Pith. sign in

REVIEW 6 minor 30 references

Minimum output p-Rényi entropy is shown nonadditive for all p>3/4 and all 0≤p<1/4, leaving only [1/4,3/4] open in 0<p<1.

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 ↗

T0 review · deepseek-v4-flash

2026-08-01 23:52 UTC pith:4FUBMAYR

load-bearing objection Solid quantitative advance: explicit uniform intervals (0,1/4) and (3/4,∞) for violation of minimum-output Rényi entropy additivity, reducing the open range to [1/4,3/4]; the proof relies on external free-probability theorems but the internal asymptotics check out.

arxiv 2607.15210 v1 pith:4FUBMAYR submitted 2026-07-16 quant-ph math-phmath.MPmath.PR

Counterexamples to additivity of minimum output p-R\'enyi entropy of quantum channels for p>3/4 and 0leq p<1/4

Debbie Leung , Benjamin Lovitz , Peixue Wu This is my paper
classification quant-ph math-phmath.MPmath.PR MSC 81P4546L5460B2094A17 PACS 03.67.-a03.67.Hk
keywords quantum channelsminimum output entropyRényi entropyadditivityrandom projectionsfree probabilityBell statenonadditivity
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 paper asks whether two independent noisy quantum channels can amplify noise more than the sum of their individual effects—the additivity question for minimum output Rényi entropy. It proves that additivity fails for every Rényi order p>3/4 and for every 0≤p<1/4: in each case there exist finite-dimensional quantum channels, built from random projections, whose combined minimum output entropy is strictly smaller than the sum of the separate minima. The proof gives explicit asymptotic entropy gaps, and because the arguments extend continuously to p=1, the von Neumann point is shown to be part of the same nonadditive mechanism rather than a singular case. If the theorem is right, the only unsettled orders in 0

Core claim

The central claim is Theorem 1.1: for every p in (0,1/4) ∪ (3/4,∞), there exist finite-dimensional projection-induced quantum channels Φ and Ψ with S_p^min(Φ⊗Ψ) < S_p^min(Φ)+S_p^min(Ψ). The paper establishes this by computing the large-dimension limits of two random-projection constructions. For p>3/4, it pairs a channel with its complex conjugate and evaluates the product on a maximally entangled input; the output converges to a known isotropic state, and comparing its entropy with the one-channel minimum gives a strict violation whenever A_p(t) > 4p(1−t)/t, which holds for small t precisely when p>3/4. For 0≤p<1/4, it pairs a half-rank projection channel with its transpose-orthogonal compl

What carries the argument

The argument rides on two limiting objects. The first is the one-channel output body K_{k,t} = {X/Tr X : 0≤X≤I_k, Tr c_t(X) ≤ 1/k}, where c_t(u) = (√(t(1−u)) − √(u(1−t)))²; this is the almost-sure Hausdorff limit of the output states of a Haar-random projection-induced channel and its minimum Rényi entropy has a known large-k expansion. The second is the isotropic Bell-state limit Z_{k,t} = r_{k,t} ψ⁺_k + (1−r_{k,t}) I_{k²}/k², with r_{k,t} = k²(1−t)/((k²−1)t + 1−t), obtained by feeding a maximally entangled state through the product of the channel and its conjugate. The rank-defect witness instead uses exact orthogonality P Q^T = 0 between a half-rank projection and its transposed complemen

Load-bearing premise

The whole asymptotic proof depends on the assumption that the random projection's block entries, after rescaling by the inverse square root of its partial trace, still converge strongly to the free-probability limit; if that normalization step destroys the convergence, the limiting output body, the Bell-state spectrum, and both entropy comparisons would not be justified.

What would settle it

Numerically evaluate the support function of the normalized Choi blocks P_A^{-1/2} S_n(a) P_A^{-1/2} for random a and compare to max_{u∈D_{k,t}} Σ a_i u_i; a stable deviation would indicate the strong block-modification theorem is not preserved under local normalization, undermining both proofs. Alternatively, for a fixed p>3/4 such as 0.8, compute the exact finite-n minimum output entropy of the product-conjugate pair and check whether the predicted asymptotic gap appears for some n; if not, the strong-convergence step is false.

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

If this is right

  • For the existence of additivity violations with 0<p<1, only the interval [1/4,3/4] remains open.
  • The von Neumann point p=1 is covered by the same high-p mechanism via continuous extension; no separate construction is needed in this ensemble, and the Bell-state criterion first detects a violation at output dimension k=182 within this model.
  • The previously unspecified neighborhood of p=0 is replaced by the explicit range 0≤p<1/4.
  • Any future universal additivity theorem for quantum channels, if it exists, must have its p-domain contained in [1/4,3/4].
  • Both violations are asymptotic with gap of order k^{-2}; hence explicit counterexamples require sufficiently large output dimensions, and the phenomena are small but non-perturbative.

Where Pith is reading between the lines

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

  • The two constructions suggest that additivity might fail throughout (0,1) except possibly a middle band; numerical continuation of the asymptotic gaps could indicate whether the endpoints 1/4 and 3/4 are sharp or artifacts of the witnesses.
  • Because the entropy gaps scale as k^{-2}, an explicit deterministic (non-random) counterexample would need a different mechanism, presumably with more structure than Haar projections.
  • The rank-defect idea might generalize to other correlated projection pairs, potentially pushing the low-p endpoint toward 1/4 from below or even beyond.
  • The connection to random-subspace geometry suggests nonadditivity is a generic high-dimensional phenomenon, so one might expect violations for 'most' projection-induced channels in these parameter regimes.

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 / 6 minor

Summary. The paper studies additivity of the minimum output p-Rényi entropy of quantum channels. It constructs, for each p in (3/4,∞) and each p in (0,1/4), finite-dimensional projection-induced channels Φ,Ψ with S_p^min(Φ⊗Ψ)<S_p^min(Φ)+S_p^min(Ψ). The high-p construction uses a Haar-random projection, locally normalizes its Choi matrix, pairs the channel with its complex conjugate, and feeds it a maximally entangled state; a free-probability limit identifies the Bell output as an isotropic state and a large-k expansion yields a threshold at p=3/4. The low-p construction pairs a half-rank projection with its transpose-orthogonal complement and uses a rank-deficit lemma to bound the joint output entropy by log(k^2−1), while a one-channel output-body limit gives 2log k−4p/k^2, yielding violation for p<1/4. The unresolved interval for 0<p<1 is thus reduced to [1/4,3/4].

Significance. Assuming correctness, this is a substantial advance: it gives the first explicit uniform intervals on both sides of the von Neumann point for which additivity of minimum output Rényi entropy fails, replacing channel-dependent neighborhoods from continuity arguments. The proof is a coherent combination of strong asymptotic freeness, support-function convergence, and asymptotic entropy expansions. The paper is honest about prior work: the p=0 endpoint and the rank-deficit lemma are attributed to CHL+08. It also reports an improvement of the output-dimension threshold for von Neumann additivity violation from 183 to 182, though this numerical claim is not load-bearing.

minor comments (6)
  1. [Abstract and Theorem 1.1] The abstract and Section 1 claim the result for 0≤p<1/4, but Theorem 1.1 as stated covers p∈(0,1/4)∪(3/4,∞). The p=0 endpoint is not part of the stated theorem, yet the text later says the unresolved part is reduced to [1/4,3/4], implying 0 is covered. Please reconcile the statement with the abstract and with the actual proof.
  2. [§5.2, Proposition 5.4] Proposition 5.4 asserts the random half-rank construction works for every 0≤p<1/4, but the p=0 case is dismissed with 'The case p=0 follows from [CHL+08].' That reference proves existence of some counterexample, not that the specific half-rank projection-induced pair works. If p=0 is intended to be covered by Theorem 1.1, supply the missing rank argument; otherwise restrict the proposition to p>0.
  3. [§1.1 and §5.2] The overview states that for fixed k≥3 and large n the one-channel outputs are almost surely of full rank. No proof is given. This fact is nontrivial: for k=2 it is false, since a random half-rank subspace of C^n⊗C^2 has a product vector in its kernel with probability one. The proof of the p=0 case (if pursued) should justify the full-rank claim, e.g., by dimension-counting the absence of product k-planes in the kernel.
  4. [Lemma 5.2] The sentence 'This limit is strictly larger than 4p precisely when p>3/4' is mathematically incorrect: the limit is 1/(1−p), and 1/(1−p)>4p holds for all p≠1/2 in (0,1). The comparison relevant to (46) is A_p(x^{-1})/(p(x−1)) > 4, i.e., 1/(1−p)>4, which is exactly p>3/4. The conclusion of the lemma is correct, but the stated comparison should be fixed.
  5. [§5.2] The proof of Proposition 5.4 says 'Almost surely, both Tr_B P_n and Tr_B Q_n are invertible for all sufficiently large n' without proof. Lemma 2.3 covers Tr_B P_n, but the invertibility of Q_n = kI_n − P_n^T needs a separate argument (e.g., that the probability of P_n containing a product k-plane is zero for k≥3 and large n). Please add a sentence or reference.
  6. [Throughout] There are small typos: 'Nechida' should be 'Nechita' in the abstract; 'developped' should be 'developed' in the proof of Lemma 2.4; the notation ⊞^k in Lemma 2.4 is unclear and should be cleaned up.

Circularity Check

0 steps flagged

No circularity: the new additivity violations are derived from independent asymptotic computations on the random-projection ensemble; overlapping-author citations supply auxiliary lemmas, not the target result.

full rationale

The central derivation is self-contained relative to external free-probability tools. The limiting one-channel output body K_{k,t} is obtained in Theorem 3.1 via Lemma 2.4, whose proof uses the strong block-modification theorem [Nec18, ANV16] and the appendix's Cauchy-transform duality Lemma A.1. The one-channel entropy expansion (10) follows from a Taylor analysis of the body in Proposition B.1; the conjugate Bell-output expansion (42) is computed separately from the second-Choi-moment limit in Proposition A.3 and Proposition B.2. The high-p result is the comparison (13) of these two independent expansions, and the threshold p>3/4 emerges from the limit t A_p(t)/(4p(1-t)) -> 1/(4(1-p)); no parameter is fitted to the target. The low-p result is the comparison (48) between the same one-channel body entropy and the rank bound log(k^2-1) from Lemma 2.2; the paper explicitly states that its contribution is the quantitative asymptotic entropy analysis, not the construction. Overlapping-author citations appear only as published auxiliary facts: Lemma 2.2 from [CHL+08] and the p=0 endpoint, also from [CHL+08]. The p=0 endpoint is cited rather than re-proved, which is a presentational gap, but it is not circular because the cited result is an independent peer-reviewed theorem and the main new ranges 0<p<1/4 and 3/4<p<1 do not reduce to it. No fitted input is renamed as a prediction, no uniqueness theorem is imported to force the answer, and the asymptotic formulas are not equivalent to the additivity-violation inequalities by construction.

Axiom & Free-Parameter Ledger

0 free parameters · 5 axioms · 0 invented entities

No constants are fitted to data; the only design parameters are the projection rank ratio t (chosen small for p>3/4 or set to 1/2 for p<1/4), the output dimension k, and the input dimension n, all chosen by the proof rather than estimated. The construction introduces no new physical or mathematical entities. The central claim depends on established but heavy free-probability theorems and the cited rank-deficit lemma.

axioms (5)
  • standard math Strong asymptotic freeness of Haar random projections and deterministic matrix units (CM14), used in Prop. A.2 to identify the limiting block family.
    This external theorem underlies the convergence of P_n's blocks and hence Theorem 3.1. The paper cites it but does not prove it.
  • standard math Strong block-modification theorem (ANV16, Nec18) for the random compression estimate, Lemma 2.4.
    This produces the spectral-edge formula λmax(Σ a_i P_ii^(n)) → max_{u∈D_{k,t}} Σ a_i u_i, the key input to the support-function proof.
  • standard math Generic rank of reduced states of Haar random vectors (ŻS01), used in Lemma 2.3 to guarantee P_A>0.
    Assures the projection-induced channels are well defined almost surely.
  • domain assumption Rank-deficit lemma from CHL+08 (Lemma 2.2): for P and Q=I-P^T, some joint input has output rank ≤k²-1.
    Cited, not reproved; it supplies the joint low-rank state for the p<1/4 construction.
  • standard math Free-probability Legendre duality and Cauchy-transform edge computations (Appendix A).
    Used to compute the support function of D_{k,t} and the maximal spectral edge.

pith-pipeline@v1.3.0-alltime-deepseek · 16348 in / 34021 out tokens · 264806 ms · 2026-08-01T23:52:24.207402+00:00 · methodology

0 comments
read the original abstract

Additivity of minimum output entropies is a central problem in quantum information theory. Nonadditivity is known for every R\'enyi order $p>1$, at the von Neumann point $p=1$, and near $p=0$, while most of the interval $0<p<1$ has remained open. We prove that for every R\'enyi order $p$ satisfying either $p>3/4$ or $0\leq p<1/4$, there exist finite-dimensional projection-induced quantum channels such that additivity of the minimum output $p$-R\'enyi entropy fails. The proof combines two correlated random-projection constructions: a product-conjugate Bell-state witness for $p>3/4$, and a transpose-complement rank-defect witness for $p<1/4$. Thus the unresolved part of $0<p<1$ is reduced to $[1/4,3/4]$. Our estimates also improve the output dimension threshold for additivity violation of minimum output von Neumann entropy, first established in Belinschi, Collins and Nechida.

discussion (0)

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

Reference graph

Works this paper leans on

30 extracted references · 16 canonical work pages

  1. [1]

    G. G. Amosov, A. S. Holevo, and R. F. Werner. ``On some additivity problems in quantum information theory'' , (2000). Available online: https://arxiv.org/abs/math-ph/0003002

    Pith/arXiv arXiv 2000

  2. [2]

    Arizmendi, I

    O. Arizmendi, I. Nechita, and C. Vargas. ``On the Asymptotic Distribution of Block-Modified Random Matrices'' . http://dx.doi.org/10.1063/1.4936925 Journal of Mathematical Physics 57 (1):\,015216 (2016)

    work page doi:10.1063/1.4936925 2016

  3. [3]

    Aubrun and S

    G. Aubrun and S. J. Szarek. Alice and Bob Meet Banach: The Interface of Asymptotic Geometric Analysis and Quantum Information Theory . volume 223 of Mathematical Surveys and Monographs , American Mathematical Society http://dx.doi.org/10.1090/surv/223 (2017)

    doi:10.1090/surv/223 2017

  4. [4]

    Aubrun, S

    G. Aubrun, S. Szarek, and E. Werner. ``Non-Additivity of R \'e nyi Entropy and D voretzky's Theorem'' . http://dx.doi.org/10.1063/1.3271044 Journal of Mathematical Physics 51 (2):\,022102 (2010)

    work page doi:10.1063/1.3271044 2010

  5. [5]

    Aubrun, S

    G. Aubrun, S. Szarek, and E. Werner. `` H astings' Additivity Counterexample via D voretzky's Theorem'' . http://dx.doi.org/10.1007/s00220-010-1172-y Communications in Mathematical Physics 305 (1):\,85--97 (2011)

    work page doi:10.1007/s00220-010-1172-y 2011

  6. [6]

    S. T. Belinschi, B. Collins, and I. Nechita. ``Eigenvectors and Eigenvalues in a Random Subspace of a Tensor Product'' . http://dx.doi.org/10.1007/s00222-012-0386-3 Inventiones Mathematicae 190 (3):\,647--697 (2012)

    work page doi:10.1007/s00222-012-0386-3 2012

  7. [7]

    S. T. Belinschi, B. Collins, and I. Nechita. ``Almost One Bit Violation for the Additivity of the Minimum Output Entropy'' . http://dx.doi.org/10.1007/s00220-015-2561-z Communications in Mathematical Physics 341 (3):\,885--909 (2016)

    work page doi:10.1007/s00220-015-2561-z 2016

  8. [8]

    F. G. S. L. Brand \ a o and M. Horodecki. ``On H astings' Counterexamples to the Minimum Output Entropy Additivity Conjecture'' . http://dx.doi.org/10.1142/S1230161210000047 Open Systems & Information Dynamics 17 (1):\,31--52 (2010)

    work page doi:10.1142/s1230161210000047 2010

  9. [9]

    Collins, M

    B. Collins, M. Fukuda, and I. Nechita. ``On the Convergence of Output Sets of Quantum Channels'' . http://dx.doi.org/10.7900/jot.2013dec04.2008 Journal of Operator Theory 73 (2):\,333--360 (2015)

    work page doi:10.7900/jot.2013dec04.2008 2008

  10. [10]

    Cubitt, A

    T. Cubitt, A. W. Harrow, D. Leung, A. Montanaro, and A. Winter. ``Counterexamples to Additivity of Minimum Output p - R \'e nyi Entropy for p Close to 0'' . http://dx.doi.org/10.1007/s00220-008-0625-z Communications in Mathematical Physics 284 (1):\,281--290 (2008)

    work page doi:10.1007/s00220-008-0625-z 2008

  11. [11]

    Collins and C

    B. Collins and C. Male. ``The Strong Asymptotic Freeness of H aar and Deterministic Matrices'' . http://dx.doi.org/10.24033/asens.2211 Annales Scientifiques de l' \'E cole Normale Sup \'e rieure 47 (1):\,147--163 (2014)

    work page doi:10.24033/asens.2211 2014

  12. [12]

    Collins and I

    B. Collins and I. Nechita. ``Random Quantum Channels I : Graphical Calculus and the Bell State Phenomenon'' . http://dx.doi.org/10.1007/s00220-010-1012-0 Communications in Mathematical Physics 297 (2):\,345--370 (2010)

    doi:10.1007/s00220-010-1012-0 2010

  13. [13]

    Collins and I

    B. Collins and I. Nechita. ``Random Quantum Channels II : Entanglement of Random Subspaces, R \'e nyi Entropy Estimates and Additivity Problems'' . http://dx.doi.org/10.1016/j.aim.2010.08.002 Advances in Mathematics 226 (2):\,1181--1201 (2011)

    work page doi:10.1016/j.aim.2010.08.002 2010

  14. [14]

    Collins and I

    B. Collins and I. Nechita. ``Random Matrix Techniques in Quantum Information Theory'' . http://dx.doi.org/10.1063/1.4936880 Journal of Mathematical Physics 57 (1):\,015215 (2016)

    doi:10.1063/1.4936880 2016

  15. [15]

    Derksen and B

    H. Derksen and B. Lovitz. ``Constructive Counterexamples to the Additivity of Minimum Output R \'e nyi Entropy of Quantum Channels for All p>1 '' . arXiv:2510.07547, (2025). Available online: https://arxiv.org/abs/2510.07547

    arXiv 2025

  16. [16]

    Fukuda, C

    M. Fukuda, C. King, and D. K. Moser. ``Comments on H astings' Additivity Counterexamples'' . http://dx.doi.org/10.1007/s00220-010-0996-9 Communications in Mathematical Physics 296 (1):\,111--143 (2010)

    work page doi:10.1007/s00220-010-0996-9 2010

  17. [17]

    Grudka, M

    A. Grudka, M. Horodecki, and T1 138 . Pankowski. ``Constructive Counterexamples to Additivity of Minimum Output R \'e nyi Entropy of Quantum Channels for All p>2 '' . http://dx.doi.org/10.1088/1751-8113/43/42/425304 Journal of Physics A: Mathematical and Theoretical 43 (42):\,425304 (2010)

    work page doi:10.1088/1751-8113/43/42/425304 2010

  18. [18]

    M. B. Hastings. ``Superadditivity of Communication Capacity Using Entangled Inputs'' . http://dx.doi.org/10.1038/nphys1224 Nature Physics 5 (4):\,255--257 (2009)

    doi:10.1038/nphys1224 2009

  19. [19]

    A. S. Holevo. ``The Capacity of the Quantum Channel with General Signal States'' . http://dx.doi.org/10.1109/18.651037 IEEE Transactions on Information Theory 44 (1):\,269--273 (1998)

    doi:10.1109/18.651037 1998

  20. [20]

    Hayden and A

    P. Hayden and A. Winter. ``Counterexamples to the Maximal p -Norm Multiplicativity Conjecture for All p>1 '' . http://dx.doi.org/10.1007/s00220-008-0624-0 Communications in Mathematical Physics 284 (1):\,263--280 (2008)

    work page doi:10.1007/s00220-008-0624-0 2008

  21. [21]

    Kukulski, I

    R. Kukulski, I. Nechita, T1 138 . Pawela, Z. Pucha a, and K. \.Z yczkowski. ``Generating Random Quantum Channels'' . http://dx.doi.org/10.1063/5.0038838 Journal of Mathematical Physics 62 (6):\,062201 (2021)

    doi:10.1063/5.0038838 2021

  22. [22]

    Mezzadri

    F. Mezzadri. ``How to generate random matrices from the classical compact groups'' , (2007). Available online: https://arxiv.org/abs/math-ph/0609050

    Pith/arXiv arXiv 2007

  23. [23]

    J. A. Mingo and R. Speicher. Free Probability and Random Matrices . volume 35 of Fields Institute Monographs , Springer http://dx.doi.org/10.1007/978-1-4939-6942-5 (2017)

    doi:10.1007/978-1-4939-6942-5 2017

  24. [24]

    I. Nechita. ``On the Separability of Unitarily Invariant Random Quantum States---The Unbalanced Regime'' . http://dx.doi.org/10.1155/2018/7105074 Advances in Mathematical Physics 2018 :\,7105074 (2018)

    work page doi:10.1155/2018/7105074 2018

  25. [25]

    P. W. Shor. ``Equivalence of Additivity Questions in Quantum Information Theory'' . http://dx.doi.org/10.1007/s00220-003-0981-7 Communications in Mathematical Physics 246 (3):\,453--472 (2004)

    work page doi:10.1007/s00220-003-0981-7 2004

  26. [26]

    Szczygielski and M

    K. Szczygielski and M. Studzi \'n ski. ``New Constructive Counterexamples to Additivity of Minimum Output R \'e nyi p -Entropy of Quantum Channels'' . http://dx.doi.org/10.1109/TIT.2024.3446191 IEEE Transactions on Information Theory 70 (10):\,7023--7035 (2024)

    arXiv 2024

  27. [27]

    Schumacher and M

    B. Schumacher and M. D. Westmoreland. ``Sending Classical Information via Noisy Quantum Channels'' . http://dx.doi.org/10.1103/PhysRevA.56.131 Physical Review A 56 (1):\,131--138 (1997)

    work page doi:10.1103/physreva.56.131 1997

  28. [28]

    R. F. Werner and A. S. Holevo. ``Counterexample to an Additivity Conjecture for Output Purity of Quantum Channels'' . http://dx.doi.org/10.1063/1.1498491 Journal of Mathematical Physics 43 (9):\,4353--4357 (2002)

    doi:10.1063/1.1498491 2002

  29. [29]

    Yu and M

    N. Yu and M. Ying. ``Non-Additivity of Minimum Output p - R \'e nyi Entropy'' . arXiv:1006.1733, (2010). Available online: https://arxiv.org/abs/1006.1733. Withdrawn on 24 December 2012 due to a crucial error

    Pith/arXiv arXiv 2010

  30. [30]

    \.Z yczkowski and H.-J

    K. \.Z yczkowski and H.-J. Sommers. ``Induced Measures in the Space of Mixed Quantum States'' . http://dx.doi.org/10.1088/0305-4470/34/35/335 Journal of Physics A: Mathematical and General 34 (35):\,7111--7125 (2001)

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