{"id":"ebaed7dd-dcbf-4d18-b99c-337497dbd2fd","arxiv_id":"1908.03917","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The paper derives lower and upper Holevo capacity bounds for generalized Pauli channels and obtains exact classical capacities for Pauli channels and symmetric two-parameter channels.","lead":"This paper calculates bounds on how much classical information can be sent through a family of noisy quantum channels called generalized Pauli channels. It shows that when the bounds meet, the exact capacity is known, and it gives explicit formulas for Pauli channels and symmetric qudit channels.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Lemma 1's proof forces the Weyl channel coefficient matrix to factor as an outer product, a step that is unproved for generic channels and on which the capacity equalities (41)-(42) depend.","rationale":"The reader's conditional verdict is appropriate. The most load-bearing point is the unproved Lemma 1: the paper's new results for d>2, and even the Pauli derivation, rely on the upper bound, and the reproduced proof contains a factorization step that is not generally valid. This is an internal gap rather than a disagreement with consensus. A numerical sanity check on random d=3 channels would settle whether the inequality itself survives. Separately, the false statement about P-divisible dynamics and monotonic capacity (Eq. 52 and the conclusions) is a real but secondary error; it does not undermine the capacity formulas. The step from additivity of the lower bound to additivity of the Holevo capacity is also under-justified: Proposition 1 proves additivity of chi_low, not of chi, so 'if bounds coincide, C equals the common value' needs an additional argument. For Pauli channels, King's additivity fills that gap, but for the d>2 equality families it is not supplied. These considerations reinforce the conditional verdict without changing its direction.","tokens_in":12232,"tokens_out":34978,"duration_ms":362247,"concrete_test":"Sample many d=3 generalized Pauli channels uniformly over probability distributions p satisfying the generalized Fujiwara-Algoet conditions (Eq. 13). For each channel, numerically minimize S(Lambda_GP[rho]) over pure states rho with a global optimizer, and compare chi = ln 3 - S_min with chi_up from Eq. (32). If any channel has chi > chi_up, Lemma 1 and Theorem 2 are false. If hundreds of samples all satisfy chi <= chi_up, the concern is weakened. Additionally, test the two-parameter family of Eq. (42) for d=3 by checking whether the numerical chi equals the claimed formula, which would also determine whether the additivity step from Proposition 1 is sufficient.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central capacity equalities (41), (42), and the Pauli formula (44) all require Theorem 2's upper bound, which is imported from Refs. [34,35] through Lemma 1. In the appendix proof of Lemma 1, after writing the output as a sum over p_jk W_jk rho* W_jk^dagger (Eq. 60), the proof selects s_j, u_j, and S_k with s_j q_k = p_jk. This forces the arbitrary Kraus coefficient matrix p_jk of a Weyl channel to factor as an outer product of two probability vectors. That is not true for generic p_jk, and no argument is given that linear dependencies among the orbit matrices W_jk rho* W_jk^dagger make the factorization unnecessary. The paper explicitly says 'we do not check whether rho* exists', so the existence of a pure optimal state realizing both the majorization and the factorization is unverified. If Lemma 1 fails for some generalized Pauli channel, the upper bound (32) is not valid, and the claimed classical capacities in the coincidence cases do not follow. This is the single load-bearing gap in the central claim.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies generalized Pauli channels on prime-power-dimensional Hilbert spaces and derives lower and upper bounds on their Holevo capacity. The lower bound is obtained from projectors onto mutually unbiased bases, and the upper bound is imported from known results for Weyl channels via a lemma. The paper proves that the lower bound is weakly additive, so that whenever the lower and upper bounds coincide, the classical capacity equals the common value. Explicit capacity formulas are given for depolarizing-type limits and for two-qubit Pauli channels, and a claim is made about monotonicity of Pauli channel capacity under P-divisible dynamics.","tokens_in":12465,"tokens_out":9300,"duration_ms":96696,"significance":"If the main results are correct, the paper provides explicit classical capacities for a natural family of covariant quantum channels, complementing earlier additivity results. The proof of weak additivity of the lower bound (Proposition 1) is self-contained and appears sound, and the specialization to Pauli channels is compact and potentially useful. The main caveat is that the upper bound and the consequent equality results depend on Lemma 1, whose proof is only recapitulated from Refs. [34,35] and contains an unjustified factorization step; in addition, the paper contains a false statement about P-divisibility and monotonicity of Pauli channel capacity. The analytical style is clear, but the central claim currently rests on an unverified imported lemma and a demonstrably incorrect side claim.","major_comments":[{"comment":"The proof of Lemma 1, specifically Eqs. (60)-(61), states that an admissible choice is s_j q_k = p_jk. This forces the Kraus coefficient matrix p_jk of an arbitrary (multipartite) Weyl channel to factor as the outer product of two probability vectors, which is not true for generic p_jk. No argument is given that linear dependencies among the orbit matrices W_jk ρ* W_jk† make the factorization unnecessary. Since the main text explicitly says 'we do not check whether ρ* exists' (after Theorem 2), the existence of a pure optimal state realizing both the majorization condition and the factorization is not established. The upper bound in Theorem 2, Eq. (32), and consequently the coinciding-bound capacity formulas in Eqs. (41), (42), and (44), all depend on Lemma 1; this gap is therefore load-bearing and must be closed.","section":"Appendix, Proof to Lemma 1; used in Theorem 2"},{"comment":"The claim that P-divisibility of a Pauli dynamical map implies monotone decrease of its classical capacity is false. Equation (52) differentiates only λ_max(t), but the capacity formula in Eq. (44) depends on λ* = max{|λ_min|, λ_max}. If the most negative eigenvalue becomes more negative, |λ_min| increases even though all eigenvalues are non-increasing. A concrete counterexample is λ1 = 0.3, λ2 = 0.2, and λ3(t) = -0.2 - 0.3t for t ∈ [0,1]; all eigenvalues are non-increasing, so the map is P-divisible by the criterion cited in the paper, and the Fujiwara-Algoet conditions remain satisfied throughout. Yet λ* increases from 0.3 to 0.5, so C(Λ_P(t)) increases. The corresponding sentence in the Conclusions must be corrected or removed.","section":"Special case: Pauli channels, Eq. (52); Conclusions"}],"minor_comments":[{"comment":"The text says 'eqs. (42) and (42) recover the classical capacity of the depolarizing channel'; this should read 'eqs. (41) and (42)'.","section":"Classical capacity, paragraph after Eq. (42)"},{"comment":"The symbols λ_max and λ_min are used before being defined; please define them explicitly, for example as λ_max = max_α λ_α and λ_min = min_α λ_α.","section":"Special case: Pauli channels, Eqs. (43)-(44)"},{"comment":"There is a typo: 'quantim channel' should be 'quantum channel'.","section":"Introduction, paragraph after Eq. (3)"},{"comment":"The set J defined by Eq. (64) as {p0, p1/(d-1), ..., p_{d+1}/(d-1)} has d+1 distinct entries only if multiplicities are ignored; since the text states |J| = d^2, it should be described as a multiset with p_α/(d-1) appearing d-1 times.","section":"Appendix, Proof to Theorem 2"}],"recommendation":"major_revision","confidential_remarks":"The upper-bound part of the paper is largely a specialization of results from Refs. [34,35]; the novelty is concentrated in the lower-bound additivity and the coincidence formulas. The referee should ask the authors to supply a complete proof or a precise citation that resolves the factorization issue in Lemma 1, and to correct the false P-divisibility monotonicity claim before resubmission."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Plainly: this is a useful paper that is not quite finished. The lower bound for the Holevo capacity of generalized Pauli channels (Theorem 1) and its weak additivity (Proposition 1) are new and solid, and the exposition is clear. The specialized upper bound in Theorem 2 is convenient, but it is inherited from the Weyl channel results of ur Rehman et al. [34,35]. The exact capacity values for the symmetric two-parameter qudit channels are a genuine addition, if the upper bound holds.\n\nThe main problem sits exactly where the stress-test note points. The proof sketch of Lemma 1 recaps a majorization argument in which the channel probabilities are forced to factor as s_j q_k = p_jk. For a generic Weyl channel that factorization is not true, and the paper does not show that linear dependencies among the operators W_jk rho* W_jk† make it unnecessary. The paper itself says 'we do not check whether rho* exists'. Since the equality cases (41), (42) and the Pauli formula (44) all depend on the upper bound, the central equality claims are conditional on this gap. For Pauli channels the capacity formula is already known via King's additivity, so that particular result survives; the two-parameter qudit formula does not have that crutch.\n\nThere is also a separate mistake in the P-divisibility discussion. The equivalence 'P-divisible iff \\dot λ_α(t) ≤ 0' is false when some λ_α are negative; the correct condition is λ_α \\dot λ_α ≤ 0, meaning each |λ_α| is non-increasing. The capacity monotonicity claim may survive once that is fixed, but Eq. (52) and the argument around it need to be redone.\n\nCitation practice is normal: the upper bound is credited to prior authors, and self-citations are for background. Nothing there looks inflated.\n\nThis paper is for people working on covariant channel capacities. A referee can get value from it. I would send it to review, not desk reject it, but I would not accept it as is: the lower-bound part is publishable now, the upper-bound gap needs to be closed or explicitly discharged via the cited papers, and the P-divisibility error has to be corrected.","headline":"Useful lower-bound results and a clean specialization, but the capacity equalities rest on an unproven factorization in the imported upper bound, and the P-divisibility remark has a sign error.","tokens_in":50,"tokens_out":18355,"would_cite":true,"duration_ms":222965,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["81P45","94A40"],"pacs":[],"model":"deepseek-v4-flash","headline":"For a generalized Pauli channel, the classical capacity is bracketed by two explicit entropy formulas, and when the bounds coincide the capacity equals their common value; qubit Pauli channels always fall in this case.","keywords":["generalized Pauli channels","classical capacity","Holevo capacity","mutually unbiased bases","Weyl-covariant channels","weak additivity","Pauli channels","quantum channel capacity"],"falsifier":"Numerically maximize the Holevo capacity over input ensembles for a $d=4$ generalized Pauli channel with mixed-sign eigenvalues, such as $\\lambda=(1/2,-1/4,-1/4,-1/4,-1/4)$; if the maximum exceeds $\\ln d - H[\\zeta(p)]$ from Eq. (32), or if any channel has a lower bound from Eq. (23) larger than its upper bound, the central theorem pair fails.","tokens_in":12045,"feed_emoji":"📡","tokens_out":11250,"duration_ms":109528,"temperature":0.7,"pith_summary":"This paper calculates usable bounds on the Holevo capacity of every generalized Pauli channel, and shows when those bounds determine the channel's classical capacity exactly. The lower bound is obtained by feeding the channel basis states from mutually unbiased bases; the paper proves this lower bound is weakly additive, so it cannot be improved by using the channel many times. The upper bound is adapted from a known result for Weyl-covariant channels. Whenever the two bounds coincide, the classical capacity equals their common value, and for qubit Pauli channels this always happens, giving a single closed formula in terms of the largest absolute eigenvalue.","feed_headline":"Exact capacity formula found for all Pauli channels","feed_subtitle":"One eigenvalue of the qubit channel, the largest in absolute value, completely fixes its classical capacity.","key_machinery":"The workhorse is the generalized Pauli channel itself, a random unitary channel built from $d+1$ mutually unbiased bases and specified by $d+1$ eigenvalues $\\lambda_\\alpha$. The lower bound in Eq. (23) is the capacity of the classical symmetric channel $T^{(\\alpha)}$ induced on each basis, and its weak additivity follows because the tensor product of two generalized Pauli channels tensors these classical maps, adding their row entropies. The upper bound in Eq. (32) is the entropy of the vector $\\zeta(p)$ of the $d$ largest sums of Kraus probabilities; the proof that this entropy dominates the true output entropy relies on a majorization relation and on a pure optimizer $\\rho_*$ inherited from the Weyl-channel setting.","core_discovery":"On the paper's own terms, the central claim is that for a generalized Pauli channel with eigenvalues $\\lambda_1,\\dots,\\lambda_{d+1}$, the Holevo capacity $\\chi$ satisfies the lower estimate of Eq. (23) and the upper estimate of Eq. (32). Because the lower bound is weakly additive, any channel for which the two estimates agree has classical capacity $C(\\Lambda_{\\rm GP}) = \\chi(\\Lambda_{\\rm GP})$ equal to that common value. The paper further claims that for $d=2$ the estimates always agree, so every Pauli channel has $C(\\Lambda_P) = \\frac{1+\\lambda_*}{2}\\ln(1+\\lambda_*) + \\frac{1-\\lambda_*}{2}\\ln(1-\\lambda_*)$ with $\\lambda_* = \\max\\{|\\lambda_{\\min}|,\\lambda_{\\max}\\}$.","pith_inferences":["If the upper bound in Eq. (32) were also weakly additive, exact capacities would follow for every channel that saturates it, not only for channels where the two bounds meet; the paper leaves this route open.","The paper's explicit caveat that an optimal $\\rho_*$ is not shown to exist suggests a numerical search over $d=4$ and $d=8$ generalized Pauli channels with mixed-sign eigenvalues; a channel whose true capacity lies strictly below Eq. (32) would confine the exact-capacity formulas to the symmetric cases.","Because the Pauli-channel formula depends only on the extreme channel fidelities, a natural testable conjecture is that generalized Pauli channel capacity is controlled by the extreme fidelities on mutually unbiased basis projectors.","Monitoring the classical capacity of a time-dependent generalized Pauli channel could give an information-theoretic probe of divisibility: a transient increase in $C$ would signal a break of P-divisibility, while a decrease alone would not certify divisibility."],"forward_implications":["Every Pauli channel has an exact classical capacity, determined by $\\lambda_*=\\max\\{|\\lambda_{\\min}|,\\lambda_{\\max}\\}$ rather than by the full noise structure.","Whenever the lower and upper Holevo bounds meet, the asymptotic classical capacity equals the single-use Holevo capacity, so no regularization over many channel uses is needed.","Highly symmetric generalized Pauli channels with all eigenvalues of one sign and at most two distinct values have closed-form capacities, Eqs. (41)-(42), with the depolarizing channel as the all-equal case.","For P-divisible Pauli dynamics, the classical capacity is non-increasing in time; the converse fails, so capacity monotonicity is only a one-way witness of divisibility."],"supporting_citations":[{"why":"Supplies the Weyl-channel bounding method and the majorization construction that the paper adapts to generalized Pauli channels.","marker":"[34]"},{"why":"Companion result for Weyl channels from which the upper-bound lemma is imported.","marker":"[35]"},{"why":"Introduces generalized Pauli channels as Pauli diagonal channels constant on axes and gives their complete-positivity conditions.","marker":"[26]"},{"why":"Gives the proportionality between Holevo capacity and minimal output entropy for irreducibly covariant channels, used to extract the lower bound from basis-projector inputs.","marker":"[11]"},{"why":"Establishes that weak additivity of the minimal output entropy makes the classical capacity equal to the Holevo capacity for covariant channels.","marker":"[12]"},{"why":"Define the Holevo capacity and prove the coding theorem making the classical capacity the regularized Holevo capacity.","marker":"[5, 6]"},{"why":"Supplies additivity for unital qubit channels and the complete-positivity inequalities used for Pauli-channel eigenvalues.","marker":"[13]"},{"why":"Supplies additivity for depolarizing channels, recovered as the all-equal eigenvalue case.","marker":"[14]"}],"fun_headline_variants":["Pauli channel capacity equals a simple formula","Exact classical capacity for every Pauli channel","Largest eigenvalue alone sets Pauli channel capacity","Generalized Pauli channels: capacity from coincident bounds","Pauli channels: one eigenvalue fixes the classical capacity"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The result stands on an imported upper-bound lemma that assumes, for every generalized Pauli channel, an optimal input $\\rho_*$ exists and that its channel output can be written as a mixture of unitary rotations of one fixed state; the paper states explicitly that it does not check whether $\\rho_*$ exists.","fun_headline_variants_meta":{"raw":{"variants":["Pauli channel capacity equals a simple formula","Exact classical capacity for every Pauli channel","Largest eigenvalue alone sets Pauli channel capacity","Generalized Pauli channels: capacity from coincident bounds","Pauli channels: one eigenvalue fixes the classical capacity"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000724,"raw_usage":{"total_tokens":3157,"prompt_tokens":764,"completion_tokens":2393,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":380,"completion_tokens_details":{"reasoning_tokens":2320}},"tokens_in":380,"tokens_out":2393,"duration_ms":18207,"temperature":1.0,"reasoning_tokens":2320,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:58:38.683956+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Numerically maximize the Holevo capacity over input ensembles for a $d=4$ generalized Pauli channel with mixed-sign eigenvalues, such as $\\lambda=(1/2,-1/4,-1/4,-1/4,-1/4)$; if the maximum exceeds $\\ln d - H[\\zeta(p)]$ from Eq. (32), or if any channel has a lower bound from Eq. (23) larger than its upper bound, the central theorem pair fails.","supporting_citations":[{"cited_title":"Siudzińska and D","cited_arxiv_id":null,"evidence_quote":"Supplies the Weyl-channel bounding method and the majorization construction that the paper adapts to generalized Pauli channels."},{"cited_title":"ur Rehman, Y","cited_arxiv_id":null,"evidence_quote":"Companion result for Weyl channels from which the upper-bound lemma is imported."},{"cited_title":"Siudzińska, Evolution of open quantum systems gov- erned by unitarily covariant quantum channels , Ph.D","cited_arxiv_id":null,"evidence_quote":"Introduces generalized Pauli channels as Pauli diagonal channels constant on axes and gives their complete-positivity conditions."},{"cited_title":"Smith, IEEE Information Theory Workshop (2010)","cited_arxiv_id":null,"evidence_quote":"Gives the proportionality between Holevo capacity and minimal output entropy for irreducibly covariant channels, used to extract the lower bound from basis-projector inputs."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies additivity for unital qubit channels and the complete-positivity inequalities used for Pauli-channel eigenvalues."},{"cited_title":"King and M","cited_arxiv_id":null,"evidence_quote":"Supplies additivity for depolarizing channels, recovered as the all-equal eigenvalue case."}],"review_version":1}