Pith. sign in

REVIEW 3 major objections 4 minor 1 cited by

On Strong Converse Bounds for the Private and Quantum Capacities of Anti-degradable Channels

T0 review · 3 major / 4 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read Anti-degradable quantum channels have a strong converse for private classical capacity: for any allowed error and privacy parameters, the largest message a code can carry is bounded by a constant independent of the number of channel uses.

desk verdict The core idea is right and the proof is mostly clean, but Theorem 2 as stated is false: the parameter domain is too weak and an explicit counterexample kills it. read the letter →

arxiv 2507.15661 v1 pith:FD2DA37I submitted 2025-07-21 quant-ph cs.ITmath.IT

Zahra Baghali Khanian , Christoph Hirche This is my paper · ORCID
classification quant-phcs.ITmath.IT
keywords anti-degradablechannelsprivateclassicalcapacityquantumstrongconverseprettysmoothmin-entropymax-entropychannel
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 proves a strong converse for the private classical capacity of anti-degradable quantum channels, meaning channels whose environment can reproduce Bob's output on its own. For any allowed error $\epsilon>0$ and privacy parameter $\delta>0$ satisfying $\delta\sqrt{1-\epsilon^2}+\epsilon\sqrt{1-\delta^2}<1$, every $n$-use code has message set bounded by $\log M(n,\epsilon,\delta)\le 2\log(1/\cos(\alpha+\beta))$, where $\alpha=\sin^{-1}\epsilon$ and $\beta=\sin^{-1}\delta$. Since the right-hand side does not depend on $n$, positive-rate private communication over such channels is impossible in the strong sense. The same smooth-entropy argument gives a pretty strong converse for the quantum capacity of the same channels for any error $\epsilon<1/\sqrt{2}$, with $\log \mathcal{N}(n,\epsilon)\le \log(1/\cos(2\alpha))$. The result replaces the previous pretty strong converse for private capacity with a full strong converse for an entire class of zero-capacity channels.

What carries the argument

The load-bearing object is the pair of smoothed conditional entropies $H^{\eta}_{\min}(A|B)_{\rho}$ and $H^{\eta}_{\max}(A|B)_{\rho}$, defined by optimizing the ordinary min- and max-entropies over states within purified distance $\eta$. The proof has three steps: (i) in the ideal private code the min-entropy of the classical reference $X'$ given the environment $E^n$ equals $\log M$, while the max-entropy of $X'$ given Bob's decoded output $X$ is $0$; (ii) data processing moves these entropies between $B^n$ and $E^n$, which is exactly where anti-degradability enters, since the channel's output can be produced from its environment; and (iii) Lemma 1, the inequality $H_{\min}^{\sin\alpha}(A|B)\le H_{\max}^{\sin\beta}(A|B)+\log(1/\cos^2(\alpha+\beta))$ for $\alpha+\beta<\pi/2$, converts the two smoothing parameters into a log-cosine constant. The quantum-capacity case is the same chain with $\beta=\alpha$ and uses the duality between smooth min- and max-entropy.

What would settle it

Exhibit any finite-$n$ private code on an anti-degradable channel (for example the erasure channel with erasure probability at least $1/2$) whose error and privacy satisfy $\delta\sqrt{1-\epsilon^2}+\epsilon\sqrt{1-\delta^2}<1$ but whose message set exceeds $2\log(1/\cos(\alpha+\beta))$; that single instance would falsify Theorem 2. Alternatively, a direct counterexample to Lemma 1 on a subnormalized bipartite state within its stated domain would break both capacity bounds.

Watch

Extended reading notes

Core claim

The central claim is that anti-degradability turns the private capacity problem into a sandwich of smooth entropies with an $n$-independent output. Namely, Theorems 2 and 3 show that for a finite-dimensional anti-degradable channel $\mathcal{N}$, every integer $n$, and any error/privacy pair with $\delta\sqrt{1-\epsilon^2}+\epsilon\sqrt{1-\delta^2}<1$, the private classical code size obeys $\log M(n,\epsilon,\delta)\le 2\log(1/\cos(\alpha+\beta))$, with $\alpha=\sin^{-1}\epsilon$ and $\beta=\sin^{-1}\delta$. Because the bound contains no $n$, any family of codes whose rates stay above zero must fail either reliability or privacy; hence the private classical capacity of anti-degradable channels is strongly zero. The same machinery gives the quantum-capacity corollary: for $\epsilon<1/\sqrt{2}$, $\log\mathcal{N}(n,\epsilon)\le \log(1/\cos(2\alpha))$, a pretty strong converse with a simpler proof than the previously known one.

Load-bearing premise

The whole bound rests on the quoted smooth min/max entropy inequality of Lemma 1; if that inequality fails for some subnormalized state or for some pair of smoothing parameters with $\alpha+\beta<\pi/2$, the constant in Theorems 2 and 3 no longer follows.

Editorial extensions

If this is right

  • For every anti-degradable channel, any sequence of private codes whose message size grows faster than a constant must violate either the error or the privacy condition in the limit; in particular, the private classical capacity is strongly zero.
  • The finite-$n$ bound holds for all codes at once, so it also restricts one-shot and short-block private communication on anti-degradable channels, not just the asymptotic rate.
  • The quantum-capacity result extends the pretty strong converse for anti-degradable channels from $\epsilon<1/2$ to $\epsilon<1/\sqrt{2}$, with an explicit $n$-independent bound $\log(1/\cos(2\alpha))$.
  • The sufficient condition $\delta\sqrt{1-\epsilon^2}+\epsilon\sqrt{1-\delta^2}<1$ covers in particular all pairs with $\delta+\epsilon<1$, showing that the privacy parameter can be arbitrarily small while the error is close to $1$, and vice versa.

Reading between the lines

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

  • An extension left implicit by the paper: because the bound is finite-$n$ and fully explicit, it automatically supplies a one-shot private-capacity upper bound for anti-degradable channels, which could be compared against achievable one-shot rates to test tightness.
  • The same smooth min/max template may prove strong converses for any channel class in which Bob's output can be simulated from the environment (a hierarchy of 'simulatable' channels), provided the corresponding data-processing chain survives; this goes beyond what the paper itself claims.
  • Optimizing the smoothing parameters or replacing Lemma 1 by a sharper min/max inequality could move the threshold toward the full boundary $\delta+\epsilon<1$ and improve the constant; the authors do not claim their constant is optimal.
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 / 4 minor

Summary. The paper studies finite-blocklength converse bounds for anti-degradable quantum channels. Its main result, Theorem 2, claims a strong converse for the private classical capacity: for any error epsilon > 0 and privacy delta > 0 with delta sqrt(1 - epsilon^2) + epsilon sqrt(1 - delta^2) < 1, every integer n satisfies log M(n, epsilon, delta) <= 2 log(1/cos(alpha + beta)), where alpha = arcsin(epsilon) and beta = arcsin(delta). The proof combines a smooth min-entropy lower bound from the privacy condition, data processing from the environment to Bob (using anti-degradability), a smooth max-entropy upper bound from the decoding condition, and the smooth min/max entropy inequality of Lemma 1. The paper also proves, in Theorem 3, a pretty strong converse for the quantum capacity of anti-degradable channels: log N(n, epsilon) <= log(1/cos(2 alpha)) for any epsilon < 1/sqrt(2). The quantum result and the proof strategy are transparent, but the domain of validity of Theorem 2 is incorrect as stated.

Significance. If corrected, the proof technique would provide a clean finite-n converse bound for anti-degradable channels using only standard smooth-entropy tools, and the quantum-capacity bound is a nice simplification of earlier work. The paper is explicit and the derivations are easy to follow. However, the central private-capacity theorem as stated is false: the condition in Theorem 2 is strictly weaker than the condition under which Lemma 1 is invoked, and a concrete constant-channel counterexample falsifies the claimed bound. The result can be repaired by replacing the condition with epsilon^2 + delta^2 < 1 (equivalently alpha + beta < pi/2), but the abstract, Theorem 2, Figure 1, and the discussion of the boundary all need substantial revision. With that correction the paper would still be a meaningful contribution, though the claimed sharp boundary would be lost.

major comments (3)
  1. [IV, Theorem 2 and Eqs. (20)-(22)] The theorem's stated condition is delta sqrt(1 - epsilon^2) + epsilon sqrt(1 - delta^2) < 1, i.e. sin(alpha + beta) < 1. For alpha, beta in [0, pi/2], this also holds when alpha + beta > pi/2, but Lemma 1 is valid only for alpha + beta < pi/2. In the regime alpha + beta > pi/2 the quantity cos(alpha + beta) is negative and the claimed bound 2 log(1/cos(alpha + beta)) is not real. The proof acknowledges this with 'it holds for alpha + beta < pi/2' but then incorrectly concludes the bound on the larger region. The correct condition is alpha + beta < pi/2, equivalently epsilon^2 + delta^2 < 1. The error is load-bearing: for the anti-degradable constant channel N(rho) = Tr(rho)|0><0| with isometry U|i> = |0>|e_i>, taking n = 1, epsilon = delta = 0.95, and M = 3, encoding all messages to the same input and decoding to a fixed output gives purified error sqrt(1 - 1/9) <= 0.95 and privacy delta = 0, so M(1, 0.95, 0.95) >= 3. The theorem as stated would give M <= 1.54, a contradiction. The proof cannot be repaired by a sign convention; the lemma genuinely does not apply when alpha + beta > pi/2.
  2. [I and Figure 1] The introduction's statement that the bound 'can be simplified to delta + epsilon < 1' is not a simplification of the stated condition sin(alpha + beta) < 1; the latter holds, for example, for epsilon = delta = 0.95, where delta + epsilon > 1. With the necessary correction epsilon^2 + delta^2 < 1, the relation delta + epsilon < 1 is merely a sufficient subregion. Figure 1 currently shades the region sin(alpha + beta) < 1, which includes the invalid region alpha + beta > pi/2; it should instead show the disk epsilon^2 + delta^2 < 1. The abstract's phrase 'sharply defining the boundary' is also overstated: the corrected theorem gives a sufficient condition for a finite-n bound, not a sharp characterization of all parameters for which private communication is impossible.
  3. [IV, first paragraph] The parenthetical assertion that the private classical capacity of anti-degradable channels is equal to 0 is made without proof or citation. This fact is used to frame the result as a strong converse for the capacity, so it should be justified, either by a reference or by a short argument from the definition of anti-degradability (the environment can simulate Bob's output).
minor comments (4)
  1. [Eq. (21)] In the application of Lemma 1, the roles of alpha and beta are exchanged relative to the lemma statement: the proof bounds H_min^{sin beta} by H_max^{sin alpha} plus the logarithmic term. This is valid after renaming the lemma's parameters, but a sentence noting the exchange would improve readability.
  2. [Theorem 3 and Eq. (25)] The statement gives log N(n, epsilon) <= log(1/cos(2 alpha)), while the proof derives 2 log N <= log(1/cos^2(2 alpha)); the equivalence is immediate but should be made explicit to avoid confusion.
  3. [Discussion and Acknowledgments] There are minor typos: 'for the later' should be 'for the latter', and the Acknowledgments contain a duplicated 'and'. These do not affect the content.
  4. [Abstract and Theorem 2] After the condition is corrected to epsilon^2 + delta^2 < 1, the abstract should state the condition explicitly; the current wording 'whenever ... satisfy ...' will otherwise continue to assert a false domain of validity.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the proofs derive the converse bounds from an external smooth-entropy inequality and data processing, not from the assumptions or conclusions.

full rationale

The derivation chain in this paper is self-contained relative to standard external results, and no circular step is present. The central bound in Theorem 2 is obtained by adding two entropy inequalities: Eq. (18) uses data processing for anti-degradable channels to go from the environment to Bob's output, and Eq. (19) uses data processing on the decoding map; Lemma 1 is then invoked to relate smoothing parameters. Lemma 1 is an external published result (Proposition 5.5 of Tomamichel's PhD thesis), not authored by the present authors, and it does not assume the strong converse or any capacity bound. Similarly, Theorem 3 uses the same external lemma together with duality and data processing. There are no fitted parameters, no quantities defined in terms of the result being proved, and no self-citations carrying the argument. The skeptical concern that Theorem 2's stated parameter region is broader than the α+β<π/2 validity condition of Lemma 1 is a possible correctness/domain error in the implication, not a circularity: the proof does not assume its conclusion, it merely applies an external inequality on a domain that may be too large. Under the reviewing rules, mathematical validity concerns of that kind belong to correctness risk rather than to the circularity score. Accordingly, the appropriate circularity finding is a clean non-finding with score 0.

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

The paper introduces no free parameters and no new entities. The proof relies on standard smooth entropy relations and the standard definition of anti-degradability. The only external input with a sharp parameter tradeoff is Lemma 1, which is cited from Tomamichel's thesis. No circular dependence on the authors' own prior results appears.

assumptions (4)
  • standard math Lemma 1: H_min^{sin alpha}(A|B) <= H_max^{sin beta}(A|B) + log(1/cos^2(alpha+beta)) for alpha+beta < pi/2, cited as Proposition 5.5 of Tomamichel's thesis.
    Used in Eq. (20)-(22) and (25) to combine min- and max-entropy bounds. All constants in the final bounds come from this lemma.
  • standard math Data processing inequality for smooth conditional min- and max-entropies under CPTP maps on the conditioning system.
    Used to move from the environment system to Bob's output via the degrading map in (18) and (24), and from Bob's output to the decoded message in (19) and (23).
  • standard math Smooth min/max entropy duality: -H_max^epsilon(A|B) = H_min^epsilon(A|C) for a purification.
    Used in Eq. (24) to convert the max-entropy bound on Bob's system into a min-entropy bound on the environment system.
  • domain assumption Anti-degradable channel definition: there exists a CPTP map M such that N = M composed with N_c, so Bob's output is a post-processing of the environment.
    This is the defining property of the channel class under study and enables the data-processing steps that transfer entropy bounds from E^n to B^n.

how reviews work

0 comments
Cite this review

Pith. "Pith review of On Strong Converse Bounds for the Private and Quantum Capacities of Anti-degradable Channels." pith.science (2026). https://pith.science/paper/FD2DA37I

@misc{pith2026250715661,
  author       = {Pith},
  title        = {Pith review of: On Strong Converse Bounds for the Private and Quantum Capacities of Anti-degradable Channels},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FD2DA37I}},
  note         = {Machine review of arXiv:2507.15661}
}
abstract

We establish a strong converse bound for the private classical capacity of anti-degradable quantum channels. Specifically, we prove that this capacity is zero whenever the error $\epsilon > 0$ and privacy parameter $\delta > 0$ satisfy the inequality $\delta (1-\epsilon^2)^{\frac{1}{2}}+\epsilon (1-\delta^2)^{\frac{1}{2}}<1$. This result strengthens previous understandings by sharply defining the boundary beyond which reliable and private communication is impossible. Furthermore, we present a ``pretty simple'' proof of the ``pretty strong'' converse for the quantum capacity of anti-degradable channels, valid for any error $\epsilon < \frac{1}{\sqrt{2}}$. Our approach offers clarity and technical simplicity, shedding new light on the fundamental limits of quantum communication.

Figures

Figures reproduced from arXiv: 2507.15661 by the authors.

Figure 1
Figure 1. δ vs ϵ: for any δ and ϵ satisfying δ √ 1 − ϵ 2 + ϵ √ 1 − δ 2 < 1 the converse bound holds. for any error ϵ and privacy δ satisfying δ √ 1 − ϵ 2 + ϵ √ 1 − δ 2 < 1 and every integer n, log M(n, ϵ, δ) ≤ 2 log 1 cos(α + β)  , where β = sin−1 (δ) and α = sin−1 (ϵ). Proof. For the target state ρ E nX′ = ω E n ⊗ P m 1 M |m⟩⟨m| X′ , the environment is decoupled from the classical reference, hence Hmin(X′ |En)ρ = log M. We… view at source ↗

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

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

  1. A strong converse for stabilizer codes over Pauli channels via the blowing-up lemma

    quant-ph 2026-07 accept novelty 7.0 of 10

    Above the coherent information of its own input, any stabilizer code over a product Pauli channel has entanglement fidelity decaying exponentially in block length.

Reference graph

Works this paper leans on

14 extracted references · 13 canonical work pages · cited by 1 Pith paper

  1. [1]

    The quantum channel capacity and coherent information,

    P. W. Shor, “The quantum channel capacity and coherent information,” inLecture Notes, MSRI Workshop on Quantum Computation (2002)

    work page 2002

  2. [2]

    On Strong Converse Bounds for the Private and Quantum Capacities of Anti-degradable Channels

    Our approach offers clarity and technical simplicity, shedding new light on the fundamental limits of quantum communication. I. INTRODUCTION The private classical capacity of a quantum channel is the maximum rate at which classical information can be transmitted reliably and privately, ensuring that an eavesdropper has no knowledge of the transmitted info...

    work page Pith review arXiv 2025

  3. [3]

    The paper is structured as follows: In Section II we introduce the necessary notation and preliminary results

    We provide a slightly stronger result and a simpler proof: a pretty strong converse holds for any errorϵ < 1√ 2. The paper is structured as follows: In Section II we introduce the necessary notation and preliminary results. In Section III we define the quantum and private capacities. Then in Sections IV and V we prove our results regarding the private and...

    work page

  4. [4]

    (6) A channel N : A → B is called degradable if it can be degraded to its complementary channel, i.e

    Then H sin(α) min (A|B)ρ ≤ H sin(β) max (A|B)ρ + log 1 cos2(α + β) . (6) A channel N : A → B is called degradable if it can be degraded to its complementary channel, i.e. if there exists a CPTP mapM such that Nc = M ◦ N. Introducing the Stinespring dilation ofM by an isometry V : B ,→ B′E′, the channel output systemB can be mapped to the composite systemB...

    work page

  5. [5]

    This relation is illustrated in Fig

    This implies that for anyδ and ϵ satisfying δ √ 1 − ϵ2 + ϵ √ 1 − δ2 < 1 the strong converse holds. This relation is illustrated in Fig. 1. ■ V. PRETTY STRONG CONVERSE FOR THE QUANTUM CAP ACITY In this section, we provide a simple proof that a pretty strong converse bound holds for the quantum capacity of anti-degradable channels (the quantum capacity of a...

    work page

  6. [6]

    Workshop on Information Theory and Related Fields: In Memory of Ning Cai

    This implies that for anyϵ < 1√ 2 the pretty strong converse holds. ■ VI. DISCUSSION In this work we report progress on the problem of proving a strong converse for the private and quantum capacity. In particular we prove a strong converse for the former and a pretty strong converse for the later for the particular choice of anti-degradable channels. Whil...

    work page

  7. [7]

    Capacity of the noisy quantum channel,

    Seth Lloyd, “Capacity of the noisy quantum channel,” Physical Review A55, 1613–1622 (1997)

    work page 1997

  8. [8]

    The private classical capacity and quantum capacity of a quantum channel,

    I. Devetak, “The private classical capacity and quantum capacity of a quantum channel,” IEEE Trans. Inf. Theory 51, 44–55 (2005)

    work page 2005

Show all 14 references
  1. [9]

    “pretty strong

    C. Morgan and A. Winter, ““pretty strong” converse for the quantum capacity of degradable channels,” IEEE Trans. Inf. Theory60, 317–333 (2014)

    2014

  2. [10]

    Resource theory of unextendibility and nonasymptotic quantum capacity,

    E. Kaur, S. Das, M. Wilde, and A. Winter, “Resource theory of unextendibility and nonasymptotic quantum capacity,” Phys. Rev. A104, 022401 (2021)

    2021

  3. [11]

    Cryptographic distinguishability measures for quantum-mechanical states,

    C. A. Fuchs and J. van de Graaf, “Cryptographic distinguishability measures for quantum-mechanical states,” IEEE Trans. Inf. Theory45, 1216–1227 (1999)

    1999

  4. [12]

    Tomamichel,Quantum Information Processing with Finite Resources (Springer Cham, 2016)

    M. Tomamichel,Quantum Information Processing with Finite Resources (Springer Cham, 2016)

    2016

  5. [13]

    Duality between smooth min- and max-entropies,

    M. Tomamichel, R. Colbeck, and R. Renner, “Duality between smooth min- and max-entropies,” IEEE Trans. Inf. Theory 56, 4674–4681 (2010)

    2010

  6. [14]

    Tomamichel,A framework for non-asymptotic quantum information theory , PhD thesis, ETH Zurich, Dept

    M. Tomamichel,A framework for non-asymptotic quantum information theory , PhD thesis, ETH Zurich, Dept. Phys., Switzerland (2011), arXiv:quant-ph/1203.2142

    2011 arXiv

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.