Optimal uniform continuity bound for conditional entropy of classical--quantum states

Mark M. Wilde

Optimal uniform continuity bound for conditional entropy of classical--quantum states

T0 review · 0 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read This paper proves that a classical–quantum state's conditional entropy satisfies the optimal uniform continuity bound $\varepsilon\log_2(d_B-1)+h_2(\varepsilon)$ whenever the two states are within trace distance $\varepsilon$, and that…

desk verdict A clean reduction to Alhejji–Smith proves the optimal cq conditional entropy bound; the proof checks out, the external dependency is honest, and the corollaries are useful — worth refereeing. read the letter →

arxiv 1909.01755 v2 pith:RQZEMCN5 submitted 2019-09-04 quant-ph cs.IThep-thmath-phmath.ITmath.MP

Mark M. Wilde This is my paper · ORCID

classification quant-phcs.IThep-thmath-phmath.ITmath.MP MSC 81P4594A17 PACS 03.67.-a

keywords conditionalentropyclassical-quantumstatesuniformcontinuityboundtracedistancebinaryentanglementofformationdataprocessingoptimal

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 an optimal uniform continuity bound for the conditional entropy of classical–quantum states: whenever two such states are within trace distance $\varepsilon$ and $\varepsilon$ lies in $(0,1-1/d_B]$, their conditional entropies differ by at most $\varepsilon\log_2(d_B-1)+h_2(\varepsilon)$, where $d_B$ is the dimension of the quantum system and $h_2$ is the binary entropy. The proof converts the problem into a classical one by dephasing each conditional quantum state in the eigenbasis of one of the states; this does not decrease the conditional entropy of the comparison state and does not increase the trace distance, so the optimal classical bound applies. The quantum bound inherits optimality from the classical one: for every $d_B$ and every allowed $\varepsilon$, there is a pair of classical–quantum states that satisfies the inequality with equality. An immediate corollary improves the known uniform continuity bound for entanglement of formation, and the argument also covers countably infinite classical conditioning alphabets.

What carries the argument

The key device is the conditional dephasing channel $\Delta^{\mathrm{cd}}_{XB}(\omega_{XB})=\sum_{x,y}(|x\rangle\langle x|_X\otimes|\varphi^{y,x}\rangle\langle\varphi^{y,x}|_B)\omega_{XB}(|x\rangle\langle x|_X\otimes|\varphi^{y,x}\rangle\langle\varphi^{y,x}|_B)$, which dephases the quantum system $B$ in the eigenbasis $\{|\varphi^{y,x}\rangle\}$ of each $\rho_x^B$ after reading the classical label $x$. This channel leaves $\rho_{XB}$ unchanged, maps $\sigma_{XB}$ to a commuting state, and by data processing never increases the normalized trace distance; being unital, it also never decreases the entropy of the full state, so the conditional entropy of the dephased state dominates the original. The reduction turns the quantum inequality into the classical bound $|H(Y|X)_r-H(Y|X)_s|\leq\varepsilon\log_2(|Y|-1)+h_2(\varepsilon)$, and the classical example that saturates that bound provides the saturating classical–quantum pair.

What would settle it

For $d_B=3$ and $\varepsilon=0.3$, the claimed bound is $0.3\log_2(2)+h_2(0.3)\approx1.181$ bits; a numerical search over classical–quantum states with $\frac12\|\rho-\sigma\|_1=0.3$ that produces any pair with a larger conditional-entropy gap would refute Proposition 1, while recovering exactly this gap from the classical saturating construction would confirm it.

Watch

Extended reading notes

Core claim

The central claim is Proposition 1: for finite-dimensional classical–quantum states $\rho_{XB}=\sum_x r(x)|x\rangle\langle x|_X\otimes\rho_x^B$ and $\sigma_{XB}=\sum_x s(x)|x\rangle\langle x|_X\otimes\sigma_x^B$, with $\varepsilon\geq \frac12\|\rho_{XB}-\sigma_{XB}\|_1$ and $\varepsilon\in(0,1-1/d_B]$, one has $|H(B|X)_\rho-H(B|X)_\sigma|\leq \varepsilon\log_2(d_B-1)+h_2(\varepsilon)$. The bound is uniform in that the right-hand side depends only on $\varepsilon$ and $d_B$, and it is optimal in the strongest sense: for every $d_B$ and every $\varepsilon$ in the stated range there exists a pair of states attaining equality. The paper also derives the corresponding uniform continuity bound for entanglement of formation and extends the classical–quantum bound to countable alphabets.

Load-bearing premise

The quantum proof inherits everything from a quoted classical bound; if that classical bound had a gap or a narrower valid range, the claimed quantum result would lose both its range and its tightness.

Editorial extensions

If this is right

For every $d_B$ and every $\varepsilon\in(0,1-1/d_B]$, there exist states meeting the bound exactly, so no uniform bound of the same form can be improved.
The entanglement-of-formation version bounds $|E_F(\rho_{AB})-E_F(\sigma_{AB})|$ by $\delta\log_2(d-1)+h_2(\delta)$ with $\delta=\sqrt{\varepsilon(2-\varepsilon)}$ and $d=\min\{d_A,d_B\}$, valid up to $\varepsilon=1-\sqrt{(2d-1)/d}$, improving the previous bound.
The bound extends without change to a countably infinite classical alphabet $X$ as long as the quantum system remains finite-dimensional.
Tight conditional-entropy estimates of this kind are the standard ingredient for converting approximate closeness of quantum channels into estimates of their communication capacities, so the optimal value sharpens those estimates whenever classical–quantum output states are involved.

Reading between the lines

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

The proof's reliance on eigenbasis dephasing suggests that the open quantum–classical and fully quantum analogues will need a different mechanism, since no single dephasing channel can preserve the entropy structure of both states in those cases.
A testable extension is to apply the same conditional-dephasing reduction to conditional mutual information or other one-sided information measures; if a matching tight classical inequality exists, the same argument would likely produce the optimal quantum bound.
The bound's dependence only on $d_B$, and not on the classical alphabet size, indicates that the conditioning side enters only through probability weights; related one-sided measures may exhibit the same collapse of dimension dependence.

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

Short version: the paper does exactly what the title promises. It proves the optimal uniform continuity bound for the conditional entropy of classical–quantum states, with the correct constant ε log(d_B − 1) + h₂(ε), by reducing the quantum problem to the recent classical equivocation bound of Alhejji and Smith. I checked the reduction line by line and it holds up.

What is new: Winter's earlier bound for this case was suboptimal; here the bound is tight for every ε in (0, 1 − 1/d_B] and every d_B, and the saturation transfer works because the trace distance between the two constructed cq states equals the total variation distance of the induced classical distributions. The entanglement of formation corollary is a real improvement over Winter, and the countable-alphabet extension answers an open question from Alhejji–Smith. The conditional dephasing trick itself is not new (it goes back to Petz, via Winter), but the application is clean and the optimality statement is the genuinely new content.

Where it is soft: the single load-bearing input is the quoted Alhejji–Smith bound; the paper does not prove it, and if that bound were wrong or its ε-range narrower than claimed, Proposition 1 would collapse. That is an honest dependency — the paper flags it explicitly — and the classical bound has since held up in the literature, so I do not treat it as a real flaw. Corollary 2 delegates its proof to Winter's method, and Corollary 3 leans on Kuznetsova's limit theorem; both are minor because the cited arguments are standard and the central claim, Proposition 1, is proved in full. One small note: in the upper half of the range, ε ∈ (1/2, 1−1/d_B], the right-hand side is still increasing in ε because (d_B−1)(1−ε)/ε ≥ 1, so applying the classical bound with the looser ε is legitimate.

Who this is for: anyone who needs uniform continuity bounds for capacities, entropy estimates, or resource-theoretic quantities in classical–quantum settings. This is a short, checkable, useful note, not a paradigm shift. I would send it to peer review — the optimality of a fundamental bound merits referee time, and a single careful referee can verify the whole proof in one sitting.

Referee Report

0 major / 4 minor

Summary. The paper proves an optimal uniform continuity bound for the conditional entropy of finite-dimensional classical–quantum (cq) states. Proposition 1 (Eq. (4)) states that for ε ∈ (0, 1 − 1/d_B], if ρ_XB and σ_XB are cq states with half trace distance ≤ ε, then |H(B|X)_ρ − H(B|X)_σ| ≤ ε log2(d_B − 1) + h2(ε), and that the bound is tight for every d_B and ε in the range. The proof uses a conditional dephasing channel in the eigenbasis of ρ's B-conditional states, exploits unitality to bound σ's conditional entropy by a classical conditional entropy, and invokes the Alhejji–Smith classical equivocation bound (Eq. (1)). Corollary 2 gives a Winter-style uniform continuity bound for entanglement of formation, and Corollary 3 extends the main bound to countably infinite classical conditioning alphabets.

Significance. If the Alhejji–Smith classical bound is valid, Proposition 1 is optimal and improves the corresponding case in Winter's Lemma 2. The reduction is elegant and the trace-distance bookkeeping is exact; there are no free parameters or fitted constants. The paper is explicit that both the range and the saturation example are inherited from [1], so the internal derivation is not circular. The entanglement-of-formation application and the countable-X extension are useful additions, although they rely on standard methods and cited results rather than new techniques.

minor comments (4)

[Proposition 1, tightness paragraph] The one-sentence reference to the saturation example in Eqs. (27)–(28) of [1] would be clearer if it explicitly said that the classical alphabet Y is encoded as diagonal states on system B, so the classical pair of distributions converts to a pair of cq states with the same trace distance and the same conditional entropies.
[Corollary 2] Because the proof is delegated to Winter's Corollary 4, please add one sentence explaining how δ = sqrt(ε(2 − ε)) arises from Uhlmann's theorem and that both directions of the entanglement-of-formation inequality follow from the same argument; as written the reader must reconstruct this step.
[Corollary 3] Please define ρ_B = Σ_x r(x)ρ_B^x explicitly before Eq. (49), where the notation first appears, so that the final identity H(B)_ρ − I(X;B)_ρ = Σ_x r(x)H(ρ_B^x) is unambiguous.
[References] Reference [1] should be updated to its published version, if one exists, rather than cited as a September 2019 preprint, because the optimality claim of Proposition 1 depends on its correctness.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the classical-quantum bound is a transparent reduction to the externally established Alhejji–Smith classical bound.

full rationale

The derivation of Proposition 1 is a direct reduction of the classical-quantum statement to the classical conditional-entropy bound of Alhejji and Smith, stated in Eq. (1) and attributed to [1]. The conditional dephasing channel in Eq. (8) is unital, preserves the X marginal, leaves rho_XB invariant, and maps the trace-distance hypothesis to the total-variation distance between the induced classical distributions r_XY and s_XY, as shown in Eqs. (19)–(21). The proof then invokes Eq. (1) with |Y| = d_B and obtains exactly the claimed bound. Saturation is inherited from the explicit classical example cited as Eqs. (27)–(28) of [1], with a commuting cq embedding preserving trace distance and conditional entropy. No parameter is fitted, no target quantity is used to define an input, and no load-bearing step depends on the author's own prior work; the self-citations present are motivational only. The paper's reliance on the unproved bound [1] is an external correctness and optimality dependency, not a circularity, because [1] is a separate result by different authors and is not derived from the claim being established. No circular step is present, so the appropriate score is 0.

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

The central claim rests on the external classical bound of Alhejji and Smith and on standard quantum information inequalities (data processing, unital entropy monotonicity). No free parameters or invented entities are introduced; the bound is parameter-free for given epsilon and d_B.

assumptions (4)

domain assumption Classical optimal uniform continuity bound (Eq. (1), from Alhejji and Smith): |H(Y|X)_p - H(Y|X)_q| <= epsilon log2(|Y|-1) + h2(epsilon) for epsilon in (0, 1-1/|Y|] and epsilon >= (1/2)||p-q||_1.
The cq proof reduces to this external classical result; the paper does not prove it. If this bound were not tight, the cq bound would also not be optimal.
standard math Data processing inequality for normalized trace distance under quantum channels.
Used in Eq. (19) to upper bound the trace distance of the dephased states.
standard math Unital quantum channels do not decrease von Neumann entropy.
Used in Eq. (14) to show H(BX)_sigma <= H(BX)_{Delta(sigma)}.
domain assumption Continuity of conditional entropy under finite-dimensional projections for infinite-dimensional cq states (Kuznetsova).
Used in Corollary 3's proof to take the limit in Eq. (53).

how reviews work

0 comments

Cite this review

Pith. "Pith review of Optimal uniform continuity bound for conditional entropy of classical--quantum states." pith.science (2026). https://pith.science/paper/RQZEMCN5

@misc{pith2026190901755,
  author       = {Pith},
  title        = {Pith review of: Optimal uniform continuity bound for conditional entropy of classical--quantum states},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RQZEMCN5}},
  note         = {Machine review of arXiv:1909.01755}
}

read the original abstract

In this short note, I show how a recent result of Alhejji and Smith [arXiv:1909.00787] regarding an optimal uniform continuity bound for classical conditional entropy leads to an optimal uniform continuity bound for quantum conditional entropy of classical--quantum states. The bound is optimal in the sense that there always exists a pair of classical--quantum states saturating the bound, and so no further improvements are possible. An immediate application is a uniform continuity bound for entanglement of formation that improves upon the one previously given by Winter in [arXiv:1507.07775]. Two intriguing open questions are raised regarding other possible uniform continuity bounds for conditional entropy, one about quantum--classical states and another about fully quantum bipartite states.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

23 extracted references · 19 canonical work pages

[1]

A Tight Uniform Continuity Bound for Equivocation

Mohammad A. Alhejji and Graeme Smith. A tight uniform con tinuity bound for equivocation. September 2019. arXiv:1909.00787 v1

work page Pith review arXiv 2019
[2]

Estimating mutual information via Kolm ogorov dis- tance

Zhengmin Zhang. Estimating mutual information via Kolm ogorov dis- tance. IEEE Transactions on Information Theory , 53(9):3280–3282, September 2007

work page 2007
[3]

Koenraad M. R. Audenaert. A sharp continuity estimate fo r the von Neumann entropy. Journal of Physics A: Mathematical and Theoretical , 40(28):8127, July 2007. arXiv:quant-ph/0610146

arXiv 2007
[4]

Continuity of quantum cha nnel ca- pacities

Debbie Leung and Graeme Smith. Continuity of quantum cha nnel ca- pacities. Communications in Mathematical Physics , 292(1):201–215, 2009

work page 2009
[5]

Approximate Degradable Quantum Channels

David Sutter, Volkher B. Scholz, Andreas Winter, and Ren ato Renner. Approximate degradable quantum channels. IEEE Transactions on In- formation Theory, 63(12):7832–7844, December 2017. arXiv:1412.0980

work page Pith review arXiv 2017
[6]

Quantum and private capacities of low-noise channels

Felix Leditzky, Debbie Leung, and Graeme Smith. Quantum and private capacities of low-noise channels. Physical Review Letters , 120(16):160503, April 2018. arXiv:1705.04335

work page Pith review arXiv 2018
[7]

Felix Leditzky, Eneet Kaur, Nilanjana Datta, and Mark M. Wilde. Approaches for approximate additivity of the Holevo inform ation of quantum channels. Physical Review A , 97(1):012332, January 2018. arXiv:1709.01111

arXiv 2018
[8]

Eneet Kaur and Mark M. Wilde. Amortized entanglement of a quantum channel and approximately teleportation-simulable channels. Journal of Physics A: Mathematical and Theoretical , July 2017. arXiv:1707.07721

work page Pith review arXiv 2017

Show all 23 references

[9]

Wilde, Sushovit Adhikari, and Masa hiro Takeoka

Kunal Sharma, Mark M. Wilde, Sushovit Adhikari, and Masa hiro Takeoka. Bounding the energy-constrained quantum and priv ate capac- ities of bosonic thermal channels. New Journal of Physics , 20:063025, June 2018. arXiv:1708.07257

2018 arXiv
[10]

Sumeet Khatri, Kunal Sharma, and Mark M. Wilde. Informa tion- theoretic aspects of the generalized amplitude damping cha nnel. March
[11]

Eneet Kaur, Saikat Guha, and Mark M. Wilde. Asymptotic s ecurity of discrete-modulation protocols for continuous-variabl e quantum key distribution. January 2019. arXiv:1901.10099

2019 arXiv
[12]

A continuity property of the entropy densi ty for spin lattices

Mark Fannes. A continuity property of the entropy densi ty for spin lattices. Communications in Mathematical Physics , 31:291, 1973

1973
[13]

Continuity of quantum co nditional in- formation

Robert Alicki and Mark Fannes. Continuity of quantum co nditional in- formation. Journal of Physics A: Mathematical and General , 37(5):L55– L57, February 2004. arXiv:quant-ph/0312081

2004 arXiv
[14]

Tight uniform continuity bounds for qu antum en- tropies: conditional entropy, relative entropy distance a nd energy con- straints

Andreas Winter. Tight uniform continuity bounds for qu antum en- tropies: conditional entropy, relative entropy distance a nd energy con- straints. Communications in Mathematical Physics , 347(1):291–313, October 2016. arXiv:1507.07775

2016 arXiv
[15]

Shirokov

Maksim E. Shirokov. Tight uniform continuity bounds fo r the quan- tum conditional mutual information, for the holevo quantit y, and for capacities of quantum channels. Journal of Mathematical Physics , 58(10):102202, 2017

2017
[16]

Shirokov

Maksim E. Shirokov. Adaptation of the Alicki-Fannes-W inter method for the set of states with bounded energy and its use. Reports on Mathematical Physics, 81(1):81–104, February 2018. arXiv:1609.07044

2018 arXiv
[17]

Shirokov

Maksim E. Shirokov. Uniform continuity bounds for info rmation char- acteristics of quantum channels depending on input dimensi on and on input energy. Journal of Physics A: Mathematical and Theoretical , 52(1):014001, December 2018. arXiv:1610.08870

2018 arXiv
[18]

Shirokov

Maksim E. Shirokov. Advanced Alicki-Fannes-Winter me thod for energy-constrained quantum systems and its use. July 2019. arXiv:1907.02458

2019 arXiv
[19]

Quantum Information Theory and Quantum Statistics

Denes Petz. Quantum Information Theory and Quantum Statistics . Springer Verlag, Berlin Heidelberg, 2008

2008
[20]

Bennett, David P

Charles H. Bennett, David P. DiVincenzo, John A. Smolin , and William K. Wootters. Mixed-state entanglement and quantum er- ror correction. Physical Review A , 54(5):3824–3851, November 1996. arXiv:quant-ph/9604024. 10

1996 arXiv
[21]

Kuznetsova

Anna A. Kuznetsova. Quantum conditional entropy for in ﬁnite- dimensional systems. Theory of Probability & Its Applications , 55(4):709–717, November 2011. arXiv:1004.4519

2011 arXiv
[22]

Inequalities of J

Harold Falk. Inequalities of J. W. Gibbs. American Journal of Physics , 38(7):858–869, July 1970

1970
[23]

Entropy, information and quantum mea surements

G¨ oran Lindblad. Entropy, information and quantum mea surements. Communications in Mathematical Physics , 33(4):305–322, December 1973. 11

1973

Pith tools

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

[1] [1]

A Tight Uniform Continuity Bound for Equivocation

Mohammad A. Alhejji and Graeme Smith. A tight uniform con tinuity bound for equivocation. September 2019. arXiv:1909.00787 v1

work page Pith review arXiv 2019

[2] [2]

Estimating mutual information via Kolm ogorov dis- tance

Zhengmin Zhang. Estimating mutual information via Kolm ogorov dis- tance. IEEE Transactions on Information Theory , 53(9):3280–3282, September 2007

work page 2007

[3] [3]

Koenraad M. R. Audenaert. A sharp continuity estimate fo r the von Neumann entropy. Journal of Physics A: Mathematical and Theoretical , 40(28):8127, July 2007. arXiv:quant-ph/0610146

arXiv 2007

[4] [4]

Continuity of quantum cha nnel ca- pacities

Debbie Leung and Graeme Smith. Continuity of quantum cha nnel ca- pacities. Communications in Mathematical Physics , 292(1):201–215, 2009

work page 2009

[5] [5]

Approximate Degradable Quantum Channels

David Sutter, Volkher B. Scholz, Andreas Winter, and Ren ato Renner. Approximate degradable quantum channels. IEEE Transactions on In- formation Theory, 63(12):7832–7844, December 2017. arXiv:1412.0980

work page Pith review arXiv 2017

[6] [6]

Quantum and private capacities of low-noise channels

Felix Leditzky, Debbie Leung, and Graeme Smith. Quantum and private capacities of low-noise channels. Physical Review Letters , 120(16):160503, April 2018. arXiv:1705.04335

work page Pith review arXiv 2018

[7] [7]

Felix Leditzky, Eneet Kaur, Nilanjana Datta, and Mark M. Wilde. Approaches for approximate additivity of the Holevo inform ation of quantum channels. Physical Review A , 97(1):012332, January 2018. arXiv:1709.01111

arXiv 2018

[8] [8]

Eneet Kaur and Mark M. Wilde. Amortized entanglement of a quantum channel and approximately teleportation-simulable channels. Journal of Physics A: Mathematical and Theoretical , July 2017. arXiv:1707.07721

work page Pith review arXiv 2017