pith. sign in

arxiv: 2501.15516 · v1 · submitted 2025-01-26 · 🪐 quant-ph

Imperfection analyses for random-telegraph-noise mitigation using spectator qubits

Y. Liu , A. Chantasri , H. Song , H. M. Wiseman This is my paper

Pith reviewed 2026-05-23 04:36 UTC · model grok-4.3

classification 🪐 quant-ph
keywords spectator qubitsrandom telegraph noisedecoherence mitigationBayesian estimationadaptive measurementquantum controlimperfection analysis
0
0 comments X

The pith

The adaptive spectator-qubit protocol for random-telegraph noise keeps its large decoherence suppression under realistic imperfections up to derived bounds.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper examines how imperfections affect a method using spectator qubits to mitigate noise on data qubits. It shows that the suppression of decoherence, which scales quadratically with sensitivity, holds approximately as in ideal cases within certain limits. This is done by generalizing Bayesian estimation and map formalisms to include parameter uncertainties, readout delays, and extra decoherence. A sympathetic reader would care because such protocols need to work in imperfect hardware to be useful for quantum devices.

Core claim

A protocol with adaptive measurement on the SQs and a Bayesian estimation-based control can suppress the data qubits' decoherence rate by a large factor with quadratic scaling in the SQ sensitivity, and this suppression remains approximately the same as under ideal conditions up to derived imperfection bounds on parameter uncertainties, readout efficiency and delay, and additional SQ decoherence.

What carries the argument

The map-based formalism generalized to non-ideal scenarios, combined with time-domain analytical Bayesian estimation for deriving control.

If this is right

  • The decoherence suppression remains approximately the same as under ideal conditions up to the derived imperfection bounds.
  • The quadratic scaling of suppression with SQ sensitivity persists under the analyzed non-ideal conditions.
  • Analytical methods yield explicit bounds on acceptable readout efficiency, delay, and additional decoherence.
  • The generalized map formalism allows performance prediction without full numerical simulation for these imperfections.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Experimental setups could prioritize meeting the derived bounds on readout delay to retain most of the ideal suppression.
  • The approach might be tested first on systems where parameter uncertainties can be calibrated to within the paper's bounds.
  • Similar imperfection analyses could be applied to extend the protocol to other noise spectra.

Load-bearing premise

The random-telegraph noise model plus the modeled imperfections fully capture the system dynamics, with no significant unmodeled effects that would invalidate the Bayesian estimation.

What would settle it

An experiment showing the observed suppression factor deviates significantly from the predicted quadratic scaling when imperfection levels remain below the derived bounds.

Figures

Figures reproduced from arXiv: 2501.15516 by A. Chantasri, H. M. Wiseman, H. Song, Y. Liu.

Figure 1
Figure 1. Figure 1: Schematic pictures to illustrate a Before-time (tB) and a time between a pair of NNRs (wτ ), which will be used in analytical calculations of coherence in Section 2.5 and Section 4. In the first row, we show a realization of the RTP, z(t), with one flip from +1 to −1 (orange line). The second row shows possible SQ’s measurement results for a scheme with no imperfection (black dots), where one NNR is observ… view at source ↗
Figure 2
Figure 2. Figure 2: Unnormalized probability functions, ℘(YN1 ∣tB, Tm) for m = 0, 1, 2 in Eq. (45), are plotted as a function of tB, where tB (Before-time) is the time between the actual flip and the detected flip of the RTP. Here τ , the measurement waiting time, is different for the three different curves, according to τ = Θ/K. The ticks of the horizontal axis are chosen to be multiples of π/(2K), which is the value of the … view at source ↗
Figure 3
Figure 3. Figure 3: Schematic diagrams showing the SQ’s measurement and decoherence mitigation with different types of imperfections we consider in this work. The imperfections we consider in this paper are labelled in the dark gray box with text explanations below. that if the dead time is longer than the optimal measurement time, τdd ≥ τ , then it is better to wait and measure the SQ at a suboptimal time instead of measurin… view at source ↗
Figure 4
Figure 4. Figure 4: The scaled decoherence rate of DQ are plotted as a function of measurement angle uncertainty, showing both analytical results (Eq. (65), solid green) and numerical results (red points). The numerical results are generated from summing over all possible paths (SOP) using the mapping matrices F y s with Θ + δΘ. To obtain the numerical data points, we generated all the possible Y strings according to their po… view at source ↗
Figure 5
Figure 5. Figure 5: The decoherence rate of DQ is plotted as a function of the uncertainty δκ, showing both analytical and numerical results. The analytical result (solid green) is from (70) and the numerical result (red points) was generated from summing over all possible paths (SOP), using the mapping matrices F y s with κ + δκ. The decoherence rate is extracted in the similar way as in [PITH_FULL_IMAGE:figures/full_fig_p0… view at source ↗
Figure 6
Figure 6. Figure 6: Schematic diagrams showing the SQ’s probing (τ ) and reset (τsr) times. After the SQ probes the noise for time τ and is measured at tn ∈ {t1, t2, t3, ...}, the SQ will be reset to its initial state. However, if the reset process is not instantaneous and takes some time, τsr, then the SQ will not be functioning to probe the RTP noise during that time, causing the degradation in decoherence of the DQ. 6. Imp… view at source ↗
Figure 7
Figure 7. Figure 7: Plots of decoherence rate as a function of θsr. The solid green line shows the analytical result in (76), and the red square points are degenerated by SOP. We use κ = 0.2, K = 100, γ↑ = γ↓ = 1,Θ ≈ 1.50055. SQ does not sense noise during this reset time, but the DQ still evolves under the RTP, a coherence vector An at time tn for this case should be updated using a combination of H(τsr, κ) and F(µn = {θn, τ… view at source ↗
Figure 8
Figure 8. Figure 8: Systematic depict of detector’s dead time. The optimal waiting time between two measurements is τ . However, when the detector is not ready to measure the SQ in τ time and can only be available in τdd > τ after each measurement, the best time to conduct next measurement is not immediately after τdd but τ ′ . compare our numerical results with the scaled decoherence rate for the perfect (no imperfection) ad… view at source ↗
Figure 9
Figure 9. Figure 9: Scaled decoherence rate with different waiting time τ ′ for both analytical and numerical results. We use the parameters κ = 0.2, K = 100, γ↑ = γ↓ = 1. The numerical results are shown in red points which are generated using SOP, and the analytical result is simply plotting the ideal HΘ for Θ = Kτ ′ . We do not expect these two results to match perfectly since they are representing different situations, and… view at source ↗
Figure 10
Figure 10. Figure 10: The decoherence rates change with measurement error rates ϵ. In the vertical axis, Γ¯0 is the decoherence rate with ϵ = 0, which also indicates the result of the perfect case. To obtain the decoherence rate, we first use the SOP in Section 3.1 to calculate the expected decoherence with time. Then the decoherence rate is extracted from the slopes by throwing away the first 10 measurement steps (of 18 measu… view at source ↗
Figure 11
Figure 11. Figure 11: The two possible cases for a fixed RTP noise. In both left and right figures, we show the readout strings of yn for the perfect case, the cases with measurement angle uncertainty δΘ and measurement error ϵ, and their corresponding estimated phases and decoherence with time. The parameters we use are κ = 0.2, K = 100, γ↓ = γ↑ = 1,Θ = Θ⋆ ≈ 1.50055, τ = Θ⋆ /K, ϵ = 0.01 and δΘ = 0.2, which gives ϵ ≈ sin2 (δΘ/… view at source ↗
read the original abstract

Spectator qubits (SQs) for random-telegraph noise mitigation have been proposed by Song et al., Phys. Rev. A, 107, L030601 (2023), where an SQ operates as a noise probe to estimate optimal noise-correction control on the hard-to-access data qubits. It was shown that a protocol with adaptive measurement on the SQs and a Bayesian estimation-based control can suppress the data qubits' decoherence rate by a large factor with quadratic scaling in the SQ sensitivity. However, the protocol's practicality in real-world scenarios remained in question, due to various sources of imperfection that could affect the performance. We therefore analyze here the proposed adaptive protocol under non-ideal conditions, including parameter uncertainties in the system, efficiency and time delay in readout and reset processes of the SQs, and additional decoherence on the SQs. We also explore analytical methods of Bayesian estimation in the time domain and generalize the map-based formalism to non-ideal scenarios. This allows us to derive imperfection bounds at which the decoherence suppression remains approximately the same as under ideal conditions.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The manuscript extends the adaptive spectator-qubit protocol of Song et al. (PRA 2023) for random-telegraph noise mitigation to non-ideal conditions. It generalizes the map-based formalism to include parameter uncertainties, finite readout efficiency and delay, reset imperfections, and additional SQ decoherence, derives analytical bounds on these imperfections, and shows that the quadratic-in-sensitivity suppression of data-qubit decoherence remains approximately intact inside those bounds.

Significance. If the derived bounds are rigorous, the work supplies concrete, experimentally usable tolerances that directly address the practicality gap left by the ideal-case analysis. The explicit time-domain Bayesian estimation and the map generalization constitute a clear technical advance over purely numerical studies.

major comments (2)
  1. [§4] §4 (map generalization): the perturbative expansion around the ideal map is used to obtain the imperfection bounds, yet the manuscript does not supply an explicit remainder estimate showing that the neglected O(ε²) terms remain smaller than the ideal η² suppression throughout the claimed regime. This is load-bearing for the central claim that suppression “remains approximately the same.”
  2. [Eq. (bound expression)] Eq. (bound expression) and surrounding text: the bound derivation assumes that the Bayesian estimator’s error scales identically under the imperfect map; no separate propagation-of-uncertainty calculation is provided to confirm that readout delay and efficiency errors do not introduce an additional linear-in-η term that would cancel the quadratic advantage.
minor comments (2)
  1. Notation for the imperfect map M_ε is introduced without an explicit comparison table to the ideal map M_0; adding one would improve readability.
  2. Figure 3 caption states “suppression factor” but the y-axis label is “decoherence rate ratio”; consistent terminology would help.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading and constructive feedback. We respond point-by-point to the major comments below.

read point-by-point responses

  1. Referee: [§4] §4 (map generalization): the perturbative expansion around the ideal map is used to obtain the imperfection bounds, yet the manuscript does not supply an explicit remainder estimate showing that the neglected O(ε²) terms remain smaller than the ideal η² suppression throughout the claimed regime. This is load-bearing for the central claim that suppression “remains approximately the same.”

    Authors: We agree that an explicit remainder bound would make the perturbative argument fully rigorous. In the revised manuscript we add a short derivation (new paragraph in §4) using the Lagrange form of the remainder for the map expansion. Under the stated regime ε < η/5 the O(ε²) contribution is shown to be at most 20 % of the leading η² term, which is sufficient to keep the suppression “approximately the same” within the tolerance already quoted in the abstract. revision: yes

  2. Referee: [Eq. (bound expression)] Eq. (bound expression) and surrounding text: the bound derivation assumes that the Bayesian estimator’s error scales identically under the imperfect map; no separate propagation-of-uncertainty calculation is provided to confirm that readout delay and efficiency errors do not introduce an additional linear-in-η term that would cancel the quadratic advantage.

    Authors: The map generalization already folds readout delay and finite efficiency into the effective transition matrix that enters the Bayesian update; the estimator variance is therefore computed under the imperfect map by construction. Nevertheless, to make the scaling explicit we have added a one-page appendix that performs a first-order propagation of the readout imperfections through the estimator. The calculation confirms that the extra terms remain O(ε) and do not generate a linear-in-η correction capable of canceling the quadratic advantage inside the derived bounds. revision: yes

Circularity Check

0 steps flagged

No circularity: analytical extension of prior map formalism to imperfections

full rationale

The paper cites the 2023 Song et al. protocol as the ideal-case starting point and performs an independent analysis by generalizing the map-based formalism to include readout delay/efficiency, parameter uncertainty, and extra SQ decoherence. Derivation of the imperfection bounds proceeds via this extension and time-domain Bayesian estimation; no step reduces a claimed prediction or bound to a fitted input, self-definition, or load-bearing self-citation chain. The central result (suppression remains approximately unchanged up to derived bounds) is obtained from the non-ideal model equations rather than by construction from the ideal inputs.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

Abstract provides limited information; no free parameters, invented entities, or detailed axioms are described.

axioms (1)
  • domain assumption Random telegraph noise accurately models the dominant decoherence source on data qubits.
    Central modeling choice for the mitigation protocol and its analysis.

pith-pipeline@v0.9.0 · 5731 in / 1097 out tokens · 56978 ms · 2026-05-23T04:36:34.572008+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

27 extracted references · 27 canonical work pages

  1. [1]

    Gupta R S, Edmunds C L, Milne A R, Hempel C and Biercuk M J 2020 npj Quantum Information 6 53

    work page 2020

  2. [2]

    Majumder S, Andreta de Castro L and Brown K R 2020 npj Quantum Information 6 19

    work page 2020

  3. [3]

    Song H, Chantasri A, Tonekaboni B and Wiseman H M 2023 Physical Review A 107 L030601

    work page 2023

  4. [4]

    Tonekaboni B, Chantasri A, Song H, Liu Y and Wiseman H M 2023 Physical Review A 107 032401

    work page 2023

  5. [5]

    Singh K, Bradley C, Anand S, Ramesh V, White R and Bernien H 2023 Science 380 1265–1269

    work page 2023

  6. [6]

    Youssry A, Paz-Silva G A and Ferrie C 2023 New Journal of Physics 25 073004

    work page 2023

  7. [7]

    Lingenfelter A and Clerk A A 2023 npj Quantum Information 9 81

    work page 2023

  8. [8]

    Itakura T and Tokura Y 2003 Phys. Rev. B 67(19) 195320 URL https://link.aps.org/doi/ 10.1103/PhysRevB.67.195320

    work page doi:10.1103/physrevb.67.195320 2003

  9. [9]

    Galperin Y M, Altshuler B L, Bergli J and Shantsev D V 2006 Phys. Rev. Lett. 96(9) 097009 URL https://link.aps.org/doi/10.1103/PhysRevLett.96.097009

    work page doi:10.1103/physrevlett.96.097009 2006

  10. [10]

    Culcer D, Hu X and Das Sarma S 2009 Applied Physics Letters 95

    work page 2009

  11. [11]

    Bergli J, Galperin Y M and Altshuler B L 2009 New Journal of Physics 11 025002 URL https://doi.org/10.1088/1367-2630/11/2/025002

    work page doi:10.1088/1367-2630/11/2/025002 2009

  12. [12]

    Viola L, Knill E and Lloyd S 1999 Physical Review Letters 82 2417

    work page 1999

  13. [13]

    Viola L and Knill E 2003 Physical Review Letters 90 037901

    work page 2003

  14. [14]

    Biercuk M, Doherty A and Uys H 2011 Journal of Physics B: Atomic, Molecular and Optical Physics 44 154002

    work page 2011

  15. [15]

    Ng H K, Lidar D A and Preskill J 2011 Physical Review A 84 012305

    work page 2011

  16. [16]

    Souza A M, Alvarez G A and Suter D 2011 Physical Review Letters 106 240501 Imperfection analyses for random-telegraph-noise mitigation using spectator qubits 37

    work page 2011

  17. [17]

    2012 Physical Review Letters 108 086802

    Medford J, Barthel C, Marcus C, Hanson M, Gossard A et al. 2012 Physical Review Letters 108 086802

    work page 2012

  18. [18]

    Paz-Silva G A and Lidar D 2013 Scientific Reports 3 1530

    work page 2013

  19. [19]

    Zhang J, Souza A M, Brandao F D and Suter D 2014 Physical Review Letters 112 050502

    work page 2014

  20. [20]

    Shor P W 1995 Phys. Rev. A 52(4) R2493–R2496 URL https://link.aps.org/doi/10.1103/ PhysRevA.52.R2493

    work page 1995

  21. [21]

    Steane A M 1996 Phys. Rev. Lett. 77(5) 793–797 URL https://link.aps.org/doi/10.1103/ PhysRevLett.77.793

    work page 1996

  22. [22]

    Terhal B M 2015 Rev. Mod. Phys. 87(2) 307–346 URL https://link.aps.org/doi/10.1103/ RevModPhys.87.307

    work page 2015

  23. [23]

    Paz-Silva G A, Norris L M and Viola L 2017 Physical Review A 95 022121

    work page 2017

  24. [24]

    2020 PRX Quantum 1 010305

    von L¨ upke U, Beaudoin F, Norris L M, Sung Y, Winik R, Qiu J Y, Kjaergaard M, Kim D, Yoder J, Gustavsson S et al. 2020 PRX Quantum 1 010305

    work page 2020

  25. [25]

    Chalermpusitarak T, Tonekaboni B, Wang Y, Norris L M, Viola L and Paz-Silva G A 2021 PRX Quantum 2 030315

    work page 2021

  26. [26]

    2010 Nature 467 687–691

    Morello A, Pla J J, Zwanenburg F A, Chan K W, Tan K Y, Huebl H, M¨ ott¨ onen M, Nugroho C D, Yang C, Van Donkelaar J A et al. 2010 Nature 467 687–691

    work page 2010

  27. [27]

    Hanson R, Kouwenhoven L P, Petta J R, Tarucha S and Vandersypen L M 2007Reviews of Modern Physics 79 1217

    work page