Suppressing Measurement Noise in Logical Qubits Through Measurement Scheduling
Xiao-Yue Xu, Chen Ding, Wan-Su Bao · 2025
Read Suppressing Measurement Noise in Logical Qubits Through Measurement Scheduling by Xiao-Yue Xu, Chen Ding, Wan-Su Bao (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2505-07173.txt.
Introduction
Xu, Ding, and Bao cut read noise in proof qubits by task shifts in time. Proof fix is a must for trust-worthy quantum work at scale. Face codes give high fail bars and lean gear in that role. Yet noise in one-shot reads caps how true proof reads can be. Their plan shifts read tasks from shaky qubits to firm nodes. Shallow tied loops do the shift work in the plan.
That shift trades gate flaws for read noise in a fair split. Picks rest on shape weight and flaw marks per node. Time caps from fade and fix needs bind the plan in real runs. A learn tool then steers the shifts to fit those caps. Sims show proof flaw rates fall by up to 34 percent. Gains span code spans from three to 11 in the sims.
Catalog record
| CREATOR | Xiao-Yue Xu, Chen Ding, Wan-Su Bao |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 36 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2505-07173
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Suppressing Measurement Noise in Logical Qubits Through Measurement Scheduling Xiao-Yue Xu, Chen Ding, Wan-Su Bao Abstract Quantum error correction is essential for reliable quantum computation, where surface codes demonstrate high fault-tolerant thresholds and hardware efficiency. However, noise in single-shot measurements limits logical readout fidelity, forming a critical bottleneck for fault-tolerant quantum computation. We propose a dynamic measurement scheduling protocol that suppresses logical readout errors by adaptively redistributing measurement tasks from error-prone qubits to stable nodes. Using shallow entangled circuits, the protocol balances gate errors and measurement noise. This is achieved by dynamically prioritizing resource allocation based on topological criticality and error metrics. When addressing realistic scenarios where temporal constraints are governed by decoherence limits and error-correction requirements, we implement reinforcement learning (RL) to achieve adaptive measurement scheduling. Numerical simulations show that logical error rates can be reduced by up to 34% across code distances for 3 to 11, with enhanced robustness in measurement-noise-dominated systems. Our protocol offers a versatile, hardware-efficient solution for high-fidelity quantum error correction, advancing large-scale quantum computing. I Introduction Quantum error correction (QEC) Cory et al. 1998 ; Knill et al. 2000 ; Chiaverini et al. 2004 ; Reichardt et al. 2024 ; Paetznick et al. 2024 is a critical framework for protecting quantum information from decoherence and operational errors, enabling reliable quantum computation. Among various quantum error-correcting (QEC) codes, the surface code has emerged as a particularly promising candidate, demonstrating superior fault-tolerant thresholds and unique compatibility with 2D lattice architectures requiring only nearest-neighbor qubit interactions Acharya et al. 2024 ; Acharya et al. 2023 ; Fowler et al. 2012 ; Zhao et al. 2022 ; Fowler et al. 2010 ; Horsman et al. 2012 . Its fault tolerance stems from topological protection against local errors, achieving a high error threshold ( \sim 1%) while requiring only nearest-neighbor parity measurements. The code distance d exhibits exponential suppression of logical error rate p_{L} (scaling as p_{L}\propto(p/p_{\text{th}})^{d/2} ) when physical error rate p remains below the threshold p_{\text{th}} . This combination of hardware-efficient geometry, robust error suppression, and low operational overhead makes it a cornerstone for scalable quantum computation. The surface code operates on a lattice with alternating smooth ( Z -type) and rough ( X -type) boundaries. Two types of stabilizers, X -type and Z -type, are measured periodically to detect errors via syndrome extraction, as illustrated in Fig. 1 . Figure 1: A rotated surface code lattice and stabilizer measurement circuits. The two-dimensional array of the rotated surface code which represents a logical qubit with code distance d=3 . Data qubits (blue) store quantum information, while measurement qubits (orange/red) execute stabilizer operations—orange for Z -type stabilizers \hat{Z}^{\otimes 4} and red for X -type stabilizers \hat{X}^{\otimes 4} , with corresponding colored blocks highlighting qubit subsets involved in each stabilizer circuit. Boundary conditions are implemented via blue blocks enclosing data qubits at the code’s smooth boundary, where direct Z -basis physical measurements resolve the logical qubit state. The surface code’s fault tolerance hinges on two complementary pillars: robust error correction based on stabilizer measurement and precise logical state readout. In planar code architectures Horsman et al. 2012 , the logical Z are evaluated by parity through horizontal data-qubit paths traversing smooth boundaries, requiring high-fidelity readout of boundary-adjacent data qubits. When considering the measurement error during these two process, stabilizer measurements effectively suppress single-shot readout errors by leveraging temporal redundancy, typically through 5–15 rounds of syndrome averaging. This redundancy allows the quantum state to be preserved with high fidelity over extended periods. However, the final readout for logical state measurement is irreversible and collapses the quantum state, permanently forfeiting the opportunity for multi-round error mitigation. To harness the surface code’s full potential, noise-resilient logical readout protocols must bridge this gap, reconciling the destructive nature of direct measurement with the stringent fidelity demands of scalable quantum computation. Figure 2: The illustration of measurement scheduling protocol in a 2D chain quantum topological structure. The left panel depicts a 2D chain-type quantum topological structure, where blue marks indicate qubits requiring measurement. Arrows outline the scheduling process across three sequential time steps (ticks). Gray circles represent qubits idle during their respective tick, while blue circles denote measurement targets in the current tick. The orange box demonstrates the measurement transfer modality: a shallow quantum circuit reroutes measurement tasks (blue) to adjacent qubits (red markers within the box). After three ticks, all qubits are measured. In this work, we propose a measurement scheduling (MS) protocol that directly targets the suppression of logical readout errors by redistributing measurement tasks across subsystems through measurement transfer modalities, as illustrated in Fig. 2 . This approach is motivated by the hardware asymmetry in modern quantum architectures. While two-qubit gate fidelities are compatible with fault-tolerant thresholds Foxen et al. 2020 ; Xu et al. 2020 ; Negîrneac et al. 2021 ; Sung et al. 2021 ; Walter et al. 2017 , single-shot readout errors remain at significantly higher levels with up to an order of magnitude discrepancy, constituting a critical performance limitation Bultink et al. 2020 ; Linke et al. 2017 ; Chen et al. 2023 . Moreover, unlike gate errors that can be mitigated through techniques such as dynamical decoupling, effective suppression of readout-induced errors remains particularly challenging Acharya et al. 2023 ; Pino et al. 2021 ; Tantivasadakarn et al. 2023 ; Smith et al. 2023 . By reframing the quantum readout task as a resource allocation problem, our MS protocol reroutes measurement tasks from error-prone qubits to stable neighboring nodes. This protocol resolves qubit allocation conflicts through dynamic prioritization, guided by real-time error metrics and topological criticality. We first introduce an easily implementable local MS scheme that does not account for the waiting times linked to qubit allocation conflicts. When considering temporal constraints such as error correction deadlines and qubit decoherence limits, the protocol integrates reinforcement learning (RL) Mnih et al. 2015 ; Silver et al. 2016 ; Silver et al. 2017 ; Sutton et al. 1999 ; Schulman et al. 2015 ; Schulman et al. 2017 to iteratively refine the allocation. This temporal-aware protocol achieves adaptive measurement scheduling while preserving conflict resolution integrity, inherently balancing computational tractability and scheduling precision. We perform extensive numerical experiments exploring the performance landscape of the proposed MS framework, validating both the practical efficacy and theoretical advantages of our approach. II Measurement Transfer Modalities We first introduce measurement transfer modalities as the schedulable units of our MS protocol, designed to work seamlessly with fault-tolerant quantum architectures. These modalities shift measurement tasks from error-prone qubits to high-fidelity subsystems using shallow entangled circuits, reducing overall readout errors on quantum processors with higher gate fidelity. II.1 General measurement transfer modalities Within our MS framework, these modalities are divided into two basic types of logical readout operations: multi-qubit parity measurement and single-qubit measurement. The parity measurement of n qubits, which assesses whether the number of the number of |1\rangle states in their measurement outcomes is even or odd, can be encoded in the expectation value of the \hat{Z}^{\otimes n} operator. Here, eigenvalues +1 and -1 of \hat{Z}^{\otimes n} correspond to even and odd parity, respectively. We define the single-shot measurement operator as \mathcal{M}[\hat{O}](\cdot)\in{0,1} , representing the single-qubit outcome of observable O on its computational basis. Then