Entanglement detection via third-order local invariants from randomized measurements

Anindita Bera; Giovanni Scala; Gniewomir Sarbicki

REVIEW 2 major objections 5 minor 43 references

Entanglement detection via third-order local invariants from randomized measurements

T0 review · 2 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read The paper reconstructs all third-order local invariants from randomized measurements and uses them to prove a separability criterion that detects Werner states in dimension 3 beyond the $p>1/2$ threshold.

desk verdict Solid third-order randomized-measurement entanglement criterion, but the printed invariant-form inequality has a factor-of-two error and the qubit limit is unhandled. read the letter →

arxiv 2506.18303 v2 pith:FPG6SQGX submitted 2025-06-23 quant-ph

Giovanni Scala , Anindita Bera , Gniewomir Sarbicki This is my paper · ORCID

classification quant-ph

keywords entanglementdetectionrandomizedmeasurementsthird-orderlocalinvariantsWeingartencalculusseparabilitycriteriaWernerstatespartialtranspose

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 entanglement can be detected from randomized measurements by using third-order local invariants, quantities built from three copies of the density matrix that are unchanged by local unitary rotations. The authors show that these invariants are fully determined by measured correlation statistics via an invertible linear system, so no full state tomography is needed. They derive a separability criterion, inequality (47), $\mathrm{Tr}(\rho^3)+\mathrm{Tr}((\rho^\Gamma)^3)\le 2\mathrm{Tr}(\rho_A\mathrm{Tr}_B\rho^2)$, which every separable bipartite state must satisfy. For Werner states in dimension 3, this detects entanglement for $p>10^{-1/3}$, improving the second-order purity criterion's $p>1/2$ threshold. If correct, the protocol offers an experimentally feasible path to entanglement detection beyond spectral criteria.

What carries the argument

The machinery is the Weingarten calculus for Haar-random unitaries applied to the outer-loop average of products of three outcome probabilities. For each subsystem, the permutation group $S_3$ acts on the three copies of $\rho$; its conjugacy classes — identity, transpositions, and 3-cycles — define linear maps $I_K$, $T_K$, $C_K$ whose traces over the diagrams in Tables I and II are the invariants $\{x_i\}$. The matrix $M=SW_A\otimes SW_B$ connects the measured statistics $\vec y$ to $\vec x$, and its explicit inversion (after reducing equal columns and using $y_4-y_5=\eta(x_4-x_5)$) yields the invariants directly from experimental data.

What would settle it

Prepare a two-qutrit Werner state with $p$ in the interval $(10^{-1/3}, 1/2)$, estimate $x_S$ and $x_8$ from randomized measurements, and compare with values from full tomography: the claim predicts $x_S>2x_8$ for every such $p$, so finding a $p$ in that range where the inequality still holds would falsify the criterion's improvement.

Watch

Extended reading notes

Core claim

The central discovery is that all third-order local invariants accessible from randomized measurements can be reconstructed from the $(3,2)$-twirling data $\vec y^{(2,3)}$ — three-point outcome correlations averaged over local Haar-random unitaries — by inverting a $10\times10$ matrix $M=SW_A\otimes SW_B$. Among the recovered invariants is $x_S=(x_9+x_{10})/2=\mathrm{Tr}(\rho^3)+\mathrm{Tr}((\rho^\Gamma)^3)$, and the paper proves $x_S\le 2x_8$ with $x_8=\mathrm{Tr}(\rho_A\,\mathrm{Tr}_B\rho^2)$ for every separable state, giving inequality (47). The proof uses positivity of the reduction map $(I\otimes R)(\rho)=\rho_A\otimes I_B-\rho\ge0$ and adds the criterion for $\rho$ and $\rho^\Gamma$. Benchmarked on Werner states, the criterion detects entanglement for $d=3$ at $p>10^{-1/3}$, compared with $p>1/2$ for second-order correlations.

Load-bearing premise

The protocol's correctness rests on the experimental estimate of the Haar-averaged three-point correlations: if the finite sample of random local unitaries does not faithfully represent the uniform average (or a unitary design), the $10\times10$ inversion will not return the true invariants; the qubit case $d=2$ is also left without an explicit nonsingular formula.

Editorial extensions

If this is right

A separable bipartite state must satisfy $x_S\le 2x_8$; observing $x_S>2x_8$ from randomized-measurement data certifies entanglement without reconstructing the full state.
In dimension 3 the third-order criterion lowers the Werner-state detection threshold from $p>1/2$ to $p>10^{-1/3}$, a concrete regime where second-order spectral criteria fail.
The same linear-inversion scheme estimates all marginal purities of $N$-partite states with arbitrary local dimensions, extending the earlier randomized-measurement purity protocol.
The criterion is built from positivity of the reduction map, so it is locally unitary invariant and requires no calibration of a specific entangled measurement basis.

Reading between the lines

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

The explicit inversion formulas are written for $d\ge3$; since the Weingarten denominators contain $d^2-4$, applying the method to qubits (the most common experimental platform) requires a separate limiting or reconstructing procedure that the paper does not spell out.
A natural next step is to search for higher-order invariants from four- and five-point correlations; the paper shows the third-order system reduces cleanly to a $10\times10$ matrix, and the pattern of equality classes suggests the linear-algebraic structure persists at higher orders.
Because the criterion combines $\rho$ and $\rho^\Gamma$, it may be worth benchmarking against bound-entangled states in $3\times3$, where partial transposition is positive; the paper does not test this, but the inequality is not manifestly restricted to NPT states.
The use of unitary designs instead of full Haar sampling could reduce the experimental overhead; the paper notes averaging can be simplified to sampling over an appropriate unitary design, so a concrete design construction would make the protocol directly implementable.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Desk Editor's Note

read the letter

Quick take: the paper is worth refereeing, but the version as printed has a factor-two error in the actionable form of its central criterion, plus an unhandled qubit limit that the abstract overclaims. The underlying derivation is sound.

What's new: the second-order purity criterion and the Haar-averaged randomized measurement scheme are known from van Enk–Beenakker and Brydges et al.; the genuinely new piece is the systematic reconstruction of third-order local invariants via Weingarten calculus and the reduction-map separability criterion Eq. (47), including the partial-transpose combination. I checked the Werner d=3 benchmark: the threshold 1/∛10 follows from Eq. (47) with the definitions in Eq. (42), and the derivation uses no fitted parameters. The 10x10 inversion is explicit and the citation pattern is clean.

Soft spots: the sentence “In terms of (42) it reads as xS ≤ 2x8” is not right. The proven inequality is Tr ρ^3 + Tr((ρ^Γ)^3) ≤ 2x8; with the intended xS = (x9 + x10)/2, this is xS ≤ x8, not xS ≤ 2x8. The printed form is strictly weaker and will miss detections that Eq. (47) guarantees. On top of that, Eq. (42)'s definition of xS as half of Tr((ρ^{T_B})^3) + Tr((ρ^{T_A})^3) does not match the left side of Eq. (47); the invariant labeling needs to be fixed so that x9 + x10 is actually the measurable combination. For what it's worth, the specific counterexample in the note I was sent doesn't hold numerically—I get Tr((ρ^Γ)^3) = 1/9 for the two-qutrit maximally entangled state, not 31/36—but the factor mismatch is real.

Second issue: the third-order Weingarten denominators contain d^2−4, so the formulas as written are singular for d=2. The abstract says arbitrary local dimensions. The qubit case needs a separate limit or a statement that the third-order criterion is vacuous or derivable there. That is a limitation, not a fatal flaw.

Bottom line: the paper is a solid subfield contribution that needs corrections before someone should implement it. The core math holds up; the mistakes are in the translation from inequality to invariants and in the domain of validity. I would send it to a referee, and I'd tell the authors to fix the factor and the d=2 caveat before posting a revised version. I wouldn't cite the current printed criterion.

Referee Report

2 major / 5 minor

Summary. The paper proposes an experimentally feasible entanglement-detection scheme based on third-order local invariants accessible via randomized measurements. The authors use Weingarten calculus to reconstruct a set of third-order invariants from averaged correlation statistics for bipartite systems, generalizing the second-order purity reconstruction. They then derive separability criteria from positivity of the reduction map, obtaining the third-order inequality Tr(ρ^3)+Tr((ρ^Γ)^3)≤2Tr(ρ_A Tr_B ρ^2). For Werner states in dimension d=3, this improves the detection threshold from p>1/2 (second-order) to p>1/∛10 (third-order). The paper claims the scheme works for arbitrary local dimensions and discusses two-qubit Bell-diagonal states as a benchmark.

Significance. If the reconstruction and criteria are valid, this is a valuable contribution: it probes third-order spectral invariants without full tomography, goes beyond second-order purity criteria, and gives a concrete falsifiable prediction for Werner states. The derivation is analytic and parameter-free, using standard Weingarten calculus, and the Werner-state threshold is a clean benchmark. However, the operational invariant form of the third-order criterion contains a factor error, and the d=2 case is not covered by the reconstruction, so the manuscript requires revision before the central claims are fully supported.

major comments (2)

[IV.A, Eq. (47) and the following sentence] The translation of inequality (47) into the invariant basis is inconsistent with the definitions in Eq. (42) and the diagram assignments in Table II. The left side of Eq. (47) is Trρ^3+Tr((ρ^Γ)^3), and evaluating the diagrams in Table II for a two-qutrit maximally entangled state gives x9=Trρ^3=1, x10=Tr((ρ^{T_B})^3)=1/9, and x8=1/3; hence the left side is 10/9 while 2x8=2/3, so Eq. (47) is violated. However, with the printed invariant form xS≤2x8 one obtains xS=(x9+x10)/2=5/9≤2/3, so the form is satisfied and the state is not detected. The correct invariant form is xS≤x8. This is load-bearing because the invariant form is the operationally used criterion, even though the Werner-state threshold in Eq. (49) appears to have been evaluated directly from Eq. (47). Please correct the factor and the definition of xS in Eq. (42).
[IV, Eqs. (7), (41), and Appendix A] The reconstruction of the third-order invariants from the measured vector y(2,3) is singular for local dimension d=2. The Weingarten functions in Eq. (7) and the matrices Q_K in Eq. (41) contain factors (d^2−4) in the denominator, so the advertised inversion x=M^{-1}y is not defined for qubit local dimensions. The abstract claims arbitrary local dimensions, and Section IV.B explicitly discusses two-qubit Bell-diagonal states, but no separate d=2 treatment or explicit limiting argument is provided. The final polynomial Eq. (49) can be evaluated at d=2 by continuity, but the central procedure for obtaining xS from randomized measurements is not justified for d=2. The authors should either restrict the third-order protocol to d_A,d_B≥3 or supply a dedicated qubit analysis.

minor comments (5)

[IV.A, Eq. (45)] The displayed equality appears to contain a misprint: positivity of the reduction map applied to ρ gives Tr(ρ(ρ_A⊗I_B−ρ))=Trρ_A^2−Trρ^2≥0, not Trρ_A^2−Trρ≥0.
[Eq. (42)] The line defining x4 and x5 lacks punctuation and is difficult to parse; please separate the definitions clearly.
[Abstract and Conclusions] There are grammatical errors: 'Werner states in d=3 the entanglement is detected' should read 'For Werner states in d=3, entanglement is detected', and the sentence 'All the higher-order criteria using reduction won’t be able to detect entangled states' appears to miss the word 'more'.
[References] Ref. [23] is cited in the text as 'Zoller et al.', but the first author is Brydges et al.; please adjust the citation style.
[Eqs. (46) and (47)] The expressions ρ_A Tr_B ρ^2 would be clearer with explicit parentheses, e.g., Tr(ρ_A (Tr_B ρ^2)), to avoid ambiguity about the operator ordering.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: the invariant reconstruction and the reduction-map separability criterion are derived from independent definitions and analytic inequalities, not from the data they predict.

full rationale

The derivation is self-contained. The second-order and third-order invariants in Eq. (42) are defined directly from powers and partial traces of the unknown state rho (e.g. x8 = Tr(TrB rho^2 TrB rho), 2xS = Tr((rho^{TB})^3)+Tr((rho^{TA})^3)); they are not defined in terms of the measured y-vector. The experimental quantities y are related to these invariants by the linear systems y = M x (Eqs. 22 and 37-38), and the invariants are recovered by inverting that system, which is a reconstruction step rather than a fitted prediction. The separability criterion in Eq. (47) follows from reduction-map positivity applied to rho and to its partial transpose (Eqs. 43-46), a necessary condition for separability that is independent of the randomized-measurement protocol; the printed invariant form xS <= 2x8 is presented as a translation of that inequality, not as the definition of xS or x8. The Werner-state threshold p > 1/cuberoot(10) is an analytic evaluation of the resulting polynomials (Eqs. 48-51), not a fit to any subset of the data. No load-bearing premise is justified solely by a self-citation: the reduction criterion is attributed to Horodecki and Horodecki [20], and the randomized-measurement framework is attributed to independent prior work [23,29], with the present paper explicitly re-deriving the generalization. Two concerns noted from the text are correctness or completeness issues rather than circularity: the claim that Eq. (47) reads as xS <= 2x8 appears to be a possible factor error in translating between the proven inequality and the invariant variables, and the Weingarten denominators d^2-4 leave the stated arbitrary-dimension claim unproved for d=2. Neither makes any derived quantity equivalent to its own input by construction, so the circularity score remains 0.

Assumptions & free parameters 0 free parameters · 4 assumptions · 0 invented entities

The central derivation relies on standard quantum information results (reduction-map positivity, partial-transpose invariance of separability, Weingarten calculus) and on the experimental assumption that Haar-random local unitaries can be sampled via unitary designs. No free parameters are fitted and no new entities are introduced.

assumptions (4)

standard math For every separable bipartite state ρ, (I⊗R)(ρ) ≥ 0, where R(X)=Tr(X)I−X is the reduction map.
Basis of criteria (29), (46), and (47); cited to Ref [20].
standard math The partial transpose of a separable state is again separable, so the reduction-map inequality also holds for ρ^Γ.
Used to add the two inequalities and obtain the measurable combination in Eq. (47).
domain assumption Outer-loop averages over local Haar unitaries can be realized experimentally via unitary designs and repeated circuit compilations.
The protocol in Section II assumes the measured vector y approximates the Haar-averaged correlators; the paper notes this can be reduced to sampling a unitary design.
standard math Weingarten formula Eq. (3) correctly evaluates Haar integrals of products of unitary matrix elements.
Central tool for computing the invariant decomposition; cited to Refs [27,28].

how reviews work

0 comments

Cite this review

Pith. "Pith review of Entanglement detection via third-order local invariants from randomized measurements." pith.science (2026). https://pith.science/paper/FPG6SQGX

@misc{pith2026250618303,
  author       = {Pith},
  title        = {Pith review of: Entanglement detection via third-order local invariants from randomized measurements},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FPG6SQGX}},
  note         = {Machine review of arXiv:2506.18303}
}

abstract

We compute all third-order local invariants accessible via randomised measurements and employ them to derive separability criteria. The reconstruction of the invariants yields experimentally accessible entanglement criteria for multipartite states with arbitrary local dimensions. The results show that third-order invariants capture inter-subsystem correlations beyond second-order spectral criteria within more feasible entanglement detection protocols than full tomography. As an example, Werner states in $d=3$ the entanglement is detected for $p>\frac 12$ at the second-order correlations, and it is improved to $p>\frac 1{\sqrt[3]{10}}$ at the third-order.

Figures

Figures reproduced from arXiv: 2506.18303 by the authors.

**Figure 2.** FIG. 2: Comparison for [PITH_FULL_IMAGE:figures/full_fig_p011_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

43 extracted references · 25 canonical work pages

[1]

Since vector⃗ x(2,3) has only 11 different entries, related to diagrams in table I and table II, we sum together columns related to the samexi, resulting in25×11matrix

work page
[2]

Among 25 equations there are only 10 different equations, roughly corresponding to pairs of conjugacy classes, except the pair([21],[21])which is represented by two non-equivalent equality types:(i A j = iA k ̸=i A ℓ,iB j =i B k ̸=i B ℓ )and(i A j =i A k ̸=i A ℓ,iB j ̸=i B k =i B ℓ )

work page
[3]

Quantum Secure Networks Partnership

Observe, that the last two columns are the same. We reduce number of columns introducing variable xS = (x9 +x 10)/2, resulting in10×10matrix:   y0 ... y9   =   iA   iB 3tB 2cB aB fB 2bB fB 3fB 2fB   3tA   iB 2tB tB 2cB aB 2fB fB 2bB fB 2fB fB 2fB   2cA   iB 3tB 2cB aB fB 2bB fB 3fB 2fB   aA   iB 3tB 2cB aB...

work page
[4]

Horodecki, P

R. Horodecki, P. Horodecki, M. Horodecki, and K. Horodecki, Reviews of modern physics81, 865 (2009)

2009
[5]

Gühne and G

O. Gühne and G. Tóth, Physics Reports474, 1 (2009)

2009
[6]

Chruściński and G

D. Chruściński and G. Sarbicki, Journal of Physics A: Mathematical and Theoretical47, 483001 (2014)

work page 2014
[7]

Aaronson, SIAM Journal on Computing49, STOC18 (2020)

S. Aaronson, SIAM Journal on Computing49, STOC18 (2020)

2020
[8]

Elben, R

A. Elben, R. Kueng, H.-Y. R. Huang, R. van Bijnen, C. Kokail, M. Dalmonte, P. Calabrese, B. Kraus, J. Preskill, P. Zoller, and B. Vermersch, Physical Review Letters125, 10.1103/physrevlett.125.200501 (2020)

doi:10.1103/physrevlett.125.200501 2020

Show all 43 references

[9]

Y. Zhou, P. Zeng, and Z. Liu, Physical Review Letters125, 10.1103/physrevlett.125.200502 (2020)

2020 doi
[10]

X.-D. Yu, S. Imai, and O. Gühne, Physical Review Letters127, 10.1103/physrevlett.127.060504 (2021)

2021 doi
[11]

S. A. Ghoreishi, G. Scala, R. Renner, L. L. Tacca, J. Bouda, S. P. Walborn, and M. Pawłowski, ArXiv preprint 2504.06350 (2025)

2025
[12]

Huang, R

H.-Y. Huang, R. Kueng, and J. Preskill, Nature Physics16, 1050 (2020)

2020
[13]

Elben, S

A. Elben, S. T. Flammia, H.-Y. Huang, R. Kueng, J. Preskill, B. Vermersch, and P. Zoller, Nature Reviews Physics5, 9 (2022)

2022
[14]

Gurvits, inProceedings of the thirty-fifth annual ACM symposium on Theory of computing, STOC03 (ACM,

L. Gurvits, inProceedings of the thirty-fifth annual ACM symposium on Theory of computing, STOC03 (ACM,
[15]

R. F. Werner, Physical Review A40, 4277 (1989)

1989
[16]

A. C. Doherty, P. A. Parrilo, and F. M. Spedalieri, Physical Review A69, 10.1103/physreva.69.022308 (2004)

2004 doi
[17]

Romero-Pallejà, J

J. Romero-Pallejà, J. Ahiable, A. Romancino, C. Marconi, and A. Sanpera, Journal of Mathematical Physics66, 10.1063/5.0240964 (2025)

2025 doi
[18]

Rico, arXiv preprint arXiv:2502.18446 (2025)

A. Rico, arXiv preprint arXiv:2502.18446 (2025)

2025
[19]

M. A. Nielsen and I. L. Chuang,Quantum computation and quantum information(Cambridge university press, 2010)

2010
[20]

M. A. Nielsen and G. Vidal, Quantum Info. Comput.1, 76–93 (2001)

2001
[21]

Hiroshima, Physical Review Letters91, 10.1103/physrevlett.91.057902 (2003)

T. Hiroshima, Physical Review Letters91, 10.1103/physrevlett.91.057902 (2003)

2003 doi
[22]

M. H. Partovi, Physical Review A86, 10.1103/physreva.86.022309 (2012)

2012 doi
[23]

Horodecki and P

M. Horodecki and P. Horodecki, Physical Review A59, 4206 (1999)

1999
[24]

Aggarwal, S

S. Aggarwal, S. Adhikari, and A. Majumdar, Physical Review A109, 012404 (2024). 13

2024
[25]

S. J. van Enk and C. W. J. Beenakker, Physical Review Letters108, 10.1103/physrevlett.108.110503 (2012)

2012 doi
[26]

Brydges, A

T. Brydges, A. Elben, P. Jurcevic, B. Vermersch, C. Maier, B. P. Lanyon, P. Zoller, R. Blatt, and C. F. Roos, Science364, 260 (2019)

2019
[27]

Mezzadri, Notices of the American Mathematical Society54, 592 (2007)

F. Mezzadri, Notices of the American Mathematical Society54, 592 (2007)

2007
[28]

P. J. Cavalcanti, G. Scala, A. Mandarino, and C. Lupo, arXiv preprint 2308.07744 (2023)

2023 arXiv
[29]

Życzkowski, K

K. Życzkowski, K. A. Penson, I. Nechita, and B. Collins, Journal of Mathematical Physics52, 062201 (2011), https://pubs.aip.org/aip/jmp/article-pdf/doi/10.1063/1.3595693/15599040/062201_1_online.pdf

2011 doi
[30]

Collins and P

B. Collins and P. Śniady, Communications in Mathematical Physics264, 773 (2006)

2006
[31]

Puchała and J

Z. Puchała and J. Miszczak, Bulletin of the Polish Academy of Sciences Technical Sciences65, 21 (2017)

2017
[32]

Cieśliński, S

P. Cieśliński, S. Imai, J. Dziewior, O. Gühne, L. Knips, W. Laskowski, J. Meinecke, T. Paterek, and T. Vértesi, Physics Reports1095, 1 (2024)

2024
[33]

Coecke and A

B. Coecke and A. Kissinger, inDiagrammatic Representation and Inference: 10th International Conference, Diagrams 2018, Edinburgh, UK, June 18-22, 2018, Proceedings 10(Springer, 2018) pp. 28–31

2018
[34]

Penroseet al., Combinatorial mathematics and its applications1, 3 (1971)

R. Penroseet al., Combinatorial mathematics and its applications1, 3 (1971)

1971
[35]

Gigena, E

N. Gigena, E. Panwar, G. Scala, M. Araújo, M. Farkas, and A. Chaturvedi, npj Quantum Information11, 10.1038/s41534-025-01029-6 (2025)

2025 doi
[36]

Tyc and J

T. Tyc and J. Vlach, The European Physical Journal D69, 1 (2015)

2015
[37]

Arnault, T

F. Arnault, T. M. Cong, D. Borivoje, L. Wieslaw, and T. Paterek, Physical Review A: Atomic, molecular, and optical physics [1990-2015]92(2015)

2015
[38]

M. C. Tran, B. Dakić, W. Laskowski, and T. Paterek, Physical Review A94, 042302 (2016)

2016
[39]

Ketterer, N

A. Ketterer, N. Wyderka, and O. Gühne, Physical Review Letters122, 120505 (2019)

2019
[40]

Grassl, M

M. Grassl, M. Rötteler, and T. Beth, Physical Review A58, 1833 (1998)

1998
[41]

Rains, IEEE Transactions on Information Theory46, 54 (2000)

E. Rains, IEEE Transactions on Information Theory46, 54 (2000)

2000
[42]

Makhlin, Quantum Information Processing1, 243 (2002)

Y. Makhlin, Quantum Information Processing1, 243 (2002)

2002
[43]

P. D. Jarvis, Journal of Physics A: Mathematical and Theoretical47, 215302 (2014). Appendix A: Local invarants from 3-point correlations Let us write down explicitly the equation (38)   y0 ... y9   = 1 DADB diag   1 dB−2 (dB−2)·(d B−1) dA−2 (dA−2)·(d B−...

2014

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.

[1] [1]

Since vector⃗ x(2,3) has only 11 different entries, related to diagrams in table I and table II, we sum together columns related to the samexi, resulting in25×11matrix

work page

[2] [2]

Among 25 equations there are only 10 different equations, roughly corresponding to pairs of conjugacy classes, except the pair([21],[21])which is represented by two non-equivalent equality types:(i A j = iA k ̸=i A ℓ,iB j =i B k ̸=i B ℓ )and(i A j =i A k ̸=i A ℓ,iB j ̸=i B k =i B ℓ )

work page

[3] [3]

Quantum Secure Networks Partnership

Observe, that the last two columns are the same. We reduce number of columns introducing variable xS = (x9 +x 10)/2, resulting in10×10matrix:   y0 ... y9   =   iA   iB 3tB 2cB aB fB 2bB fB 3fB 2fB   3tA   iB 2tB tB 2cB aB 2fB fB 2bB fB 2fB fB 2fB   2cA   iB 3tB 2cB aB fB 2bB fB 3fB 2fB   aA   iB 3tB 2cB aB...

work page

[4] [4]

Horodecki, P

R. Horodecki, P. Horodecki, M. Horodecki, and K. Horodecki, Reviews of modern physics81, 865 (2009)

2009

[5] [5]

Gühne and G

O. Gühne and G. Tóth, Physics Reports474, 1 (2009)

2009

[6] [6]

Chruściński and G

D. Chruściński and G. Sarbicki, Journal of Physics A: Mathematical and Theoretical47, 483001 (2014)

work page 2014

[7] [7]

Aaronson, SIAM Journal on Computing49, STOC18 (2020)

S. Aaronson, SIAM Journal on Computing49, STOC18 (2020)

2020

[8] [8]

Elben, R

A. Elben, R. Kueng, H.-Y. R. Huang, R. van Bijnen, C. Kokail, M. Dalmonte, P. Calabrese, B. Kraus, J. Preskill, P. Zoller, and B. Vermersch, Physical Review Letters125, 10.1103/physrevlett.125.200501 (2020)

doi:10.1103/physrevlett.125.200501 2020