{"id":"1812b7e3-d44a-4b2c-b3e3-aadc94dea31b","arxiv_id":"2608.07694","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":1.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A review of Calabi-Yau constructions via smoothing normal-crossing unions of quasi-Fano manifolds, including non-Kähler examples in every dimension ≥4 and 6518 mirror pairs.","lead":"This paper surveys a decade and a half of work on building Calabi-Yau manifolds by gluing and smoothing pairs of simpler pieces. It reports non-Kähler Calabi-Yau examples in every dimension above three and thousands of mirror pairs.","discovery_kind":"review","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The non-Kählerity proof for the first fourfold relies on an unverified group-theoretic computation: the invariant Néron-Severi intersection of two involutions must contain no ample class, and the paper gives no supporting matrices or proof.","rationale":"I considered two candidate concerns: first, that the stated b2 formula for Sano's higher-dimensional examples might be wrong, but that formula is directly cited from [13] and is not argued in the paper; second, that the algebraic-dimension non-Kählerity argument for the Sano family could fail, but that too is attributed to Sano. The fourfold non-Kählerity argument is the paper's own and its conclusion is a headline result; it is also the point where an internally checked computation is missing. The reader's weakest assumption already identifies the same invariant-Néron–Severi computation, so I agree. The concern does not change the conditional verdict: the paper is a survey, and the missing verification is not supplied in the text, but the reader should treat the fourfold example as conditional pending an independent check of the invariant lattice computation.","tokens_in":7320,"tokens_out":10127,"duration_ms":101990,"concrete_test":"Re-derive the two involutions from [8] as matrices in GL(3, Z[ζ]) with no translation part, and compute the action on NS(E_ζ^3), the rank-9 lattice of Hermitian forms whose imaginary part is integral on Z[ζ]^3. Let L = NS(E_ζ^3)^{σ1} ∩ NS(E_ζ^3)^{σ2}; test whether L contains a positive-definite Hermitian form, for instance by computing the Gram matrix of L and applying lattice reduction or a semidefinite-feasibility check. If such a form exists, the proof's key assertion fails; if none exists and ⟨σ1, σ2⟩ is infinite, the contradiction stands.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 3's non-Kählerity proof for the fourfold M hinges on the unproved assertion that the involution-generated subgroup ⟨σ1, σ2⟩ of Aut(E_ζ^3) is infinite and that NS(E_ζ^3)^{σ1} ∩ NS(E_ζ^3)^{σ2} contains no ample class. If an ample class were invariant, the big class pulled back from a hypothetical Kähler metric on M would be invariant, the stated contradiction would evaporate, and M could be projective. The paper does not display the matrices defining σ1 and σ2, nor any computation of the invariant Néron–Severi lattice; it refers to the author's own [8]. Since this is the only place the paper proves a new non-Kählerity result (the higher-dimensional Sano examples use a separate algebraic-dimension argument), an unchecked group-theoretic computation is load-bearing for the claim that the first non-Kähler fourfold exists.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper is a survey-style update of the author's construction of Calabi-Yau manifolds by smoothing normal crossing unions of two varieties with a common anticanonical divisor, a method introduced to the physics community in [7]. Section 1 sets out the smoothing setup and the d-semistability condition. Section 2 reviews Hashimoto-Sano's simply connected non-Kähler Calabi-Yau threefolds with unbounded second Betti number and Lee's threefolds with h^{1,1}=h^{1,2}=1. Section 3 describes Lee's construction of a non-Kähler Calabi-Yau fourfold from Beauville's rigid Calabi-Yau threefold and Sano's higher-dimensional non-Kähler Calabi-Yau manifolds with arbitrarily large second Betti number. Section 4 explains a scheme for mirror pairs of Calabi-Yau threefolds from quasi-Fano threefold pairs, reporting 6518 mirror pairs from three-dimensional reflexive polytopes, and Section 5 discusses the open problem of whether Kähler Calabi-Yau threefolds have finitely many topological types. The paper gives no new proofs; its results are drawn from [2], [13], and the author's own publications and preprints [6]-[10].","tokens_in":7488,"tokens_out":10016,"duration_ms":103696,"significance":"If the underlying constructions are valid, the paper is a useful concise compilation of results showing that the smoothing method produces Calabi-Yau manifolds outside the standard complete-intersection and toric families, including non-Kähler examples in arbitrarily high dimension and threefolds with extremely small Hodge numbers. The clear statements of the d-semistability condition, the explicit formulas for b2 and Euler characteristics, and the careful caveat that the mirror pairs are established only at the level of Hodge numbers are strengths. However, the paper is not self-contained at exactly the points where its advertised novelty is largest: the non-Kählerity proof for the fourfold in Section 3 relies on an unverified group-theoretic assertion, and two of the main quantitative claims rest on an unpublished preprint [6] and an author-hosted appendix [10].","major_comments":[{"comment":"The proof that the fourfold M is non-Kähler hinges on the assertion that 'the subgroup generated by these two involutions is infinite, and the intersection of their invariant Néron–Severi subgroups contains no ample class.' No matrices defining σ1 and σ2 are displayed, no computation of NS(E_ζ^3)^{σ1} ∩ NS(E_ζ^3)^{σ2} is given, and no precise statement from [8] is reproduced. Since this assertion is the only step that excludes a Kähler metric, it is load-bearing for the paper's central claim of a first non-Kähler Calabi-Yau fourfold. Please either include the explicit involutions and the invariant-lattice computation, or restate the argument as a theorem whose full proof is in [8] and state exactly which facts from [8] are being used. In addition, the sentence 'Pulling this class back to the abelian threefold E_ζ^3 would give a big, and therefore ample, divisor class invariant under both σ1 and σ2' appeals to a nontrivial fact about big line bundles on abelian varieties; this fact should be stated explicitly rather than passed over.","section":"Section 3"},{"comment":"The construction of Calabi-Yau threefolds with h^{1,1}=h^{1,2}=1 is one of the paper's advertised highlights, but it rests entirely on reference [6], an unpublished preprint. The Hodge-number formulas h^{1,1}=r-rk⟨γ_i⟩+1 and h^{1,2}=21+Σ g(γ_i)-rk⟨γ_i⟩ are stated without derivation, and the existence of twenty rational curves on the Fermat quartic with the required independence and linear-equivalence properties is asserted without a concrete configuration. If the paper is intended as a self-contained announcement, the missing configuration and the derivation of the formulas are essential; if it is intended as a survey, the dependence on [6] should be made prominent and the preprint should be published or uploaded to a stable repository before the claim is repeated as established.","section":"Section 2"},{"comment":"The statement that applying the quasi-Fano mirror construction to all three-dimensional reflexive polytopes produces 6518 mirror pairs of Calabi-Yau threefolds, including 79 self-mirror examples, is supported only by the author-hosted appendix [10]. The reader cannot verify the enumeration or the claimed Hodge-number exchange from the material in this paper. Please provide a machine-readable table, a reproducible algorithm, or an independently published source; otherwise, the sentence should be rephrased as a report of unpublished enumerations in [10] and distinguished from the peer-reviewed results of [9].","section":"Section 4"},{"comment":"The non-Kählerity of Sano's higher-dimensional examples is argued through the algebraic dimension: 'For a very general smoothing, every meromorphic function on X(m) essentially comes from the base T. Consequently, its algebraic dimension is a(X(m))=dim T=N−2.' The paper does not explain why the projection to T extends through the smoothing, nor why very general smoothings have no additional meromorphic functions. Since this is the step that proves non-Kählerity for these manifolds, it should be stated as a theorem with a precise reference to [13], or a proof sketch should be supplied.","section":"Section 3"}],"minor_comments":[{"comment":"The bibliographic data for reference [13] appear inconsistent: 'Geom. Topol. 14 (2021), no. 4, 1448–1460' pairs a 2010 volume number with a 2021 year. Please check the volume, year, and page numbers against the published version.","section":"References"},{"comment":"The displayed Hodge diamond for the h^{1,1}=h^{1,2}=1 threefold would be easier to read if the entries were labeled with h^{p,q} or if the unlabeled rows were accompanied by a one-sentence explanation of the convention for orienting the diamond.","section":"Section 2"},{"comment":"The quantity α_X is used to state the numerical mirror relation α_X+α_Y=20, but the reason 20 is the relevant bound for K3 lattices (the rank of the transcendental lattice of a K3 surface or the Picard rank bound) is not explained; adding a sentence would make the relation understandable to the intended hep-th readership.","section":"Section 4"},{"comment":"The phrase 'Landau–Ginzburg models' is invoked in the abstract and in Section 4 without a definition or a reference to the specific LG framework being used; a brief definition or a pointer to the relevant physics literature would improve accessibility.","section":"Section 1"},{"comment":"The final paragraph of Section 4 correctly notes that the constructed pairs are only conjectural as full mirror pairs because only Hodge numbers are exchanged; this caveat is valuable and should perhaps be printed more prominently, for example at the first occurrence of the word 'mirror pair' in the section.","section":"Section 4"}],"recommendation":"major_revision","confidential_remarks":"This manuscript is essentially a summary of the author's own previous work, with the notable additions being the non-Kähler fourfold and the higher-dimensional examples. The main technical gap is the unverified Néron-Severi computation in Section 3, which is the only new-looking proof in the paper. The editor may also wish to consider whether the journal's review format accepts claims based on an unpublished preprint [6] and an author-hosted appendix [10]; the manuscript should at minimum make these dependencies explicit and offer a route to verification. There is no indication of any deliberate misrepresentation, but the verifiability of the central claims is currently insufficient for the paper to stand alone."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThis is a review, not a research paper. Nam-Hoon Lee summarizes his own program for constructing Calabi-Yau manifolds by smoothing normal-crossing unions of quasi-Fano or non-compact CY pieces, and he does it clearly. The genuinely new material here is zero—the threefold families, the non-Kähler fourfold, the unbounded b2 in higher dimensions, and the 6,518 mirror pairs all appeared in earlier papers, several of them his own. What the paper does well is synthesis: it lays out the common smoothing mechanism, the d-semistability condition, and the Tyurin/Landau-Ginzburg context in a way that a reader new to the area can follow.\n\nThe soft spots are in proportion. First, the paper leans on two items the reader cannot check: the preprint [6] for the h^{1,1}=h^{1,2}=1 examples, and an author-hosted appendix [10] for the 6,518 mirror pairs. That's acceptable in a survey, but it limits verification. Second, the proof sketch for the non-Kähler fourfold in Section 3 omits the load-bearing group-theoretic computation. The claim that ⟨σ1,σ2⟩⊂Aut(E_ζ^3) is infinite and that the intersection of the invariant Néron-Severi groups contains no ample class is exactly what rules out a Kähler metric. The paper gives no matrices, no invariant-lattice computation, just a pointer to [8]. The stress-tester is right that a reader of this survey cannot verify the flagship example from the text. I don't count that as a flaw in the mathematical claim, because [8] was peer-reviewed; but it is a flaw in the survey's self-containedness, and irritating in a paper whose main purpose is to advertise the result.\n\nThe mirror-pair section is honest: it states clearly that only Hodge-number exchange is established, not mirror symmetry in full. Good.\n\nWho should read this? Physicists or mathematicians who want a map of the smoothing construction and its recent outputs, with minimal technical overhead. It is not a place to learn proofs. I would send it to peer review if the venue publishes expository reviews; for a research journal, the lack of new content argues for desk rejection. If it goes to review, the referee should ask the author to either include the σ1/σ2 computation or state explicitly that the verification is in [8] and why it can't be reproduced here.","headline":"A clear, self-citing survey of the author's smoothing constructions; no new results, and the non-Kähler fourfold proof is asserted rather than demonstrated.","tokens_in":7977,"tokens_out":2812,"would_cite":false,"duration_ms":27253,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14J32","14J28","32Q25"],"pacs":[],"model":"deepseek-v4-flash","headline":"Smoothing two non-compact Calabi-Yau pieces yields Calabi-Yau manifolds with unbounded second Betti numbers and the first non-Kähler examples in every dimension above three.","keywords":["Calabi-Yau manifolds","normal crossing variety","smoothing","non-Kähler manifolds","second Betti number","quasi-Fano manifolds","mirror symmetry","Tyurin degeneration"],"falsifier":"Compute the subgroup generated by the two involutions $\\sigma_1$ and $\\sigma_2$ acting on $E_\\zeta^3$ and the intersection of their invariant Néron-Severi groups; an ample class in that intersection would give the big divisor whose absence is needed, and would overturn the non-Kähler fourfold example. Conversely, exhibiting a Kähler metric on the smoothed fourfold would also falsify the claim.","tokens_in":7101,"feed_emoji":"🪞","tokens_out":11619,"duration_ms":96300,"temperature":0.7,"pith_summary":"This review develops one idea: a compact Calabi-Yau manifold can be built from two non-compact Calabi-Yau pieces by gluing them along a common anticanonical divisor and smoothing the resulting normal crossing union. The paper claims this smoothing method produces Calabi-Yau threefolds with second Betti number growing without bound, a simply connected Calabi-Yau threefold with $h^{1,1}=h^{1,2}=1$, and the first non-Kähler Calabi-Yau manifolds in dimensions four and higher, with unbounded second Betti numbers in every fixed dimension $N\\ge 4$. It also organizes 6,518 mirror pairs of Calabi-Yau threefolds, including 79 self-mirror examples, through Landau-Ginzburg mirror symmetry for pairs of quasi-Fano threefolds. If these constructions are correct, the topology of Calabi-Yau manifolds is much less constrained than complete-intersection examples suggest, and smoothing degenerations becomes a systematic source of new Calabi-Yau geometries for string compactification.","feed_headline":"Non-Kähler Calabi-Yau manifolds exist in every dimension above three","feed_subtitle":"Gluing two non-compact Calabi-Yau pieces yields unbounded Betti numbers and 6,518 mirror pairs.","key_machinery":"The central object is the normal crossing variety $X_0=Y_1\\cup Y_2$ with common smooth divisor $D$, where $D$ belongs to the anticanonical system of each component, together with the d-semistability condition $N_{D/Y_1}\\otimes N_{D/Y_2}\\simeq\\mathcal{O}_D$; the Kawamata-Namikawa smoothing theorem then turns $X_0$ into a smooth Calabi-Yau manifold. The unbounded Betti numbers are produced by infinite-order automorphisms of the common divisor: involutions on a K3 surface in $\\mathbb{P}^1\\times\\mathbb{P}^1\\times\\mathbb{P}^1$ in the threefold case, and Cremona transformations on rational elliptic surfaces in the higher-dimensional case, which twist the gluing before smoothing. Non-Kählerness is established by showing that a hypothetical Kähler metric would force a big line bundle on the normal crossing variety, whose restriction to the common divisor would pull back to an ample class on an abelian threefold invariant under two involutions; the chosen involutions generate an infinite group whose invariant Néron-Severi intersection contains no ample class, producing the contradiction. For the mirror construction, the operative mechanism is a quasi-Fano threefold with an anticanonical K3 fibration, whose complement carries a Landau-Ginzburg superpotential; mirror pairs are defined by inducing mirror lattice-polarized K3 structures on the common fibers.","core_discovery":"The paper's central claim is that the smoothing of a d-semistable normal crossing union $X_0=Y_1\\cup Y_2$, where each $Y_i$ is a smooth projective variety and $D=Y_1\\cap Y_2$ is an anticanonical divisor in both, produces a smooth compact Calabi-Yau manifold, and this mechanism realizes previously unknown families. In three dimensions it yields simply connected non-Kähler Calabi-Yau threefolds with $b_2=a+3$ for every positive integer $a$, and a simply connected Calabi-Yau threefold with $h^{1,1}=h^{1,2}=1$, self-intersection $\\xi^3=2$, and $\\xi\\cdot c_2=44$. In dimension four, the smoothing of two fourfolds glued along a rigid Calabi-Yau threefold gives a simply connected non-Kähler Calabi-Yau fourfold with Euler number 108. Extending the construction with rational elliptic surfaces and Cremona transformations gives, for every dimension $N\\ge 4$, simply connected non-Kähler Calabi-Yau manifolds with $b_2=m+10$ when $N=4$ and $b_2=m+2$ when $N\\ge 5$, so the second Betti number is unbounded in each fixed dimension. The paper further claims that mirror pairs of quasi-Fano threefolds, combined with lattice-polarized K3 mirror symmetry, yield 6,518 mirror pairs of Calabi-Yau threefolds.","pith_inferences":["A direct extension the paper leaves implicit is to use the same two-component smoothing to produce Calabi-Yau manifolds with prescribed Hodge numbers beyond the threefold case; the threefold formulas express $h^{1,1}$ and $h^{1,2}$ in terms of curve classes blown up on a K3 surface, and analogous formulas in higher dimensions would make the construction a general machine for building exotic Calabi","If the smoothing method can be made to preserve projectivity as well as d-semistability, it would likely generate infinitely many topological types of Kähler Calabi-Yau threefolds, directly attacking the open question the paper poses about finiteness of topological types.","The mirror construction suggests a physical test: the Landau-Ginzburg superpotentials defined by the two quasi-Fano components should determine the quantum cohomology of the smoothed Calabi-Yau threefold, so computing Gromov-Witten invariants for a smoothing of a known normal crossing union would test the mirror correspondence beyond Hodge numbers."],"forward_implications":["For every fixed dimension $N\\ge 4$, the construction gives infinitely many simply connected non-Kähler Calabi-Yau manifolds, with second Betti numbers $m+10$ (when $N=4$) and $m+2$ (when $N\\ge 5$) as $m$ ranges over positive integers.","There exist simply connected Calabi-Yau threefolds with $h^{1,1}=h^{1,2}=1$, intersection form $\\xi^3=2$, and $\\xi\\cdot c_2=44$, providing extremal small-Hodge-number examples for classification and mirror symmetry.","The smoothing method produces 6,518 mirror pairs of Calabi-Yau threefolds from three-dimensional reflexive polytopes, including 79 self-mirror examples, so mirror pairs are plentiful outside the toric complete-intersection setting.","The higher-dimensional non-Kähler examples have algebraic dimension $N-2$, exactly two below the maximal possible value, which shows how close non-Kähler Calabi-Yau manifolds can come to being algebraic.","The paper's open question is whether the same flexibility can create infinitely many topological types of Kähler Calabi-Yau threefolds, which would settle a long-standing finiteness problem."],"supporting_citations":[{"why":"supplies the Kawamata-Namikawa smoothing theorem that converts a d-semistable normal crossing union into a smooth Calabi-Yau manifold.","marker":"[12]"},{"why":"gives the construction of simply connected non-Kähler Calabi-Yau threefolds with arbitrarily large second Betti number, the threefold prototype for unbounded $b_2$.","marker":"[2]"},{"why":"contains the first non-Kähler Calabi-Yau fourfold, built from a rigid Calabi-Yau threefold and two involutions, which the review reports.","marker":"[8]"},{"why":"provides the higher-dimensional analogue producing non-Kähler Calabi-Yau manifolds with arbitrarily large $b_2$ in every dimension $N\\ge 4$.","marker":"[13]"},{"why":"constructs Calabi-Yau threefolds with $h^{1,1}=h^{1,2}=1$ by blowing up a quartic K3 surface along twenty rational curves.","marker":"[6]"},{"why":"establishes the mirror-pair construction from mirror pairs of quasi-Fano threefolds, the basis for the 6,518 mirror pairs.","marker":"[9]"},{"why":"collects the 6,518 mirror pairs and 79 self-mirror examples obtained by the quasi-Fano construction.","marker":"[10]"},{"why":"supplies the lattice-polarized toric K3 mirror symmetry used to recover divisor classes missing from the ambient toric threefold.","marker":"[11]"},{"why":"supplies the Tyurin-degeneration and Landau-Ginzburg mirror principle that motivates constructing mirror fibrations from quasi-Fano pairs.","marker":"[1]"}],"fun_headline_variants":["Smoothing glues give non-Kähler CYs in every dimension above 3","6,518 mirror pairs and unbounded Betti numbers from CY smoothing","Non-Kähler Calabi-Yau manifolds for all dimensions ≥4","Gluing two non-compact pieces yields CYs with unbounded invariants","From quasi-Fano pairs: 6,518 mirror pairs and non-Kähler CYs"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The fourfold's non-Kähler conclusion rests on the group-theoretic assertion that two involutions of the elliptic-threefold quotient generate an infinite automorphism group whose common invariant divisor classes contain no ample class; if that assertion fails, the contradiction excluding a Kähler metric disappears.","fun_headline_variants_meta":{"raw":{"variants":["Smoothing glues give non-Kähler CYs in every dimension above 3","6,518 mirror pairs and unbounded Betti numbers from CY smoothing","Non-Kähler Calabi-Yau manifolds for all dimensions ≥4","Gluing two non-compact pieces yields CYs with unbounded invariants","From quasi-Fano pairs: 6,518 mirror pairs and non-Kähler CYs"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000799,"raw_usage":{"total_tokens":3521,"prompt_tokens":955,"completion_tokens":2566,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":571,"completion_tokens_details":{"reasoning_tokens":2458}},"tokens_in":571,"tokens_out":2566,"duration_ms":17651,"temperature":1.0,"reasoning_tokens":2458,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T00:22:42.661440+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the subgroup generated by the two involutions $\\sigma_1$ and $\\sigma_2$ acting on $E_\\zeta^3$ and the intersection of their invariant Néron-Severi groups; an ample class in that intersection would give the big divisor whose absence is needed, and would overturn the non-Kähler fourfold example. Conversely, exhibiting a Kähler metric on the smoothed fourfold would also falsify the claim.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"supplies the Kawamata-Namikawa smoothing theorem that converts a d-semistable normal crossing union into a smooth Calabi-Yau manifold."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"gives the construction of simply connected non-Kähler Calabi-Yau threefolds with arbitrarily large second Betti number, the threefold prototype for unbounded $b_2$."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"contains the first non-Kähler Calabi-Yau fourfold, built from a rigid Calabi-Yau threefold and two involutions, which the review reports."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"provides the higher-dimensional analogue producing non-Kähler Calabi-Yau manifolds with arbitrarily large $b_2$ in every dimension $N\\ge 4$."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"constructs Calabi-Yau threefolds with $h^{1,1}=h^{1,2}=1$ by blowing up a quartic K3 surface along twenty rational curves."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"establishes the mirror-pair construction from mirror pairs of quasi-Fano threefolds, the basis for the 6,518 mirror pairs."},{"cited_title":"Mirror pairs of Calabi–Yau threefolds from mirror pairs of quasi-Fano threefolds","cited_arxiv_id":null,"evidence_quote":"collects the 6,518 mirror pairs and 79 self-mirror examples obtained by the quasi-Fano construction."},{"cited_title":"Lattice polarized toric K3 surfaces","cited_arxiv_id":"hep-th/0409290","evidence_quote":"supplies the lattice-polarized toric K3 mirror symmetry used to recover divisor classes missing from the ambient toric threefold."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"supplies the Tyurin-degeneration and Landau-Ginzburg mirror principle that motivates constructing mirror fibrations from quasi-Fano pairs."}],"review_version":1}