{"id":"57dad0ef-aafc-4d81-842c-73bb21cf6230","arxiv_id":"2607.06033","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":0,"one_line_summary":"The initialization-free Bernstein-Vazirani algorithm's optimal success probability is derived in closed form, with a necessary and sufficient condition for maximal performance and a proof of advantage over the standard probabilistic version.","lead":"This paper derives an exact formula for the success probability of a variant of the Bernstein-Vazirani quantum algorithm that tolerates arbitrary initial states. It matters because it clarifies which quantum resources drive algorithmic advantage and broadens the class of useful initial states.","discovery_kind":"unclear","skeptic_critique":{"model":"glm-5.2","headline":"Main results (Theorems 2, 6) are mathematically sound. The comparison theorem (Theorem 8) compares a two-query algorithm against a one-query algorithm without isolating the contribution of the extra query, but this is an interpretation gap, not a correctness flaw.","rationale":"The paper's central contributions — the explicit performance formula (Theorem 2) and the complete characterization of maximal-performance states (Theorem 6) — are mathematically correct, with clean proofs using standard SDP duality and Cauchy-Schwarz. The reader's ACCEPT verdict with HIGH confidence is appropriate for these results.\n\nThe reader correctly identified the relabeling convention as the weakest assumption in Theorem 8, and the authors handle it with full transparency (explicit statement, counterexample in Remark 9, and correction of prior work in Appendix A). My additional concern — that the comparison in Theorem 8 does not control for the 2-vs-1 query asymmetry — is about interpretation rather than correctness. The mathematical statement P_W ≥ P (under the convention) is true; the question is whether the advantage reflects the initialization-free property or merely the extra query. This concern does not affect Theorems 2 or 6, which are the load-bearing results cited in the reader's strongest_claim.\n\nThe paper also gives credit to independent support: the constructive procedure in Remark 7 provides a parameter-free method for generating optimal states, and the examples in Examples 4 are explicitly verifiable. The correction of the implicit ordering assumption in Ref. [21] (Proposition 11, Remark 12) is a valuable contribution.\n\nGiven that the main results are sound and the concerns about Theorem 8 are secondary and interpretive, the verdict should remain ACCEPT. The query-complexity concern would be worth addressing in future work (e.g., comparing against a 2-query standard BV baseline), but it does not undermine the paper's core claims.","tokens_in":14835,"tokens_out":9282,"duration_ms":507997,"concrete_test":"For the first example in Examples 4 (N=2, the state achieving P_W = 1 with neither |ϕ₀⟩ nor |ϕ₁⟩ maximally coherent), compute P(|µ⟩) under the relabeling convention and then compute the optimal 2-query performance of the standard BV algorithm for the same state — e.g., by applying U_ā twice with an optimized intermediate CPTP map on the oracle qubit and solving the resulting channel discrimination SDP. If the 2-query standard BV performance also reaches 1 (or ≥ P_W), the IF-BV advantage in Theorem 8 is attributable to the extra query rather than the initialization-free structure, weakening the practical significance of the comparison.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central results — the performance formula (Theorem 2) and the necessary-and-sufficient condition for maximal performance (Theorem 6) — are correct. The proof of Theorem 2 cleanly reduces the IF-BV discrimination problem to a robustness-of-coherence computation on the subspace H = span{|ψ''_x⟩} via standard SDP duality (Eqs. 3–5), and Theorem 6 follows directly from Cauchy-Schwarz on the vector (S_x)_x with Σ S_x² = 1. I verified both directions of the SDP argument and the equality condition; they hold.\n\nThe reader's identified concern — the relabeling convention |C_{1,0}| = max_x |C_{1,x}| in Theorem 8 — is valid but well-handled: the authors state it explicitly, prove its necessity via Remark 9, and extend the same analysis to correct Ref. [21] in Appendix A (Proposition 11, Remark 12). This is transparent scholarship, not a hidden assumption.\n\nThe more substantive concern about Theorem 8 is different: the IF-BV oracle W_ā = (σ_z ⊗ I)U_ā(σ_z ⊗ I)U_ā uses two oracle calls, while the standard BV algorithm uses one. Theorem 8 proves P_W ≥ P, but this inequality conflates two effects: (i) the initialization-free modification (the σ_z insertions) and (ii) the doubling of query count. The paper acknowledges the query difference (Section I: \"the modified protocol requires two oracle queries\") but the comparison theorem and its framing as \"outperforms\" do not control for it. If one applied U_ā twice in the standard BV framework (with some intermediate operation), the extra query alone might close or exceed the gap P_W − P, making the initialization-free property irrelevant to the advantage. This does not affect Theorems 2 or 6, which stand independently of the comparison.","agreement_with_reader":"partial"},"referee_report":{"model":"glm-5.2","summary":"This paper studies a variant of the Bernstein-Vazirani (BV) algorithm called the initialization-free (IF) BV algorithm, in which the oracle register may be initialized in an arbitrary state. The modified oracle W_ā = (σ_z ⊗ I)U_ā(σ_z ⊗ I)U_ā uses two oracle queries but removes the initialization constraint. The authors derive an explicit formula for the performance (optimal average success probability) of the probabilistic IF-BV algorithm for arbitrary pure initial states |μ⟩ = α|+⟩|ϕ₀⟩ + β|−⟩|ϕ₁⟩, obtaining P_W(|μ⟩) = (1/2^N)[1 + |α|²R(|ϕ₀⟩) + |β|²R(|ϕ₁⟩)] + D/2^N, where D is a nonnegative cross-term. They prove a necessary and sufficient condition for maximal performance (Theorem 6: |α|²|C_{0,x}|² + |β|²|C_{1,x}|² = 1/2^N for all x), provide a constructive procedure for optimal states (Remark 7), and prove that under a relabeling convention the IF-BV algorithm outperforms the standard probabilistic BV algorithm (Theorem 8). Appendix A corrects and clarifies a result from Ref. [21] regarding the same relabeling convention.","tokens_in":15588,"tokens_out":1166,"duration_ms":137542,"significance":"The paper provides a clean, self-contained, and parameter-free derivation of the performance formula for the probabilistic IF-BV algorithm. The connection between channel discrimination and robustness of coherence via SDP duality (Theorem 2) is standard but correctly and elegantly applied. The necessary and sufficient condition for maximal performance (Theorem 6) is a clean Cauchy-Schwarz argument that yields a constructive characterization of all optimal states, which is a valuable addition to the resource-theoretic analysis of oracle algorithms. The correction to Ref. [21] in Appendix A (Proposition 11, Remark 12) demonstrates careful scholarship. The results are falsifiable and verifiable by direct computation, as illustrated by the explicit examples.","major_comments":[{"comment":"Section IV, Theorem 8: The comparison P_W(|μ⟩) ≥ P(|μ⟩) compares a two-query algorithm (IF-BV, using W_ā which calls U_ā twice) against a one-query algorithm (standard BV). The paper acknowledges this query difference in Section I but the framing of Theorem 8 as 'outperforms' does not control for the extra query. This is an interpretation gap rather than a mathematical error—the inequality is correctly proven under the stated assumptions—but it would strengthen the paper to explicitly acknowledge this caveat in the theorem statement or its immediate discussion, and to note that the comparison is between protocols with different query costs. As stated, a reader could infer that the IF-BV modification itself (the σ_z insertions) is the sole source of the advantage, when the doubled query count may also contribute.","section":null}],"minor_comments":[{"comment":"Section I: The phrase 'the modified protocol requires two oracle queries' is stated but the implications for the comparison in Theorem 8 are not discussed there. A forward reference to Section IV noting the query-count caveat would help the reader.","section":null},{"comment":"Theorem 2 proof: The notation C''_x and |ψ''_x⟩ uses double primes without a single-prime precursor. Consider simplifying to C'_x and |ψ'_x⟩, or introducing the notation more gradually.","section":null},{"comment":"Examples 4, first example: It would help to explicitly verify that the balancing condition |α|²|C_{0,x}|² + |β|²|C_{1,x}|² = 1/2^N holds for all x, rather than only stating P_W = 1, to connect the example directly to Theorem 6.","section":null},{"comment":"Remark 7: The constraint p_x ≤ 1/(|α|² 2^N) is stated but the derivation of this bound is not shown. A brief justification would be helpful (it appears to follow from requiring q_x ≥ 0).","section":null},{"comment":"Corollary 5: The bound P_W ≤ 2/2^N for x₀ ≠ x₁ is derived, but for N = 1 this gives P_W ≤ 1, which does not rule out maximal performance. The corollary correctly excludes N = 1, but the reader may benefit from an explicit note that the N = 1 case with both states incoherent is covered by the last example in Examples 4.","section":null},{"comment":"Reference [3]: Listed as 'arXiv preprint quant-ph/0504163' but was published in New Journal of Physics. Consider updating to the full journal reference.","section":null}],"recommendation":"minor_revision","confidential_remarks":"The paper is mathematically sound and the central results (Theorems 2 and 6) are correct and well-proven. The only substantive concern is the interpretation of Theorem 8's comparison between a two-query and one-query protocol, which the authors should address more explicitly but does not require new mathematics. The relabeling convention issue is handled transparently and well. I recommend minor revision."},"author_rebuttal":{"model":"glm-5.2","summary":"We thank the referee for the careful reading and the constructive feedback. The referee correctly identifies that Theorem 8 compares a two-query algorithm (IF-BV) against a one-query algorithm (standard BV), and that the theorem statement and its immediate discussion should explicitly acknowledge this query-cost asymmetry to avoid potential misinterpretation. We agree with this point and will revise the manuscript accordingly.","responses":[{"response":"We agree with the referee that the query-cost asymmetry between the two algorithms should be explicitly acknowledged in the statement of Theorem 8 and in its immediate discussion, not only in Section I. The referee is correct that the mathematical content of Theorem 8 is unaffected—the inequality P_W(|μ⟩) ≥ P(|μ⟩) is proven under the stated relabeling convention—but the framing could be misread as attributing the advantage solely to the σ_z insertions without accounting for the doubled oracle query count. We will revise the manuscript to add an explicit caveat in the discussion surrounding Theorem 8, noting that the comparison is between protocols with different query costs (two queries for IF-BV vs. one query for standard BV) and that the advantage shown by the inequality reflects the combined effect of the modified oracle structure and the additional query. We believe this clarification strengthens the paper without altering any results.","revision_made":"yes","referee_comment":"Section IV, Theorem 8: The comparison P_W(|μ⟩) ≥ P(|μ⟩) compares a two-query algorithm (IF-BV, using W_ā which calls U_ā twice) against a one-query algorithm (standard BV). The paper acknowledges this query difference in Section I but the framing of Theorem 8 as 'outperforms' does not control for the extra query. This is an interpretation gap rather than a mathematical error—the inequality is correctly proven under the stated assumptions—but it would strengthen the paper to explicitly acknowledge this caveat in the theorem statement or its immediate discussion, and to note that the comparison is between protocols with different query costs. As stated, a reader could infer that the IF-BV modification itself (the σ_z insertions) is the sole source of the advantage, when the doubled query count may also contribute."}],"tokens_in":14642,"tokens_out":459,"duration_ms":46162,"standing_objections":[]},"desk_editor":{"model":"glm-5.2","letter":"The main results are solid and correct. The paper gives an explicit closed-form performance formula for the probabilistic initialization-free Bernstein-Vazirani algorithm (Theorem 2), a clean necessary-and-sufficient condition for maximal performance (Theorem 6), and a constructive procedure for optimal states (Remark 7). The SDP duality argument reducing the discrimination problem to a robustness-of-coherence computation is standard and correctly applied. The Cauchy-Schwarz equality condition for Theorem 6 is straightforward. Appendix A corrects an implicit ordering assumption in Ref. [21] (co-authored by a present author), which is good scholarship done transparently rather than buried. The examples in Section III are genuinely illuminating—they show that maximal performance is governed by a balancing condition, not by maximal coherence, which is the most interesting conceptual point in the paper. The relabeling convention in Theorem 8 (|C_{1,0}| = max_x |C_{1,x}|) is explicitly stated, and Remark 9 provides a counterexample showing its necessity. This is transparent. The more substantive concern is about Theorem 8's framing. The IF-BV oracle W_ā uses two oracle calls (U_ā twice, interleaved with σ_z), while the standard BV algorithm uses one. Theorem 8 proves P_W ≥ P, but this inequality conflates two effects: the initialization-free modification and the doubling of query count. The paper acknowledges the query difference in the introduction but the comparison theorem and its framing as 'outperforms' do not control for it. If one used U_ā twice in the standard framework, the extra query alone might close or exceed the gap. This does not affect Theorems 2 or 6, which stand independently. It is an interpretation issue, not a correctness flaw, but it does mean the 'outperforms' claim is weaker than it sounds. This is a well-executed paper for researchers working on quantum resource theory and oracle-based algorithms. The core results are correct, self-contained, and parameter-free. It deserves a serious referee who should press the authors to address the query-count confound in the framing of Theorem 8, even if only in the discussion.","headline":"Clean performance formula and tight characterization for probabilistic IF-BV; the comparison theorem has an interpretation gap","tokens_in":15707,"tokens_out":517,"would_cite":true,"duration_ms":54532,"reading_group":"no","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"glm-5.2","headline":"Perfect quantum guessing needs balance, not coherence","keywords":["Bernstein-Vazirani algorithm","quantum query complexity","robustness of coherence","channel discrimination","initialization-free quantum algorithms","quantum resource theory","probabilistic quantum algorithms","oracle-based computation"],"falsifier":"A pure initial state that satisfies the uniform balancing condition |alpha|^2 |C_{0,x}|^2 + |beta|^2 |C_{1,x}|^2 = 1/2^N for all x but fails to achieve P_W = 1 would falsify the central theorem. Conversely, a state achieving P_W = 1 without satisfying this condition would also falsify it.","tokens_in":15116,"feed_emoji":"","tokens_out":1018,"duration_ms":100170,"temperature":0.7,"pith_summary":"The paper studies a variant of the Bernstein-Vazirani quantum algorithm in which the oracle register can start in any state rather than a fixed one. This variant uses a modified oracle built from two sequential queries with local phase operations sandwiched between them. The authors derive a closed-form formula for the optimal success probability of this algorithm when the initial state is an arbitrary pure state, and they prove that perfect performance is achieved if and only if a specific balancing condition holds: the weighted combination of coefficient magnitudes from the two components of the initial state must be uniform across all computational basis elements. This condition is a distribution-level constraint rather than a coherence requirement, meaning states with little or no coherence can still achieve perfect performance if the balance is right. The authors further prove that, under a coefficient-ordering convention, this variant always performs at least as well as the standard single-query Bernstein-Vazirani algorithm for any pure initial state.","feed_headline":"Perfect quantum guessing needs balance, not coherence","feed_subtitle":"An initialization-free variant of the Bernstein-Vazirani algorithm hits maximal success whenever coefficient weights are uniform—no entangl","key_machinery":"The proof machinery centers on rewriting the modified oracle's action on the initial state so that it becomes equivalent to a phase-flip operation on an effective incoherent basis. This reduces the channel-discrimination problem to a robustness-of-coherence calculation, which is formulated as a semidefinite program. The Cauchy-Schwarz inequality then provides both the upper bound on performance and the equality condition characterizing optimal states. The comparison with the standard algorithm uses concavity of the relevant functional over the probability simplex, with the minimum attained at an extreme point determined by a monotonicity argument.","core_discovery":"The central result is that maximal performance in the probabilistic initialization-free Bernstein-Vazirani algorithm is governed by a uniform balancing condition on the weighted coefficient distributions of the initial state's two components, not by coherence or entanglement. Specifically, for an initial state decomposed as a superposition of two branches, the weighted sum of squared coefficient magnitudes must equal 1/2^N at every basis index. This is both necessary and sufficient, and it follows from a Cauchy-Schwarz equality condition. The proof works by showing that the algorithm's performance reduces to a robustness-of-coherence calculation on an effective subspace, after which the Caqu","pith_inferences":["The fact that the comparison theorem requires a relabeling convention suggests that the standard algorithm's performance formula implicitly privileges a specific basis element, and a basis-independent comparison would require a symmetrized version of the standard algorithm's performance measure.","The balancing condition resembles a flat-spectrum requirement in the effective basis, which could connect to results on minimum-error discrimination of unitary channels where flatness of the input state's spectrum in the relevant eigenbasis determines optimality.","If mixed initial states were considered, the balancing condition would likely generalize to a condition on the eigenvalue-weighted sum of the eigenvectors' coefficient distributions, though the clean Cauchy-Schwarz characterization may not survive."],"forward_implications":["The balancing condition provides a constructive recipe for generating optimal initial states: pick any valid pair of probability distributions whose weighted sum is uniform, then assign phases freely, yielding a large family of perfect-performance states.","The result separates the resource of coherence from algorithmic performance in this setting, adding to the body of evidence that quantum advantage can arise without maximal entanglement or coherence.","The comparison theorem suggests that spending an extra oracle query to remove initialization constraints can be worthwhile even in probabilistic settings, broadening the design space for near-term quantum algorithms.","The characterization of equality cases between the two algorithm variants identifies precisely when the extra query provides no benefit, which could guide resource-cost tradeoffs in oracle-based protocols."],"fun_headline_variants":["Uniform coefficient balance drives maximal quantum guessing success","Balanced weights, not coherence, maximize initialization-free BV algorithm","Init-free Bernstein-Vazirani peaks when coefficient weights stay uniform","Maximal quantum success hinges on coefficient balance, not entanglement","Uniform weight condition guarantees peak performance in IF-BV algorithm"],"cache_read_input_tokens":0,"weakest_assumption_plain":"The proof that the initialization-free variant outperforms the standard algorithm requires relabeling the computational basis so that the largest coefficient of one component of the initial state aligns with a specific basis element used in the standard algorithm's performance formula. Without this alignment, the advantage can reverse, as the authors demonstrate with an explicit counterexample.","fun_headline_variants_meta":{"raw":{"variants":["Uniform coefficient balance drives maximal quantum guessing success","Balanced weights, not coherence, maximize initialization-free BV algorithm","Init-free Bernstein-Vazirani peaks when coefficient weights stay uniform","Maximal quantum success hinges on coefficient balance, not entanglement","Uniform weight condition guarantees peak performance in IF-BV algorithm","Coefficient balance, not coherence, sets the ceiling for quantum guessing","IF-BV algorithm reaches maximal success under uniform coefficient weights","Balanced initial states unlock peak performance in initialization-free BV","Uniform coefficients are necessary and sufficient for maximal IF-BV success","Weight balance alone governs success in initialization-free Bernstein-Vazirani"]},"model":"glm-5.2","effort":"high","cost_usd":0.0,"raw_usage":{"total_tokens":863,"prompt_tokens":473,"completion_tokens":390,"prompt_tokens_details":null},"tokens_in":473,"tokens_out":390,"duration_ms":12618,"temperature":1.0,"reasoning_tokens":301,"cache_read_input_tokens":0,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-08T18:33:44.835578+00:00","model_set":{"reader":"glm-5.2"},"falsifier":"A pure initial state that satisfies the uniform balancing condition |alpha|^2 |C_{0,x}|^2 + |beta|^2 |C_{1,x}|^2 = 1/2^N for all x but fails to achieve P_W = 1 would falsify the central theorem. Conversely, a state achieving P_W = 1 without satisfying this condition would also falsify it.","supporting_citations":[],"review_version":1}