{"id":"a8ff843f-d72c-4686-a0b0-46b569170e82","arxiv_id":"2608.06226","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"The energy of the one-dimensional fractional Allen-Cahn layer is continuous and strictly decreasing in the fractional exponent s, with an explicit pole at s=1/2 and an explicit linear expansion at s=1.","lead":"This paper computes how the energy of a one-dimensional fractional transition layer depends on the range of the interaction, showing it falls continuously from infinity at the most nonlocal endpoint to the classical value at the local endpoint. It is a rigorous result with explicit asymptotic formulas at both ends and a computer-assisted proof of strict monotonicity in between.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Interior strict decrease rests on the interval-arithmetic certificate (C1)-(C4) of Proposition 5.5; until the attached code is independently executed, Theorem 1.2 is conditional on that certificate being correct.","rationale":"I read the paper looking for a load-bearing gap or internal inconsistency in the central claim. The endpoint analyses are explicit and self-contained: Theorem 3.1 at s=1 gives E(s)=2*sqrt(2)/3+kappa1(1-s)+kappa2(1-s)^2+O((1-s)^3) with explicit constants and E'(s)<-0.03 on [0.980,1], and Lemma 4.5 plus Proposition 4.1 control the pole and give E'(s)<0 on (1/2,0.530]. The Pohozaev identity, the Fourier-side residue computation, and the transfer from approximate to true layer are all coherent, and no fitted parameters are hidden in the asymptotics. The interior proof is also logically sound once conditions (C1)-(C4) hold: semiconcavity propagates a negative derivative from the grid, and the fixed-point argument identifies the constructed profile with the Cabre-Sire layer via uniqueness. The only place where the argument depends on an unverified resource is the computer-assisted certificate. The text states that all bounds are computed in interval arithmetic at 200-bit precision, but the certificate is not reproduced as tables, and I could not execute the code. Therefore the strict decrease on the interior interval, and hence Theorem 1.2, is conditional on the correctness of that code. This is exactly the reader's weakest assumption, so I agree with the reader's conditional verdict and recommend no change.","tokens_in":50329,"tokens_out":12246,"duration_ms":122311,"concrete_test":"Obtain the supplementary code from GitHub [11], run the six modules in order at 200-bit precision, and parse the certificate file to verify for all 56 grid points: (C1) theta_core<1 and theta_tail<1, (C2) alpha(s_i)<1 and tau(s_i)<=1, (C3) J(s_i,Phi_ap)<0 and eta_i>0, and (C4) 2 M*_sc([a_i,b_i]) (s_{i+1}-s_i)/eta_i<1. For a stronger check, independently re-implement the Schur certificate for s=0.530 and s=0.980 with a different interval library and confirm both bounds; if the certificate is not reproduced, strict decrease on I0 is not established.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The decisive step for the central claim is Proposition 5.5, which reduces strict decrease on I0=[0.530,0.980] to conditions (C1)-(C4) verified only by the attached interval-arithmetic code. The most fragile condition is (C1), the Schur inequality (85) with theta_i<1 at each grid point, because it yields the coercivity bound K0(s_i)<=9 used to obtain the contraction radius tau(s_i) and the margin eta_i in (C3). A failure at even one grid point would break the monotonicity chain. The certificate in Appendix A.2 is a long finite computation involving Taylor jets, rigorous kernel enclosures, cutoffs (delta, Lambda, K0, X), and tail asymptotic bounds, with many implementation choices; no numerical table of the certified values theta_i, tau(s_i), eta_i appears in the text. The analytic endpoint arguments and the fixed-point/semiconcavity structure are otherwise explicit and coherent, with explicit constants, so this is not an internal inconsistency but an external verification gap: the theorem is as reliable as the correctness of that code.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the energy E(s)=E_s[Phi_s] of the one-dimensional fractional Allen-Cahn layer as a function of s in (1/2,1]. It claims that E is continuous and strictly decreasing, with an explicit first-order expansion at s=1 (coefficient kappa_1 given as a convergent integral), and a simple pole 1/(pi(s-1/2)) with bounded remainder at s=1/2. The proof combines analytic endpoint expansions (a corrector argument near s=1 and a Fourier-side pole computation near s=1/2), a semiconcavity argument on the interior interval, and a computer-assisted certificate that reduces strict decrease on [0.530,0.980] to four finite interval-arithmetic conditions (C1)-(C4) at a 56-point grid. The analytic parts are presented in detail with explicit constants; the interior part is the main delicate point.","tokens_in":50579,"tokens_out":17626,"duration_ms":178402,"significance":"If the result is correct, it provides the first sharp quantitative description of the energy of one-dimensional fractional Allen-Cahn layers across the full range s in (1/2,1], including the exact pole residue and the linear correction at s=1. The endpoint arguments are valuable and are carried out with explicit, checkable constants; the pole residue is derived from a Pohozaev identity and the identity that the integral of the layer derivative equals 2, so no fitting parameter enters. The computer-assisted part is also a worthwhile methodological contribution, since it combines rigorous interval arithmetic with a fixed-point validation for a nonlocal semilinear equation on the whole line. The main caveat is that the central theorem depends on a finite interval-arithmetic certificate whose numerical output is not reproduced in the text; until that certificate data and code are independently auditable, the interior monotonicity claim remains conditional on the correctness of the attached computation.","major_comments":[{"comment":"The strict decrease on I0=[0.530,0.980] is exactly the content of conditions (C1)-(C4), but the paper does not present the numerical certificate: no grid values s_i, no values of theta_i from (85), no tau(s_i), eta_i, alpha(s_i), or the cell widths are listed. A reader cannot verify or even audit these load-bearing inequalities without executing the attached code and reproducing the 2h8m run. Please include a table (or a machine-readable data file) of all certified values and margins for every grid point, together with the exact code version or checksum, and state explicitly that Theorem 1.2 rests on those computations. A single failed enclosure at one grid point would invalidate the monotonicity chain.","section":"Section 5.3, Proposition 5.5; Appendix A"},{"comment":"The displayed cell-mass bound 'integral over I_k of K_s(x,y) dy <= integral over I_k of R_s(x-y) dy - inf_{x in I_j} integral over I_k of R_s(x+y) dy' is not, as written, an upper bound uniform in x in I_j: the first term on the right must be the supremum over x in I_j of the integral of R_s(x-y), not an evaluation at an arbitrary or fixed x. If the implementation computes this term at a fixed representative (for example a cell midpoint) and not a rigorous enclosure of the supremum, the Schur certificate could fail. Please clarify the quantifier and confirm that the code encloses the maximum over x in I_j.","section":"Appendix A.2, compact-part estimate"},{"comment":"The fixed-point radius tau(s) in Proposition 5.4 depends on omega(s)=6 C*(2s)(||Phi_ap||_Linf + C*(2s)), so a rigorous upper bound on ||Phi_ap||_Linf is needed. The appendix states that the approximate layer is bounded and that the profile is used in the contraction certificate, but it does not record the certified L-infinity enclosure used for ||Phi_ap||_Linf. Since condition (C1) only implies the weak bound |Phi_ap| <= sqrt(5/3), the tighter value used in omega(s) is part of the missing certificate data and should be reported.","section":"Proposition 5.4 and Appendix A.1"},{"comment":"The evaluation of J(s,Phi_ap) and the weighted seminorm is a modal exact sum and is a strength of the paper. However, the final sign margin in (C3), eta_i = |J(s_i,Phi_ap)| - L_J(s_i) tau(s_i), is only declared positive; no numerical values of eta_i or of the transfer constant L_J(s_i) are printed. Because C4 uses eta_i in the denominator, the size of the margin controls the admissible cell widths. Please include these values in the certificate table required by the first major comment.","section":"Appendix A.3"}],"minor_comments":[{"comment":"The title page and figure captions contain obvious OCR artifacts ('LA YERS', 'JA VIER', '©ap', '∙1'); these should be cleaned before publication.","section":"Title page and figures"},{"comment":"The 56-point grid is only specified in the accompanying code. It would improve readability to print the grid values and the chosen cells [a_i,b_i] in an appendix table.","section":"Equation (95) and grid"},{"comment":"The approximate value kappa_1 approx 1.073 is stated, but the proof only establishes the lower bound kappa_1 > 119/900. A rigorous enclosure of kappa_1 would make the expansion fully quantitative and is easy to compute from the explicit integral.","section":"Theorem 1.1(i)"},{"comment":"In Lemma 3.9, the constants A1, A2, A3 are estimated by splitting at |xi|=1; the presentation would be clearer if the displayed values 67, 230, 125 were labeled as upper bounds rather than exact evaluations, since they are rounded. No mathematical issue is involved.","section":"Section 3.3"}],"recommendation":"major_revision","confidential_remarks":"The analytic endpoint portions are strong and, in my reading, correct. The only reason for major revision is that the interior strict-decrease claim, which is the centerpiece of Theorem 1.2, is verified only by an interval-arithmetic certificate whose numerical output is absent from the manuscript. I could not run the 2h8m computation during refereeing. I would urge the editor to require the authors to supply the certificate data (grid, theta_i, tau(s_i), eta_i, alpha(s_i), cell widths) in the supplement and, ideally, to have the code independently executed by a third party. If the certificate is confirmed, the paper should be acceptable for publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague — The genuinely new content here is the first sharp quantitative account of the layer energy: strict decrease, continuity, the endpoint expansion E(s)=2√2/3+κ1(1−s)+o(1−s) with explicit κ1, and E(s)=1/(π(s−1/2))+O(1). As far as I can tell from the cited literature, these results are not in prior work. The endpoint proofs are careful, with explicit constants: corrector expansion, spectral gap for the odd linearized operator, fixed-point contraction, Pohozaev identity, and the low-frequency residue. I read through the algebra and did not find a circular step or hidden fitting; the approximate layer is matched to the PDE residual, not to the energy sign.\n\nThe strongest part is the analytic treatment at s=1 and s→1/2+. The explicit κ1 integral and the pole residue coming from the identity ∫Φ'_s=2 are clean. The interior strict decrease is the weak spot, in the ordinary sense: it depends on Proposition 5.5, and conditions (C1)–(C4) are verified only by attached interval-arithmetic code at 56 grid points. The paper describes the certificate in detail, including kernel enclosures, Taylor jets, weight design, and tail bounds, but does not print the certified values θ_i, τ(s_i), η_i. So the central claim \"strictly decreasing on [0.530,0.980]\" is as reliable as the code. That is not a fatal defect; computer-assisted proofs are legitimate, and the authors have shipped code. But the claim is conditional until someone independently executes it. This is an external verification gap, not an internal inconsistency.\n\nI also note conditions (C2)–(C4) are quantitative and would be quick to check once (C1) holds; the Schur inequality is the load-bearing piece.\n\nOverall: good paper, honest about the computer assistance, and the analytic results are valuable on their own. The right endpoint expansion and pole expansion deserve citation even if the interior monotonicity were weakened. I would send it to a serious referee, with a request that the referee or editor arrange an independent run of the certificate.","headline":"A genuinely new quantitative description of E(s), with careful endpoint analysis and a conditional computer-assisted core; referee it, but make sure the code gets independently run.","tokens_in":51083,"tokens_out":1718,"would_cite":true,"duration_ms":19506,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":null,"created_at":"2026-08-07T12:09:47.191415+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":null,"supporting_citations":[],"review_version":1}