{"id":"072c17d5-fd43-4315-a797-9b74675a2d15","arxiv_id":"2607.19770","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"Under latency constraints, longer quantum error-correcting codes can reduce teleportation reliability because acquiring more entangled pairs delays and degrades them; adaptive puncturing of a base code selects the best code length per regime.","lead":"This paper models how slow, unreliable entanglement generation and memory decay affect quantum teleportation when quantum error-correcting codes are used. It shows that longer codes can actually hurt under tight latency, and that adaptively choosing a punctured code improves reliability.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The unvalidated waiting-time approximation in Eq. (9) is the load-bearing weak point; without an exact or Monte-Carlo check, the common-latency comparison and decision regions in Figs. 6-9 are unproven.","rationale":"The reader's weakest_assumption identifies Eq. (9) precisely, and this is indeed the most load-bearing concern. The paper's central claim is comparative: it must show that, under the same average latency constraint, different punctured code lengths are optimal in different regimes. That comparison is constructed by solving E[L(w,n)] = L for each n, so the approximation used for E[L(w,n)] directly determines the operating point p for each tier. If the approximation error is not uniform in n, the comparison is biased and the decision regions may be artifacts. The authors assert conservatism and accuracy as p approaches 1 but provide no evidence, and the exact evaluation is infeasible for the largest tiers they consider. This is an internally unverified premise rather than a disagreement with external consensus, so it does not by itself overturn the qualitative latency-reliability tradeoff, but it does prevent the quantitative decision maps from being accepted as established. I considered the product formula in Eq. (6) for logical error probability as an alternative concern; it is also an approximation under depolarizing noise because X and Z branch successes are correlated through Y errors. However, a common bias across all tiers is less likely to change the relative ordering that determines n*, whereas Eq. (9) injects a per-tier bias into the latency constraint itself. The proposed Monte Carlo test is straightforward and would settle whether the approximation error is material. The verdict CONDITIONAL remains appropriate: the framework is coherent, but the key numerical comparison needs validation before the claimed decision regions can be trusted.","tokens_in":11868,"tokens_out":7056,"duration_ms":66589,"concrete_test":"Run a Monte Carlo simulation of the discrete-time Bernoulli generation/storage process for w=20, using the p values that Eq. (9) yields for n=1,8,13,17 across the latency range L=20..400. Estimate E[L(w,n)] to high precision (e.g., 10^6 realizations) and compare with Eq. (9). If the relative error is non-monotonic in n or exceeds a few percent at any operating point used in Figs. 6-9, recompute the decision regions using the empirical p that satisfies the exact constraint and check whether any crossover point or n* assignment changes.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central comparison under a shared latency constraint is made by inverting Eq. (9): for each code length n, p is chosen so that E[L(w,n)] = L, and then initial fidelity follows from Eq. (7). Eq. (9) is an approximation whose error is never quantified. The text states that 'Comparison with exact evaluations shows that (9) conservatively overestimates' the expected waiting time, but no such comparison is shown. For n=13 and n=17, the exact linear system of dimension binomial(w-1,n-1) is intractable, so those tiers cannot be validated by the exact method described. If the approximation error grows with n, then at fixed nominal L the different tiers are actually operated at different true average waiting times, and the 'common latency constraint' is violated. Since p enters F0(p) in Eq. (7) and the heterogeneous fidelity distribution, a p bias that varies with n can alter the relative ordering of the PL curves and shift the crossover points in Figs. 6-9, thereby changing the optimizing tier n*(L,p_d,eta) in Eq. (11). This is not a disagreement with external consensus; it is an internally unverified numerical premise that the decision regions depend on.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper studies latency-constrained encoded quantum teleportation. The authors model stochastic entanglement generation as Bernoulli trials with success probability p, relate p to initial fidelity through Eq. (7), and apply exponential memory decoherence within a storage window of size w. For a target average latency L, they invert the approximate waiting-time expression Eq. (9) to obtain the generation probability p for each code length n. They then simulate packet assembly, obtain heterogeneous fidelities, map them to Pauli error probabilities, and compute logical error probability PL for CSS codes using the Poisson-binomial expressions Eqs. (4)–(6). The puncturing tiers considered are n=17, 13, 8, and uncoded n=1. The central claim is that longer codes do not uniformly help because acquiring more pairs increases waiting time and decoherence, so the optimal code length depends on latency and link parameters via Eq. (11). Numerical results in Figs. 6–9 show decision regions for symmetric and asymmetric noise.","tokens_in":12173,"tokens_out":6321,"duration_ms":61653,"significance":"If the quantitative results are reliable, the paper offers a useful engineering framework for adaptive code selection in quantum networks and highlights a genuine latency–reliability tradeoff. The model is clearly stated, and the simulation methodology is transparent, with all parameters identified. The paper builds on established components—Poisson-binomial decoding, the generation-fidelity relation from [8], and punctured codes from [7]—rather than introducing ad hoc entities. The main issue is that the waiting-time approximation in Eq. (9) is load-bearing and is not validated in the manuscript. Because this approximation is the only mechanism for mapping the latency constraint L to the generation probability p, and hence to all fidelities and error probabilities, the quantitative decision regions and the argmin in Eq. (11) are unproven until the approximation error is characterized. The central qualitative insight—that resource-acquisition latency should influence code selection—is plausible and likely correct, but the specific crossover points and the claimed 'substantial gains' rely on numerical accuracy.","major_comments":[{"comment":"Eq. (9) is the only way the latency constraint enters the model. It is used to solve for p such that E[L(w,n)]=L for each n, and the resulting p determines F0(p) via Eq. (7). The text asserts that 'Comparison with exact evaluations shows that (9) conservatively overestimates the expected waiting time,' but no such comparison is shown. For n=13 and n=17 the exact linear system is intractable, so the claim cannot be checked in the operating regime relevant to the paper's main comparison. If the approximation error varies with n, then at a fixed nominal L the different tiers are actually operated at different true average waiting times, violating the 'common latency constraint.' This could alter the relative ordering of the PL curves and shift the decision regions in Figs. 6–9 and the optimal tier n*(L,pd,η) in Eq. (11). Please add a derivation of Eq. (9), a validation against exact enumera","section":"III-A, Eq. (9)"},{"comment":"After solving Eq. (9) for p, the simulation samples the actual generation process. The paper does not report whether the simulated average waiting time per tier equals the nominal L. This is a directly checkable consequence of the model and is essential to verify the 'common average latency constraint' that the comparison is premised on. Please report the empirical mean (and, if possible, the distribution) of the waiting time for each tier in the regimes of Figs. 6–9, and discuss any deviation from L. If the approximation systematically biases the true average latency, the comparison is not fair.","section":"IV, numerical procedure"},{"comment":"The authors claim the overestimate in Eq. (9) is conservative and leads to a 'conservative bias.' This is only conservative with respect to meeting the latency requirement; it is not automatically conservative for code selection. An overestimate of waiting time increases p, which decreases the initial fidelity F0(p). Because longer codes require more pairs and hence larger p, a frequency-dependent error in Eq. (9) could penalize longer codes more and bias the decision rule toward shorter codes. The paper should quantify the sensitivity of the decision regions in Figs. 6–9 to the approximation error, for example by repeating the analysis with a perturbed Eq. (9) or using an exact/Monte Carlo wait-time constraint for the small-n cases.","section":"III-A, conservative bias"}],"minor_comments":[{"comment":"The relation F0(p) is taken from [8] but no derivation is given; a sentence explaining its origin would help the reader understand the parameter M and the detection probability pd.","section":"II-B, Eq. (7)"},{"comment":"The symbol L is used both for the latency constraint and, in E[L(w,n)], for the waiting time. This is confusing; consider using a different symbol (e.g., W) for the waiting-time random variable.","section":"III-A"},{"comment":"The legends read ' =1', ' =8', etc.; the missing variable (presumably n) should be included. Adding gridlines or error bars would improve readability.","section":"Figs. 6–9"},{"comment":"Monte Carlo results are presented without confidence intervals. Given that some PL differences between tiers are small, error bars or a statistical significance statement would be helpful.","section":"IV"},{"comment":"The text states that T=10000 time slots 'corresponds to a coherence time on the order of one second,' but no time-slot duration is specified. Please state the assumed slot duration or phrase this as an example.","section":"IV"},{"comment":"The caption says 'Minimum fidelity' but the text describes it as the oldest stored pair. Clarify in the caption that the minimum is over the qubits in the packet at the time of completion.","section":"Fig. 4"}],"recommendation":"major_revision","confidential_remarks":"The paper is within scope and the idea is timely. The main technical obstacle is the unvalidated approximation in Eq. (9). If the authors can supply the requested validation and demonstrate that the decision regions are robust, I would be happy to recommend acceptance. I would also encourage them to make the simulation code available to strengthen reproducibility."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: the paper couples stochastic entanglement acquisition with encoded teleportation reliability and uses punctured codes to adapt code length. The qualitative tradeoff is real and the framework is clearly built. The soft spot is the waiting-time approximation in Eq. (9), which carries the comparison and is never actually validated.\n\nWhat's new: prior work [7] fixed entanglement availability and quality. Here, generation probability, memory decoherence, and code selection interact under an average latency constraint. The decision regions in Figs. 6–9 are new, and the idea of discrete puncturing tiers is sensible. The model is internally consistent, parameters are explicit, and the Monte Carlo average over 10^5 realizations is stated.\n\nThe load-bearing issue is Eq. (9). For each n, they solve E[L(w,n)] = L to get the generation probability p, then feed p into F0(p) in (7). The text says a comparison with exact evaluations shows (9) conservatively overestimates, but no comparison is shown. For n=13 and n=17 the exact linear system is intractable, so those tiers can't be checked that way. If the approximation error grows with n, the 'common latency constraint' is violated: tiers are actually run at different true average waiting times. Since p sets initial fidelity, this can shift the PL curves and change the optimal tier in (11). This is a real, fixable weakness. A Monte Carlo estimate of the true expected waiting time, or an exact solve for the smaller tiers, would settle the bias.\n\nMinor: the punctured-code construction is imported from [7] without details, and no code or data is shipped, which makes reproduction harder than it should be.\n\nThe central qualitative conclusion — longer codes can lose when resource acquisition degrades fidelity — is robust and doesn't hinge on the approximation. The paper is useful for quantum-network designers working on resource-adaptive coding. I'd send it to peer review, with the request that reviewers push for validation of Eq. (9) and, ideally, a code release.","headline":"A coherent simulation study with a useful qualitative tradeoff, but the common-latency comparison rests on an unvalidated approximation.","tokens_in":12679,"tokens_out":2453,"would_cite":false,"duration_ms":22245,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["81P68","81P70"],"pacs":["03.67.-a","03.67.Hk"],"model":"deepseek-v4-flash","headline":"Under a latency constraint, the best quantum teleportation code is the one matching the time it takes to gather entangled pairs, and puncturing a single base code supplies that flexibility.","keywords":["quantum teleportation","entanglement generation","quantum error correction","code puncturing","latency constraint","memory decoherence","logical error probability","adaptive coding"],"falsifier":"Compute the exact expected waiting time for a moderate packet size (e.g., n=8) by solving the stated linear system of dimension (w−1 choose n−1), and re-run the latency-constrained comparison; if the crossover latencies where the optimal code changes shift, Eq. (9) is not trustworthy enough to support the decision rule.","tokens_in":11708,"feed_emoji":"⚛️","tokens_out":4411,"duration_ms":41961,"temperature":0.7,"pith_summary":"Quantum teleportation needs shared entangled pairs, but in a real network those pairs arrive one by one and decay while waiting in memory. The paper argues that the logical error probability of encoded teleportation therefore depends not just on the code's error-correction power but on how long it takes to acquire the required number of pairs. It develops a unified model that couples stochastic entanglement generation, memory decoherence, and CSS-code decoding, and uses it to compare uncoded teleportation with a family of punctured codes derived from a single length-17 base code. The central result is a decision rule: for each latency budget, link quality, and noise asymmetry, there is an optimal code length, and no fixed code wins everywhere. A sympathetic reader should care because this identifies a resource-aware adaptation principle for quantum networks, where code choice is a network-layer decision, not just a coding-theoretic one.","feed_headline":"Latency, not code strength, picks the winning teleportation code","feed_subtitle":"A framework shows the optimal code length shifts with entanglement-acquisition time, so networks should puncture a shared base code.","key_machinery":"The central object is the latency-constrained decision rule n*(L,p_d,η)=argmin_n PL(n;L,p_d,η), which selects the puncturing tier (effective code length) that minimizes logical error probability for a given latency budget, link quality, and noise asymmetry. It is evaluated through a pipeline: the waiting-time approximation E[L(w,n)]≈(1/Pe − 1)(w + n)/(2p) + n/p in Eq. (9) is inverted to find the generation probability p satisfying the latency constraint; p sets the initial fidelity F0(p) via Eq. (7); memory decoherence Eq. (8) produces heterogeneous per-qubit fidelities; and CSS decoding with Poisson-binomial error distributions yields PL via Eq. (6). The puncturing tiers — [[17,1,5,5]], [[1","core_discovery":"The paper's central claim is that under a common average latency constraint, the reliability of encoded teleportation is governed by an entanglement-acquisition tradeoff: longer codes provide stronger error correction but require larger entanglement packets, which means higher generation probabilities (and hence lower initial fidelity) and longer storage times (and hence more decoherence). The authors show numerically, using [[17,1,5,5]], [[13,1,5,3]], [[8,1,3,3]] puncturing tiers and uncoded teleportation, that the logical error probability PL(n; L,p_d,η) exhibits crossover behavior, so the optimal code length n*(L,p_d,η)=argmin_n PL depends on the latency budget and link conditions. In sym","pith_inferences":["The latency-fidelity coupling implies that in early quantum access networks, a single network-wide base code with per-user puncturing could be a simpler deployment model than per-link code selection; this is an architectural inference beyond the paper's numerical scope.","The waiting-time approximation's conservative bias means the reported decision regions may shift if exact waiting times are used; a natural extension is to quantify the gap for moderate n where the linear system is still solvable.","The framework could be extended to dynamic online code selection based on instantaneous packet availability, as the authors themselves mention in future work; the decision regions suggest hysteresis effects when latency budgets fluctuate.","Because puncturing reduces code distance, the framework implies a direct tradeoff between resource availability and error-correction strength that could also inform entanglement purification decisions."],"forward_implications":["If true, a quantum network should not fix a single code; it should precompute a policy map and switch puncturing tiers per request based on latency budget and link quality.","Longer codes are not universally better: under tight latency budgets or high-quality links, shorter codes or uncoded transmission can beat long codes, so resource-aware adaptation is essential.","The same base code can serve many operating regimes via puncturing, which reduces implementation overhead compared to switching among unrelated code families.","Under asymmetric noise (phase errors dominating), a specially punctured code like [[13,1,5,3]] offers the best reliability over wide latency and link ranges, suggesting asymmetry should guide code design.","The framework provides a way to set entanglement generation probability from a latency requirement, linking application-level quality-of-service to physical-layer entanglement parameters."],"fun_headline_variants":["Punctured codes flex teleportation for latency-constrained nets","Quantum teleportation codes adapt to entanglement delays","Latency rules code choice in encoded teleportation","Adaptive puncturing boosts teleportation under latency"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The paper's quantitative comparison relies on an approximate formula for how long it takes to collect a packet of entangled pairs; if that approximation is wrong for longer codes, the claimed optimal code choices could change.","fun_headline_variants_meta":{"raw":{"variants":["Punctured codes flex teleportation for latency-constrained nets","Quantum teleportation codes adapt to entanglement delays","Latency rules code choice in encoded teleportation","Adaptive puncturing boosts teleportation under latency"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000634,"raw_usage":{"total_tokens":2778,"prompt_tokens":777,"completion_tokens":2001,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":521,"completion_tokens_details":{"reasoning_tokens":1936}},"tokens_in":521,"tokens_out":2001,"duration_ms":16044,"temperature":1.0,"reasoning_tokens":1936,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T11:45:50.135066+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the exact expected waiting time for a moderate packet size (e.g., n=8) by solving the stated linear system of dimension (w−1 choose n−1), and re-run the latency-constrained comparison; if the crossover latencies where the optimal code changes shift, Eq. (9) is not trustworthy enough to support the decision rule.","supporting_citations":[],"review_version":1}