Pith. sign in

REVIEW 3 major objections 5 minor 44 references

Graph States as a Resource for Quantum Metrology

T0 review · 3 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Any graph state's quantum Fisher information is a sum of squared twin-class sizes; bundled graph states from this identity approach the Heisenberg limit and handle dephasing and small erasures.

desk verdict The QFI formula and bundled graph-state construction are solid and useful; the abstract's unqualified erasure-robustness claim is average-case only and should be fixed before publication. read the letter →

arxiv 1908.05047 v3 pith:ADTE6E3B submitted 2019-08-14 quant-ph

Nathan Shettell , Damian Markham This is my paper · ORCID
classification quant-ph
keywords quantummetrologygraphstatesFisherinformationHeisenberglimitphaseestimationstabilizerdephasingnoiseerasurerobustness
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper claims that the usefulness of a graph state for phase estimation is encoded in its symmetries: for any graph state with no isolated vertices, the quantum Fisher information equals the sum of squared sizes of the groups of qubits with identical neighbourhoods. Using this identity, the authors construct bundled graph states by replacing each vertex of a seed graph with several qubits that all share the same neighbourhood, reaching a QFI of at least $n^2/k$, close to the Heisenberg limit. They show that under independent dephasing the QFI keeps the form $(1-2p)^2 Q(G)+4np(1-p)$, so bundled states beat the standard quantum limit for dephasing probability up to about $0.2$. For erasures, they compute the QFI averaged over all patterns of $e$ lost qubits and find that bundled cyclic graphs retain an advantage on average for up to three erasures, while bundled star graphs do not survive one. Because graph states are already produced and distributed on many platforms, these claims would make robust quantum-enhanced sensing a near-free add-on to existing graph-state technology.

What carries the argument

The machinery is the partition of graph vertices into twin classes: sets of vertices with identical neighbourhoods. For a graph state with no isolated vertices, a class of size $v_l$ contributes $v_l^2$ to the quantum Fisher information because every pair in the class corresponds to a stabilizer $X_iX_j$; summing classes gives $Q(G)=\sum_l v_l^2$. The bundled graph state construction amplifies this effect by replacing each vertex with $n_i$ mutually twin qubits, which pushes the QFI to $\sum_i n_i^2$. A second mechanism, true twins ($N(i)\cup\{i\}=N(j)\cup\{j\}$) together with local Clifford operations, produces the extended formula $Q=\sum_l v_l^2+\sum_m u_m^2-n$ and explains why complete graphs reach $Q=n^2$. The same twin-class bookkeeping drives the dephasing and erasure formulas, where the QFI decomposes over classes $V_l$ and their shared neighbourhoods $N_l$.

What would settle it

Compute the exact quantum Fisher information of a bundled cyclic graph after erasing one entire bundle together with its neighbours, for a fixed pattern rather than an average; if that value falls below the standard quantum limit for $e\le 3$ while the average over all patterns stays above it, the unqualified claim of robustness to finite erasures fails. Similarly, for a bundled graph whose shared neighbourhood size $N_l$ is small, evaluate the exact dephasing QFI at $p=0.15$ and compare with $(1-2p)^2Q(G)+4np(1-p)$; a large deviation would show the approximation's boundary.

Watch

Extended reading notes

Core claim

For phase estimation with Hamiltonian $H=\sum_{j=1}^n X_j$, the paper proves that an $n$-qubit graph state with no isolated vertices has $Q(G)=\sum_l v_l^2$, where $V_l$ are the disjoint sets of vertices sharing the same neighbourhood; the identity follows because $X_iX_j$ stabilizes the graph state exactly when $N(i)=N(j)$, so the QFI counts twin pairs. Extending the same counting to true twins and local Clifford operations gives $Q(G_{\mathrm{LC}})=\sum_l v_l^2+\sum_m u_m^2-n$, recovering $Q=n^2$ for complete graphs. The bundled construction replaces vertex $i$ by $n_i$ qubits with a common neighbourhood, yielding $Q(G_{\mathrm{bundle}})\ge \sum_i n_i^2 \ge n^2/k$ for a $k$-vertex seed graph, i.e. near-Heisenberg scaling. Under iid dephasing the derived closed form is approximately $(1-2p)^2Q(G)+4np(1-p)$ when shared neighbourhoods are large, and the paper shows bundled cyclic and star graphs beat the SQL for $p\le 0.2$. For $e$ erasures the paper computes the average QFI over all $\binom{n}{e}$ loss patterns; bundled cyclic graphs stay above the SQL on average for $e\le 3$, whereas bundled star graphs drop below after one erasure. Finally, the authors show the optimal $1/Q$ precision can be attained by measuring a stabilizer built from $Y$ and $Z$ operators, or by appending one extra qubit connected to a suitable set, for small phases.

Load-bearing premise

The robustness claims for erasures and dephasing rest on averaging over all erasure patterns and on shared neighbourhoods being large enough that the correction term $(2p(1-p)+1/2)^{N_l}$ is negligible; a fixed bad loss pattern or a small shared neighbourhood can remove the advantage.

Editorial extensions

If this is right

  • Cluster states have QFI $n$, confirming that unmodified cluster states give no scaling advantage, while star (GHZ-type) states have QFI $(n-1)^2+1$ and complete graphs reach $n^2$ after local Clifford operations.
  • Bundling any $k$-qubit seed graph with no isolated vertices produces an $n$-qubit state with $Q\ge n^2/k$, so a constant number of bundles gives near-Heisenberg scaling for arbitrarily large $n$.
  • Under iid dephasing with probability $p$, bundled graph states have QFI approximately $(1-2p)^2Q(G)+4np(1-p)$, keeping them above the standard quantum limit for $p\lesssim 0.2$ when shared neighbourhoods are large.
  • On average over all patterns of up to three erased qubits, bundled cyclic graphs retain a quantum advantage; bundled star graphs do not survive even a single erasure.
  • For small phases, the optimal precision $\Delta\theta^2=1/Q$ is achievable with a fixed stabilizer measurement, or with one extra appended qubit for arbitrary graph states.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Because $Q(G)$ is fixed entirely by the sizes of twin classes, any graph state can be screened for metrological usefulness by counting neighbourhood-equivalence classes alone, without simulating the state or optimizing measurements.
  • The erasure results are averages over all $\binom{n}{e}$ loss patterns; for adversarial or correlated losses, a worst-case analysis could give a very different picture, and the paper does not provide that guarantee.
  • One natural extension, not pursued here, is to protect the twin-class structure with graph-state quantum error correction so the stabilizer pairings survive after correction; this could turn the average erasure advantage into a guaranteed one.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper studies graph states as resources for single-parameter phase estimation under the Hamiltonian H = (1/2)∑X_i. It derives a closed-form expression for the quantum Fisher information (QFI) of any graph state with no isolated vertices, Q(G) = ∑_l v_l^2, where v_l are the sizes of disjoint groups of vertices with identical neighborhoods (Eq. 7). It then introduces a "bundling" construction that converts any k-qubit graph state into an n-qubit graph state with Q ≥ n^2/k, approximately Heisenberg-limited (Eq. 15). The paper also analyzes robustness under iid dephasing and finite erasures, derives an approximate QFI formula for dephasing (Eq. 20) and an average-QFI formula for single erasures (Eq. 23), provides a protocol for saturating the QFI with single-qubit measurements under stated conditions, and gives a lower bound on the number of stabilizer states useful for metrology (Appendix B).

Significance. The characterization Q(G) = ∑_l v_l^2 is elegant, correct, and immediately useful for identifying metrologically valuable graph states; the bundled construction is a simple and effective way to achieve near-Heisenberg scaling. The paper ships explicit derivations in the appendices, including exact erasure formulas and a dephasing approximation, and it proposes a concrete single-qubit measurement protocol. These are genuine strengths. If the noise-robustness claims are restricted to the regimes actually proven (average-case erasure robustness and dephasing in the regime where Eq. 19 holds), the paper will be a solid contribution to the practical quantum-metrology literature. As written, however, the abstract and conclusion overstate the robustness results.

major comments (3)
  1. [Abstract; §3, after Eq. (23); Appendix D] The abstract and conclusion state that bundled graph states "maintain a quantum advantage after being subjected to ... finite erasures" without qualification. The only derivation in the manuscript is for the QFI averaged uniformly over all erasure patterns, defined after Eq. (23) and evaluated in Appendix D (Eqs. D.7–D.9). For a fixed erasure pattern the exact formula Eq. (D.5)–(D.6) can give values far below this average: for a bundled star with equal bundles of size n/k, losing one leaf qubit gives Q = n/k, below the standard quantum limit n−1 for k > 1 (as Fig. 2 concedes for stars), and for a bundled cyclic graph with k = 4, erasing one vertex in one bundle yields Q = n/2, also below n−1. Thus the advertised robustness is an average-case property, not a worst-case guarantee. The abstract, conclusion, and any further statements should either explicitly say "on average over erasure patterns" or restrict the claim to graphs and erasure patterns for which a worst-case bound is proven.
  2. [§3, Eqs. (19)–(21)] The dephasing result is an approximation whose validity is controlled by Eq. (19), (2p(1−p)+1/2)^{N_l} ≈ 0. When the shared neighborhoods N_l are small or when p is not very small, the correction term in Eq. (C.5) is non-negligible; the paper itself notes a change of slope for bundled star graphs with large k after Eq. (22). Therefore Eq. (21) and the abstract's claim of robustness under iid dephasing should be presented as holding in the parameter regime where Eq. (19) is valid, rather than as an unconditional property of bundled graph states.
  3. [§4, Eqs. (24)–(25)] The single-qubit measurement protocol is derived for small θ and for states possessing a stabilizer consisting only of Y and Z operators. The derivation is sound in that regime, but the text should state more explicitly that the error-propagation formula Δθ² ≈ 1/Q(G) is a small-θ approximation and that for states failing the stabilizer condition the appended-qubit construction changes the effective measurement; the current presentation is slightly too terse to make the domain of validity of Eq. (25) unambiguous.
minor comments (5)
  1. [Affiliation line] The affiliation contains missing accents: "Laboratoire dInformatique" and "Universit" should be "Laboratoire d'Informatique" and "Université".
  2. [Appendix D, first paragraph] The phrase "To obtain To obtain any sort of meaningful value" contains a duplicated "To obtain".
  3. [Eq. (24)] The summation index in Eq. (24) runs from i = 0 in two places; it should run from i = 1 to n consistently with Eq. (1).
  4. [References and text] The author name "Ozmenaic" in the main text (near Eq. (12)) is a misspelling of "Oszmaniec" (reference [7]).
  5. [Fig. 2 caption] The notation "lognQ" in the caption is ambiguous; the base of the logarithm and the argument should be made explicit, e.g., "log_{10}(Q)" or "ln Q".

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the derivation is self-contained, starting from standard QFI and stabilizer formulas.

full rationale

The core result Q(G)=sum_l v_l^2 (Eq. 7) follows directly from the standard QFI formula (Eq. 3) and the graph-state stabilizer generators (Eq. 6): the only Pauli stabilizers of the form X_iX_j are those with identical neighborhoods, and the QFI counts them. The bundled-graph construction (Eqs. 13-14) is an explicit graph operation, and Eq. 15 is a direct application of Eq. 7 rather than a fitted or assumed prediction. The dephasing and erasure results are derived from the same QFI formula in Appendices C and D; no parameter is fitted to a subset of data and then 'predicted'. The paper does cite the authors' own previous work [8] and [14], but only as context or comparison (e.g., 'These results compare favorably with other resource states for tolerating noise [7,8]'), and the central argument does not depend on any self-cited uniqueness theorem or unverified ansatz. The caveat that the erasure robustness claim uses QFI averaged over all erasure patterns, stated explicitly after Eq. (23), is a limitation on the abstract's unqualified claim, not a circularity. Overall, no load-bearing step reduces by construction to its inputs.

Assumptions & free parameters 0 free parameters · 7 assumptions · 0 invented entities

No fitted parameters. The derivations rely on standard QFI and stabilizer machinery plus two explicit modeling assumptions: the small-noise approximation in Eq. (19) and the average-over-erasures metric after Eq. (23). These are axioms of the analysis, not free parameters.

assumptions (7)
  • standard math Quantum Fisher information formula for general states and pure states, cited to Hyllus et al. and Petz-Ghinea, invoked in Eqs. (2)-(3).
    The central tool for quantifying precision; accepted background result.
  • standard math Stabilizer state formalism and the Aaronson-Gottesman count N_n of stabilizer states, invoked in Eqs. (4)-(5) and Appendix B.
    Used to characterize stabilizer states and to count useful states.
  • standard math Graph state generators g_i = X_i tensor product Z over the neighborhood, Eq. (6).
    Standard definition linking graphs to stabilizer states.
  • domain assumption Dephasing noise model G -> sum_k p^k(1-p)^(n-k) Z^k G Z^k, Eq. (18).
    Assumed noise model; standard Pauli dephasing channel.
  • domain assumption Erasure model: tracing out lost qubits yields an equally weighted mixture over the erasure set and its neighborhood, Eq. (D.2).
    Derived from stabilizer properties but is a modeling assumption about the channel.
  • domain assumption Approximation (2p(1-p)+1/2)^(N_l) close to 0 when N_l is large and p small, used to obtain Eq. (20).
    This is an approximation, not an exact equality; its range of validity is central to the dephasing claims.
  • ad hoc to paper Use of average QFI over erasure patterns as the robustness metric, after Eq. (23).
    The paper switches from state-specific QFI to an average over erasure configurations, which changes the meaning of the robustness claim.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Graph States as a Resource for Quantum Metrology." pith.science (2026). https://pith.science/paper/ADTE6E3B

@misc{pith2026190805047,
  author       = {Pith},
  title        = {Pith review of: Graph States as a Resource for Quantum Metrology},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ADTE6E3B}},
  note         = {Machine review of arXiv:1908.05047}
}
read the original abstract

By using highly entangled states, quantum metrology guarantees precision impossible with classical measurements. Unfortunately such states can be very susceptible to noise, and it is a great challenge of the field to maintain quantum advantage in realistic conditions. In this study we investigate the practicality of graph states for quantum metrology. Graph states are a natural resource for much of quantum information, and here we characterize their quantum Fisher information (QFI) for an arbitrary graph state. We then construct families of graph states which approximately achieves the Heisenberg limit, we call these states bundled graph states. We demonstrate that bundled graph states maintain a quantum advantage after being subjected to iid dephasing or finite erasures. This shows that these graph states are good resources for robust quantum metrology. We also quantify the number of n qubit stabilizer states that are useful as a resource for quantum metrology.

Figures

Figures reproduced from arXiv: 1908.05047 by the authors.

Figure 1
Figure 1. FIG. 1: Construction of an [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2: Robustness of [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

44 extracted references · 30 canonical work pages

  1. [1]

    Begin with anyk qubit graph stateG = (V,E ) with no isolated vertices

    work page

  2. [2]

    Vertex i is replaced with ni qubits, labelled i(1),...,i (ni), such that ∑k i=1ni =n

    work page

  3. [3]

    The resulting n qubit graph state Gbundle = (V′,E′) has vertices V′ ={1(1),..., 1(n1),...,k (1),...,k (nk)}, (13) and edges E′ ={(i(a),j (b))∀a,b| (i,j )∈E}

    If (i,j )∈E then (i(a),j (b))∈E′∀a,b . The resulting n qubit graph state Gbundle = (V′,E′) has vertices V′ ={1(1),..., 1(n1),...,k (1),...,k (nk)}, (13) and edges E′ ={(i(a),j (b))∀a,b| (i,j )∈E}. (14) 3 We illustrate the construction of a bundled graph state in Fig. 1 by transforming a 3 qubit graph state into an n = 10 qubit bundled graph state. We chos...

    work page

  4. [4]

    The phase we are trying to estimate is small

    work page

  5. [5]

    4 (a) Bundled Graphs subjected to iid Dephasing (b) Bundled Graphs subjected to e Erasures FIG

    There exists a stabilizer, SM, for the graph which consists entirely of Y and Z operators. 4 (a) Bundled Graphs subjected to iid Dephasing (b) Bundled Graphs subjected to e Erasures FIG. 2: Robustness of n = 120 qubit bundled graphs subjected to iid dephasing (a) and a small number of erasures (b). In each scenario, the graphs are divided into k bundles o...

    work page

  6. [6]

    T´ oth and I

    G. T´ oth and I. Apellaniz, Journal of Physics A: Mathe- matical and Theoretical 47, 424006 (2014)

    2014

  7. [7]

    Giovannetti, S

    V. Giovannetti, S. Lloyd, and L. Maccone, Physical re- view letters 96, 010401 (2006)

    2006

  8. [8]

    Giovannetti, S

    V. Giovannetti, S. Lloyd, and L. Maccone, Science 306, 1330 (2004)

    2004

Show all 44 references
  1. [9]

    Demkowicz-Dobrza´ nski, J

    R. Demkowicz-Dobrza´ nski, J. Ko lody´ nski, and M. Gut ¸˘ a, Nature communications 3, 1063 (2012)

    2012

  2. [10]

    Gross, Journal of Physics B: Atomic, Molecular and Optical Physics 45, 103001 (2012)

    C. Gross, Journal of Physics B: Atomic, Molecular and Optical Physics 45, 103001 (2012)

    2012

  3. [11]

    Friis, D

    N. Friis, D. Orsucci, M. Skotiniotis, P. Sekatski, V. Dun- jko, H. J. Briegel, and W. D¨ ur, New Journal of Physics 19, 063044 (2017)

    2017

  4. [12]

    Oszmaniec, R

    M. Oszmaniec, R. Augusiak, C. Gogolin, J. Ko lody´ nski, A. Acin, and M. Lewenstein, Physical Review X 6, 041044 (2016)

    2016

  5. [13]

    Ouyang, N

    Y. Ouyang, N. Shettell, and D. Markham, arXiv e-prints arXiv:1908.02378 (2019), 1908.02378

    2019 arXiv

  6. [14]

    Nichols, T

    R. Nichols, T. R. Bromley, L. A. Correa, and G. Adesso, Physical Review A 94, 042101 (2016)

    2016

  7. [15]

    Albarelli, M

    F. Albarelli, M. A. Rossi, D. Tamascelli, and M. G. Genoni, in Quantum Information and Measurement (Op- tical Society of America, 2019), pp. S1A–5

    2019

  8. [16]

    Hyllus, W

    P. Hyllus, W. Laskowski, R. Krischek, C. Schwemmer, W. Wieczorek, H. Weinfurter, L. Pezz´ e, and A. Smerzi, Physical Review A 85, 022321 (2012)

    2012

  9. [17]

    Ma, Y.-x

    J. Ma, Y.-x. Huang, X. Wang, and C. Sun, Physical Re- view A 84, 022302 (2011)

    2011

  10. [18]

    Zhang, X

    Y. Zhang, X. Li, W. Yang, and G. Jin, Physical Review A 88, 043832 (2013)

    2013

  11. [19]

    Markham and B

    D. Markham and B. C. Sanders, Phys. Rev. A78, 042309 (2008)

    2008

  12. [20]

    Y. Qian, Z. Shen, G. He, and G. Zeng, Physical Review A 86, 052333 (2012)

    2012

  13. [21]

    Pirker, J

    A. Pirker, J. Walln¨ ofer, and W. D¨ ur, New Journal of Physics 20, 053054 (2018)

    2018

  14. [22]

    Meignant, D

    C. Meignant, D. Markham, and F. Grosshans, arXiv preprint arXiv:1811.05445 (2018)

    2018 arXiv

  15. [23]

    Raussendorf and H

    R. Raussendorf and H. J. Briegel, Phys. Rev. Lett. 86, 5188 (2001)

    2001

  16. [24]

    M. A. Nielsen, Reports on Mathematical Physics 57, 147 (2006)

    2006

  17. [25]

    Schlingemann and R

    D. Schlingemann and R. F. Werner, Phys. Rev. A 65, 012308 (2001)

    2001

  18. [26]

    M. Hein, W. D¨ ur, J. Eisert, R. Raussendorf, M. Nest, and H.-J. Briegel, arXiv preprint quant-ph/0602096 (2006)

    2006 arXiv

  19. [27]

    J. T. Barreiro, M. M¨ uller, P. Schindler, D. Nigg, T. Monz, M. Chwalla, M. Hennrich, C. F. Roos, P. Zoller, and R. Blatt, Nature 470, 486 (2011)

    2011

  20. [28]

    C. Song, K. Xu, W. Liu, C.-p. Yang, S.-B. Zheng, H. Deng, Q. Xie, K. Huang, Q. Guo, L. Zhang, et al., Phys. Rev. Lett. 119, 180511 (2017)

    2017

  21. [29]

    Cramer, N

    J. Cramer, N. Kalb, M. A. Rol, B. Hensen, M. S. Blok, M. Markham, D. J. Twitchen, R. Hanson, and T. H. Taminiau, Nature communications 7, 11526 (2016)

    2016

  22. [30]

    Clark, C

    S. Clark, C. M. Alves, and D. Jaksch, New Journal of Physics 7, 124 (2005)

    2005

  23. [31]

    J. C. Adcock, S. Morley-Short, J. W. Silverstone, and M. G. Thompson, Quantum Science and Technology 4, 015010 (2018)

    2018

  24. [32]

    X. Gu, M. Erhard, A. Zeilinger, and M. Krenn, Pro- ceedings of the National Academy of Sciences 116, 4147 (2019)

    2019

  25. [33]

    van Loock, C

    P. van Loock, C. Weedbrook, and M. Gu, Physical Re- view A 76, 032321 (2007)

    2007

  26. [34]

    Walschaers, S

    M. Walschaers, S. Sarkar, V. Parigi, and N. Treps, Phys- ical review letters 121, 220501 (2018)

    2018

  27. [35]

    Petz and C

    D. Petz and C. Ghinea, in Quantum probability and re- lated topics (World Scientific, 2011), pp. 261–281

    2011

  28. [36]

    Aaronson and D

    S. Aaronson and D. Gottesman, Physical Review A 70, 052328 (2004)

    2004

  29. [37]

    Cogis and E

    O. Cogis and E. Thierry, Discrete Optimization 2, 185 (2005)

    2005

  30. [38]

    S. L. Braunstein and C. M. Caves, Physical Review Let- ters 72, 3439 (1994)

    1994

  31. [39]

    Chiribella, G

    G. Chiribella, G. M. DAriano, P. Perinotti, and M. F. Sacchi, Physical Review A 70, 062105 (2004)

    2004

  32. [40]

    S. I. Knysh, E. H. Chen, and G. A. Durkin, arXiv preprint arXiv:1402.0495 (2014)

    2014 arXiv

  33. [41]

    S. Zhou, M. Zhang, J. Preskill, and L. Jiang, Nature communications 9, 78 (2018)

    2018

  34. [42]

    Demkowicz-Dobrza´ nski, J

    R. Demkowicz-Dobrza´ nski, J. Czajkowski, and P. Sekatski, Physical Review X 7, 041009 (2017)

    2017

  35. [43]

    Sekatski, M

    P. Sekatski, M. Skotiniotis, J. Ko lody´ nski, and W. D¨ ur, Quantum 1, 27 (2017)

    2017

  36. [44]

    H. L. Van Trees, Detection, estimation, and modulation theory, part I: detection, estimation, and linear modula- tion theory (John Wiley & Sons, 2004). Appendix A: Fundamentals of Estimation Theory Consider an observable random variable X which is dependent on some unknown par...

    2004

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.