{"id":"49d1312e-c3d3-4d7d-b3f8-b5f3b7c30469","arxiv_id":"1908.04227","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"The paper constructs the SYZ mirror Landau-Ginzburg model of a genus 2 curve and proves a cohomology-level homological mirror symmetry embedding of line bundles into a new Fukaya-Seidel category of a non-exact fibration.","lead":"A mathematician pairs a genus 2 curve with a mirror space built from its tropical geometry and shows that some algebraic data on the curve matches symplectic data on the mirror. This is one of the first working mirror statements for a curve of general type, with a newly defined Fukaya category of a non-exact symplectic fibration.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The target category H^0FS(Y,v0) is defined through unpublished Abouzaid-Seidel localization lemmas in a non-exact, compact-fiber setting; if those lemmas fail or do not apply, the central embedding theorem is not defined.","rationale":"The reader's weakest assumption correctly identifies the target category's dependence on unpublished localization machinery. This is load-bearing because the main theorem is an embedding into H^0FS(Y,v0); if the category is not well-defined or its morphism spaces do not match the directed model, the right vertical arrow has no truth value. The paper's detailed left-arrow proof, monodromy computation, and differential computation are independent support for parts of the construction, and no outright error was found, so a reject is not warranted. A conditional/unchanged verdict is appropriate: the concern is concrete and repairable by supplying the missing proofs or references.","tokens_in":68666,"tokens_out":7604,"duration_ms":88880,"concrete_test":"In the local model of (Y,v0) with compact torus fiber and the symplectic form of Definition 3.28, reproduce the proof of Lemma 4.8: construct the quasi-unit e_L for a U-shaped parallel transport, prove it is a cycle, and check that localizing at e_L gives the same morphism spaces and compositions as the directed category, including contributions from sphere bubbles. If the argument requires exactness, monotonicity, or a Lefschetz-type singular fiber, the target category H^0FS(Y,v0) is not established.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Theorem 1.2 asserts a fully faithful embedding into H^0FS(Y,v0). Definition 4.6 constructs this category by categorical localization, and Lemma 4.8 (localized morphisms coincide with directed ones) and Lemma 4.9 (composition of roofs is a roof) are assigned to unpublished Abouzaid-Seidel and Abouzaid-Auroux work; Lemma 4.54 (quasi-invariance) is marked 'References for proof'. The paper explicitly notes in Remark 4.11 that this is not the exact setting of Seidel's Lefschetz construction: fibers are compact tori, so sphere bubbles occur. The standard localization proof for quasi-units in exact or monotone settings does not automatically cover this case, and the paper does not supply the missing verification. No internal contradiction was found in the left-arrow computation or the monodromy/differential calculation, but those results alone do not establish the existence or invariance of the target category.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a cohomological mirror symmetry statement for a complex genus 2 curve realized as a theta divisor H in an abelian surface V. It constructs the SYZ mirror as a Landau-Ginzburg model (Y,v0), where Y is a Gamma_B-quotient of an infinite-type toric variety and v0 = xyz, equipped with a non-exact symplectic fibration structure. Theorem 1.2 asserts two fully faithful embeddings: D^b_LCoh(V) into H^0Fuk(V^vee) (left vertical arrow) and D^b_LCoh(H) into H^0FS(Y,v0) (right vertical arrow). The left arrow is established by explicit computation of morphisms between linear Lagrangians and a triangle count whose weights match theta-function multiplication in Lemma 2.20. The right arrow rests on a definition of a DFS-type Fukaya-Seidel category via categorical localization (Definition 4.6), a monodromy computation (Lemma 4.20), and a computation of the differential as proportional to the theta function after incorporating disc and sphere counts. The main structural gap is that the target category for the right arrow is not constructed self-containedly: its localized morphism and quasi-invariance lemmas are postponed to unpublished work.","tokens_in":68781,"tokens_out":9051,"duration_ms":107545,"significance":"If the category foundations can be supplied, the paper would be a significant contribution: a concrete, computable HMS statement for a general-type curve with the curve on the complex side, and one of the first Fukaya-Seidel categories for a non-exact, non-monotone symplectic fibration with compact fibers. The explicit triangle count in Lemma 2.5 and its matching with theta multiplication in Lemma 2.20 are strong internal evidence for the left arrow; the monodromy computation is detailed; and the differential computation is ambitious and, conditional on the standard tools, internally consistent. The paper also gives a clean description of how the mirror is forced by the tropicalization of the defining theta function. The manuscript is not self-contained in a load-bearing way; however, the gaps are of the missing-proof or unpublished-reference type rather than obvious contradictions, so the result is credible as a research announcement and worthy of a major revision.","major_comments":[{"comment":"The category H^0FS(Y,v0) in Theorem 1.2 is the target of the main embedding, but its definition is not self-contained: Definition 4.6 defines morphisms by categorical localization, and Lemmas 4.8 and 4.9, which identify the localized morphisms and compositions with directed ones, are assigned to unpublished work of Abouzaid-Seidel and Abouzaid-Auroux; Lemma 4.54 (quasi-invariance on regular choices) is tagged \"References for proof.\" Because Remark 4.11 explicitly notes that the fibration is not exact (compact torus fibers) and Remark 4.13 notes it is not monotone, the standard localization and quasi-invariance arguments do not automatically apply, and no replacement argument is supplied. This is load-bearing: without these lemmas the right vertical arrow is not defined.","section":"Section 4.1, Definition 4.6; Lemmas 4.8, 4.9, 4.54"},{"comment":"Quasi-invariance of H^0FS on regular choices is needed for the computation in Section 5 to be an invariant of the category: the paper computes the differential at J0 with an admissible perturbation and uses a cobordism from a generic J, but the cobordism and the independence of the count from the perturbation must be established within the non-exact setting. The proof reference to unpublished work is not enough for a foundational lemma of this kind, since the category used in the main theorem is defined only after this invariance is known.","section":"Section 4.6, Lemma 4.54"},{"comment":"The sphere-bubble contributions to the differential are imported from [KL19] for the infinite-type toric cover ~Y, while the target Y is the quotient by Gamma_B with compact fibers. The paper does not prove that the relevant counts, homology classes, and weights descend to the quotient, or that the open Gromov-Witten invariants used in Corollary 5.3 are Gamma_B-invariant. Without such a descent statement, the proportionality of the differential to the theta function is not established on Y itself.","section":"Section 5.3, Theorem 5.8"}],"minor_comments":[{"comment":"There is a typo in the Highlights: \"sympectic side\" should be \"symplectic side.\"","section":"Highlights"},{"comment":"The paper uses H^0Fuk(V^vee) and H^0FS(Y,v0) for cohomological categories, but the Maslov grading conventions for the linear Lagrangians are not spelled out; since the grading determines which part of Floer cohomology is taken, a sentence specifying the grading would help the reader.","section":"Theorem 1.2 and Section 2"},{"comment":"The bump functions alpha3,...,alpha6 are constrained only by inequalities in the text; their precise domains, boundary conditions, and derivative bounds should be collected in one place, since Appendix A is invoked repeatedly and the reader cannot easily check the claimed smallness of derivative contributions.","section":"Definition 3.28"}],"recommendation":"major_revision","confidential_remarks":"The paper is clearly a substantial piece of work, and the concrete computations are likely to be correct. The main obstacle is foundational: the target category H^0FS(Y,v0) is defined through unpublished Abouzaid-Seidel localization, and the paper does not provide enough detail to verify that the non-exact, non-monotone setting is covered. If the author can replace the deferred lemmas by published references with precise statements, or include proofs in an appendix, the result would be much stronger. I recommend major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here is my take on Cannizzo.\n\nWhat is actually new: the right vertical arrow of Theorem 1.2—an AAK-style generalized SYZ mirror (Y,v0) for the genus 2 curve as a hypersurface in its Jacobian, where the superpotential is turned into a symplectic fibration. The left arrow is a re-derivation of known HMS for abelian varieties, done with explicit linear Lagrangians and a clean triangle count. The genuinely new parts are the Gamma_B-quotient construction of (Y,v0), the monodromy computation in Lemma 4.20, and the reduction of the differential to the theta function in Theorem 5.4. These are substantial, detailed, and internally consistent; the paper deserves credit for actually doing the tropicalization and the quotient rather than leaving it at the level of speculation.\n\nThe soft spot is exactly the one the stress-test note names. The target category H^0FS(Y,v0) is defined by categorical localization in Definition 4.6, and the key lemmas—localized morphisms equal directed ones, roofs compose, and quasi-invariance—are assigned to unpublished Abouzaid-Seidel and Abouzaid-Auroux work or marked \"References for proof.\" In a Seidel-type exact Lefschetz fibration these lemmas would be standard. Here Remark 4.11 says the fibration is non-exact, the fibers are compact tori, and sphere bubbles contribute. The paper does not supply the missing verification that the localization machinery extends to this setting. That is a load-bearing gap, not a stylistic one: if the localized category is not well-defined, the embedding theorem has no target. The paper is transparent about the dependence, which is to its credit, but transparency does not close the gap.\n\nTwo smaller caveats. I could not verify the appendices on non-degeneracy of the symplectic form, and the application of the KL19 open mirror theorem to the quotient is asserted rather than fully proved. I have also not seen Section 6, so the final embedding argument itself is not before me. If the computations in Sections 2 and 5 are right, the construction is still a useful concrete example of a non-exact, non-Lefschetz symplectic fibration. The headline theorem, however, remains conditional on foundations that are not in the paper.\n\nWho should read it: people working on SYZ mirrors for hypersurfaces of abelian varieties, and anyone collecting explicit examples of Fukaya categories for non-exact symplectic fibrations. It is not yet a proof of HMS for the genus 2 curve in the strong category-level sense, but the geometric and Floer-theoretic computations are worth serious referee time. I would send it to peer review, with the expectation that the localization lemmas either be supplied or the theorem be reformulated to match what is actually established.","headline":"A serious, detailed construction of the first non-exact SYZ mirror for a curve of general type, whose main theorem is conditional on unpublished localization machinery for the target category.","tokens_in":69553,"tokens_out":3354,"would_cite":true,"duration_ms":39279,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["53D37","14F08","53D40","14K25"],"pacs":[],"model":"deepseek-v4-flash","headline":"A genus 2 curve is shown to satisfy mirror symmetry on cohomology: line bundles on the curve correspond to Lagrangians in a symplectic fibration, and products are computed by holomorphic triangle counts.","keywords":["homological mirror symmetry","genus 2 curve","Fukaya category","derived category of coherent sheaves","Landau-Ginzburg model","symplectic fibration","theta functions","SYZ mirror symmetry"],"falsifier":"Compute the morphism space between $L_k$ and its upward pushoff in the localized category of Definition 4.6 by both the directed bigon count and by roofs; if the two answers differ for some regular almost complex structure, Lemma 4.8 fails and the right vertical embedding has no well-defined target.","tokens_in":68208,"feed_emoji":"🪞","tokens_out":13195,"duration_ms":122269,"temperature":0.7,"pith_summary":"This paper proves a cohomological form of homological mirror symmetry for a complex genus 2 curve $\\Sigma_2$, realized as a $\\theta$ divisor $H=s^{-1}(0)$ inside the abelian surface $V=(\\mathbb{C}^*)^2/\\Gamma_B$ with $\\Gamma_B=\\mathbb{Z}\\langle\\gamma',\\gamma''\\rangle$, $\\gamma'=(2,1)$, $\\gamma''=(1,2)$. It constructs the mirror as a Landau-Ginzburg model $(Y,v_0)$, a $\\Gamma_B$-quotient of an infinite-type toric variety with superpotential $v_0=xyz$, equipped with the structure of a symplectic fibration. The main theorem states that the natural restriction diagram commutes and that the vertical maps are fully faithful embeddings: $D^b_{\\mathrm{LCoh}}(V)\\hookrightarrow H^0\\mathrm{Fuk}(V^\\vee)$ and $D^b_{\\mathrm{LCoh}}(H)\\hookrightarrow H^0\\mathrm{FS}(Y,v_0)$. A sympathetic reader should care because the product in the mirror category is computed by counting holomorphic triangles, so the canonical ring of $\\Sigma_2$ becomes a symplectic invariant: multiplication of sections of powers of the canonical bundle is reproduced by triangle counts in the mirror. The paper is also one of the first to define and use a Fukaya category for a non-exact, non-Lefschetz symplectic fibration with compact torus fibers.","feed_headline":"Line bundles on a genus 2 curve become triangle counts in a mirror","feed_subtitle":"Cohomological mirror symmetry: Floer products reproduce multiplication in the canonical ring of the genus 2 curve.","key_machinery":"The central object is the family of linear Lagrangians $\\ell_k$ in the SYZ dual abelian variety $V^\\vee=T_B\\times T_F$, defined in action-angle coordinates by $\\theta\\equiv-k\\lambda\\xi\\pmod{\\mathbb{Z}^2}$ with $\\lambda=\\begin{pmatrix}2&1\\\\1&2\\end{pmatrix}^{-1}$; these are Lagrangian graphs whose intersections with $\\ell_j$ and $\\ell_i$ number $(j-i)^2$, matching the dimension of $H^0(V,L^{j-i})$. The same Lagrangians, parallel transported over U-shaped curves in the base of the superpotential $v_0=xyz$, generate the Fukaya-Seidel-type category on $(Y,v_0)$. The symplectic fibration itself is built from a Kähler potential patched from the toric potentials of $\\mathbb{C}^3$ charts and of the singular fiber $\\mathbb{CP}^2(3)/\\Gamma_B$, with the symplectic form chosen so that $v_0$ is a symplectic fibration. The computational engine is a Leibniz rule reducing the differential on all Lagrangians to the differential for the moment-map fiber Lagrangian $t_x$, followed by a cobordism argument that identifies the desired count with $J_0$-holomorphic discs; that disc count equals the $\\theta$ function defining the line bundle, while disc-with-sphere configurations are included through an open mirror theorem for infinite-type toric Calabi-Yau manifolds.","core_discovery":"On its own terms, the discovery is Theorem 1.2: choose $L\\to V$ to be the ample line bundle whose factor of automorphy is $s(\\gamma,x)=x^{\\lambda(\\gamma)}\\tau^{\\kappa(\\gamma)}$, with $\\lambda=\\begin{pmatrix}2&1\\\\1&2\\end{pmatrix}^{-1}$; then $H=s^{-1}(0)$ is a complex genus 2 curve. The paper proves that the diagram with $D^b_{\\mathrm{LCoh}}(V)\\to D^b_{\\mathrm{LCoh}}(H)$ on the complex side and $H^0\\mathrm{Fuk}(V^\\vee)\\to H^0\\mathrm{FS}(Y,v_0)$ on the symplectic side commutes, with fully faithful vertical embeddings. The objects on the symplectic side are explicit: $L^k$ maps to the linear Lagrangian $\\ell_k$ in $V^\\vee$, and $L^k|_H$ maps to the fibered Lagrangian obtained by parallel transporting $\\ell_k$ around a U-shaped curve in the base of $v_0$. The core numerical fact is that the $\\mu_2$ product in the torus, which counts holomorphic triangles, matches the product of $\\theta$ functions computing the multiplication of sections; since $L|_H$ is the canonical bundle of $\\Sigma_2$, the ring structure on $\\bigoplus_{i\\ge0}H^0(\\Sigma_2,L^i)$ is determined by those triangle counts. The proof of the right vertical arrow uses a computation of the differential as the $\\theta$ function, with disc counts contributing in the leading order and sphere-bubble corrections handled separately.","pith_inferences":["Editorial inference: the same pattern -- linear Lagrangians mirroring powers of an ample line bundle, with products counted by triangles -- should extend to any smooth curve in an abelian variety; the honeycomb tropicalization of the corresponding theta function would determine the mirror quotient and the triangle weights.","Editorial inference: the equality between theta-function multiplication and the $\\mu_2$ count gives a testable numerical prediction: for low powers $L$, $L^2$, $L^3$ on $V$, the coefficients of the product of theta functions should reproduce the weighted triangle sums term by term as $\\tau\\to0$.","Editorial inference: if the unpublished localization results used for Definition 4.6 are replaced by an explicit construction, the cohomological Fukaya-Seidel category of $(Y,v_0)$ may admit a description as a wrapped-type Fukaya category on the smooth part of the fibration, which would make the right vertical embedding computable by standard wrapped-Floer methods."],"forward_implications":["If the embeddings of Theorem 1.2 hold, the canonical ring $\\bigoplus_{i\\ge0}H^0(\\Sigma_2,L^i)$ is isomorphic to the cohomological endomorphism ring of the mirror Lagrangians, so the projective embeddings of $\\Sigma_2$ are determined by Lagrangian intersection data.","The product in that ring is computable, in principle, by counting holomorphic triangles in $(Y,v_0)$; the paper writes the $\\mu_2$ coefficient as a sum of weights $\\tau^{-(l/l'l'')\\kappa(\\cdots)}$ over lattice elements $\\gamma_A$.","For the abelian surface $V$, the embedding $D^b_{\\mathrm{LCoh}}(V)\\hookrightarrow H^0\\mathrm{Fuk}(V^\\vee)$ gives a direct verification in this example that multiplication of sections of a line bundle matches the triangle count in the mirror torus.","Because the line bundles and their shifts generate the subcategories considered, an $A_\\infty$ enhancement would extend these cohomological embeddings to the whole derived category $D^b\\mathrm{Coh}(H)$, a step the paper identifies as the natural next one."],"supporting_citations":[{"why":"Supplies the generalized SYZ mirror construction for hypersurfaces in toric varieties and the Lagrangian torus fibration on the blow-up that underlies $(Y,v_0)$.","marker":"[AAK16]"},{"why":"Introduces mirror pairs as dual torus fibrations (SYZ), the geometric motivation for taking $V^\\vee$ as the mirror abelian variety.","marker":"[SYZ96]"},{"why":"Formulates homological mirror symmetry, the equivalence whose cohomological version is proved here.","marker":"[Kon95]"},{"why":"Provides the Lefschetz-fibration Fukaya category framework -- strip-like ends, moduli spaces, regularity -- adapted to the non-Lefschetz fibration.","marker":"[Sei08]"},{"why":"Unpublished work cited for the categorical localization defining morphisms in the Fukaya-Seidel category and for the roof composition lemmas.","marker":"[AS]"},{"why":"Introduces the U-shaped non-compact Lagrangians used to define the objects $L_k$ in the fibration.","marker":"[AS19]"},{"why":"Supplies the weighted count of holomorphic discs in toric varieties that identifies the differential with the theta function.","marker":"[CO06]"},{"why":"Gives the open mirror theorem for infinite-type toric Calabi-Yau manifolds used to count disc-with-sphere configurations.","marker":"[KL19]"},{"why":"Proves the reduction of open to closed Gromov-Witten counts by adding a ray to the fan, used in the sphere-bubble correction.","marker":"[Cha11]"},{"why":"Provides the localized-category propositions that the paper cites for the equality of directed and localized morphisms.","marker":"[Gan16]"}],"fun_headline_variants":["Mirror symmetry on cohomology for genus 2 curve","Genus 2 curve: line bundles match mirror triangle counts","Cohomological mirror symmetry proven for genus 2 curve","Fukaya-Seidel vs derived category: genus 2 curve match","Triangle counts reproduce canonical ring of genus 2 curve"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the unpublished categorical-localization technology for defining morphisms in a Fukaya-Seidel category works for this non-exact symplectic fibration with compact torus fibers and sphere bubbles; if it does not, the category $H^0\\mathrm{FS}(Y,v_0)$ used in the main theorem is not defined.","fun_headline_variants_meta":{"raw":{"variants":["Mirror symmetry on cohomology for genus 2 curve","Genus 2 curve: line bundles match mirror triangle counts","Cohomological mirror symmetry proven for genus 2 curve","Fukaya-Seidel vs derived category: genus 2 curve match","Triangle counts reproduce canonical ring of genus 2 curve"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000374,"raw_usage":{"total_tokens":2172,"prompt_tokens":1296,"completion_tokens":876,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":912,"completion_tokens_details":{"reasoning_tokens":791}},"tokens_in":912,"tokens_out":876,"duration_ms":8384,"temperature":1.0,"reasoning_tokens":791,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:51:09.425198+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the morphism space between $L_k$ and its upward pushoff in the localized category of Definition 4.6 by both the directed bigon count and by roofs; if the two answers differ for some regular almost complex structure, Lemma 4.8 fails and the right vertical embedding has no well-defined target.","supporting_citations":[],"review_version":1}