLs corresponding to different orders of robustness. Numerical simulations for single-qubit systems show that the obtained time-optimal control pulses can effectively suppress gate errors (to the prescribed robustness order) caused by qubit frequency and field amplitude uncertainties. These results provide a practical guide for selecting pulse lengths in the pulse-level compilation of quantum circuits.

quant-ph↗

arXiv2023

Fast superconducting qubit control with sub-harmonic drives

Increasing the fidelity of single-qubit gates requires a combination of faster pulses and increased qubit coherence. However, with resonant qubit drive via a capacitively coupled port, these two objectives are mutually contradictory, as higher qubit quality factor requires a weaker coupling, necessitating longer pulses for the same applied power. Increasing drive power, on the other hand, can heat the qubit's environment and degrade coherence. In this work, by using the inherent nonlinearity of the transmon qubit, we circumvent this issue by introducing a new parametric driving scheme to perform single-qubit control. We achieve rapid gate speed by pumping the transmon's native Kerr term at approximately one third of the qubit's resonant frequency. Given that transmons typically operate within a fairly narrow range of anharmonicity, this technique is applicable to all transmons. In both theory and experiment, we show that the Rabi rate of the process is proportional to applied drive amplitude cubed, allowing for rapid gate speed with only modest increases in applied power. In addition, we demonstrate that filtering can be used to protect the qubit's coherence while performing rapid gates, and present theoretical calculations indicating that decay due to multi-photon losses, even in very strongly coupled drive lines, will not limit qubit lifetime. We demonstrate single qubit gates as short as 37.4ns with fidelity as high as 99.91%. We also present calculations indicating that this technique could reduce cryostat heating for fast gates, a vital requirement for large-scale quantum computers.

quant-ph↗

arXiv2022

Experimental Quantum End-to-End Learning on a Superconducting Processor

Machine learning can be substantially powered by a quantum computer owing to its huge Hilbert space and inherent quantum parallelism. In the pursuit of quantum advantages for machine learning with noisy intermediate-scale quantum devices, it was proposed that the learning model can be designed in an end-to-end fashion, i.e., the quantum ansatz is parameterized by directly manipulable control pulses without circuit design and compilation. Such gate-free models are hardware friendly and can fully exploit limited quantum resources. Here, we report the first experimental realization of quantum end-to-end machine learning on a superconducting processor. The trained model can achieve 98% recognition accuracy for two handwritten digits (via two qubits) and 89% for four digits (via three qubits) in the MNIST (Mixed National Institute of Standards and Technology) database. The experimental results exhibit the great potential of quantum end-to-end learning for resolving complex real-world tasks when more qubits are available.

quant-ph↗