A new design for quantum gates fights noise in circuits up to eight qubits

Researchers from the University of Chicago, Johns Hopkins University Applied Physics Laboratory, and Northwestern University have designed new quantum gate sets that suppress noise in circuits of up to eight qubits. The work, tested on the IBM Quantum Platform, demonstrates a scalable approach to mitigating quantum crosstalk, a key obstacle to building larger, more reliable quantum computers. The team’s noise-informed gates enhanced median algorithmic performance by a factor of four in simulations of a four-qubit system, and numerical evidence suggests these optimized sets could halve two-qubit gate times by enabling larger qubit-qubit coupling strengths.

Scalable Gate Sets Mitigate Qubit Crosstalk

Optimized gate sets allow for increased qubit-qubit coupling strengths, potentially halving two-qubit gate times, a result demonstrating that coordinated hardware and software design can unlock new quantum computing architectures. The work centers on crosstalk-robust gate sets, designed using a quantum optimal control protocol that globally suppresses errors arising from concurrent gate operations. This approach uses the concept of coloring gates, assigning each a unique identifier to ensure orthogonality and minimize interference, a strategy particularly relevant given the heavy-hex layout of qubits on IBM devices.

These processors are susceptible to coherent ZZ crosstalk due to always-on static coupling between neighboring qubits, a challenge the new gate sets directly address. The design of these crosstalk-robust gate sets (CRGS) relies on a principle similar to graph coloring, where each gate represents a vertex and potential crosstalk errors form the edges.

By carefully selecting gate combinations, with two colors proving sufficient for IBM devices, the system minimizes the impact of unwanted interactions. The research demonstrates a scalable solution, moving beyond previous error suppression methods by successfully implementing the gate sets on circuits of up to eight qubits on the IBM Quantum Platform.

Optimal Control Framework Using Piccolo.jl

This improvement stems from a novel framework for constructing crosstalk-robust gate sets, or CRGS, which uses the open-source Piccolo.jl software to address the challenges of quantum noise at scale. Unlike previous crosstalk suppression methods, this approach achieves linear scaling by capturing the local structure of interference within the optimization process.

The framework introduces a coordinated quantum optimal control problem, allowing for the creation of collections of single-qubit gates that actively suppress pairwise crosstalk when executed simultaneously on a quantum device. This is achieved by solving for gates, treated as vertices, subject to crosstalk errors, which function as edges in the system. The design mirrors the principles of graph coloring, ensuring orthogonal gate operations.

Researchers found that identical X gates suffer from ZZ crosstalk when played in parallel, but distinct, optimized X gates can operate without interference, a critical distinction for complex calculations. Piccolo.jl’s capabilities extend beyond simply finding robust gate sets. The software package facilitates the integration of these sets into existing calibration and compilation workflows. The framework uses piecewise-constant acceleration as a control input, guaranteeing continuously differentiable pulse shapes without additional restrictions.

The software’s ability to introduce fidelity constraints and set bounds on control amplitudes further refines the optimization process. While IBM cloud-based calibration initially limited the demonstration of single qubit gate performance, the team does not view this as a fundamental limitation, citing successful calibration on APL hardware. Incorporation of improved crosstalk metrics into the optimal control framework, they suggest, can further enhance crosstalk suppression.

Heavy-Hex Layout Testing on IBM Brisbane

IBM’s 127-qubit Eagle processor, using a heavy-hex layout, served as a testing ground for the newly developed crosstalk-robust gate sets, demonstrating their functionality on a commercially available quantum computing architecture. Researchers generated and executed the sets on IBM Brisbane, alongside a similar in-house device featuring a ring layout and six qubits, allowing for comparative analysis of performance across different qubit arrangements.

This metric provides a baseline for evaluating the effectiveness of the crosstalk suppression methods. This approach allows for the suppression of all crosstalk for a given qubit layout if the number of colors is sufficient for a vertex coloring of that layout, a concept central to the optimization process.

Crosstalk Reduction in Randomized Eight-Qubit Circuits

The suppression of quantum crosstalk now extends to multi-qubit circuits thanks to newly designed gate sets, demonstrated on devices from the IBM Quantum Platform. These sets were tested using randomized parallel single-qubit circuits of up to eight qubits. This testing employed existing error suppression protocols to evaluate the gate sets’ efficacy, providing a benchmark for performance improvements. A Hamiltonian simulation of a four-qubit transverse field Ising model revealed a substantial performance gain with these noise-informed gates.

The optimized design achieved a decay parameter 30% of that found for Gaussian pulses in the full eight-qubit demonstration, indicating a marked increase in stability. Increasing qubit-qubit coupling strength, enabled by the optimized gate sets, offers a pathway to faster two-qubit gates.

While complete assessment via standard randomized benchmarking was limited by hardware access, initial results suggest the potential for reconciliation with coherence limits given more extensive testing. The dominance of two-qubit gates in circuit runtime and their impact on coherence limits emphasise the importance of this speedup, as faster gates directly translate to improved fidelity and more complex computations.

Four-Qubit TFIM Simulation Performance Enhancement

Using a Hamiltonian simulation of a four-qubit transverse field Ising model, noise-informed gates enhance median algorithmic performance by a factor of four over baseline Gaussian gates using the same calibration procedures. Simulations using Qiskit Dynamics suggest that these gates maintain performance even with increased coupling, opening opportunities for hardware-software co-design.

The work demonstrates a trade-off between interaction strength and ZZ errors can be overcome, allowing for faster gate operations without sacrificing fidelity. “We also leverage purpose-built simulation tools based on Qiskit Dynamics to explore stronger coupling regimes and show that our method supports higher gate speeds without performance loss,” the researchers state, emphasizing the potential for faster quantum computations.

ZZ Error Suppression and Gate Fidelity Improvement

The optimization process focuses on the connectivity graph of a quantum processor, formulating crosstalk suppression as a scalable optimal control problem. “We see that the ceiling of attainable fidelity is lifted by increasing the coupling strength,” the researchers report, indicating a path toward more efficient quantum computations. This result highlights the potential of these gates to enhance the performance of complex quantum algorithms.

Hardware-Software Co-Design for Faster Gate Speeds

This advancement stems from a co-design strategy that balances control software with hardware parameters, specifically the coupling strength parameter, J. The work demonstrates that crosstalk-robust gate sets mitigate susceptibility to parallel operations, allowing for increased J values without performance loss. Two-qubit gates, which dominate circuit runtime and define coherence limits, benefit most from this approach. The study highlights that increasing coupling strength, and thus gate speed, typically exacerbates crosstalk, but the crosstalk-robust gate sets effectively counter this effect. The researchers state.

Robust Gates Integrated into Existing Workflows

A framework achieving linear scaling in crosstalk suppression distinguishes this approach from previous methods. These gate sets are generated using the open-source Piccolo.jl framework and integrate directly into existing calibration and compilation procedures, offering a path toward practical implementation. The optimization process, formulated as a constrained problem within Piccolo.jl, balances error reduction with gate complexity, represented by the equation arg minZ R(Z;H_(error)) + 1/2||Z||^2R, subject to fidelity and control constraints.

This allows for the creation of gates specifically tailored to mitigate the effects of qubit frequency misalignment, a common source of error, by setting the error Hamiltonian, Herror, to Δ σZ for a given detuning value, Δ. Analytical solutions demonstrate that pairs of X gates using 2π-offset square waves can eliminate ZZ crosstalk susceptibility when executed in parallel, a contrast to identical X gates which experience parasitic effects. While IBM’s cloud-based calibration procedures presented limitations, the researchers do not view these as fundamental, citing successful calibration on hardware from the Applied Physics Laboratory.

👉 More information
🗞 Quantum noise suppression at scale with crosstalk-robust gate sets
✍️ Andy J. Goldschmidt, Emilio Peláez Cisneros, Ryan Sitler, Kevin Olsson, Kaitlin N. Smith and Gregory Quiroz
🧠 DOI: http://link.aps.org/doi/10.1103/psqn-dt4d

Stay current

See today’s quantum computing news on Quantum Zeitgeist for the latest breakthroughs in qubits, hardware, algorithms, and industry deals.

Avatar of Ivy Delaney

Ivy Delaney

Ivy Delaney has been working with neural networks and machine learning since the mid-nineties, back when a couple of hidden layers and a long afternoon of training counted as ambitious. She has watched the field go from academic curiosity to the thing quietly running underneath everything, and she brings that long view to quantum computing. For Quantum Zeitgeist she covers the ground where the two fields meet. That means quantum machine learning and the variational algorithms it leans on, and it also means the less glamorous but more interesting story of classical machine learning already doing real work inside quantum machines, decoding error-correcting codes, calibrating noisy hardware and learning the error models that simulators depend on. She writes about the hardware those algorithms have to run on too, and about the post-quantum cryptography scramble that the same hardware has set off. Her stories typically start with the paper, whether that is peer-reviewed work, conference proceedings or an arXiv preprint, with the source linked so you can hold a claim up against the research it came from. She is unimpressed by benchmarks that will not say what they beat, and by demonstrations that only work in the press release.

Latest Posts by Ivy Delaney: