Pith. sign in

REVIEW 1 major objections 6 minor 19 references

Upper bounds on the Holevo quantity arising from the fundamental entropic inequality

T0 review · 1 major / 6 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read For any ensemble of quantum states, the Holevo quantity is bounded by a weighted combination of two auxiliary Holevo quantities plus a mean binary-entropy term.

desk verdict The main Holevo upper bound (Theorem 1) is clean, correct, and useful, but Proposition 2's claimed improvement has a genuine misweighting error in the Pinsker lower bound and should not be cited as stated. read the letter →

arxiv 2506.05335 v2 pith:DYYWPH6X submitted 2025-06-05 quant-ph cs.ITmath-phmath.ITmath.MP

M.E.Shirokov This is my paper
classification quant-phcs.ITmath-phmath.ITmath.MP MSC 81P4594A17 PACS 03.67.-a
keywords Holevoquantityfundamentalentropicinequalityquantumensembletrace-normdistancevonNeumannentropybinaryupperboundsinformation
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 aims to prove that the Holevo quantity of any discrete or continuous ensemble of quantum states—the standard upper bound on the classical information extractable by measurement—can be bounded from above using two auxiliary ensembles built from the positive and negative parts of the deviations of each state from the ensemble average. The main inequality states that $\chi(\mu) \le \varepsilon_{\rm av}(\chi(\mu_+)-\chi(\mu_-)) + \bar h(\mu)$, where $\varepsilon_{\rm av}$ is the mean trace-norm distance from the states to the average state and $\bar h(\mu)$ is the mean binary entropy of those distances. Because the auxiliary Holevo quantities are themselves bounded by Shannon entropies, the paper derives easy-to-compute upper bounds expressed only through the probabilities and metric characteristics of the ensemble. A sympathetic reader cares because the exact Holevo quantity usually requires full entropy calculations, while the new bounds are elementary and, as the examples show, reasonably tight.

What carries the argument

The central object is the fundamental entropic inequality $S(\rho)+\varepsilon S(\tau_-) \le S(\sigma)+\varepsilon S(\tau_+) + h(\varepsilon)$, applied termwise to each state $\rho_i$ in the ensemble with $\sigma = \bar\rho(\mu)$. The construction of $\mu_+$ and $\mu_-$ is the second load-bearing element: defining $\varepsilon_i = \tfrac12\|\rho_i-\bar\rho(\mu)\|_1$ and $\tau_i^{\pm} = \varepsilon_i^{-1}[\rho_i-\bar\rho(\mu)]_{\pm}$ gives normalized states, and the identity $\sum_i p_i\varepsilon_i(\tau_i^- - \tau_i^+) = 0$ forces $\bar\rho(\mu_+)=\bar\rho(\mu_-)$, so the two auxiliary Holevo quantities share one average state. That common-average identity is what converts the entropy differences into $\chi(\mu_+)-\chi(\mu_-)$.

What would settle it

Compute both sides of the fundamental entropic inequality (1) for a pair of finite-dimensional states with known entropies, for example two noncommuting qubit states with a small overlap; a single pair for which the right-hand side falls below the left-hand side would invalidate the derivation of (7). For a direct test of the application, compute $\chi(\mu)-[\varepsilon_{\rm av}H(\{p_i\varepsilon_i/\varepsilon_{\rm av}\})+\sum_i p_i h(\varepsilon_i)]$ over a random search of three-state qubit ensembles, since any positive difference would refute Proposition 1.

Watch

Extended reading notes

Core claim

For an ensemble $\mu$ with finite-entropy average state, the paper proves the bound $\chi(\mu) \le \varepsilon_{\rm av}(\chi(\mu_+)-\chi(\mu_-)) + \bar h(\mu)$, with equality attainable for equiprobable mutually orthogonal pure states. The two auxiliary ensembles $\mu_+$ and $\mu_-$ are obtained by normalizing the positive and negative parts of $\rho_i - \bar\rho(\mu)$ with weights $p_i\varepsilon_i/\varepsilon_{\rm av}$, and they share the same average state, which is what makes the cancellation in the proof possible. From this, the paper derives Proposition 1, giving $\chi(\mu) \le \varepsilon_{\rm av} H(\{p_i\varepsilon_i/\varepsilon_{\rm av}\}) + \sum_i p_i h(\varepsilon_i)$ under only finite average entropy or finite Shannon entropy; Corollary 1 restricts this to $\chi \le \varepsilon_{\rm av}\log m + h(\varepsilon_{\rm av})$ for $m$-state ensembles; and Proposition 2 refines the bound using the deviation-operator diameter $C(\mu)$ and a relative-entropy lower bound $D(\mu)$ for $\mu_-$. Theorem 2 extends the same inequality to continuous ensembles under continuity of the state family, using an approximation technique.

Load-bearing premise

The main bound inherits all its force from the fundamental entropic inequality (1), which the paper uses as a black box from [2] and does not prove itself; if that inequality fails for some pair of states, or if $S(\bar\rho(\mu))$ is infinite, the proof of (7) does not go through.

Editorial extensions

If this is right

  • If Theorem 1 is correct, computing an upper bound on the Holevo quantity of an ensemble requires only trace-norm distances and binary or Shannon entropies, not full von Neumann entropies of the member states.
  • The finite-dimensional corollary yields the simple estimate $\chi \le \varepsilon_{\rm av}\log m + h(\varepsilon_{\rm av})$, refining earlier distance-dependent upper bounds.
  • The refined bound (18) shows that knowledge of the shapes of the deviation operators $[\rho_i-\bar\rho]_{\pm}$ improves the estimate, with the improvement measured by $C(\mu)$ and $D(\mu)$.
  • The continuous-ensemble Theorem 2 extends the same control to generalized ensembles, so the bound applies to infinite-dimensional settings when the state family is continuous.
  • Equality in (7) is achieved by equiprobable mutually orthogonal pure states, showing that the inequality is tight rather than merely asymptotic.

Reading between the lines

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

  • The paper does not state this, but the bound suggests a continuity-type estimate: if all states of an ensemble lie within trace distance $\varepsilon$ of the average state, then $\chi(\mu)$ is at most about $\varepsilon\log m + h(\varepsilon)$, so small metric fluctuations force the Holevo quantity toward zero in a quantified way.
  • One could test the bound as a fast numerical diagnostic for channel output ensembles: compare the exact $\chi$ with the bound (12) under coarse-graining, and check whether the bound remains close under perturbations of the ensemble.
  • The equality case suggests that the bound is loosest for highly non-orthogonal or highly mixed ensembles; a natural extension is to characterize the ensembles for which the gap between $\chi(\mu)$ and the bound is maximal, which would sharpen its use in finite-blocklength information analysis.
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

1 major / 6 minor

Summary. The paper derives, from the fundamental entropic inequality of Audenaert et al., a relation for the Holevo quantity of an ensemble μ: χ(μ) ≤ ε_av(χ(μ+)−χ(μ−)) + h̄(μ) (Theorem 1). The auxiliary ensembles μ± are built from the positive and negative parts of the operators ρ_i − ρ̄(μ), weighted by ε_i = ½||ρ_i−ρ̄(μ)||₁. It then uses this inequality to obtain explicit upper bounds (12), (17), and an 'improved' bound (18) depending on additional metric data. A continuous version (Theorem 2) is stated for generalized ensembles over a metric space.

Significance. If established, the main inequality is a clean and broadly applicable tool: it gives parameter-free, computable upper bounds on the Holevo quantity in terms of trace distances and Shannon entropies, and it recovers tight equality for ensembles of orthogonal pure states. The derivation of Theorem 1 is a short, correct application of the cited fundamental inequality, and Proposition 1 and Corollary 1 follow from it. However, the announced improvement in Proposition 2 currently contains a scaling error in the Pinsker-type lower bound, so that advertised part of the paper is not yet valid as written. Since the flaw is local and fixable, the paper is worth major revision.

major comments (1)
  1. [Section 3.2, Eqs. (18)–(19)] The proof of Proposition 2 asserts χ(µ−) ≥ D(µ) with D(µ) = (log e / 2) Σ_i p_i ||ε_i^{-1}[ρ_i−ρ̄(µ)]_− − ρ̄(µ−)||_1². This is not the lower bound delivered by the standard relative-entropy/trace-distance inequality. The ensemble µ− has probabilities q_i = p_i ε_i / ε_av, so χ(µ−) = Σ_i q_i D(τ_i^- || ρ̄(µ−)) ≥ (log e/2) Σ_i q_i ||τ_i^- − ρ̄(µ−)||_1² = (log e/(2ε_av)) Σ_i p_i ε_i ||τ_i^- − ρ̄(µ−)||_1². The factor ε_i/ε_av is missing from D(µ). Consequently ε_av D(µ) is not in general a lower bound on ε_av χ(µ−), and the subtraction in (18) can make the claimed upper bound smaller than the true χ(µ). Example 2 does not expose the problem because all ε_i are equal there. Please replace D(µ) by the correctly weighted quantity (or adjust the subtracted term accordingly) and update the example; with this correction the proof strategy appears valid.
minor comments (6)
  1. [Section 2.1] The states τ_i^± are not defined when ε_i = 0; since such indices have zero weight in µ±, please state that they may be set to an arbitrary fixed state (as is done in the continuous case) or simply omitted.
  2. [Section 2.2] The sentence 'By the obvious modification of the arguments from Section 3.1' should refer to Section 2.1.
  3. [Theorem 2] The proof of Theorem 2 is only a pointer to the approximation technique in [17, Proposition 9]; please expand it with enough detail to verify that the finiteness assumptions and the ε_x=0 convention are preserved under the approximation.
  4. [Eq. (6)] The notation in Eq. (6) defines ε_i inside the displayed summation, which is easy to misread; please define ε_i before the sum and give the analogous definition for ε_x before Eq. (10).
  5. [References] Reference [11] contains the typo 'strong suadditivity'; it should read 'strong subadditivity'.
  6. [Proposition 1] In the statement of Proposition 1, the case H({p_i})<+∞ should clarify that the Holevo quantity is then understood via the relative-entropy sum, since the entropy-difference formula may not be well defined when S(ρ̄(µ)) is infinite.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the main bound follows from an external inequality with no fitted parameters or assumed target result.

full rationale

Theorem 1 derives inequality (7) by substituting the cited fundamental entropic inequality (1) from [2] into the identity for the Holevo quantity, then identifying the resulting terms as the Holevo quantities of the auxiliary ensembles μ+ and μ− using (5). This is a direct implication, not a reformulation of the conclusion. Proposition 1 and Corollary 1 use only standard bounds (χ ≤ H and concavity of binary entropy). The self-citations to [16] and [17] are used for comparison and approximation, not as load-bearing proofs of the target bound. The tightness example is independently computed. The only notable technical issue is in Proposition 2, where inequality (19) appears to use unweighted p_i in D(μ) while the μ− ensemble has weights q_i = p_i ε_i / ε_av; if correct, this would be a correctness gap in an auxiliary improvement, not circularity. Consequently, the paper's central derivation is self-contained relative to an external lemma and receives a circularity score of 0.

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

The only external input is the fundamental entropic inequality from [2]; no new entities, no fitted parameters. The auxiliary ensembles μ± are derived from the given ensemble, not postulated independently.

assumptions (4)
  • domain assumption Fundamental entropic inequality (Eq. (1)): S(ρ) + ε S(τ-) ≤ S(σ) + ε S(τ+) + h(ε) for ε = ½∥ρ-σ∥₁, with τ± = (1/ε)[ρ-σ]±, valid for any states ρ, σ (possibly with +∞ values).
    Stated as proved in [2]; the paper inherits this result without independent proof. If this inequality is false or requires extra conditions, Theorem 1 fails.
  • domain assumption Finiteness conditions: S(ρ̄(μ)) < +∞ (Theorems 1, 2) or H({p_i}) < +∞ (Proposition 1).
    The Holevo quantity is expressed as S(ρ̄) - Σ p_i S(ρ_i); this requires the finite entropy condition or is handled by approximation under finite Shannon entropy.
  • standard math Standard facts: lower semicontinuity of relative entropy, concavity of binary entropy, and Simon's dominated convergence theorem for Shannon entropy.
    Used in the approximation argument of Proposition 1; these are classical results in quantum information and functional analysis.
  • domain assumption For the continuous case: X is a metric space, ρ_x is continuous, and the approximation technique of Proposition 9 in [17] applies.
    Theorem 2's proof is delegated to this technique; the assumptions ensure the discrete approximations converge to the continuous ensemble.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Upper bounds on the Holevo quantity arising from the fundamental entropic inequality." pith.science (2026). https://pith.science/paper/DYYWPH6X

@misc{pith2026250605335,
  author       = {Pith},
  title        = {Pith review of: Upper bounds on the Holevo quantity arising from the fundamental entropic inequality},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DYYWPH6X}},
  note         = {Machine review of arXiv:2506.05335}
}
abstract

We show how the fundamental entropic inequality proved recently in [arXiv:2408.15306] can be applied to obtain a useful relation for the Holevo quantity of discrete and continuous ensembles of quantum states. This relation gives a tight upper bound on the Holevo quantity of a given ensemble $\mu$ expressed in terms of the Holevo quantities of two auxiliary ensembles $\mu_+$ and $\mu_-$ produced by $\mu$. Among others, this implies quite accurate upper bounds on the Holevo quantity of a discrete ensemble of quantum states expressed via the probabilities and the metric characteristics of an ensemble.

Figures

Figures reproduced from arXiv: 2506.05335 by the authors.

Figure 1
Figure 1. The values of χ(µ(N)) and χb(µ(N)) as functions of N number of quanta N and, hence, the Holevo quantity χ(µ(N)) is equal to g(N) = (N + 1) log(N + 1) − N log(N). Since 1 2 ∥γ(N) − |n⟩⟨n|∥1 = 1 − (1 − qN )q n N , the upper bound (12) corresponding to the above ensemble is equal to χb(µ(N)) = a(N)H  λn(1 − λn) a(N) +∞ n=0! + X +∞ n=0 λnh(1 − λn), a(N) = 2qN 1 + qN . Both quantities χ(µ(N)) and χb(µ(N)) are easily ca… view at source ↗

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

19 extracted references · 18 canonical work pages

  1. [1]

    Quantum Skew Divergence

    K.M.R. Audenaert ”Quantum Skew Divergence”, J. Math. Phys. 55, 112202 (2014); arXiv:1304.5935

    work page Pith review arXiv 2014

  2. [2]

    Audenaert, B.Bergh, N.Datta, M.G.Jabbour, ´A.Capel, P.Gondolf, ”Con- tinuity bounds for quantum entropies arising from a fundamental entropic inequal- ity”, arXiv:2408.15306

    K.M.R. Audenaert, B.Bergh, N.Datta, M.G.Jabbour, ´A.Capel, P.Gondolf, ”Con- tinuity bounds for quantum entropies arising from a fundamental entropic inequal- ity”, arXiv:2408.15306

    arXiv

  3. [3]

    Briet and P

    J. Briet and P. Harremoes, ”Properties of classical and quantum Jensen-Shannon divergence”, Phys. Rev. A 79, 052311 (2009)

    work page 2009

  4. [4]

    N.Datta, T.Dorlas, R.Jozsa, F.Benatti ”Properties of subentropy”, Journal of Mathematical Physics, 55, 062203 (2014); arXiv:1310.1312

    work page Pith review arXiv 2014

  5. [5]

    Pure Appl

    G.F.Dell’Antonio, ”On the limits of sequences of normal states”, Commun. Pure Appl. Math. 20, 413-430, 1967

    work page 1967

  6. [6]

    de Melo, W

    M.Fannes, F. de Melo, W. Roga, and K.Zyczkowski, ”Matrices of fidelities for ensembles of quantum states and the Holevo quantity”, Quantum Inf. Comput. 12(5–6), 472–489 (2012)

    work page 2012

  7. [7]

    A.S.Holevo, ”Bounds for the quantity of information transmitted by a quantum communication channel”, Probl. Inf. Transm. (USSR) V.9, 177-183 (1973)

    work page 1973

  8. [8]

    A mathematical introduc- tion”, Berlin, DeGruyter, 2012

    A.S.Holevo ”Quantum systems, channels, information. A mathematical introduc- tion”, Berlin, DeGruyter, 2012

    work page 2012

Show all 19 references
  1. [9]

    A.S.Holevo, ”Accessible information of a general quantum Gaussian ensemble”, J. Math. Phys., 62(9), 092201 , 13 pp. (2021); arXiv: 2102.01981

    2021 arXiv

  2. [10]

    A.S.Holevo, M.E.Shirokov, ”Continuous ensembles and theχ-capacity of infinite dimensional channels”, Theory of Probability and Applications, V.50, N.1, 86-98 (2005); arXiv:quant-ph/0408176

    2005 arXiv

  3. [11]

    E.H.Lieb, M.B.Ruskai, ”Proof of the strong suadditivity of quantum mechanical entropy”, J.Math.Phys. 1973. V.14, P.1938. 10

    1973

  4. [12]

    M.A.Nielsen, I.L.Chuang ”Quantum Computation and Quantum Information”, Cambridge University Press, 2000

    2000

  5. [13]

    Berlin: Springer-Verlag, 1993

    M.Ohya,D.Petz ”Quantum Entropy and Its Use”, Texts and Monographs in Physics. Berlin: Springer-Verlag, 1993

    1993

  6. [14]

    W.Roga, M.Fannes, and K.Zyczkowski, ”Universal bounds for the Holevo quantity, coherent information and the Jensen-Shannon divergence”, Phys. Rev. Lett. 105, 040505 (2010)

    2010

  7. [15]

    Lu, H.Kinkawa, T.Kuwahara, ”Operator Spreading and Informa- tion Propagation: Equivalence and Beyond”, arXiv:2505.07955

    C.Shang, Z.-G. Lu, H.Kinkawa, T.Kuwahara, ”Operator Spreading and Informa- tion Propagation: Equivalence and Beyond”, arXiv:2505.07955

    arXiv

  8. [16]

    Transmission, 55:3 (2019), 201–217

    M.E.Shirokov, ”Upper bounds for the Holevo quantity and their use”, Problems Inform. Transmission, 55:3 (2019), 201–217

    2019

  9. [17]

    Math., 215:11 (2024), 1549–1581; arXiv:2302.04809

    M.E.Shirokov, ”Lower semicontinuity of relative entropy disturbance and its con- sequences”, Sb. Math., 215:11 (2024), 1549–1581; arXiv:2302.04809

    2024 arXiv

  10. [18]

    Wehrl, A., ”General properties of entropy”, Rev. Mod. Phys.50, 221-250 (1978)

    1978

  11. [19]

    Press, (2013)

    Wilde, M.M., ”Quantum Information Theory”, Cambridge, UK: Cambridge Univ. Press, (2013). 11

    2013

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.