{"id":"05f5042a-99db-404e-98a3-af2bb5da9570","arxiv_id":"2611.17073","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Discrete sets of the form ±n^α and ±n^β form Schrödinger uniqueness pairs for linear Schrödinger equations with Gevrey potentials and for the cubic NLS, under conditions on α,β and the Gevrey order.","lead":"This paper proves that if a solution to a Schrödinger equation vanishes on two specific discrete sets of points at two different times, then it must be identically zero. The result applies to linear Schrödinger equations with Gevrey-class potentials and to the cubic nonlinear Schrödinger equation, extending earlier Fourier uniqueness results to the PDE setting.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek/deepseek-v4-flash","headline":"The induction in Proposition 5.2 for the NLS case uses the H1 algebra property but does not account for the fact that the nonlinear term |u|^2 u involves complex conjugation, which breaks the direct Leibniz rule for Γ_t when the operator is not real-linear.","rationale":"The reader's verdict is ACCEPT with high confidence, but the stress-test reveals a potentially fatal technical flaw in the proof of the NLS case (Theorem 1.2). The central inductive estimate (Proposition 5.2) depends on an expansion of Γ_t^j(|u|^2 u) that treats the nonlinearity as if it were a polynomial in u alone, ignoring the complex conjugation. Because Γ_t is not real-linear (its coefficient depends on t and is not real), it does not commute with conjugation, so the Leibniz rule for products does not apply directly to |u|^2 u. The paper provides no justification for the claimed identity. If this step is invalid, the entire proof of Theorem 1.2 collapses. The concern is concrete, technical, and not addressed by the authors. The reader did not flag this issue; the reader's weakest_assumption focused on the Gevrey condition for the potential, which is a different matter. Hence the verdict should be REJECT unless the authors can supply a corrected argument. The test is straightforward: compute the expansion for low j explicitly to see if the claimed structure holds.","tokens_in":32252,"tokens_out":786,"duration_ms":8897,"concrete_test":"Re-derive the action of Γ_t^j on |u|^2 u explicitly for j=1 and j=2 using the definition Γ_t = x + 2it ∂_x and the fact that |u|^2 = u \\bar{u}. Check whether the claimed product formula holds. If it fails, determine whether a corrected expansion still yields the same leading-order structure (2|u|^2 Γ_t^j u + ...) or whether additional terms involving conjugates of Γ_t^j u appear that break the induction.","verdict_should_be":"REJECT","load_bearing_attack":"In the proof of Proposition 5.2 (Section 5.2), the authors expand Γ_t^j(|u|^2 u) using the Leibniz rule for Γ_t acting on products. However, Γ_t = x + 2it ∂_x is a complex-linear differential operator, but the nonlinearity |u|^2 u = u \\bar{u} u involves complex conjugation, which is not complex-linear. The derivation in the text (around equation (5.20)) treats Γ_t^j(|u|^2 u) as if it were a sum of products of Γ_t^{j_i} u, implicitly assuming that Γ_t commutes with complex conjugation. In fact, Γ_t does not commute with complex conjugation because the coefficient 2it is not real when t ≠ 0. The claimed identity Γ_t^j(|u|^2 u) = Σ (j!/(j_1! j_2! j_3!)) (-1)^{j_3} (Γ_t^{j_1} u)(Γ_t^{j_2} u)(Γ_t^{j_3} u) is therefore not justified. This error is load-bearing because the entire induction for the NLS case (Theorem 1.2) relies on this expansion to separate the leading term and to apply the inductive hypothesis. If the expansion is invalid, the energy estimates for Γ_t^j u in H^1 are unsupported, and the Gaussian decay needed for the Morgan uncertainty principle cannot be established. The authors do not address this issue or provide an alternative argument.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek/deepseek-v4-flash","summary":"The paper studies discrete unique continuation for one-dimensional Schrödinger equations. It proves that if a solution vanishes at points ±n^α at time 0 and at points ±n^β at time 1, then the solution must be identically zero, under certain conditions. Two main cases are considered: (1) linear Schrödinger equation with a potential uniformly of Gevrey class σ, with condition max{1,σ}(α/(1-α)+β/(1-β))<1; (2) cubic nonlinear Schrödinger equation, with condition α/(1-α)+β/(1-β)<1. The proof strategy involves first showing that the zeros force the solution to be Schwartz class at all times (using Fourier uniqueness results of Kulikov-Nazarov-Sodin), then converting zeros into Gaussian decay via energy estimates for iterations of the vector field Γ = x+2it∂_x, and finally applying the Morgan-type uncertainty principle of Escauriaza-Kenig-Ponce-Vega.","tokens_in":32720,"tokens_out":2271,"duration_ms":20034,"significance":"The results represent a significant advance in the understanding of unique continuation for Schrödinger equations from discrete sets of zeros. Previous work by Kehle-Ramos (2022) only handled logarithmically spaced zeros using complex analysis. The present paper obtains polynomial spacing (n^α, n^β) and uses purely real-variable methods, which is a substantial improvement. The techniques developed, combining Fourier uniqueness theory with energy estimates for the operator Γ, are likely to be influential. The paper also provides a clean separation of the linear and nonlinear cases, and the conditions on α,β are essentially sharp in the nonlinear case (matching the Fourier uniqueness threshold). The proofs appear rigorous and well-structured, with careful tracking of constants.","major_comments":[{"comment":"The derivation of the Leibniz rule for Γ_t^j(|u|^2 u) treats Γ_t as commuting with complex conjugation, which is not justified. Γ_t = x + 2it ∂_x has an imaginary coefficient when t≠0, so Γ_t(conj(u)) ≠ conj(Γ_t(u)). The expansion (5.20) writes Γ_t^j(|u|^2 u) as a sum of products of Γ_t^{j_i} u with signs (-1)^{j_3}, implicitly assuming Γ_t acts linearly on the complex conjugate. This is a load-bearing step for the entire NLS induction (Proposition 5.2). The authors need to either justify this expansion by working with real and imaginary parts separately, or provide an alternative argument that does not rely on this identity. Without this fix, the proof of Theorem 1.2 is incomplete.","section":"§5.2, equation (5.20) and surrounding text","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The proof of Claim 4.1 contains a notational inconsistency: the parameter ρ is introduced but the text uses both ρ and κ (apparently interchangeably). In the statement of Claim 4.1, the bound uses ρ, but in the proof the constant κ appears (e.g., 'for κ < κ_2'). The final line says '0<κ<κ_0 = min{κ_1, κ_2}' but κ_1 is not defined (only ρ_1 appears earlier). This makes the argument difficult to follow. The authors should clarify the notation and ensure the constants are consistently named.","section":"§4.2, Proposition 4.1 and Claim 4.1","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"In the estimate (4.6), the term for l=0 in the second sum (the one involving ||Γ_t^k u||) is included, but the sum over l=1 to k-1 excludes l=0. However, in the subsequent inequality (4.7), the second sum is written as sum_{l=1}^{k-1} but the term for l=0 from the first sum is already accounted. This is fine, but the derivation of the bound for the first sum in (4.7) uses the estimate for ||B_x^l \\tilde{u}||_{L^2} that includes a term with ||Γ_t^k u|| for l=0? Actually, for l=0, the estimate would be ||\\tilde{u}||_{L^2} ≤ ||u_0||_{L^2} (since no derivative). The splitting with θ seems to be applied for l≥1 only. The authors should clarify the handling of the l=0 term in the interpolation estimate to avoid confusion.","section":"§4.2, equation (4.6) and (4.7)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The proof of Lemma 5.1 uses the bound (5.16) to derive (5.17), but the step 'using twice (5.16)' is not fully explained. The sum over j_1, j_2, j_3 with j_1+j_2+j_3=j is reduced to a double sum, but the application of (5.16) requires careful handling of the ranges. The authors should provide a more detailed derivation or cite a standard convolution estimate.","section":"§5.2, Lemma 5.1","authors_acknowledge":null,"authors_acknowledge_p":null}],"minor_comments":[{"comment":"The definition of D_σ in (1.5) for 0≤σ≤1 uses the condition α/(1-α)+β/(1-β)<1, while (1.6) for σ>1 uses α/(1-α)+β/(1-β)<1/σ. This is clear, but the notation D_σ is reused for both cases, which could be confusing. Consider using D_σ^1 and D_σ^2 or a unified definition with a case distinction.","section":"§1, equation (1.5) and (1.6)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The proof of Lemma 2.1 uses the change of variable r = (t/τ)^{1/s}, but the integration limits are not explicitly transformed. The statement 'I_{j,s} = s^{-1} τ^{-(j+1)/β} ...' contains a typo: β should be s. Also, the constant c_{j,s,τ} is not explicitly bounded; the final estimate ||B_x^j f||_{L^∞} ≤ c_{j,s,τ} (j!)^{1/s} is claimed but the dependence on j is not fully tracked. The authors should provide a more precise bound or cite a standard reference.","section":"§2.2, Lemma 2.1","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"In the energy estimate (3.3), the term Im ∫ V |B_x^k \\tilde{u}_m|^2 dx is claimed to vanish because V is real-valued. This is correct, but the justification could be made explicit: the imaginary part of a real number times a real quantity is zero. The authors might add a brief remark.","section":"§3, equation (3.3)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The term (2πA^{-1}θ)^{-(k-l)} appears in the second sum, but earlier the bound for ||B_x^l \\tilde{u}||_{L^2} used (2πθ)^{-(k-l)}. The factor A^{-1} seems to come from the Gevrey bound on V. This is correct, but the derivation could be clearer: the constant A from (1.4) is absorbed into the definition of θ. The authors should check that the factor A^{-1} is correctly accounted for in the subsequent estimates.","section":"§4.2, equation (4.6)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The decomposition of Γ_s^j(|u|^2 u) into three terms (2|u|^2 Γ_s^j u + (-1)^j u^2 Γ_s^j \\bar{u} + remainder) is stated without proof. The factor (-1)^j comes from the complex conjugation in the Leibniz rule. However, as noted in the major comments, this expansion is not justified when Γ_s does not commute with conjugation. The authors should either provide a rigorous derivation or modify the argument.","section":"§5.2, equation (5.21)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The choice of l = ⌊k/\\tilde{b}⌋ with \\tilde{b} > b is used to apply the induction hypothesis. The condition 'k ≥ k_1' is stated but k_1 is defined as 2\\tilde{b}/(\\tilde{b}-1)+1. This seems to ensure that l+1 < k, so the induction hypothesis applies. The authors should verify that this condition is sufficient and that the induction step works for all k.","section":"§5.3, equation (5.26)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The paper uses the notation H^1 to mean H^1(R) ∩ L^2(x^2 dx), which is nonstandard. While this is defined in the text, it could cause confusion. Consider using a different notation, e.g., H^1_1 or W^1_2, or explicitly state the definition at first use.","section":"General","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The formula (2.5) for Γ_t^j involves (2it)^{|j|} but j is a scalar, so |j| = j. The notation |j| suggests multi-index, but here j is an integer. This is a minor inconsistency.","section":"§2.3, equation (2.5)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"Reference [27] (Kehle-Ramos) is listed as 'Annals of PDE 8, 2 (2022), 21' but the title is 'Uniqueness of solutions to nonlinear Schrödinger equations from their zeros'. The reference should include the full title for completeness.","section":"References","authors_acknowledge":null,"authors_acknowledge_p":null}],"recommendation":"major_revision"},"author_rebuttal":null,"desk_editor":{"model":"deepseek/deepseek-v4-flash","letter":"The paper extends Fourier uniqueness pairs (power-law spaced zeros) to Schrödinger equations with Gevrey potentials and to the cubic NLS. The main advance over Kehle–Ramos (2022) is replacing logarithmically spaced zeros with power-law spaced zeros and removing complex analysis from the proof. The linear case (Theorem 1.1) is clean: the energy estimates for Γ_t combined with the Gevrey control on V are carefully tracked, and the induction on k is handled with Stirling bounds. The Schwartz-class bootstrap (Theorems 1.3–1.4) is a nice application of the Kulikov–Nazarov–Sodin gap argument and works in the whole lower triangle α+β<1. The paper is well written and the constants are tracked honestly.\n\nNow the soft spot. The stress-test note about the NLS induction (Proposition 5.2) is correct. The expansion of Γ_t^j(|u|^2 u) around (5.20) treats the nonlinearity as if Γ_t commutes with complex conjugation. It does not: Γ_t = x + 2it ∂_x, and when t ≠ 0 the coefficient 2it is not real, so Γ_t(\\bar{u}) ≠ \\overline{Γ_t u}. The claimed Leibniz-type formula for Γ_t^j(|u|^2 u) is therefore not justified. This is load-bearing: the whole induction for the NLS case (Theorem 1.2) relies on separating the leading term and applying the inductive hypothesis. If the expansion is invalid, the energy estimates for Γ_t^j u in H^1 are unsupported, and the Gaussian decay needed for the Morgan principle cannot be established. The authors do not address this issue or provide an alternative argument. This is a genuine gap, not a minor oversight.\n\nWho is this for? Specialists in PDEs and Fourier uniqueness. The linear result (Theorem 1.1) is likely correct and valuable. The NLS result (Theorem 1.2) is suspect until the conjugation issue is resolved. The paper deserves a serious referee, but the referee should be asked to check the NLS induction carefully. I would not cite the NLS result in its current form.\n\nRecommendation: send to peer review, but flag the conjugation problem prominently. The linear part is solid; the NLS part needs a fix or a different argument.","headline":"A technically solid extension of Fourier uniqueness to Schrödinger equations, but the NLS induction has a load-bearing gap with complex conjugation.","tokens_in":33213,"tokens_out":584,"duration_ms":7754},"pith_extraction":null,"created_at":"2026-11-13T10:18:37.163543+00:00","model_set":{"reader":"deepseek/deepseek-v4-flash"},"pipeline_version":"pith-pipeline@v1.4.1-daily-deepseek-or-medium","input_kind":null,"input_chars":null,"input_total_chars":null,"passes_missing":null,"published_standing":null,"published_confidence":null,"objection_audit":null,"run_record":null,"basis":{"kind":"not_recorded","source":null,"chars_given":null,"chars_total":null},"receipt":null,"falsifier":null,"supporting_citations":[],"lane_accuracy":{"status":"not yet measured","lane":{"model":"deepseek/deepseek-v4-flash","input_kind":null},"url":"https://pith.science/accuracy"},"review_version":1}