{"id":"d1a74fd2-c2d6-48e2-9782-6e4d2675edd9","arxiv_id":"2412.03742","paper_version":1,"verdict":"CONDITIONAL","confidence":"LOW","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"A gluing theorem approximates the collapsing Calabi-Yau metrics on affine quadrics by scaled warped QAC Calabi-Yau pieces as singular fibers collide, with a bubble tree limit structure.","lead":"The paper constructs a detailed gluing description of collapsing Calabi-Yau manifolds modeled on affine quadrics, showing how the metric near colliding singular fibers is approximated by scaled warped QAC Calabi-Yau pieces. It also verifies a conjecture by Yang Li about these collapsing metrics and describes a bubble tree structure for the limits.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The §3 coordinate change is inconsistent with the claimed model Xi: with the printed scaling z_{n+2}-s_i=a_t v_{n+2}, the balancing exponent is not x=-θ_i; the model requires z_{n+2}-s_i=a_t^{-θ_i}v_{n+2}.","rationale":"The reader identified the imported mapping properties of the Laplacian on the model geometries as the weakest assumption. That is a legitimate concern, but I found a more immediate and concrete problem internal to the manuscript: the coordinate change used to exhibit the local models Xi around the colliding critical points is inconsistent with the stated family and with the claimed exponent x=-θ_i. If taken literally, the gluing maps Fi,t do not produce the warped-QAC models Xi near the points S_i, so the central convergence statement in Theorem 1.7 is not established by the written proof. The issue is very likely a typographical omission of the exponent -θ_i on the base coordinate: with z_{n+2}-s_i=a_t^{-θ_i}v_{n+2}, the computation of x=-θ_i is consistent, and the t^{2θ_i} scaling in Theorem 1.7 makes sense. Because the fix is transparent and the surrounding strategy is coherent, I would not reject the paper outright; the conditional verdict remains appropriate, with the coordinate correction as an explicit condition. I therefore recommend UNCHANGED relative to the reader's CONDITIONAL verdict, while noting that the reason for conditionality should include this internal inconsistency rather than only the imported analytic facts.","tokens_in":34067,"tokens_out":28826,"duration_ms":269038,"concrete_test":"Independently re-derive the §3 scaling: substitute (z~, z_{n+2}-s_i)=(a_t^x v~, a_t^{-θ_i}v_{n+2}) into X_t and impose the limiting equation P(v~)+Q_i(v_{n+2})=0; verify that this forces x=-θ_i. Then repeat the same computation with the printed scaling z_{n+2}-s_i=a_t v_{n+2} and check whether the exponent equation has any solution matching x=-θ_i. If the printed scaling does not produce the model Xi, the coordinate change in §3 must be corrected before the gluing statement in Theorem 1.7 can be accepted as written.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The gluing setup rests on the claim in §3 that the coordinate change (z~, z_{n+2}-s_i)=(a_t^x·v~, a_t v_{n+2}) converts X_t into the model P(v~)+Q_i(v_{n+2})=0 with the unique choice x=-θ_i=-(p-q)/(p-q_i). Substituting the printed scaling into X_t: P(z~)+a^{q-p}Q_t(z_{n+2})=0, the i-th cluster factors are a_t v_{n+2}-a_t^{-θ_i}s_{i,j}≈a_t v_{n+2} for fixed v_{n+2}≠0, because θ_i>0. Hence Q_t grows like a_t^{q_i} times a constant times v_{n+2}^{q_i}, while P(a_t^x v~)=a_t^{px}P(v~). Balancing exponents gives px=q-p+q_i, so x=(q_i+q-p)/p, not -θ_i. With the printed scaling the zeros of the i-th cluster all collapse to v_{n+2}=0, so the limiting equation is not Xi with distinct roots s_{i,j}. The claimed model Xi and the t^{2θ_i} scaling in Theorem 1.7 are instead consistent with the base scaling z_{n+2}-s_i=a_t^{-θ_i}v_{n+2}; under that scaling one indeed obtains x=-θ_i. Thus, as printed, the derivation of the local warped-QAC models around the colliding critical points is internally inconsistent. This is likely a missing exponent in the coordinate change, but it is load-bearing because the biholomorphisms Fi,t and the scaled convergence in Theorem 1.7 depend on exactly this identification.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a gluing construction for collapsing warped quasi-asymptotically-conical (QAC) Calabi-Yau manifolds. It first proves existence of warped QAC Calabi-Yau metrics on affine hypersurfaces P(z~)+Q(z_{n+2})=0 under a degree hypothesis p>q, improving the Ricci-potential decay estimate relative to earlier work. It then considers a one-parameter family X'_t in which q singular fibers of the fibration z_{n+2} collide, at a rate controlled by exponents θ_i, and claims that for large t the unique Calabi-Yau metric is obtained by gluing scaled copies of the warped QAC models X_i near the collision points. A bubble-tree convergence result is also stated in Section 5.","tokens_in":34468,"tokens_out":13590,"duration_ms":135744,"significance":"If correct, the paper would verify a conjecture of Yang Li in a noncompact setting and produce new examples of warped QAC Calabi-Yau metrics, including cases not covered by Conlon-Rochon's numerical constraints. The explicit numerical conditions in Proposition 3.2 and the gluing statement in Theorem 4.8 are concrete and falsifiable, and the bubble-tree discussion suggests a useful framework. The paper also makes a real technical contribution in Proposition 2.3, where the Ricci potential estimate is derived in a way that is optimized for the gluing argument. However, the central identification of the local model Xi contains a coordinate-scaling inconsistency that currently undermines the proof of Theorem 1.7, and several other load-bearing hypotheses and cut-off identities need correction.","major_comments":[{"comment":"The local model Xi is not obtained with the scaling printed in the manuscript. Substituting (z~, z_{n+2}-s_i)=(a_t^x v~, a_t v_{n+2}) into X_t gives cluster factors a_t v_{n+2}-a_t^{-θ_i}s_{i,j}=a_t(v_{n+2}-a_t^{-θ_i-1}s_{i,j}), so for fixed v_{n+2} the q_i roots collapse to 0. Balancing P(a_t^x v~) ~ a_t^{px}P(v~) with a_t^{q-p}Q_t(s_i+a_t v) ~ a_t^{q-p+q_i} v_{n+2}^{q_i} yields px=q-p+q_i, i.e. x=(q_i+q-p)/p, not x=-θ_i. The model Xi with the distinct shifts s_{i,j} and the t^{2θ_i} scaling in Theorem 1.7 follow only if z_{n+2}-s_i=a_t^{-θ_i}v_{n+2}; with that replacement the balancing computation indeed gives x=-θ_i. As printed, the biholomorphisms F_{i,t} and the scaled convergence in Theorem 1.7 are therefore not justified.","section":"Section 3, coordinate change before 'Computation shows x=-θ_i'"},{"comment":"The theorem statement assumes only p>2q_i, but the construction requires the stronger hypothesis p>q. The exponents θ_i=(p-q)/(p-q_i) are positive only when p>q, the gluing intervals in Section 3 are nonempty only when (q-p)/p<σ_i<0, and if q>p then the singular points s_i+a_t^{-θ_i}s_{i,j} do not accumulate at s_i. Since p>2q_i for all i does not imply p>q when d>1, this hypothesis must be added to Theorem 1.7 (and to Proposition 3.2 if stated independently).","section":"Theorem 1.7 and Section 3, hypotheses after (3.3)"},{"comment":"For d>2 the cutoffs β3=β'_3(1-β_{2,1}-β_{2,2}) and β4=β'_4(1-β_{2,1}-β_{2,2}) do not subtract the neighborhoods of S_3,...,S_d. Since β'_3 is approximately 1 near those points, the decomposition f=f1+f3+f4+Σ f_{2,i} is not a partition of unity, and Lemma 4.5's domain U'_3 excludes exactly those neighborhoods. Consequently the parametrix identity in Proposition 4.7 has not been proved for d>2. The displayed formulas should contain a factor ∏_{k=1}^d(1-β_{2,k}) or the analogous expression.","section":"Section 4.1, definitions of β3 and β4"},{"comment":"The bounded inverses P0, P1, P2,i, P3 are load-bearing: the parametrix in Definition 4.2 is assembled from them. The paper imports these mappings with comments such as 'proof identical' or 'very similar' from [9] and [15]. However, [9] works in dimension 2n while the present manifolds have dimension 2n+2, so the indicial-root and weight-range computations are not literally identical, and Proposition 3.3 of [15] concerns a compact fibration rather than the noncompact warped-QAC setting. Please either state precisely which results in [9] and [15] imply the stated mapping properties, including the admissible δ and τ values, or provide the model computations in the paper. Without this, Definition 4.2 is not fully supported.","section":"Section 2.2 and Section 4.2, Propositions 2.5-2.8 and 2.11"}],"minor_comments":[{"comment":"The proof cites 'Lemma 3.12' for the region IV estimate, but the lemma containing that estimate is Lemma 3.10; there is no Lemma 3.12 in the paper.","section":"Proof of Proposition 3.2"},{"comment":"The index condition 'i>1' in the second assertion of Lemma 3.4 appears arbitrary; the same metric approximation should hold for every singular ray i=1,...,d.","section":"Lemma 3.4"},{"comment":"References [6] and [7] are the same arXiv preprint (Conlon-Rochon, 'Warped quasi-asymptotically conical Calabi-Yau metrics') and should be merged or updated to the published version.","section":"References"},{"comment":"There are numerous presentation slips, including 'paramatrix' for 'parametrix', the mixed indices j and i in the definition of P'_1 in (4.3), and the repeated use of κρ'_1 instead of κρ'_i in (3.5); these should be corrected before resubmission.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The coordinate-scaling issue is the main mathematical obstacle: it affects the exact statement of the theorem and the proof of convergence to the local models. It appears fixable by replacing a_t v_{n+2} with a_t^{-θ_i}v_{n+2}, and the rest of the gluing strategy is coherent enough to merit a revised version. I would also ask the editor to require that the author either prove or precisely quote the model-space Laplacian mapping properties, since the parametrix construction depends on them and the present manuscript only sketches them."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dashen Yan's paper is a serious gluing construction with real novelty, but the coordinate change in §3 that produces the local models is printed with the wrong exponent. The stress-test note is right in substance: with the stated scaling z_{n+2} - s_i = a_t v_{n+2}, the model Xi is not what comes out. Fixable, but load-bearing.\n\nWhat is genuinely new: Theorem 1.1 extends Székelyhidi's warped-QAC construction with an optimized decay estimate for the Ricci potential (Proposition 2.3); the t-dependent weighted Hölder spaces are real machinery; Theorem 1.7 is an actual gluing theorem, and the bubble tree in Section 5 goes beyond [9] and [15]. The paper states its numerical constraints (p > 2q_i plus the extra balancing condition) and openly acknowledges the C^{0,α}-versus-C^{2,α} limitation in the gluing region. No circularity: Li's conjecture is external, and the free parameters are honest bookkeeping.\n\nI checked the algebra. With (z̃, z_{n+2}-s_i) = (a^x ṽ, a_t v_{n+2}), the l≠i cluster factors are s_i - s_l + a_t v — they diverge instead of tending to s_i - s_l, the i-th cluster roots collapse to v = 0, and balancing gives x = (2q-p)/p, not -θ_i. The paper's own formulas show the intended scaling: ρ'_i = t^{θ_i}ρ_i and the t^{2θ_i} convergence in Theorem 1.7 force z_{n+2} - s_i = a_t^{-θ_i} v, which gives x = -θ_i and exactly Xi. The strategy is coherent; the identification underlying F_{i,t} is wrong as printed. The stress test's intermediate x = (q_i+q-p)/p counts the other clusters as constants, which they are not under the printed scaling; its conclusion, not its arithmetic, is what matters. There are sibling slips in §3: the z̃ → z̃' direction at the top is reciprocated (the normalization of ~φ' requires a^{(q-p)/p}), and (3.2) does not type-check — F_{i,t} maps into X_t, so F^*_{i,t}Ω_{X_i} is a form on X_i, and a direct residue computation for the corrected scaling gives t^{(n+2-p)θ_i}, not t^{(n+1)θ_i}. A referee should have the author write out §3 cleanly first.\n\nThe other soft spots are the reader's. The Laplacian inverses in Propositions 2.5-2.8 and 2.11 are quoted from [9] and [15] rather than reproved; acceptable division of labor, but the gluing rests on those mapping properties and their compatibility with the new t-dependent spaces is asserted. Several load-bearing estimates are compressed (\"proof is very similar\", \"direct computation\"): Proposition 2.8, Lemma 3.4, Lemma 4.4, and the inequality chase in Proposition 3.2 is the linchpin but is summarized. These are verification gaps, not evident falsehoods.\n\nThis is for specialists in Kähler collapse and gluing. It deserves a serious referee — send it out, with the instruction to have the §3 scalings and volume-form normalizations corrected first. I would not cite it in its current form; after the fixes, Theorem 1.7 would be the first explicit verification of Li's conjecture in the slower-than-colliding regime. Section 3 would make a good reading-group exercise in exponent-chasing.","headline":"Serious gluing paper with real novelty; the §3 coordinate change that yields the local models is printed with the wrong exponent — fixable, but load-bearing.","tokens_in":35017,"tokens_out":39260,"would_cite":false,"duration_ms":322202,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["53C25","53C55","32Q25","35J60"],"pacs":[],"model":"deepseek-v4-flash","headline":"Under stated degree conditions, collapsing Calabi-Yau metrics are glued from scaled warped QAC models near colliding singular fibers, with error decaying as a negative power of the collapsing parameter.","keywords":["Calabi-Yau manifolds","gluing construction","collapsing metrics","warped quasi-asymptotically conical metrics","Monge-Ampère equation","Gromov-Hausdorff convergence","bubble tree","affine quadrics"],"falsifier":"Check Proposition 2.6 by computing the indicial roots of the Laplacian on C × V_0 with its singular link: if any indicial root lies inside the interval (2(2q-p)/p, 2q/p), the asserted invertibility fails and the parametrix bound in Proposition 4.7 cannot hold. Alternatively, run the construction for degrees violating the inequality $\\frac{p+4q_1}{5p-4q_1} > \\frac{2q_d}{3p-2q_d}$ and test whether the Ricci-potential estimate in Proposition 3.2 still decays; a concrete counterexample would delimit the theorem.","tokens_in":33824,"feed_emoji":"🧩","tokens_out":7067,"duration_ms":65679,"temperature":0.7,"pith_summary":"This paper proves a gluing theorem for collapsing Calabi-Yau manifolds along a holomorphic fibration when several singular fibers collide. It shows that the global Calabi-Yau metric is asymptotically assembled from a semi-Ricci-flat ansatz away from the collision points and scaled warped quasi-asymptotically-conical (QAC) Calabi-Yau metrics near each collision point. The gluing error decays as a negative power of the collapsing parameter, so the model pieces are not merely heuristic but asymptotically exact. This verifies a conjecture posed in [15] and also yields a bubble-tree description of pointed Gromov-Hausdorff limits.","feed_headline":"Gluing theorem assembles collapsing Calabi-Yau metrics","feed_subtitle":"Near colliding singular fibers the metric is a scaled model, with error decaying as a power of t.","key_machinery":"The central object is the parametrix P: an approximate right inverse for the scalar Laplacian on the collapsing space, assembled by gluing the Green operators P_0, P_1, P_{2,i}, and P_3 provided by the Laplacian theory on the model geometries C × V_0, C × V_1, and X_i. The relevant function spaces are t-dependent double-weighted Hölder spaces $C^{{k,α}}$_{δ,τ,t}, whose weights are adapted to four geometric regions: near infinity, near each colliding singular fiber, the semi-Ricci-flat middle region, and the transition between them. The Monge-Ampère equation is then solved by a Banach fixed-point argument, with the crucial input being an optimal decay estimate for the Ricci potential of the approximate metric, obtained by expanding the volume form and using the adjunction formula.","core_discovery":"The central claim is Theorem 1.7. For a family of hypersurfaces X'_t defined by P(z~) + ∏_i ∏_j (z_{n+2} - s_i - $a_t^{{-θ_i}}$ s_{i,j}) = 0, with the degree conditions p > 2q_i and, when q_1 < p/4 < q_d, the additional inequality (p+4q_1)/(5p-4q_1) > (2q_d)/(3p-2q_d), the unique Calabi-Yau Kähler form is obtained by gluing the approximate semi-Ricci-flat metric to scaled warped QAC Calabi-Yau metrics on the local models X_i: P + Q_i = 0 near each collision point S_i. The construction produces biholomorphisms F_{i,t} whose images exhaust the models, and the scaled pulled-back Kähler forms $t^{{2θ_i}}$(F_{i,t}^{-1})^* i∂∂̄φ_t converge to the warped QAC Calabi-Yau forms on X_i. The same machinery gives a bubble-tree convergence theorem in which nonsingular pointed warped QAC Calabi-Yau manifolds converge to minimal bubbles modeled on C × V_0.","pith_inferences":["The numerical inequalities in Theorem 1.7 are likely sufficient rather than necessary; a similar gluing construction may hold in a larger parameter range, and computing the relevant indicial roots would locate the true boundary of the method.","The parametrix framework used here should adapt to higher-depth bubble trees and to fibrations whose model fibers are smoothings of more general Calabi-Yau cones, not just the affine quadrics treated in this paper.","The explicit decay rate suggests a quantitative stability statement: the global Calabi-Yau metric is determined up to the stated negative power of t by the semi-Ricci-flat ansatz and the local warped QAC models, which could be probed numerically on the affine quadric examples."],"forward_implications":["The gluing theorem verifies the conjecture in [15] in the regime where colliding is slower than collapsing, and it gives the explicit error rate t^{-(2-τ)(p-q)/p-a} for the difference between the true Calabi-Yau metric and the glued approximate metric.","The construction produces warped QAC Calabi-Yau metrics on hypersurfaces X_{k,l}: z_1^2 + z_2^2 + z_3^k + z_4^l = 1 with 2k > l, including parameter ranges excluded by the numerical constraints in [7].","It yields explicit non-collapsing complete Calabi-Yau manifolds whose pointed Gromov-Hausdorff limits are minimal bubbles C × V_0, with the limit depending sensitively on the chosen marked points and rescaling factors.","The bubble-tree proposition provides a hierarchy of pointed limits indexed by a tree of colliding zeros of the fibration, with each node modeled on a warped QAC Calabi-Yau space.","The uniqueness statement in Theorem 1.1 implies that any two Calabi-Yau potentials in the same asymptotic class that solve the same equation differ at most by a pluriharmonic function of slower-than-quadratic growth."],"supporting_citations":[{"why":"Supplies the construction scheme for warped QAC Calabi-Yau metrics and the weighted Hölder mapping properties of the Laplacian used throughout the paper.","marker":"[9]"},{"why":"Poses the conjecture being verified and provides the gluing framework and Proposition 2.11 for the harmonic analysis on the collapsing space.","marker":"[15]"},{"why":"Gives the model complete Calabi-Yau metric on C^3 whose decay rate the Ricci-potential estimate in this paper recovers.","marker":"[14]"},{"why":"Defines warped quasi-asymptotically conical metrics and gives earlier examples whose numerical constraints Theorem 1.1 extends.","marker":"[7]"},{"why":"Provides the asymptotically conical Calabi-Yau metrics on the smooth fiber V_1 used to build the semi-Ricci-flat form ω_SRF.","marker":"[5]"},{"why":"Provides the Monge-Ampère solving technique and the function-space framework used to correct the approximate metric to an exact Calabi-Yau metric.","marker":"[11]"},{"why":"Supplies the weighted Sobolev inequality used in Lemma 2.10 to turn compactly supported Ricci potential into decay estimates for the solution.","marker":"[12]"},{"why":"Used in the uniqueness argument to conclude that subquadratic harmonic functions on maximal-volume-growth Calabi-Yau manifolds are pluriharmonic.","marker":"[3]"}],"fun_headline_variants":["Gluing theorem assembles collapsing Calabi-Yau metrics","Bubble tree structure for collapsing Calabi-Yau manifolds","Collapsing Calabi-Yau metrics glued via scaled warped models","Gluing resolves Yang Li conjecture on collapsing metrics","Scaled gluing constructs collapsing Calabi-Yau metrics"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The entire gluing argument rests on the quoted existence of bounded right inverses for the Laplacian on the model spaces C × V_0, C × V_1, and X_i for the stated weight ranges; if any of those mapping properties fails, the parametrix P and with it both main theorems collapse.","fun_headline_variants_meta":{"raw":{"variants":["Gluing theorem assembles collapsing Calabi-Yau metrics","Bubble tree structure for collapsing Calabi-Yau manifolds","Collapsing Calabi-Yau metrics glued via scaled warped models","Gluing resolves Yang Li conjecture on collapsing metrics","Scaled gluing constructs collapsing Calabi-Yau metrics"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000327,"raw_usage":{"total_tokens":1806,"prompt_tokens":903,"completion_tokens":903,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":519,"completion_tokens_details":{"reasoning_tokens":819}},"tokens_in":519,"tokens_out":903,"duration_ms":8502,"temperature":1.0,"reasoning_tokens":819,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T22:08:18.155408+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Check Proposition 2.6 by computing the indicial roots of the Laplacian on C × V_0 with its singular link: if any indicial root lies inside the interval (2(2q-p)/p, 2q/p), the asserted invertibility fails and the parametrix bound in Proposition 4.7 cannot hold. Alternatively, run the construction for degrees violating the inequality $\\frac{p+4q_1}{5p-4q_1} > \\frac{2q_d}{3p-2q_d}$ and test whether the Ricci-potential estimate in Proposition 3.2 still decays; a concrete counterexample would delimit the theorem.","supporting_citations":[{"cited_title":"Sz´ ekelyhidi, Degenerations of Cn and Calabi-Yau metrics, Duke Math","cited_arxiv_id":null,"evidence_quote":"Supplies the construction scheme for warped QAC Calabi-Yau metrics and the weighted Hölder mapping properties of the Laplacian used throughout the paper."},{"cited_title":"Li, A gluing construction of collapsing Calabi-Yau m etrics on K3 fibred 3-folds, Geom","cited_arxiv_id":null,"evidence_quote":"Poses the conjecture being verified and provides the gluing framework and Proposition 2.11 for the harmonic analysis on the collapsing space."},{"cited_title":"Li, A new complete Calabi-Yau metric on C3, Invent","cited_arxiv_id":null,"evidence_quote":"Gives the model complete Calabi-Yau metric on C^3 whose decay rate the Ricci-potential estimate in this paper recovers."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the asymptotically conical Calabi-Yau metrics on the smooth fiber V_1 used to build the semi-Ricci-flat form ω_SRF."},{"cited_title":"Hein, On gravitational instantons , ProQuest LLC, Ann Arbor, MI, 2010; MR2813955","cited_arxiv_id":null,"evidence_quote":"Provides the Monge-Ampère solving technique and the function-space framework used to correct the approximate metric to an exact Calabi-Yau metric."},{"cited_title":"Hein, Weighted Sobolev inequalities under lower Ricci curvature bounds, Proc","cited_arxiv_id":null,"evidence_quote":"Supplies the weighted Sobolev inequality used in Lemma 2.10 to turn compactly supported Ricci potential into decay estimates for the solution."},{"cited_title":"Chiu, Subquadratic harmonic functions on Calabi- Yau manifolds with maximal volume growth, Comm","cited_arxiv_id":null,"evidence_quote":"Used in the uniqueness argument to conclude that subquadratic harmonic functions on maximal-volume-growth Calabi-Yau manifolds are pluriharmonic."}],"review_version":1}