Fast and Robust Remote Two-Qubit Gates on Distributed Qubits

Y. Li, Xi Zhang, Weixin Zhang, R. P. Guo, Yu Zhang, Xinsheng Tan, Yang YuarXiv 2025 · DOI · arXiv

TL;DR

Researchers have developed a method to perform ultra-fast (30 ns) quantum gates between superconducting qubits located on separate chips connected by a coaxial cable. By combining the inherent robustness of geometric phases with deep-learning optimization (the Adam optimizer) to shape control pulses, they achieved gate fidelities of over 99%, paving the way for modular, scalable quantum computers.

What Problem Does This Paper Address?

Superconducting quantum processors are currently limited by their "flat" architecture. Most designs rely on nearest-neighbor coupling, where qubits can only talk to their immediate neighbors on a single chip. This leads to "frequency crowding" (where signals interfere with each other) and massive overhead for error correction.

To scale up, we need distributed quantum computing—connecting multiple small chips together. However, connecting chips via long cables introduces a major headache: "population leakage." Because a long cable has many resonant modes (harmonics) spaced closely together, a quantum signal intended for a specific operation can easily "leak" into the wrong cable mode, ruining the computation. Existing remote gates are often too slow or too fragile to be practical for fault-tolerant systems.

Methods

The team utilized a 15cm aluminum coaxial cable to link two Transmon qubits. Their approach stands out through two main innovations:

  1. Holonomic (Geometric) Gates: Instead of relying on the precise timing of a pulse (dynamic gates), they used geometric phases. In this scheme, the quantum state follows a specific "loop" or trajectory in its state space. The resulting change (the gate) depends on the area of that loop rather than the exact time spent, making it naturally robust against certain types of noise.
  2. Deep Learning Optimization (Adam): To solve the leakage problem, the researchers treated the pulse-shaping task as an optimization problem. They used the Adam optimizer—a staple in training neural networks—to design specific control waveforms. These optimized pulses ensure the quantum state moves between qubits without getting "stuck" or lost in the various modes of the connecting cable.

The hardware setup involved parametric modulation, where flux pulses were used to tune the qubit frequencies and "activate" the coupling to the cable modes at specific times.

Datasets

As an experimental physics paper, this study does not use traditional machine learning datasets. Instead, the "data" consists of:

  • Experimental Measurements: Population dynamics and state tomography of two Transmon qubits connected via a 15cm Al coaxial cable.
  • Numerical Simulations: The researchers simulated cable lengths ranging from 12cm to 172cm (corresponding to Free Spectral Ranges of 500MHz down to 35MHz) to prove the scalability of their waveform optimization technique.

Key Insights and Ideas

  • Cables as Auxiliary Levels: Rather than viewing the cable as a simple wire, the researchers treated the cable's internal modes as auxiliary energy levels to facilitate the holonomic gate.
  • The "Adam" Advantage: Standard cosine or Gaussian pulses fail as cables get longer because the cable modes get closer together (smaller FSR). The Adam-optimized waveforms suppressed leakage by an order of magnitude compared to standard shapes, even when the cable was simulated at lengths of 1.5 meters.
  • Speed vs. Fidelity: They achieved a gate time of ~30 ns. In the world of remote gates, this is exceptionally fast, minimizing the time the system is vulnerable to decoherence (energy loss).

Key Results

  • High Fidelity: The experimental SWAP gate achieved a fidelity of 99.91% (after accounting for decoherence).
  • Low Error Rates: After excluding the effects of natural energy relaxation, the intrinsic gate error for SWAP was 1.16% and for $\sqrt{\text{SWAP}}$ was 0.91%.
  • Robustness: When subjected to frequency detuning (a common experimental error), the holonomic gate showed significantly higher stability than traditional "dynamic" gates, maintaining lower population loss under stress.
  • Distance Scaling: Simulations demonstrated that this protocol remains effective for "refrigerator-to-refrigerator" distances (over 1 meter), which is critical for building large-scale quantum networks.

Limitations and Future Work

  • Two-Level Systems (TLS): The researchers noted that they couldn't push for even faster gates experimentally because of interference from "unwanted two-level systems"—microscopic defects in the hardware that absorb energy.
  • Residual Leakage: While the Adam optimizer significantly reduced leakage, it remains the dominant source of error after decoherence. Further refinement of the pulse-shaping algorithms or better cable materials could improve this.
  • Scaling Up: The next step is to move from a two-qubit link to a multi-node network, testing how these geometric gates perform when multiple chips are interconnected simultaneously.

Summary updated

Abstract

Distributed quantum computing offers a potential solution to the complexity of superconducting chip hardware layouts and error correction algorithms. High-quality gates between distributed chips enable the simplification of existing error correction algorithms. This article proposes and demonstrates a remote quantum geometric gate scheme via parametric modulation. Our scheme inherits the intrinsic robustness of geometric phases. Meanwhile, by employing gradient-based optimization algorithms(Adaptive Moment Estimation) from deep learning, we design control waveforms that significantly suppress population leakage. We experimentally realize the rapid remote SWAP and $\sqrt{\text{SWAP}}$ gates with high fidelity, completing operation in about 30 ns. The gate error of SWAP ($\sqrt{\text{SWAP}}$) is 1.16\% (0.91\%) after excluding the effect of energy relaxation. The simulation demonstrate that this scheme can be implemented in the distributed chips connected by cables extending several meters. Our results highlight the effectiveness of the proposed protocol in enabling modular quantum processors, offering a promising path toward the realization of fault-tolerant quantum computation.
BibTeX
@misc{li2025fast,
  author = {Y. Li and Xi Zhang and Weixin Zhang and R. P. Guo and Yu Zhang and Xinsheng Tan and Yang Yu},
  title = {{Fast and Robust Remote Two-Qubit Gates on Distributed Qubits}},
  year = {2025},
  eprint = {2511.01418},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2511.01418},
  howpublished = {arXiv},
}
Figure