Open questions and potential solutions

Insights, fruitful questions and potential solutions arising from Monolithic Quantum Error Correction in the Presence of Distinguishability.

1. Research insights beyond the article

Research insight 01 · Jump-time variance controls recoverable phase loss

connectionproposed

Let $S$ have a Borel probability law supported on a bounded interval $[a, b]$ , let $S^{'}$ be an independent copy, and set $v = Var (S)$ and $ϕ (δ) = E e^{i δ S}$ for real $δ$ . On a logical qubit define $D_{ϕ}$ to preserve populations and multiply the $01$ coherence by $ϕ$ . For $0 \leq p \leq 1$ , consider the orthogonally flagged channel $N (X) = (1 - p) X \oplus p D_{ϕ} (X)$ . Its support-completed transpose recovery, with whole-input code and maximally mixed input $I / 2$ , has entanglement infidelity $L_{t r} = p (1 - ∣ ϕ ∣^{2}) / 2$ . It satisfies $0 \leq L_{t r} \leq \min {p / 2, p δ^{2} v / 2}$ and the explicit remainder bound $0 \leq p δ^{2} v / 2 - L_{t r} \leq p δ^{4} E (S - S^{'})^{4} / 48$ . Consequently, for a fixed timing law and fixed $p$ , $L_{t r} = p δ^{2} v / 2 + O (δ^{4})$ . The quadratic coefficient is sharp, deterministic shifts of $S$ do not change the loss, and $L_{t r} \leq p δ^{2} (b - a)^{2} / 8$ . Conditional on the article's four-level decay postulate, the dwell-time density is $γ e^{- γ (τ - s)} / (1 - e^{- γ τ})$ on $[0, τ]$ , with variance $γ^{- 2} - τ^{2} e^{- γ τ} / (1 - e^{- γ τ})^{2}$ for $γ, τ > 0$ . At fixed $γ, δ$ , this gives $L_{t r} = γ δ^{2} τ^{3} / 24 + O (τ^{4})$ .

Why it matters

The loss separates into a branch probability and uncertainty in the phase accumulated within that branch. This gives a quantitative bound for arbitrary bounded timing laws, explains why a deterministic delay is harmless to transpose fidelity, and identifies the variance underlying the four-level cubic coefficient.

Originality

The inspected original passages, PDF pp. 2–3 and S1 pp. 8–9, already give the coherence factor, exact transpose fidelity, and cubic coefficient. An article-wide extracted-text search found no occurrence of variance or characteristic function. The contribution here is the arbitrary-law variance inequality, its explicit remainder, and its connection to that existing calculation. Unobserved jump-time dephasing was already discussed by Michael et al., end of §V. Two targeted web searches were also performed; their results were not exhaustively reviewed. Neither the elementary probability argument nor the connection is claimed to be worldwide new.

Current knowledge

Steps 1–9 give a local proof of the stated mathematical identities, bounds, and conditional model specialization. The connection remains proposed for portal inclusion, with physical applicability explicitly conditional; it is not an open mathematical conjecture or an independently accepted library result. Fresh reading covered the five individually listed original-origin item bodies, the listed prerequisite bodies, original PDF images pp. 2, 3, 8 and 9, targeted article-text searches, and Michael et al.'s selected end-of-§V paragraphs. The two source records below are reproduced verbatim from the frozen report: their reading_limits describe that report's historical reading, not additional reading performed in this author task.

Argument and next steps

For an independent copy $S^{'}$ , bounded product integration gives $∣ ϕ ∣^{2} = E \cos (δ (S - S^{'}))$ and $E (S - S^{'})^{2} = 2 v$ . The elementary inequality $0 \leq x^{2} / 2 - (1 - \cos x) \leq x^{4} / 24$ yields both the variance bound and its remainder. Independently, support normalization makes transpose recovery act as the identity on the first positive-weight flag and as $D_{\overline{ϕ}}$ on the second. Composition therefore leaves coherence $∣ ϕ ∣^{2}$ on that branch, giving the stated entanglement infidelity. For the four-level postulate, direct matrix evolution produces the normalized dwell-time mixture; integration of its first two moments supplies the variance and cubic coefficient without using the article's fidelity claim as proof.

Proof sketch or overall strategy

proof_strategy: direct. Given: A bounded Borel timing law, real splitting, a state-independent flag probability, and the explicitly defined qubit phase-mixture channel. Compute its transpose recovery and Bell-state fidelity, then use an independent-copy identity and a scalar cosine estimate. Finally derive the four-level specialization directly from its postulated generator. The general mathematical argument has no physical premise; the physical postulate enters only steps 7–9.

Numbered argument

1.1 Put $U_{s} = diag (1, e^{- i δ s})$ . Entrywise integration gives $E (U_{S} X U_{S}^{†}) = D_{ϕ} (X)$ , where $ϕ = E e^{i δ S}$ , and the integral triangle inequality gives $r = ∣ ϕ ∣ \leq 1$ . For $r > 0$ write $z = ϕ / r$ ; for $r = 0$ choose any unit-modulus $z$ . With $U = diag (1, \overline{z})$ and $Z = diag (1, - 1)$ , direct multiplication gives $D_{ϕ} (X) = ((1 + r) / 2) U X U^{†} + ((1 - r) / 2) U Z X Z U^{†}$ . These are nonnegative unitary-mixture weights, so the phase map and its orthogonally flagged extension are channels. [given, algebra] [timing-law, flagged-channel, def-qec-prereq-finite-operator-conventions, math:thm-linearity-of-the-lebesgue-integral-on-l-one, math:thm-integral-triangle-inequality]
1.2 Write $q = 1 - p$ . Since $D_{ϕ} (I) = I$ , the noise sends the code identity to $A = q I \oplus p I$ . Its adjoint on diagonal output blocks is $N^{†} (Y_{0} \oplus Y_{1}) = q Y_{0} + p D_{\overline{ϕ}} (Y_{1})$ . Substituting the support inverse square root of $A$ into the transpose formula cancels each positive branch weight, giving branch maps $Y_{0} \mapsto Y_{0}$ and $Y_{1} \mapsto D_{\overline{ϕ}} (Y_{1})$ . Hence $T N = q id + p D_{r^{2}}$ . At $p = 0$ or $p = 1$ , omit the zero-weight support block; completion on that block never affects a noise output. [step 1.1, algebra] [step:1, flagged-channel, thm-qec-prereq-transpose-support-completion]
1.3 The recovered channel preserves populations and has coherence multiplier $λ = q + p r^{2}$ . Expand the Bell projector for $∣ Φ ⟩ = (∣ 00 ⟩ + ∣ 11 ⟩) / \sqrt{2}$ . Its two diagonal matrix units contribute $1 / 4$ each to entanglement fidelity, and its two off-diagonal units contribute $λ / 4$ each. Thus $F_{e} = (1 + λ) / 2$ , so $L_{t r} = p (1 - r^{2}) / 2$ . In particular $0 \leq L_{t r} \leq p / 2$ . [step 1.2, algebra] [step:2, def-qec-prereq-entanglement-fidelity]
1.4 On the product probability space carrying independent copies $S, S^{'}$ , all phase functions and moments used here are bounded and integrable. Fubini therefore gives $E e^{i δ (S - S^{'})} = (E e^{i δ S}) (E e^{- i δ S^{'}}) = ∣ ϕ ∣^{2}$ . Taking real parts yields $1 - ∣ ϕ ∣^{2} = E [1 - \cos (δ (S - S^{'}))]$ . Expanding the square and factoring the mixed expectation gives $E (S - S^{'})^{2} = 2 (E S^{2} - (E S)^{2}) = 2 v$ . [given, algebra] [timing-law, math:thm-fubini-theorem-for-l-one-on-sigma-finite-product-spaces, math:thm-linearity-of-the-lebesgue-integral-on-l-one]
1.5 For $x \geq 0$ , twice integrating by parts gives $Q (x) : = x^{2} / 2 - (1 - \cos x) = \int_{0}^{x} (x - t) (1 - \cos t) d t$ . Since $1 - \cos t \geq 0$ , this first proves $Q (x) \geq 0$ and hence $1 - \cos t \leq t^{2} / 2$ . Substituting that upper bound into the integral gives $Q (x) \leq \int_{0}^{x} (x - t) t^{2} / 2 d t = x^{4} / 24$ . Evenness extends both inequalities to every real $x$ . Apply them to $x = δ (S - S^{'})$ , integrate, and use steps 3–4: $L_{t r} \leq p δ^{2} v / 2$ and $0 \leq p δ^{2} v / 2 - L_{t r} \leq p δ^{4} E (S - S^{'})^{4} / 48$ . Boundedness makes the last moment finite, proving the fixed-law $O (δ^{4})$ assertion. [step 1.3, step 1.4, algebra] [step:3, step:4, math:thm-integration-by-parts, math:prop-order-and-scalar-rules-for-the-nonnegative-integral]
1.6 Replacing $S$ by $S + c$ multiplies $ϕ$ by $e^{i δ c}$ and leaves its modulus unchanged. If $m = E S$ , the inequality $(S - a) (b - S) \geq 0$ gives $v \leq (m - a) (b - m) \leq (b - a)^{2} / 4$ , proving the interval-width bound. For sharpness, take equal mass at $- h$ and $h$ , with $h > 0$ : then $v = h^{2}$ , $ϕ = \cos (δ h)$ and $L_{t r} = p \sin^{2} (δ h) / 2 = p δ^{2} h^{2} / 2 + O (δ^{4})$ . Thus the variance coefficient cannot be decreased uniformly. This example can be translated to a nonnegative timing interval. [step 1.3, step 1.5, algebra] [step:3, step:5, timing-law, lem-qec-prereq-finite-taylor-control]
1.7 Now assume only the four-level postulate, with $Δ = δ$ and initial logical operator $X$ in the upper code block. Direct substitution of its Hamiltonian and jump matrix into the generator gives upper block $e^{- γ t} X$ and lower populations $(1 - e^{- γ t}) X_{j j}$ . Writing the lower $23$ coherence as $c (t) X_{01}$ gives $c^{'} (t) = i δ c (t) + γ e^{- γ t}$ and $c (0) = 0$ . Multiplication by $e^{- i δ t}$ and integration yield $c (τ) = γ \int_{0}^{τ} e^{- γ t} e^{i δ (τ - t)} d t$ . For $γ, τ > 0$ , put $p = 1 - e^{- γ τ}$ and change variables to $s = τ - t$ . Then $c (τ) / p = \int_{0}^{τ} e^{i δ s} f_{τ} (s) d s$ , where $f_{τ} (s) = γ e^{- γ (τ - s)} / p$ integrates to one. The two output subspaces are orthogonal, so the restricted model channel has exactly the stipulated flagged form. No article fidelity formula is used. [given, algebra] [post-qec-distinguishability-four-level-model, def-qec-prereq-finite-operator-conventions, flagged-channel, math:thm-integration-by-parts]
1.8 For the density in step 7, integration by parts gives $E S = τ / p - 1 / γ$ and $E S^{2} = τ^{2} / p - 2 τ / (γ p) + 2 / γ^{2}$ . Subtracting the square of the first moment yields $v_{τ} = 1 / γ^{2} - τ^{2} e^{- γ τ} / p^{2}$ . Consequently the general variance inequality gives the exact upper bound $L_{t r} \leq (p δ^{2} / 2) [1 / γ^{2} - τ^{2} e^{- γ τ} / p^{2}]$ . The apparent cancellation in this expression near zero is handled by the next step, not by substituting into an undefined zero denominator. [step 1.5, step 1.7, algebra] [step:5, step:7, math:thm-integration-by-parts]
1.9 Set $u = S / τ$ and $η = γ τ$ . Its density on $[0, 1]$ is $g_{η} (u) = η e^{η u} / (e^{η} - 1)$ . Cancel $η$ in the denominator before expansion. Controlled exponential expansions give $g_{η} (u) = 1 + η (u - 1 / 2) + O (η^{2})$ uniformly in $u$ . Hence $E u = 1 / 2 + η / 12 + O (η^{2})$ and $E u^{2} = 1 / 3 + η / 12 + O (η^{2})$ , so $v_{τ} = τ^{2} / 12 + O (τ^{4})$ at fixed $γ$ . Also $p = γ τ + O (τ^{2})$ and $∣ S - S^{'} ∣ \leq τ$ . The remainder in step 5 is therefore $O (p δ^{4} τ^{4}) = O (τ^{5})$ at fixed parameters, giving $L_{t r} = γ δ^{2} τ^{3} / 24 + O (τ^{4})$ . When $γ τ = 0$ , the decay branch has zero weight and the loss is zero; no conditional dwell-time law is assigned. [step 1.5, step 1.7, step 1.8, algebra] ∎ [step:5, step:7, step:8, lem-qec-prereq-finite-taylor-control]

Immediate dependencies

timing-law: A Borel probability measure supported on a finite interval [a,b], a real parameter δ, and a fixed p in [0,1]. Expectation is integration against that measure; S and S' are coordinate variables on its product with itself, and $v = E$ (S−ES)^2. Source: self-contained; Insight statement and proof steps 1, 4–6. Reading limits: Explicit mathematical hypotheses and definitions supplied here; no physical timing law is inferred from them.
flagged-channel: $H = C$ ² with its ordered orthonormal basis; $K = H$ ⊕H with orthogonal coordinate inclusions. $U_{s}$ =diag(1,exp(−iδs)); Dφ(X)=E( $U_{S}$ X $U_{S}$ †), and N(X)=(1−p)X⊕pDφ(X). The flag probabilities are independent of the logical input. Source: self-contained; Insight statement and proof steps 1–2. Reading limits: Construction defined here; its finite Kraus representation and channel property are verified in step 1.
def-qec-prereq-finite-operator-conventions: Finite complex coordinate spaces use specified orthonormal bases, conjugate-transpose adjoints, outer products u v†, and ordered tensor-product bases. Source: def-qec-prereq-finite-operator-conventions; Definition and Justification. Reading limits: Read the complete frozen item body. Its deeper suppliers were not independently audited.
thm-qec-prereq-transpose-support-completion: For a finite-dimensional channel with Kraus operators $E_{a}$ and code projector P, put $A = N$ (P). The transpose operators are P $E_{a}$ † A^(−1/2), using the inverse on the support and zero on the kernel. The support map has a channel completion whose choice does not affect noise outputs of code-supported inputs. Source: thm-qec-prereq-transpose-support-completion; Statement; Proof steps 1.1–3.1. Reading limits: Read the complete frozen item, including its proof. Only the stated support/completion result is used; no library acceptance is claimed.
def-qec-prereq-entanglement-fidelity: For a qubit channel E and maximally mixed input, $F_{e}$ (E)=⟨Φ|(id⊗E)(|Φ⟩⟨Φ|)|Φ⟩, with |Φ⟩=(|00⟩+|11⟩)/√2. Source: def-qec-prereq-entanglement-fidelity; Definition and Justification. Reading limits: Read the complete frozen item. The required coherence-channel overlap is calculated directly in step 3.
math:thm-fubini-theorem-for-l-one-on-sigma-finite-product-spaces: For sigma-finite measure spaces and a complex L¹ function on their product, the product integral equals either iterated integral, with the stated almost-everywhere section convention. Source: math:thm-fubini-theorem-for-l-one-on-sigma-finite-product-spaces; Statement, clauses 1–4; Proof steps 1.1–3.1. Reading limits: Read the complete supplier file. Applied to bounded measurable functions on a product of probability spaces, which satisfies its finiteness and integrability hypotheses. Referenced books were not read.
math:thm-linearity-of-the-lebesgue-integral-on-l-one: L¹ is a complex vector space and integration is complex-linear on it. Source: math:thm-linearity-of-the-lebesgue-integral-on-l-one; Statement and Proof. Reading limits: Read the complete supplier file. All expectations to which it is applied are bounded-law integrals; underlying literature was not checked.
math:thm-integral-triangle-inequality: For a complex integrable function f, |∫f|≤∫|f|. Source: math:thm-integral-triangle-inequality; Statement; Proof steps 1.1–2.1. Reading limits: Read the complete supplier file. Used with the unit-modulus phase function; underlying literature was not checked.
math:prop-order-and-scalar-rules-for-the-nonnegative-integral: The integral of nonnegative measurable functions is monotone and homogeneous for nonnegative finite scalars, with the zero case handled separately. Source: math:prop-order-and-scalar-rules-for-the-nonnegative-integral; Statement, clauses 1–2; Proof. Reading limits: Read the complete supplier file. Used to integrate the nonnegative cosine bounds; no prerequisite-source audit was undertaken.
math:thm-integration-by-parts: For real functions differentiable on a compact interval, with integrable derivatives, ∫uv′=[uv]−∫u′v. The hypotheses hold for the polynomial, exponential, sine and cosine functions used here. Source: math:thm-integration-by-parts; Statement and Proof steps 1.1–5.1. Reading limits: Read the complete supplier file. Complex exponential calculations use real and imaginary components. All applications are smooth on compact intervals.
lem-qec-prereq-finite-taylor-control: Finite exponential expansions have controlled remainders; products and quotients preserve fixed-order expansions when the limiting denominator is nonzero. A vanishing denominator must first be canceled or extended. Source: lem-qec-prereq-finite-taylor-control; Statement; Proof steps 1.2, 2.1–2.2 and 3.1. Reading limits: Read the complete frozen item. Used for scalar exponentials and normalized densities at fixed parameters; no uniform claim over unbounded parameter families is made.
post-qec-distinguishability-four-level-model: Physical assumption: with ħ=1, H=ω₀(|0⟩⟨0|+|1⟩⟨1|)+δ|3⟩⟨3| and L=√γ(|2⟩⟨0|+|3⟩⟨1|), γ≥0, the state obeys the specified GKSL generator. The code is span{|0⟩,|1⟩}, orthogonal to the decay span. Source: post-qec-distinguishability-four-level-model; Postulate and Explanation; original attribution: PDF p. 2, Eqs. (2)–(3), and S1 opening on p. 8. Reading limits: Read the complete frozen postulate and inspected original PDF images pp. 2 and 8. Used solely as a physical assumption in steps 7–9; the original article supplies no independent proof of this assumption or of the variance result.

Independent review

Decision: accept

All nine steps are valid under the stated hypotheses. Independently checked the channel construction, support-normalized transpose recovery, Bell-state overlap, independent-copy identity, cosine remainder, sharpness, dwell-time moments and fixed-parameter asymptotics. The result is a proved scoped mathematical consequence with a conditional physics-theorem specialization. Its proposed portal status and qualified novelty description are appropriate.

Dependency review: Checked all twelve immediate dependency entries, including the two self-contained constructions. Read the complete four local mathematical supplier bodies, the four-level postulate and all five cited mathematical library files; the library files match their frozen hashes. Their hypotheses cover the bounded probability integrals, smooth compact-interval integrations and controlled expansions used here. Support completion handles $p = 0$ and $p = 1$ ; the argument also handles zero phase modulus and zero variance. The declared mathematical dependency closure contains no cycle or open-problem premise. The physical postulate enters only the model specialization; neither the article's fidelity formula nor a physical claim proves the general mathematical result. Michael et al., §V, p. 10 supports the qualitative jump-time dephasing connection. Nielsen, pp. 1–2 supports the entanglement-fidelity convention.

Review reading limits: Read the assigned frozen direction completely and all 113 original extraction-item bodies from its hash-addressed predecessor context; their individual IDs are listed above. This inventory reading is not a proof audit of unrelated article claims. Inspected the article's relevant extracted passages on pp. 1–3 and 8–9 and rendered PDF images pp. 8–9; searched the full extracted text for variance and characteristic-function terminology. Read Michael et al.'s selected end-of-§V paragraphs and Nielsen's cited definition passages. Did not read the suppliers' underlying textbooks, inspect every article page image, reproduce simulations or conduct an exhaustive novelty search. Historical reading claims reproduced in the draft are not adopted as this reviewer's reading. Empirical applicability of the stipulated dynamics remains unestablished, as the direction explicitly acknowledges.

Remaining gaps

The mathematical result assumes state-independent branch weights, orthogonal output flags, unchanged populations, and a classical mixture of relative phases. Its physical specialization is a physics-theorem conditional on the stipulated four-level postulate, not an experiment. No claim is made for amplitude-changing jumps, logical-path-revealing measurements, imperfect recovery, or noisy timing readout. Variance controls an upper bound and the small-splitting coefficient; it does not determine finite-splitting loss or order arbitrary timing laws. The coefficient concerns transpose entanglement fidelity at $I / 2$ , not every recovery or fidelity measure.

Author and reviewer gaps

Physical applicability remains conditional on the stated four-level GKSL postulate, its fixed coherent jump, and ideal transpose recovery. Their empirical or microscopic validity is not established here. The general flagged-channel identities and inequalities have no unresolved inference under their explicit mathematical hypotheses.

Sources and reading limits · 2

Monolithic Quantum Error Correction in the Presence of Distinguishability: https://arxiv.org/abs/2609.26946; Frozen PDF pp. 1–19, all text, equations, figures, captions and bibliography; S1 pp. 8–9; S2 pp. 9–11; S3 pp. 11–13; S4 pp. 14–16; S5 pp. 16–17. Reading limits: Read all 113 original extraction records and every rendered frozen PDF page, including all arguments. Read selected current authored replacements for the rank-one obstruction, ULE static/equation/error bound and vacuum density. The original is attribution/context, never independent proof of a prerequisite. No full simulation replay.
New Class of Quantum Error-Correcting Codes for a Bosonic Mode — Michael et al.: https://journals.aps.org/prx/pdf/10.1103/PhysRevX.6.031006; Printed p. 10, end of §V before §VI, recovery-stage infidelity and self-Kerr/jump-time paragraphs. Reading limits: Read the selected extracted paragraphs on recovery error per check and unobserved jump-time dephasing. No full-paper proof or experimental validation is claimed.

2. Potentially fruitful questions

Research question 01

For each displayed AE (J0=5) and Aydin–Barg (J0=4) code with the fixed jumps and lift Δ0 Jm, what is the first nonzero logical generator and entanglement-infidelity coefficient of Sτ^(T/τ), where Sτ=Raug exp(τL) V and Raug measures C, JzC, detected decay and the remaining complement? Specify the complement completion and keep Δ0 bounded but unknown.

Why it matters

A convergence bound does not identify which processes dominate the residual error. Fixing the complete recovery makes a sharp calculation possible and distinguishes errors from multiple decays from those caused by the unknown lift. It also prevents coefficients belonging to one completion from being attributed to every augmented check.

Originality

The frozen article leaves accumulated uncalibrated performance for future analysis. Frequent-measurement error suppression already appears in Erez et al. and Dominy et al.; no new general Zeno principle is claimed. Searches on 2026-11-14 for the article identifier, title and uncalibrated recovery did not locate a verified calculation of this particular completed-channel coefficient. That limited search establishes neither priority nor worldwide novelty.

Current knowledge

The inspected arXiv landing page lists v1, and the frozen PDF p. 16 identifies the broader extension as future work. The local argument below determines the requested leading terms for one fully specified reset completion and positive decay rate. This is a conditional physics-theorem calculation within an adopted finite Markov model, not an experimental result, library acceptance or independent review. Other completions and higher coefficients remain unresolved here.

Potential solution

Full solutionProposed approach

For the explicit recovery in F3, with every unused outcome reset to $I_{2} / 2$ , fix $γ > 0$ , $T > 0$ and $∣ Δ_{0} ∣ \leq D < \infty$ . For either displayed code and $τ = T / N$ , let $Q (X) = tr (X) I_{2} / 2$ and $A = (γ^{2} / 2) (Q - i d)$ . Then $S_{τ} = i d + τ^{2} A + O_{D} (τ^{3})$ and $S_{τ}^{N} = \exp (T τ A) + O_{T, D} (τ^{2})$ in diamond norm. Thus the first nonzero per-physical-time logarithmic generator is $τ^{- 1} \log S_{τ} = τ A + O_{D} (τ^{2})$ , and the maximally mixed logical entanglement infidelity is $1 - F_{e} (S_{τ}^{N}) = 3 γ^{2} T τ / 8 + O_{T, D} (τ^{2})$ . Both leading coefficients are independent of the unknown lift and identical for the two codes.

Argument and next steps

Direct codeword sums give the required orthogonal Hamiltonian and single-decay syndromes. Inverting those syndromes makes $R L V = 0$ . At second order, the corrected Hamiltonian-square term cancels, and the mixed Hamiltonian–dissipator terms cancel after summing all three polarizations. Two successive decays enter the reset complement, giving $R L^{2} V = γ^{2} (Q - i d)$ . The finite exponential remainder and a channel telescoping estimate yield the repeated-channel expansion. Finally, $F_{e} (Q) = 1 / 4$ gives the coefficient $3 γ^{2} T / 8$ . All these identities are derived below rather than assumed from the article.

Proof sketch or overall strategy

Compute the completed one-cycle channel through second order on every logical matrix. Verify the syndrome relations directly from the displayed coefficients, separate Hamiltonian, mixed and double-decay contributions, and control all remaining histories by the finite exponential series. Compare repeated cycles with an explicit depolarizing semigroup using diamond-norm telescoping. The physical evolution and instantaneous instrument are postulates; the resulting asymptotic deduction is conditional.

Numbered argument

Use the explicit coefficients in F1. Their squares satisfy $\sum_{δ} c_{k, m, δ}^{2} = 1$ and $\sum_{δ} δ c_{k, m, δ}^{2} = - m / k$ by expanding the three quadratic numerators. Consequently $\sum_{δ} L_{δ}^{†} L_{δ} = γ P_{e x c}$ , where $P_{e x c} = \sum_{k = 1}^{j} P_{k}$ . The evolution preserves matrices that are block diagonal in the spin label: each jump lowers both matrix indices by one and the Hamiltonian is block diagonal. Starting from the code, the scalar rotor energy therefore has zero commutator at every stage and can be omitted throughout the calculation. These are local algebraic deductions from the stipulated finite model. [F1, F2]
The two encodings in F2 are isometries because their columns are normalized and have disjoint magnetic supports. Direct weighted sums give $V^{†} Z V = 0$ and $V^{†} Z^{2} V = v I_{2}$ , with $v = 10$ for AE and $v = 8$ for Aydin–Barg. Substitution of the jump coefficients gives $V^{†} L_{δ}^{†} L_{ϵ} V = γ a_{δ} δ_{δ ϵ} I_{2}$ , where $(a_{- 1}, a_{0}, a_{+ 1}) = (1 / 3, 1 / 3, 1 / 3)$ and $(5 / 14, 2 / 7, 5 / 14)$ respectively. For AE, distinct support labels differ by at least three, eliminating all off-diagonal pair products. For Aydin–Barg, shifts of one have no overlapping support; the only possible shift-two logical overlap has two contributions $1 / \sqrt{126}$ and $- 1 / \sqrt{126}$ , which cancel. The diagonal products follow by inserting the squared coefficients. Thus $V$ , $B_{z} = Z V / \sqrt{v}$ and the three $B_{δ} = L_{δ} V / \sqrt{γ a_{δ}}$ have mutually orthogonal ranges and individually orthonormal columns. [F1, F2, step:1]
Define $R$ exactly as in F3. Step 2 makes $P_{⊥}$ an orthogonal projector. The five inverse-syndrome terms have Kraus operators $V^{†}$ , $B_{z}^{†}$ and $B_{δ}^{†}$ . For an orthonormal basis $(r_{b})$ of $im P_{⊥}$ , the reset term has Kraus operators $∣ a ⟩ ⟨ r_{b} ∣ / \sqrt{2}$ , with $a = 0, 1$ . Their adjoint products sum to $I_{H}$ , proving complete positivity and trace preservation directly. This realizes the four coarse outcomes in F3, followed by the specified decoder within the detected-decay outcome. In particular $R V = i d$ . The normalization removes $γ$ from each $B_{δ}$ , so the recovery uses neither $Δ_{0}$ nor a calibrated rotation angle. [F3, step:2]
Write $L = K + D$ , with $K (Y) = - i [H_{Δ}, Y]$ , $H_{Δ} = Δ_{0} \sum_{k, m} k m ∣ k, m ⟩ ⟨ k, m ∣$ , and $D = J - \frac{1}{2} {γ P_{e x c}, \cdot}$ , where $J (Y) = \sum_{δ} L_{δ} Y L_{δ}^{†}$ . For $Y = V (X)$ , the Hamiltonian commutator has only cross terms between the ranges of $V$ and $B_{z}$ , which the recovery measurement removes. Thus $R K V = 0$ . The syndrome inverses give $R J V (X) = γ \sum_{δ} a_{δ} X = γ X$ , while the anticommutator contributes $- γ X$ . Hence $R L V = 0$ on all logical matrices. [F1, F3, step:1, step:2, step:3]
The second-order Hamiltonian term also cancels. With $h = Δ_{0} j$ , projecting $- [h Z, [h Z, V (X)]]$ onto the code and decoding gives $- 2 h^{2} v X$ ; the $B_{z}$ branch gives $+ 2 h^{2} v X$ . Every other branch gives zero because $Z V$ lies in the $B_{z}$ range. For the mixed terms, scalar anticommutator losses reduce to multiples of the vanished first-order Hamiltonian output. The remaining terms are $R J K V (X) = - i [\sum_{δ} V^{†} L_{δ}^{†} L_{δ} H_{Δ} V, X] = 0$ and $R K J V (X) = - i [\sum_{δ} V^{†} L_{δ}^{†} H_{Δ} L_{δ} V, X] = 0$ . For the first equality use the scalar total rate, commutation of the diagonal jump products with $H_{Δ}$ , and $V^{†} Z V = 0$ . For the second, step 1 gives $\sum_{δ} (m + δ) c_{j, m, δ}^{2} = (j - 1) m / j$ , so the matrix inside the commutator is $γ Δ_{0} (j - 1)^{2} V^{†} Z V / j = 0$ . Therefore neither a $Δ_{0}^{2}$ term nor a $γ Δ_{0}$ term survives in $R L^{2} V$ . [F1, F3, step:1, step:2, step:3, step:4]
Both the host block $j$ and the single-decay block $j - 1$ have total rate $γ$ . Consequently $D^{2} V (X) = J^{2} V (X) - 2 γ J V (X) + γ^{2} V (X)$ . The first term is supported entirely in block $j - 2$ , hence in the reset complement. Its trace is $γ^{2} tr X$ , by applying the scalar total-rate identity successively on blocks $j$ and $j - 1$ . Therefore $R J^{2} V = γ^{2} Q$ and $R D^{2} V = γ^{2} (Q - i d)$ . Together with step 5 this proves $\frac{1}{2} R L^{2} V = A$ . The reset completion is used essentially in this step. [F1, F3, step:1, step:3, step:4, step:5]
Let $M_{D}$ be a finite uniform upper bound on $∥ L ∥_{⋄}$ for $∣ Δ_{0} ∣ \leq D$ . Such a bound exists directly from $L = L_{0} + Δ_{0} K_{1}$ and the norm triangle inequality. The exponential series, submultiplicativity and channel norm one give $∥ S_{τ} - i d - τ^{2} A ∥_{⋄} \leq M_{D}^{3} τ^{3} e^{M_{D} τ} / 6$ . This includes all higher jump histories and all higher powers of the unknown lift; no short-time Kraus truncation is treated as an exact channel. [F4, F5, step:3, step:4, step:6]
Put $E_{τ} = e^{τ^{2} A} = e^{- γ^{2} τ^{2} / 2} i d + (1 - e^{- γ^{2} τ^{2} / 2}) Q$ , which is a channel. Its Taylor remainder after $i d + τ^{2} A$ is bounded by $∥ A ∥_{⋄}^{2} τ^{4} e^{τ^{2} ∥ A ∥_{⋄}} / 2$ . Telescoping two products of channels yields $∥ S_{τ}^{N} - E_{τ}^{N} ∥_{⋄} \leq N ∥ S_{τ} - E_{τ} ∥_{⋄}$ . With $N = T / τ$ , this is at most $T M_{D}^{3} τ^{2} e^{M_{D} τ} / 6 + T ∥ A ∥_{⋄}^{2} τ^{3} e^{τ^{2} ∥ A ∥_{⋄}} / 2$ . Thus $S_{τ}^{N} = e^{T τ A} + O_{T, D} (τ^{2})$ . The convergent logarithm series near the identity also gives $τ^{- 1} \log S_{τ} = τ A + O_{D} (τ^{2})$ ; no assertion that the exact logarithm is a GKSL generator at every finite interval is needed. [F4, F5, step:7]
Under F6, $F_{e} (i d) = 1$ and $F_{e} (Q) = 1 / 4$ , since applying $Q$ to one half of the Bell state gives $I_{4} / 4$ . Therefore $F_{e} (e^{T τ A}) = (1 + 3 e^{- γ^{2} T τ / 2}) / 4$ . Bell-projector testing changes by at most the diamond distance, so step 8 gives $1 - F_{e} (S_{τ}^{N}) = 3 γ^{2} T τ / 8 + O_{T, D} (τ^{2})$ . As a separate exact consistency check, at $Δ_{0} = 0$ the probabilities of zero and one decay are $e^{- γ τ}$ and $γ τ e^{- γ τ}$ , their outputs are corrected exactly, and every history with at least two decays is reset. Hence $S_{τ} = (1 + γ τ) e^{- γ τ} i d + [1 - (1 + γ τ) e^{- γ τ}] Q$ , whose repeated fidelity has the same leading coefficient. [F6, step:3, step:4, step:6, step:8]

Immediate dependencies

F1: Physical model postulate, not an observation: fix $j \in {4, 5}$ and $H = ⨁_{k = 0}^{j} C^{2 k + 1}$ with orthonormal basis $∣ k, m ⟩$ . Adopt the finite Markov channel $e^{t L}$ with $L (Y) = - i [H_{r o t} + H_{Δ}, Y] + \sum_{δ = - 1}^{1} (L_{δ} Y L_{δ}^{†} - {L_{δ}^{†} L_{δ}, Y} / 2)$ , $H_{r o t} ∣ k, m ⟩ = ω_{0} k (k + 1) ∣ k, m ⟩$ , $H_{Δ} ∣ k, m ⟩ = Δ_{0} k m ∣ k, m ⟩$ , $γ > 0$ , and constant unknown $∣ Δ_{0} ∣ \leq D$ . Set $L_{δ} ∣ k, m ⟩ = \sqrt{γ} c_{k, m, δ} ∣ k - 1, m + δ ⟩$ , annihilating $k = 0$ and inadmissible outputs. The real normalized coefficients are $c_{k, m, 0} = - \sqrt{(k^{2} - m^{2}) / (k (2 k - 1))}$ and $c_{k, m, \pm 1} = \sqrt{(k \mp m) (k \mp m - 1) / (2 k (2 k - 1))}$ . These specify the equal-rate fixed-jump benchmark explicitly. Each cycle uses the same noise channel and starts from the re-encoded recovered logical state. Source: self-contained; Adopted finite effective model for this direction; coefficient representatives for the benchmark jumps.. Reading limits: The original model choices were checked against frozen PDF page images 3 and 14. They are adopted physical assumptions, not independently established microscopic facts. No bath derivation, Gaussianity, ultraviolet regularity or laboratory applicability is inferred. The displayed coefficient formulas can also be checked without importing a Clebsch–Gordan theorem: dividing by sqrt((2k+1)/(2k-1)), assembling the coefficients at fixed total weight, and substituting into the standard finite spin ladder equations gives normalized spin-(k-1) coupled vectors. The proof uses the explicit matrices themselves.
F2: Mathematical definitions: $Z ∣ j, m ⟩ = m ∣ j, m ⟩$ and $V (X) = V X V^{†}$ . For AE, $j = 5$ , $V ∣ 0 ⟩ = \sqrt{5 / 7} ∣ 5, - 2 ⟩ + \sqrt{2 / 7} ∣ 5, 5 ⟩$ , $V ∣ 1 ⟩ = \sqrt{2 / 7} ∣ 5, - 5 ⟩ + \sqrt{5 / 7} ∣ 5, 2 ⟩$ . For Aydin–Barg, $j = 4$ , $V ∣ 0 ⟩ = \sqrt{1 / 3} ∣ 4, - 4 ⟩ + \sqrt{2 / 3} ∣ 4, 2 ⟩$ , $V ∣ 1 ⟩ = \sqrt{2 / 3} ∣ 4, - 2 ⟩ - \sqrt{1 / 3} ∣ 4, 4 ⟩$ . Adjoints and inner products use the ordinary complex Euclidean convention. Source: self-contained; Explicit numerical encoding data; normalization, moments and required pair products proved in step 2.. Reading limits: The code data and negative amplitude were checked in frozen PDF page image 6. The article supplies attribution for the chosen vectors, not proof of their correction properties. No full absorption–emission theorem is imported or needed for this restricted downward-error calculation.
F3: Recovery definition and ideal-instrument postulate: put $v = 10$ for AE or $8$ for Aydin–Barg, $B_{z} = Z V / \sqrt{v}$ , and $B_{δ} = L_{δ} V / \sqrt{γ a_{δ}}$ , with $(a_{- 1}, a_{0}, a_{+ 1}) = (1 / 3, 1 / 3, 1 / 3)$ or $(5 / 14, 2 / 7, 5 / 14)$ respectively. Let $P_{⊥} = I - V V^{†} - B_{z} B_{z}^{†} - \sum_{δ} B_{δ} B_{δ}^{†}$ . Define $R (Y) = V^{†} Y V + B_{z}^{†} Y B_{z} + \sum_{δ} B_{δ}^{†} Y B_{δ} + tr (P_{⊥} Y) I_{2} / 2$ . Operationally measure the four orthogonal coarse outcomes $V V^{†}$ , $B_{z} B_{z}^{†}$ , $P_{j - 1}$ and their remaining complement. Within $P_{j - 1}$ invert the three orthogonal jump syndromes and reset its unused subspace. Reset the entire remaining coarse complement, including the unused host subspace and every block at or below $j - 2$ . All operations are exact and instantaneous; re-encode after decoding. Source: self-contained; Specified completion for the assigned question; well-definedness and CPTP property proved in steps 2–3.. Reading limits: The proposed host Jz-syndrome inverse was checked against frozen PDF page image 16 and the current augmented-check body. The complement reset and fully specified decay decoder are local choices, not operations claimed to have been specified or implemented by the original authors.
F4: For finite-dimensional linear maps, the diamond norm is the induced trace norm after adjoining an identity register of the input dimension. Channels have diamond norm one, and composition satisfies $∥ Ψ Φ ∥_{⋄} \leq ∥ Ψ ∥_{⋄} ∥ Φ ∥_{⋄}$ . These facts apply to encoding, decoding and noise channels of different finite dimensions. Source: https://cs.uwaterloo.ca/~watrous/TQI/TQI.pdf; Section 3.3.2, Definition 3.43 and Proposition 3.44(2), printed pp. 170–171; Proposition 3.48(1) and its proof, printed p. 174.. Reading limits: Freshly read the retrieved PDF text containing these definitions, statements and the composition proof. These are independent mathematical inputs. The book's broader channel theory and its foundational proofs of every preceding norm fact were not reconstructed.
F5: Finite-dimensional series calculus: $e^{B} = \sum_{n \geq 0} B^{n} / n!$ and, for $∥ C ∥ < 1$ , $\log (I + C) = \sum_{n \geq 1} (- 1)^{n + 1} C^{n} / n$ . Submultiplicativity gives absolute convergence. In particular the exponential tail after degree two is at most $∥ B ∥^{3} e^{∥ B ∥} / 6$ , and $\log (I + C) = C + O (∥ C ∥^{2})$ . The tail inequality follows from $(n + 3)! \geq 6 n!$ ; the logarithm remainder is bounded by $\sum_{n \geq 2} ∥ C ∥^{n}$ . Source: self-contained; Elementary norm-series estimates used in steps 7–8.. Reading limits: Local estimates supplied explicitly; no external perturbation, Zeno-limit or effective-generator theorem is assumed.
F6: Mathematical convention: $∣ Φ ⟩ = (∣ 00 ⟩ + ∣ 11 ⟩) / \sqrt{2}$ and $F_{e} (E) = ⟨ Φ ∣ (i d \otimes E) (∣ Φ ⟩ ⟨ Φ ∣) ∣ Φ ⟩$ . This is entanglement fidelity for maximally mixed logical input and is linear in the channel. For a unit vector $u$ , $∣ ⟨ u ∣ X u ⟩ ∣ \leq ∥ X ∥_{1}$ : expand a singular-value decomposition and bound each scalar product by one. Therefore Bell-projector testing of two channel outputs differs by at most their diamond distance. Source: self-contained; Definition and elementary testing inequality for step 9.. Reading limits: Definition fixed here and the required inequality justified locally. No article fidelity formula or imported optimal-recovery assertion is used.

Independent review

Decision: accept

All nine argument steps support the stated scoped solution. The recovery is CPTP and independent of the unknown lift. Direct calculation gives R L $V = 0$ and R L² V = gamma²(Q−id): the Hamiltonian-square and summed mixed contributions cancel, while two decays reach the reset complement. Independent exact radical arithmetic confirmed the syndrome relations and these identities on all four logical matrix units for both codes. The exponential remainder is uniform for bounded lift, and channel telescoping gives the claimed repeated-channel expansion. Consequently A=(gamma²/2)(Q−id) and 1−Fe=3 gamma² T tau/8+O(tau²). Acceptance concerns this conditional physics-theorem for the specified completion.

Dependency review: Reviewed every immediate premise F1–F6. F1 explicitly postulates the finite physical channel; its rate and polarization-weight identities follow from the displayed coefficients. F2 fixes the numerical encodings; direct moments and pair products establish exactly the syndrome properties used, without importing the broader absorption–emission correction claims. F3 specifies the entire recovery, including unused outcomes, and its Kraus completeness proves CPTP. F4 is an applicable independent mathematical input: Watrous Definition 3.43, Proposition 3.44(2) and Proposition 3.48 provide the norm properties needed for maps between different finite dimensions. F5's series estimates and F6's Bell-state convention and testing inequality are justified locally. The article supplies attribution and model context, not proof of these prerequisites. Erez and Dominy supply historical context only; their Zeno results are not applied to this Lindbladian. No open conclusion or circular dependency is used. The uniform asymptotics hold for fixed positive gamma and T, bounded lift, and $t a u = T$ /N.

Review reading limits: Read the assigned direction in full, including its nine steps, six immediate dependencies and limitations. Inspected all 113 original extraction bodies individually; this inventory does not signify acceptance of those unrelated claims. Read the current equal-rate, lift, code-definition and augmented-check bodies relevant to the direction, and inspected frozen PDF images 3, 6, 14 and 16. Fresh external reading covered the arXiv landing page, the identified Watrous norm passages, Dominy's abstract/introduction and Theorem 1 assumptions and bounds, and Erez's abstract/introduction. Did not read every frozen PDF page, every current replacement or the full library closure. The exact arithmetic check supplements the algebraic review; no experimental validation, microscopic derivation or full simulation replay was performed. No proposal file was written or hashed.

Remaining gaps

Full refers to the requested first nonzero terms for this specified completion and $γ > 0$ . It does not determine the next, lift-dependent coefficient, optimize the complement decoder, or settle the article's broader higher-order problem. A decoder retaining information from two-decay outputs can change the leading generator. Finite-duration checks, memory effects, changing jump amplitudes and the zero-decay boundary are outside the result.

Author and reviewer gaps

No unresolved algebraic inference remains in the stated leading-order result for the specified reset completion. The finite Markov evolution, fixed jump amplitudes, instantaneous measurement, exact syndrome inverses and re-encoding are conditional physical assumptions, not verified observations.
The next lift-dependent coefficient has not been derived. Equality of the leading terms does not establish equal finite-interval performance or equal higher-order robustness of the two codes.
The reset completion discards all two-decay information. Other completions can change the leading generator; no optimality or completion-independent coefficient is established.
The positive-rate hypothesis is essential to describing the displayed term as first nonzero. At $g a m m a = 0$ it vanishes, and the leading purely Hamiltonian higher-order behavior requires a separate calculation.
Fresh reading covered all 113 original extraction bodies, the relevant current code/model/augmented-check bodies, and frozen PDF page images 3, 6, 14 and 16. This was not a fresh reading of all current approved replacements, all their prerequisite proofs, or every PDF page. The items_read inventory records inspected original extraction bodies without granting acceptance.
The four source records below are reused verbatim from the frozen report. Their reading_limits describe that earlier source inventory, not additional reading performed by this author. Fresh external reading here was the arXiv landing page, the identified Dominy abstract/introduction and Theorem 1 passages, the Erez abstract/introduction, and the specified Watrous norm passages. No unavailable source is used as a proof premise.
A read-only floating-point check evaluated syndrome orthogonality and the two generator identities on all four logical matrix units for both codes at three lift values; residuals were below 5e-13. This was a sanity check, not exact-arithmetic proof, independent review or simulation replay. The argument rests on the displayed algebra.
The status search was targeted and did not establish exhaustive literature coverage or worldwide novelty. The original article is used for question and model attribution, never as independent proof of its own correction or convergence claims.
No unresolved inference remains in the stated leading-order result. Its finite Markov evolution, fixed jump amplitudes, instantaneous checks, exact decoding and re-encoding remain physical postulates.
The next lift-dependent coefficient and finite-interval comparison between the codes remain undetermined.
Other complement recoveries can change the leading generator; optimality and completion independence are not established.
The $g a m m a = 0$ boundary, finite-duration operations, bath memory and lift-dependent jump amplitudes are outside the proved scope.
The targeted status search does not establish exhaustive literature coverage or worldwide novelty.

Sources and reading limits · 4

Monolithic Quantum Error Correction in the Presence of Distinguishability: https://arxiv.org/abs/2609.26946; Frozen PDF pp. 1–19, all text, equations, figures, captions and bibliography; S1 pp. 8–9; S2 pp. 9–11; S3 pp. 11–13; S4 pp. 14–16; S5 pp. 16–17. Reading limits: Read all 113 original extraction records and every rendered frozen PDF page, including all arguments. Read selected current authored replacements for the rank-one obstruction, ULE static/equation/error bound and vacuum density. The original is attribution/context, never independent proof of a prerequisite. No full simulation replay.
Analysis of the quantum Zeno effect for quantum control and computation — Dominy et al.: https://arxiv.org/html/1207.5880v2; Abstract and §I; §II.3, Theorem 1, assumptions and Eqs. (10)–(13). Reading limits: Read these HTML passages for established related Zeno bounds and their stabilizer/bounded-Hamiltonian scope. The theorem is not applied to the atomic Lindbladian; the channel bound below is derived separately.
Quantum Error Correction in the Zeno Regime — Erez et al.: https://arxiv.org/html/quant-ph/0309162v2; Abstract and §1, first-order error correction and frequent syndrome projections. Reading limits: Read abstract and introduction only for prior conceptual context. No claim of a new general Zeno principle or reconstruction of that paper’s code.
The Theory of Quantum Information — John Watrous, author’s 2018 draft: https://cs.uwaterloo.ca/~watrous/TQI/TQI.pdf; §3.3.2, Definition 3.43 and Proposition 3.44(2), printed pp. 170–171; Proposition 3.48 and proof, printed p. 174 (PDF p. 182). Reading limits: Read the extracted definition/properties of the completely bounded trace norm, channel norm one and composition inequality. These finite-dimensional mathematical facts support the local telescoping bound; the rest of the book was not read.