Pith. sign in

REVIEW 5 minor 1 cited by

This paper proves that constant-precision spectrum estimation of a d-dimensional quantum state requires at least d^{2-o(1)} copies, nearly matching the cost of full tomography.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

Spectrum estimation, von Neumann entropy estimation, and rank testing of d-dimensional quantum states each require d^{2−o(1)} copies at constant precision — nearly as many as full tomography.

T0 review reviewed 2026-08-03 challenge →

load-bearing objection First superlinear lower bound for spectrum estimation, and the argument holds up better than the reader's one flagged worry suggests.

arxiv 2607.29680 v1 pith:VRGCINN6 submitted 2026-07-31 quant-ph cs.DS

Spectrum Estimation is Almost as Hard as Tomography

Marco Fanizza , Ryan O'Donnell , Chirag Wadhwa This is my paper
classification quant-ph cs.DS MSC 81P4560B2005E10 PACS 03.67.-a
keywords quantum spectrum estimationsample complexityJucys–Murphy elementsrandom projectionsvon Neumann entropyrank testingmoment matchingProuhet–Tarry–Escott
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

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 establishes that learning the eigenvalue spectrum of an unknown d-dimensional quantum state is almost as costly as learning the entire state: at least d^{2-o(1)} copies are needed for any constant-precision spectrum estimate. The same near-quadratic lower bound is proved for von Neumann entropy estimation and for rank testing. Prior to this work, only a linear lower bound was known for spectrum estimation, while the best algorithms achieved only sub-polynomial savings over full tomography. The proof builds a pair of unitarily invariant mixtures of states that are statistically almost indistinguishable with fewer than ~d² copies, yet have typically separated spectra, entropies, and ranks. This closes the gap to near-optimality for the standard Empirical Young Diagram estimation strategy.

Core claim

For every even k≥2 there is a dimension-independent precision ε_k such that distinguishing the two constructed ensembles requires Ω(d^{2−4/(k+4)}) copies, giving the d^{2−o(1)} lower bound for spectrum estimation (Theorem 1.2), entropy estimation (Theorem 1.3), and rank testing (Theorem 1.4). The hard ensembles are tilted versions of sandwiched products of Haar-random projections; the tilt makes the n-fold average state proportional to an explicit rational function of Jucys–Murphy elements. Matching the first k−1 power sums of the two parameter sequences (via a Prouhet–Tarry–Escott construction) makes the log-likelihood ratio begin at degree k, and bounding the resulting triangular discrimin

What carries the argument

The central objects are the Jucys–Murphy elements J_t=Σ_{i<t}(i t), commuting transposition sums in the symmetric group algebra; their normalized versions J̃_t=J_t/d index the common eigenbasis of unitarily invariant n-copy operators. For a Haar-random rank-r projector Π, E[Π^{⊗n}] = ∏_{t=1}^n (r+J_t)/(d+J_t), and products of random projections have tensor moments ∏ f_a(J̃_t) with f_a(z)=∏(1+a_i z)/(1+z). A tilt with density ∝ Tr(X)^n makes the average n-copy state proportional to these moments. Matching the first k−1 power sums of the two parameter sequences makes the log-likelihood ratio start at degree k, leaving only high-order Jucys–Murphy power sums to control.

Load-bearing premise

The tilted hard states are assumed to inherit the typical spectrum, entropy, and rank concentration of the underlying untitled random-projector products; if the expected normalized trace E tr(X_e) or the Lipschitz constants of tr(X^j) had a hidden d-dependence, the constant separation between the two ensembles would not hold with high probability.

What would settle it

For one fixed even k, simulate or analytically compute the sorted-total-variation distance between spectra of states drawn from the two tilted ensembles at dimension d (using the matched-power-sum construction). Proposition 5.8 predicts that separation by a constant occurs with failure probability at most exp(−Ω(d²)); observing failure probability that is not exponentially small, or observing E tr(X_e) that decays with d, would refute the load-bearing concentration step.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

If this is right

  • If the lower bound is correct, the constant-precision regime of spectrum estimation is characterized up to a d^{o(1)} factor, with the standard Θ(d²)-copy Empirical Young Diagram algorithm near-optimal.
  • Two-stage quantum state learning—first eigenvalues, then eigenvectors—cannot avoid a near-quadratic first stage; the eigenvalue stage alone is as expensive as full tomography.
  • von Neumann entropy estimation, despite being a scalar functional, also requires d^{2−o(1)} copies, matching the upper bound up to sub-polynomial factors.
  • Rank testing with two-sided error becomes almost as hard as rank testing with one-sided error, refining the prior Ω(r) bound to near-optimal Ω(d^{2−γ}).
  • Any algorithm that estimates the spectrum to constant sorted-total-variation error must use nearly the same number of copies as a full tomographic reconstruction, even allowing fully entangled measurements.

Where Pith is reading between the lines

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

  • Editorial inference: The tilting mechanism is likely reusable for other unitarily invariant properties; it supplies a quantum analogue of classical Poissonization/moment-matching and may yield tight lower bounds under restricted (non-entangled) measurements, where a d^{3/2}-to-d^3 gap remains.
  • Editorial inference: The ε-dependence is left open; the warm-up instance already yields Ω(d^{4/3}/ε^{2/3}) for ε ≳ d^{-1/4}, and the framework suggests the conjectured Θ(d²/ε²) dependence might be attainable by extending the moment-matching order to scale with 1/ε.
  • Editorial inference: Because the hard instances are mixtures of Haar-randomized states, the indistinguishability step is measurement-independent, so the d^{2−o(1)} barrier likely applies to any test, not only to the estimators considered here; a testable extension would be to prove the analogous lower bound for incoherent/LOCC measurements.
  • Editorial inference: The rank-separation proof exploits the atom at zero in the free multiplicative convolution of the projector laws; similar free-probability reasoning could give lower bounds for estimating other spectral statistics, such as Rényi entropies or entanglement spectra.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

0 major / 5 minor

Summary. The paper proves that constant-precision spectrum estimation, von Neumann entropy estimation, and rank testing of d-dimensional quantum states require nearly as many copies as full tomography. For each even k≥2 it constructs a pair of unitarily invariant mixtures built from tilted products of Haar-random projectors (PRP_d(a), PRP_d(b)), with parameters a,b obtained from the Prouhet–Thue–Morse construction. The n-copy averaged states are shown to be indistinguishable for n=o(d^{2−4/(k+4)}) by expanding the log-likelihood ratio in Jucys–Murphy elements, matching low-order moments, and bounding high-order terms with symmetric-group combinatorics. The same pairs are shown to have separated spectra, entropies (after depolarization), and ranks with probability 1−exp(−Ω_k(d^2)), yielding the claimed Ω(d^{2−o(1)}) lower bounds. The warmup section gives a self-contained Ω(d^{4/3}) proof for rank-d/2 projectors versus Haar-random marginals.

Significance. If correct, this is a major advance: it resolves the long-standing gap between spectrum estimation and full tomography in the constant-precision regime, shows that the EYD algorithm is near-optimal, and improves prior Ω(d) and Ω(d/log d) bounds dramatically. The technical contribution is substantial and largely self-contained: explicit tensor-moment formulas for product-of-random-projections ensembles (Prop. 3.5), the tilted-law normalization trick (Def. 5.4), exact trace computations in the warmup (Thm. 4.4), and a detailed combinatorial analysis of high-order Jucys–Murphy moments (Lemmas 6.5–6.7). I checked the warmup trace computations (Eqs. 4.33–4.35), the tilting identity, and the exponent bookkeeping in Theorem 5.7 and Lemmas 6.3–6.4; they are consistent. The potential weakness in transferring concentration of spectra from the untilted to the tilted law does not materialize: ν_e is a product of dimension-independent constants, the Lipschitz constants in Lemma 7.2 are O_k(j√K/d), and the cost in Prop. 5.5 is only exp(O(n)), which is dominated by exp(−Ω(d^2)) when n=o(d^2). The rank separation relies on standard free-probability results ([CM14], [Bel03]), which is an appropriat

minor comments (5)
  1. [Abstract / Theorem 1.2] The phrase 'constant-precision' should be qualified: the precision ε_k is dimension-independent but may depend on k (equivalently, on γ). Theorem 1.2 states this precisely, but the abstract could be read as claiming a universal fixed ε for every γ. A one-sentence clarification would remove this ambiguity.
  2. [Section 1.2 (Outlook)] The unreported Ω(d^{4/3}/ε^{2/3}) bound and the ChatGPT 5.6 candidate proof are promises rather than results included in this paper. They should be removed or placed in a clearly marked 'Remark (not part of the technical content)' to avoid any ambiguity about what is being claimed and verified.
  3. [Proposition 5.2] The sentence 'One can verify that this preserves the moment-matching guarantees using, say, the binomial theorem' is a missing one-line proof. After shifting by 1, the difference of j-th power sums becomes Σ_{s=0}^j C(j,s)(S_s(E)−S_s(O)), which vanishes for j<k and retains the k-th mismatch. Adding this would make the construction fully self-contained.
  4. [Section 7.2, Eq. (7.16)] The displayed chain in the proof of Lemma 7.2 contains typographical artifacts: stray brackets and a duplicated superscript in the expression for the projection-difference norm. Please proofread this display.
  5. [Section 9 (Lemma 9.1)] The phrase 'Asymptotic freeness of independent Haar conjugates implies...' could state explicitly that the mode of convergence is almost sure, matching the lemma statement. The citation to [CM14] is appropriate, but a brief sentence on the strong asymptotic freeness formulation would improve readability.

Circularity Check

0 steps flagged

No significant circularity: the lower-bound derivation is self-contained, with hard instances adversarially engineered and each load-bearing claim proved in-paper or supported by independent external results.

full rationale

The central derivation chain—PRP tensor-moment formula (Prop. 3.5), Prouhet–Tarry–Escott moment matching (Prop. 5.2), the Taylor expansion of log(f_b/f_a) (Prop. 5.3), statistical indistinguishability (Thm. 5.7), spectral/entropy/rank separation (Props. 5.8–5.10), and the reductions to Theorems 1.2–1.4—does not reduce to its own inputs by construction. The tilted law is deliberately defined so that E[ρ^{⊗n}] ∝ E[X^{⊗n}]; this is a designed property of the hard instance, not a fitted parameter, and the subsequent indistinguishability and separation statements are proved rather than assumed. The moment-matching constants come from the Thue–Morse/PTE construction, not from tuning to the target lower bound. The Ψ_j separation in Lemma 7.1 is derived from Proposition 5.3; the concentration transfer (Prop. 5.5) uses Jensen plus the explicit ν_e = Θ_k(1) from Eq. (3.10); the Lipschitz bound in Lemma 7.2 is proved in-paper and invokes an external concentration theorem [Mec19]. Rank separation invokes external free-probability results [CM14, Bel03], which are independent evidence and not supplied by the authors' own prior work. Self-citations such as [OW21] and [OW26] are contextual or superseded by the new bound, not load-bearing. The self-flagged asides (the ChatGPT 5.6 candidate proof and the unreported Ω(d^{4/3}/ε^{2/3}) bound) are explicitly non-load-bearing. No equation or fitted parameter is renamed as a prediction; no uniqueness theorem is imported from the authors; no ansatz is smuggled in via self-citation.

Axiom & Free-Parameter Ledger

3 free parameters · 13 axioms · 2 invented entities

The construction introduces no fitted parameters: the PTE/Thue–Morse sequences are exact combinatorial objects; the constants (ε_k, δ_spec, δ_ent, δ_rank, p_k, B, C_0, δ) are proven to exist as dimension-independent values, not numerically fitted; and the n-dependent tilt is part of the hard-instance design, mirroring Poissonization in classical distribution testing. The mathematics rests on standard theorems (Jucys identity, Collins–Śniady twirl, Meckes concentration, Collins–Male asymptotic freeness, Belinschi atom formula, Marchenko–Pastur/Page asymptotics) plus the paper's own proved lemmas. The invented entities are mathematical devices (the tilted law; the PRP ensembles) that are fully specified and internally justified, with no external falsifiable handle — hence independent_evidence = false for both.

free parameters (3)
  • Depolarizing strength p_k (Prop 5.9) = existence only; chosen small enough that p_k·max{1, B_k−1} < 1 (not numerically specified)
    Hand-chosen to make the centered-moment entropy series (Lemma 8.1) converge and to keep the O_k(p_k^{k+2}) error below the leading −κ_k p_k^{k+1}/k(k+1) mismatch term; dimension-independent; sets the constant precision ε_k in Theorem 1.3.
  • PTE/Thue–Morse sequences a, b ∈ Z_+^K, K = 2^{k−1} = Thue–Morse partition of {0,…,2^k−1}, shifted by +1 (Prop 5.2)
    Design parameters of the hard instances: they set the matched moment degree k−1 and hence the lower-bound exponent 2−4/(k+4); exact combinatorial construction, not fitted to data.
  • Good-set radius δ (Eq 6.18) = any constant < ½ min_i{a_i^{−1}, b_i^{−1}}
    Hand-chosen threshold separating eigenbasis indices where the log-likelihood Taylor expansion (Prop 5.3) is controlled; all constants absorb δ into C_k.
axioms (13)
  • standard math Jucys identity: ∏_{t=1}^n (z+J_t) = Σ_{π∈S_n} z^{#cyc(π)}π (Eq 2.8, [Juc74])
    Used to compute E[Π^{⊗n}] (Prop 3.2), ρ_Haar^{(n)} (Lemma 4.2), and all PRP moment formulas (Prop 3.5).
  • standard math Collins–Śniady unitary twirl formula E[U^{⊗n} A U^{†⊗n}] = Φ(A)Φ(1)^{−1} ([CŠ06, Prop 2.3])
    Upstream of Proposition 3.2, the paper's central moment formula.
  • standard math Schur–Weyl duality / commutant structure: {U^{⊗n}: U ∈ U(d)}′ is the permutation algebra ([GW09])
    Justifies the Jucys–Murphy polynomial representation of the averaged n-fold states (Section 1.1.2).
  • standard math Meckes concentration for Lipschitz functions of independent Haar unitaries ([Mec19, Thm 5.17], Eq 7.14)
    Load-bearing for Lemmas 7.2–7.3, Lemma 4.1, and all typical-spectrum/entropy/rank separation statements.
  • standard math Prouhet–Tarry–Escott / Thue–Morse: existence of K = 2^{k−1} integers with matched power sums to degree k−1 (Prop 5.2)
    Provides the exact moment matching that makes the log-likelihood ratio start at degree k.
  • standard math Asymptotic freeness of independent Haar-unitarily conjugated projections ([CM14])
    Yields the limiting law and the atom at zero used in the rank separation (Lemma 9.1).
  • standard math Atom formula for free multiplicative convolution (μ⊠ν)({0}) = max{μ({0}), ν({0})} ([Bel03])
    Computes μ_b({0}) = 1 − 1/(2^k − 1), the key input for rank separation.
  • standard math Marchenko–Pastur law and Page-curve entropy asymptotics for Haar-induced states ([Nec07], [Wei17], [VPO16], [Sen96])
    Used in the warmup (Lemma 4.1) to show rank-d/2-projector and Haar-marginal ensembles have separated spectra and entropies.
  • domain assumption Density-matrix model, n-copy access with arbitrary (entangled) measurements, worst-case copy complexity (Sections 1–2)
    Defines the tasks and the quantity being lower-bounded; the standard model for quantum property testing.
  • domain assumption The tester may implement the depolarizing channel 𝒩_p (Eq 8.1) on its copies
    Needed for the entropy lower-bound reduction (Prop 5.9 → Thm 1.3); standard and implementable.
  • domain assumption Dimension divisibility: d divisible by every a_i, b_i; otherwise embed in d′ ≤ d with constant-factor loss (footnote to Def 3.4)
    Keeps the PRP ensembles well-defined for all d up to constant factors.
  • ad hoc to paper The n-tilted law ]PRP_d^{(n)}(a) (Def 5.4) is a valid probability measure whose n-fold moments are exactly E[X^{⊗n}]/E[Tr(X)^n]
    The paper's core methodological device; proven by direct computation (Eq 1.3/5.5), but everything downstream depends on it.
  • ad hoc to paper Sandwiched product form X = Π_1···Π_K···Π_1 yields factorized moments ∏_t f_e(J̃_t) (Prop 3.5)
    The design choice making the log-likelihood analysis tractable; proved but specific to this construction.
invented entities (2)
  • The n-tilted distribution ]PRP_d^{(n)}(a) (Definition 5.4) no independent evidence
    purpose: Produces valid quantum states ρ = X̃/Tr(X̃) whose n-fold tensor moments are exactly proportional to E[X^{⊗n}], decoupling global normalization from the random matrix so that closed-form Jucys–Murphy moment formulas apply
    A new probabilistic object introduced by this paper; all properties used later (Eq 1.3, Prop 5.5, Lemma 7.3) are derived internally, not corroborated externally — but they are fully derived, so this is a specified device rather than a black-box postulate.
  • Sandwiched-product-of-random-projections ensembles PRP_d(a) (Definition 3.4) no independent evidence
    purpose: Hard instances whose n-fold moments factor as ∏_t f_a(J̃_t), enabling moment-matching indistinguishability and, under the tilt, valid state mixtures
    Built from standard constituents (Haar-random projections, [ZS01], [Col05], [CN16]) but the sandwiched form and its use here are novel; its behavior is proven in Prop 3.5.

reviewed 2026-08-03 · how reviews work

0 comments
Cite this review

Pith. "Pith review of Spectrum Estimation is Almost as Hard as Tomography." pith.science (2026). https://pith.science/paper/VRGCINN6

@misc{pith2026260729680,
  author       = {Pith},
  title        = {Pith review of: Spectrum Estimation is Almost as Hard as Tomography},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VRGCINN6}},
  note         = {Machine review of arXiv:2607.29680}
}
Share X Bluesky LinkedIn Reddit HN
abstract

We study the sample complexity of estimating and testing fundamental unitarily invariant properties of unknown quantum states; namely, the tasks of spectrum estimation, von Neumann entropy estimation, and rank-testing. For $d$-dimensional states, and for every $\gamma>0$, we prove a sample complexity lower bound of $\Omega(d^{2-\gamma})$ for spectrum estimation to constant sorted total-variation error, entropy estimation to constant additive error, and rank-testing to constant trace distance. Our hard instances are constructed from sandwiched products of Haar-random projectors, suitably normalized using a novel technique that lets us derive explicit expressions for high-order tensor moments of the resultant states. These moments can be expressed as symmetric functions of Jucys--Murphy elements of the symmetric group algebra. To show that two such mixtures are indistinguishable, we analyze the log-likelihood ratio and perform moment-matching, i.e., we set its low-order Jucys--Murphy components to zero. Indistinguishability is then obtained by bounding an $f$-divergence through the high-order components; the non-zero high-order terms and concentration of functions of Haar-random unitaries also imply separations in typical spectra, entropies, and ranks, proving all our lower bounds.

discussion (0)

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

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Nearly tight lower bounds for estimating quantum functionals: Uhlmann fidelity, trace distance, and von Neumann entropy

    quant-ph 2026-08 conditional novelty 7.0

    Estimating Uhlmann fidelity, trace distance, or von Neumann entropy of a d-dimensional quantum state requires Ω̃(d²) samples, matching known upper bounds up to polylog factors.

Reference graph

Works this paper leans on

60 extracted references · 3 linked inside Pith · cited by 1 Pith paper

  1. [1]

    arXiv preprint arXiv:2511.15806 , year=

    Mixed state tomography reduces to pure state tomography , author=. arXiv preprint arXiv:2511.15806 , year=

    arXiv

  2. [2]

    The American Mathematical Monthly , volume=

    Prouhet's 1851 solution of the Tarry-Escott problem of 1910 , author=. The American Mathematical Monthly , volume=. 1959 , publisher=

    1910

  3. [3]

    IEEE Transactions on Information Theory , volume=

    Minimax estimation of functionals of discrete distributions , author=. IEEE Transactions on Information Theory , volume=. 2015 , publisher=

    2015

  4. [4]

    Proceedings of the fortieth annual ACM symposium on Theory of computing , pages=

    Testing symmetric properties of distributions , author=. Proceedings of the fortieth annual ACM symposium on Theory of computing , pages=

  5. [5]

    Journal of Physics A: Mathematical and General , abstract =

    Karol Zyczkowski and Hans-Jürgen Sommers , title =. Journal of Physics A: Mathematical and General , abstract =. 2001 , month =. doi:10.1088/0305-4470/34/35/335 , url =

    doi:10.1088/0305-4470/34/35/335 2001

  6. [6]

    Journal of Mathematical Physics , volume=

    Random matrix techniques in quantum information theory , author=. Journal of Mathematical Physics , volume=. 2016 , publisher=

    2016

  7. [7]

    Probability theory and related fields , volume=

    Product of random projections, Jacobi ensembles and universality problems arising from free probability , author=. Probability theory and related fields , volume=. 2005 , publisher=

    2005

  8. [8]

    The Strong Asymptotic Freeness of

    Beno. The Strong Asymptotic Freeness of. Annales Scientifiques de l'. 2014 , doi =

    2014

  9. [9]

    Integral Equations and Operator Theory , volume =

    Serban Belinschi , title =. Integral Equations and Operator Theory , volume =. 2003 , doi =

    2003

  10. [10]

    2009 , doi =

    Roe Goodman and Nolan Wallach , title =. 2009 , doi =

    2009

  11. [11]

    Proceedings of the 31st Conference on Learning Theory , series =

    Yanjun Han and Jiantao Jiao and Tsachy Weissman , title =. Proceedings of the 31st Conference on Learning Theory , series =

  12. [12]

    IEEE Transactions on Information Theory , volume =

    Yihong Wu and Pengkun Yang , title =. IEEE Transactions on Information Theory , volume =. 2016 , doi =

    2016

  13. [13]

    Journal of the ACM , volume =

    Gregory Valiant and Paul Valiant , title =. Journal of the ACM , volume =. 2017 , doi =

    2017

  14. [14]

    arXiv preprint arXiv:2604.07460 , year =

    Chirag Wadhwa and Sitan Chen , title =. arXiv preprint arXiv:2604.07460 , year =

    Pith/arXiv arXiv

  15. [15]

    Proceedings of the 2022 IEEE 63rd Annual Symposium on Foundations of Computer Science (FOCS) , pages =

    Sitan Chen and Jerry Li and Brice Huang and Allen Liu , title =. Proceedings of the 2022 IEEE 63rd Annual Symposium on Foundations of Computer Science (FOCS) , pages =. 2022 , doi =

    2022

  16. [16]

    Proceedings of the 58th Annual ACM Symposium on Theory of Computing , pages =

    Ryan O'Donnell and Chirag Wadhwa , title =. Proceedings of the 58th Annual ACM Symposium on Theory of Computing , pages =. 2026 , doi =

    2026

  17. [17]

    Entanglement is necessary for optimal quantum property testing , booktitle =

    S. Entanglement is necessary for optimal quantum property testing , booktitle =. 2020 , doi =

    2020

  18. [18]

    Quantum state certification , booktitle =

    Costin B. Quantum state certification , booktitle =. 2019 , doi =

    2019

  19. [19]

    Physical Review Letters , volume =

    Hui Li and Duncan Haldane , title =. Physical Review Letters , volume =. 2008 , doi =

    2008

  20. [20]

    Reviews of Modern Physics , volume =

    Luigi Amico and Rosario Fazio and Andreas Osterloh and Vlatko Vedral , title =. Reviews of Modern Physics , volume =. 2008 , doi =

    2008

  21. [21]

    Physical Review A , volume =

    Charles Bennett and Herbert Bernstein and Sandu Popescu and Benjamin Schumacher , title =. Physical Review A , volume =. 1996 , doi =

    1996

  22. [22]

    Physical Review A , volume =

    Benjamin Schumacher , title =. Physical Review A , volume =. 1995 , doi =

    1995

  23. [23]

    arXiv preprint arXiv:quant-ph/0409113 , year =

    Alexander Klyachko , title =. arXiv preprint arXiv:quant-ph/0409113 , year =

    Pith/arXiv arXiv

  24. [24]

    Science , volume =

    Michael Walter and Brent Doran and David Gross and Matthias Christandl , title =. Science , volume =. 2013 , doi =

    2013

  25. [25]

    2013 , doi =

    Mark Wilde , title =. 2013 , doi =

    2013

  26. [26]

    Communications in Mathematical Physics , volume =

    Matthias Christandl and Graeme Mitchison , title =. Communications in Mathematical Physics , volume =. 2006 , doi =

    2006

  27. [27]

    Physical Review A , volume =

    Masahito Hayashi and Keiji Matsumoto , title =. Physical Review A , volume =. 2002 , doi =

    2002

  28. [28]

    Physical Review Letters , volume =

    Michael Nielsen , title =. Physical Review Letters , volume =. 1999 , doi =

    1999

  29. [29]

    Nature Reviews Physics , volume =

    Anurag Anshu and Srinivasan Arunachalam , title =. Nature Reviews Physics , volume =. 2024 , doi =

    2024

  30. [30]

    Andrew Childs and Aram Harrow and Pawe. Weak. Proceedings of the 24th Annual Symposium on Theoretical Aspects of Computer Science , series =. 2007 , doi =

    2007

  31. [31]

    IEEE Journal on Selected Areas in Information Theory , volume =

    Jayadev Acharya and Ibrahim Issa and Nirmal Shende and Aaron Wagner , title =. IEEE Journal on Selected Areas in Information Theory , volume =. 2020 , doi =

    2020

  32. [32]

    John Wright , title =

  33. [33]

    Proceedings of the 2026 Annual ACM-SIAM Symposium on Discrete Algorithms (SODA) , pages =

    Angelos Pelecanos and Xinyu Tan and Ewin Tang and John Wright , title =. Proceedings of the 2026 Annual ACM-SIAM Symposium on Discrete Algorithms (SODA) , pages =. 2026 , doi =

    2026

  34. [34]

    Communications in Mathematical Physics , volume =

    Ryan O'Donnell and John Wright , title =. Communications in Mathematical Physics , volume =. 2021 , doi =

    2021

  35. [35]

    Proceedings of the 58th Annual ACM Symposium on Theory of Computing , pages =

    Angelos Pelecanos and Jack Spilecki and John Wright , title =. Proceedings of the 58th Annual ACM Symposium on Theory of Computing , pages =. 2026 , doi =

    2026

  36. [36]

    IEEE Transactions on Information Theory , volume =

    Jeongwan Haah and Aram Harrow and Zhengfeng Ji and Xiaodi Wu and Nengkun Yu , title =. IEEE Transactions on Information Theory , volume =. 2017 , doi =

    2017

  37. [37]

    Proceedings of the Forty-Eighth Annual ACM Symposium on Theory of Computing , pages =

    Ryan O'Donnell and John Wright , title =. Proceedings of the Forty-Eighth Annual ACM Symposium on Theory of Computing , pages =. 2016 , doi =

    2016

  38. [38]

    Reviews in Mathematical Physics , volume =

    Michael Keyl , title =. Reviews in Mathematical Physics , volume =. 2006 , doi =

    2006

  39. [39]

    Physical Review A , volume =

    Michael Keyl and Reinhard Werner , title =. Physical Review A , volume =. 2001 , doi =

    2001

  40. [40]

    arXiv preprint arXiv:2607.27117 , year =

    Angelos Pelecanos and Jack Spilecki and Ewin Tang and John Wright , title =. arXiv preprint arXiv:2607.27117 , year =

    Pith/arXiv arXiv

  41. [41]

    2019 , doi =

    Elizabeth Meckes , title =. 2019 , doi =

    2019

  42. [42]

    Typical entanglement entropy in the presence of a center:

    Eugenio Bianchi and Pietro Don. Typical entanglement entropy in the presence of a center:. Physical Review D , volume =. 2019 , doi =

    2019

  43. [43]

    Physical Review E , volume =

    Lu Wei , title =. Physical Review E , volume =. 2017 , doi =

    2017

  44. [44]

    Physical Review Letters , volume =

    Siddhartha Sen , title =. Physical Review Letters , volume =. 1996 , doi =

    1996

  45. [45]

    Physical Review E , volume =

    Pierpaolo Vivo and Mauricio Pato and Gleb Oshanin , title =. Physical Review E , volume =. 2016 , doi =

    2016

  46. [46]

    Introduction to quantum

    D. Introduction to quantum. Quantum Probability and Related Topics , series =. 2011 , doi =

    2011

  47. [47]

    Journal of Algebra , volume =

    Michel Lassalle , title =. Journal of Algebra , volume =. 2013 , doi =

    2013

  48. [48]

    Physical Review Letters , volume =

    Don Page , title =. Physical Review Letters , volume =. 1993 , doi =

    1993

  49. [49]

    Annales Henri Poincar

    Ion Nechita , title =. Annales Henri Poincar. 2007 , doi =

    2007

  50. [50]

    Reports on Mathematical Physics , volume =

    Algimantas Jucys , title =. Reports on Mathematical Physics , volume =. 1974 , doi =

    1974

  51. [51]

    Communications in Mathematical Physics , volume =

    Seamus Albion and Eric Rains and Ole Warnaar , title =. Communications in Mathematical Physics , volume =. 2021 , doi =

    2021

  52. [52]

    Bulletin of the American Mathematical Society , volume =

    Peter Forrester and Ole Warnaar , title =. Bulletin of the American Mathematical Society , volume =. 2008 , doi =

    2008

  53. [53]

    Compositio Mathematica , volume =

    Kevin Kadell , title =. Compositio Mathematica , volume =. 1993 , url =

    1993

  54. [54]

    Free states of the canonical anticommutation relations , journal =

    Robert Powers and Erling St. Free states of the canonical anticommutation relations , journal =. 1970 , doi =

    1970

  55. [55]

    2001 , doi =

    Michel Ledoux , title =. 2001 , doi =

    2001

  56. [56]

    Essential Mathematics for Convex Optimization , publisher =

    Fatma K. Essential Mathematics for Convex Optimization , publisher =. 2026 , doi =

    2026

  57. [57]

    2010 , doi =

    Greg Anderson and Alice Guionnet and Ofer Zeitouni , title =. 2010 , doi =

    2010

  58. [58]

    Michel Ledoux , title =. S. 1999 , doi =

    1999

  59. [59]

    Integration with respect to the

    Beno. Integration with respect to the. Communications in Mathematical Physics , volume =. 2006 , doi =

    2006

  60. [60]

    Eug. M. Comptes rendus hebdomadaires des s

This paper was first reviewed by deepseek-v4-flash on August 3, 2026.