pith. sign in

arxiv: 1907.11947 · v1 · pith:2B6TYYOJnew · submitted 2019-07-27 · 🪐 quant-ph

Repetitive Readout Enhanced by Machine Learning

Genyue Liu , Mo Chen , Yi-Xiang Liu , David Layden , Paola Cappellaro This is my paper

Pith reviewed 2026-05-24 14:31 UTC · model grok-4.3

classification 🪐 quant-ph
keywords repetitive readoutmachine learningquantum non-demolition measurementqubit readout fidelitymeasurement back-actionsolid-state qubitstime trace analysisancilla readout
0
0 comments X

The pith

Machine learning on the time traces of repetitive qubit readouts improves state discrimination by detecting back-action events.

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

The paper shows that machine learning applied to the full sequence of photon detections in repetitive quantum-non-demolition measurements extracts more information than simply counting total photons. The method learns to recognize the timing signatures of back-action disturbances on the qubit and still recovers the original state. This matters for solid-state qubits that lack single-shot readout, because the extra fidelity comes from data already recorded during standard repetitive protocols and requires no additional experimental time. The approach therefore applies directly to state preparation and quantum metrology tasks that rely on repeated measurements.

Core claim

The central claim is that a machine learning classifier trained on the time-resolved photon counts from repetitive readout sequences can identify when back-action occurred during the measurement chain and correctly assign the initial qubit state, yielding higher fidelity than the conventional total-photon threshold method on the same data.

What carries the argument

A machine learning classifier that processes the full time trace of photon arrival events to detect back-action signatures and classify the original qubit state.

If this is right

  • Readout fidelity increases for qubits that require multiple ancilla measurements because the original state information is recovered even after back-action.
  • No extra experimental repetitions or measurement time are needed since the improvement uses only the already-recorded photon timing data.
  • The technique extends to preparation-by-measurement protocols that rely on repetitive readout.
  • Quantum metrology schemes that use repeated measurements gain higher effective sensitivity without added overhead.

Where Pith is reading between the lines

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

  • The same timing information might allow real-time feedback to adjust subsequent measurements and reduce the average number of repetitions required.
  • Similar classifiers could be tested on other measurement records where timing or sequence data is available but currently discarded, such as in optical or superconducting systems.
  • Combining the approach with faster hardware processing could enable closed-loop control in larger quantum processors.

Load-bearing premise

The time traces contain learnable patterns that distinguish the timing of back-action events from the initial state in a way that simple total counts cannot capture.

What would settle it

Apply the trained machine learning model to a new set of repetitive readout time traces from the same qubit and find that its state assignment accuracy is no higher, or is lower, than the accuracy obtained by thresholding the total photon number alone.

Figures

Figures reproduced from arXiv: 1907.11947 by David Layden, Genyue Liu, Mo Chen, Paola Cappellaro, Yi-Xiang Liu.

Figure 1
Figure 1. Figure 1: FIG. 1. (a) Quantum circuit for repetitive quantum-non-demolition readout of the nuclear spin state [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2. (a) Readout fidelity as a function of repetition number [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3. Cumulative number of photons as a function of read out repetitions. Each trace corresponds to one input to the neural [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: FIG. 4. More efficient state preparation-by-measurement. [PITH_FULL_IMAGE:figures/full_fig_p005_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: FIG. 5. The 33-level NV model used in our simulation, consisting of 11 electronic spin levels times 3 nuclear spin levels (level [PITH_FULL_IMAGE:figures/full_fig_p007_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: FIG. 6. More efficient state preparation-by-measurement. [PITH_FULL_IMAGE:figures/full_fig_p008_6.png] view at source ↗
read the original abstract

Single-shot readout is a key component for scalable quantum information processing. However, many solid-state qubits with favorable properties lack the single-shot readout capability. One solution is to use the repetitive quantum-non-demolition readout technique, where the qubit is correlated with an ancilla, which is subsequently read out. The readout fidelity is therefore limited by the back-action on the qubit from the measurement. Traditionally, a threshold method is taken, where only the total photon count is used to discriminate qubit state, discarding all the information of the back-action hidden in the time trace of repetitive readout measurement. Here we show by using machine learning (ML), one obtains higher readout fidelity by taking advantage of the time trace data. ML is able to identify when back-action happened, and correctly read out the original state. Since the information is already recorded (but usually discarded), this improvement in fidelity does not consume additional experimental time, and could be directly applied to preparation-by-measurement and quantum metrology applications involving repetitive readout.

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 claims that machine learning applied to time-resolved photon-count traces from repetitive quantum-non-demolition readout can identify back-action events on the qubit, yielding higher readout fidelity than conventional total-count thresholding while using only already-recorded data.

Significance. If the quantitative improvement is reproducible and exceeds what an optimized time-resolved threshold achieves, the method offers a low-overhead route to better fidelity in solid-state systems that lack single-shot readout, directly benefiting preparation-by-measurement and metrology protocols.

major comments (2)
  1. [Abstract] Abstract: the claim of improved fidelity is stated without any numerical values for the fidelity gain, training-set size, validation protocol, or baseline comparison; this prevents assessment of whether the result is load-bearing.
  2. [Results] The manuscript must demonstrate that the ML classifier extracts information beyond what a properly designed time-resolved threshold on the same traces can achieve; without this control the central claim that ML uniquely identifies back-action signatures remains untested.
minor comments (2)
  1. [Methods] Provide the ML architecture, hyper-parameters, and loss function in a dedicated methods subsection so the procedure is reproducible.
  2. [Figures] Include error bars and statistical tests on all reported fidelities; state the number of experimental repetitions used for each data point.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading and constructive comments. We address each major comment below and indicate the corresponding revisions to the manuscript.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the claim of improved fidelity is stated without any numerical values for the fidelity gain, training-set size, validation protocol, or baseline comparison; this prevents assessment of whether the result is load-bearing.

    Authors: We agree that the abstract would benefit from these quantitative details to allow proper evaluation. The revised abstract now reports the fidelity improvement, training-set size, cross-validation protocol, and explicit baseline comparison to the total-count threshold method. revision: yes

  2. Referee: [Results] The manuscript must demonstrate that the ML classifier extracts information beyond what a properly designed time-resolved threshold on the same traces can achieve; without this control the central claim that ML uniquely identifies back-action signatures remains untested.

    Authors: The manuscript's primary baseline is the conventional total photon-count threshold, which is the standard method referenced in the literature for repetitive QND readout. To address the request for an additional control, the revised manuscript includes a direct comparison to an optimized time-resolved threshold applied to the same traces, confirming that the ML approach yields further improvement by detecting back-action patterns beyond simple time-binned discrimination. revision: yes

Circularity Check

0 steps flagged

No significant circularity detected

full rationale

The paper describes an empirical application of standard machine learning classifiers to experimental time-trace photon-count data from repetitive QND readout. No derivation chain, equations, fitted parameters renamed as predictions, or self-citation load-bearing steps are present in the provided text. The central claim—that ML can exploit back-action signatures in the traces to outperform total-count thresholding—is an experimental performance result, not a self-referential construction or ansatz smuggled via prior work. The method is self-contained against external benchmarks (experimental data) and does not reduce to its inputs by definition.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Review based on abstract only; no free parameters, axioms, or invented entities are described in the provided text.

pith-pipeline@v0.9.0 · 5706 in / 949 out tokens · 19502 ms · 2026-05-24T14:31:59.047559+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

49 extracted references · 49 canonical work pages · 1 internal anchor

  1. [1]

    D. P. DiVincenzo, Fortschr. Phys. 48, 771 (2000)

    work page 2000

  2. [2]

    Raussendorf, D

    R. Raussendorf, D. E. Browne, and H. J. Briegel, Phys. Rev. A 68, 022312 (2003)

    work page 2003

  3. [3]

    M. A. Nielsen and I. L. Chuang, Quantum computation and quantum information (Cambridge University Press, Cambridge; New York, 2000)

    work page 2000

  4. [4]

    W. S. Bakr, J. I. Gillen, A. Peng, S. Folling, and M. Greiner, Nature 462, 74 (2009)

    work page 2009

  5. [5]

    Endres, H

    M. Endres, H. Bernien, A. Keesling, H. Levine, E. R. Anschuetz, A. Krajenbrink, C. Senko, V. Vuletic, M. Greiner, and M. D. Lukin, Science 354, 1024 (2016)

    work page 2016

  6. [6]

    Cooper, J

    A. Cooper, J. P. Covey, I. S. Madjarov, S. G. Porsev, M. S. Safronova, and M. Endres, Phys. Rev. X 8, 041055 (2018)

    work page 2018

  7. [7]

    A. H. Myerson, D. J. Szwer, S. C. Webster, D. T. C. Allcock, M. J. Curtis, G. Imreh, J. A. Sherman, D. N. Stacey, A. M. Steane, and D. M. Lucas, Phys. Rev. Lett. 100, 200502 (2008)

    work page 2008

  8. [8]

    Jeffrey, D

    E. Jeffrey, D. Sank, J. Y. Mutus, T. C. White, J. Kelly, R. Barends, Y. Chen, Z. Chen, B. Chiaro, A. Dunsworth, A. Megrant, P. J. J. O’Malley, C. Neill, P. Roushan, A. Vainsencher, J. Wenner, A. N. Cleland, and J. M. Martinis, Phys. Rev. Lett. 112, 190504 (2014)

    work page 2014

  9. [9]

    Morello, J

    A. Morello, J. J. Pla, F. A. Zwanenburg, K. W. Chan, K. Y. Tan, H. Huebl, M. Mottonen, C. D. Nugroho, C. Yang, J. A. van Donkelaar, A. D. C. Alves, D. N. Jamieson, C. C. Escott, L. C. L. Hollenberg, R. G. Clark, and A. S. Dzurak, Nature 467, 687 (2010)

    work page 2010

  10. [10]

    J. M. Elzerman, R. Hanson, L. H. Willems van Beveren, B. Witkamp, L. M. K. Vandersypen, and L. P. Kouwen- hoven, Nature 430, 431 (2004)

    work page 2004

  11. [11]

    Hanson, L

    R. Hanson, L. H. W. van Beveren, I. T. Vink, J. M. Elz- erman, W. J. M. Naber, F. H. L. Koppens, L. P. Kouwen- hoven, and L. M. K. Vandersypen, Phys. Rev. Lett. 94, 196802 (2005)

    work page 2005

  12. [12]

    Neumann, J

    P. Neumann, J. Beck, M. Steiner, F. Rempp, H. Fedder, 9 P. R. Hemmer, J. Wrachtrup, and F. Jelezko, Science 5991, 542 (2010)

    work page 2010

  13. [13]

    P. C. Maurer, G. Kucsko, C. Latta, L. Jiang, N. Y. Yao, S. D. Bennett, F. Pastawski, D. Hunger, N. Chisholm, M. Markham, D. J. Twitchen, J. I. Cirac, and M. D. Lukin, Science 336, 1283 (2012)

    work page 2012

  14. [14]

    Dr´ eau, P

    A. Dr´ eau, P. Spinicelli, J. R. Maze, J.-F. Roch, and V. Jacques, Phys. Rev. Lett. 110, 060502 (2013)

    work page 2013

  15. [15]

    Waldherr, Y

    G. Waldherr, Y. Wang, S. Zaiser, M. Jamali, T. Schulte- Herbruggen, H. Abe, T. Ohshima, J. Isoya, J. F. Du, P. Neumann, and J. Wrachtrup, Nature 506, 204 (2014)

    work page 2014

  16. [16]

    G.-Q. Liu, J. Xing, W.-L. Ma, P. Wang, C.-H. Li, H. C. Po, Y.-R. Zhang, H. Fan, R.-B. Liu, and X.-Y. Pan, Phys. Rev. Lett. 118, 150504 (2017)

    work page 2017

  17. [17]

    Schmidt, T

    P. Schmidt, T. Rosenband, C. Langer, W. Itano, J. Bergquist, and D. Wineland, Science 309, 749 (2005)

    work page 2005

  18. [18]

    D. B. Hume, T. Rosenband, and D. J. Wineland, Phys. Rev. Lett. 99, 120502 (2007)

    work page 2007

  19. [19]

    Jiang, J

    L. Jiang, J. S. Hodges, J. R. Maze, P. Maurer, J. M. Taylor, D. G. Cory, P. R. Hemmer, R. L. Walsworth, A. Yacoby, A. S. Zibrov, and M. D. Lukin, Science 326, 267 (2009)

    work page 2009

  20. [20]

    K. S. Cujia, J. M. Boss, K. Herb, J. Zopes, and C. L. Degen, Nature 571, 230 (2019)

    work page 2019

  21. [21]

    Pfender, P

    M. Pfender, P. Wang, H. Sumiya, S. Onoda, W. Yang, D. B. R. Dasari, P. Neumann, X.-Y. Pan, J. Isoya, R.-B. Liu, and J. Wrachtrup, Nat. Commun. 10, 594 (2019)

    work page 2019

  22. [22]

    Krizhevsky, I

    A. Krizhevsky, I. Sutskever, and G. E. Hinton, Commun. ACM 60, 84 (2017)

    work page 2017

  23. [23]

    A. Seif, K. A. Landsman, N. M. Linke, C. Figgatt, C. Monroe, and M. Hafezi, J. Phys. B: At., Mol. Opt. Phys. 51, 174006 (2018)

    work page 2018

  24. [24]

    Santagati, A

    R. Santagati, A. A. Gentile, S. Knauer, S. Schmitt, S. Paesani, C. Granade, N. Wiebe, C. Osterkamp, L. P. McGuinness, J. Wang, M. G. Thompson, J. G. Rarity, F. Jelezko, and A. Laing, Phys. Rev. X 9, 021019 (2019)

    work page 2019

  25. [25]

    H. T. Dinani, D. W. Berry, R. Gonzalez, J. R. Maze, and C. Bonato, Phys. Rev. B 99, 125413 (2019)

    work page 2019

  26. [26]

    Lian, S.-T

    W. Lian, S.-T. Wang, S. Lu, Y. Huang, F. Wang, X. Yuan, W. Zhang, X. Ouyang, X. Wang, X. Huang, L. He, X. Chang, D.-L. Deng, and L. Duan, Phys. Rev. Lett. 122, 210503 (2019)

    work page 2019

  27. [27]

    Robledo, H

    L. Robledo, H. Bernien, T. van der Sar, and R. Hanson, New J. Phys. 13, 025013 (2011)

    work page 2011

  28. [28]

    Tetienne, L

    J.-P. Tetienne, L. Rondin, P. Spinicelli, M. Chipaux, T. Debuisschert, J.-F. Roch, and V. Jacques, New J. Phys. 14, 103033 (2012)

    work page 2012

  29. [29]

    Gupta, L

    A. Gupta, L. Hacquebard, and L. Childress, J. Opt. Soc. Am. B 33, B28 (2016)

    work page 2016

  30. [30]

    N. B. Manson, J. P. Harrison, and M. J. Sellars, Phys. Rev. B 74, 104303 (2006)

    work page 2006

  31. [31]

    Poggiali, P

    F. Poggiali, P. Cappellaro, and N. Fabbri, Phys. Rev. B 95, 195308 (2017)

    work page 2017

  32. [32]

    Gali, Phys

    A. Gali, Phys. Rev. B 79, 235210 (2009)

    work page 2009

  33. [33]

    Marseglia, J

    L. Marseglia, J. P. Hadden, A. C. Stanley-Clarke, J. P. Harrison, B. Patton, Y.-L. D. Ho, B. Naydenov, F. Jelezko, J. Meijer, P. Dolan, J. Smith, J. Rarity, and J. O’Brien, App. Phys. Lett 98, 133107 (2011)

    work page 2011

  34. [34]

    Robledo, L

    L. Robledo, L. Childress, H. Bernien, B. Hensen, P. F. A. Alkemade, and R. Hanson, Nature 477, 574 (2011)

    work page 2011

  35. [35]

    N. H. Wan, B. J. Shields, D. Kim, S. Mouradian, B. Lien- hard, M. Walsh, H. Bakhru, T. Schroeder, and D. En- glund, Nano Lett. 18, 2787 (2018)

    work page 2018

  36. [36]

    J. J. H. Shim, B. Nowak, I. Niemeyer, J. Zhang, F. D. Brandao, and D. Suter, arXiv:1307.0257 (2013)

    work page internal anchor Pith review Pith/arXiv arXiv 2013

  37. [37]

    K. R. K. Rao and D. Suter, Phys. Rev. B 94, 060101 (2016)

    work page 2016

  38. [38]

    Smeltzer, L

    B. Smeltzer, L. Childress, and A. Gali, New Journal of Physics 13, 025021 (2011)

    work page 2011

  39. [39]

    Dr´ eau, J.-R

    A. Dr´ eau, J.-R. Maze, M. Lesik, J.-F. Roch, and V. Jacques, prb 85, 134107 (2012)

    work page 2012

  40. [40]

    Magesan, J

    E. Magesan, J. M. Gambetta, A. D. C´ orcoles, and J. M. Chow, Phys. Rev. Lett. 114, 200501 (2015)

    work page 2015

  41. [41]

    Aslam, M

    N. Aslam, M. Pfender, P. Neumann, R. Reuter, A. Zappe, F. F´ avaro de Oliveira, A. Denisenko, H. Sumiya, S. On- oda, J. Isoya, and J. Wrachtrup, Science 357, 67 (2017)

    work page 2017

  42. [42]

    Lovchinsky, A

    I. Lovchinsky, A. O. Sushkov, E. Urbach, N. P. de Leon, S. Choi, K. De Greve, R. Evans, R. Gertner, E. Bersin, C. M¨ uller, L. McGuinness, F. Jelezko, R. L. Walsworth, H. Park, and M. D. Lukin, Science 351, 836 (2016)

    work page 2016

  43. [43]

    C. L. Degen, F. Reinhard, and P. Cappellaro, Rev. Mod. Phys. 89, 035002 (2017)

    work page 2017

  44. [44]

    Aslam, G

    N. Aslam, G. Waldherr, P. Neumann, F. Jelezko, and J. Wrachtrup, New J. Phys. 15, 013064 (2013)

    work page 2013

  45. [45]

    Chen, C.-L

    X.-D. Chen, C.-L. Zou, F.-W. Sun, and G.-C. Guo, App. Phys. Lett 103, 013112 (2013)

    work page 2013

  46. [46]

    Hacquebard and L

    L. Hacquebard and L. Childress, Phys. Rev. A97, 063408 (2018)

    work page 2018

  47. [47]

    Smeltzer, J

    B. Smeltzer, J. McIntyre, and L. Childress, Phys. Rev. A 80, 050302 (2009)

    work page 2009

  48. [48]

    Neumann, R

    P. Neumann, R. Kolesov, V. Jacques, J. Beck, J. Tisler, A. Batalov, L. Rogers, N. B. Manson, G. Balasubrama- nian, F. Jelezko, and J. Wrachtrup, New J. Phys. 11, 013017 (2009)

    work page 2009

  49. [49]

    Born and R

    M. Born and R. Oppenheimer, Annalen der Physik 389, 457 (1927)

    work page 1927