{"id":"16102afd-338b-4108-88a2-505942b85b39","arxiv_id":"2506.10593","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A closed-form virtual count of maps from a fixed genus-g curve to a degree-ℓ hypersurface in a Grassmannian, with Schubert incidence conditions, expressed via the Vafa-Intriligator formula, plus a large-degree enumerativity theorem.","lead":"Using a standard compactification called the Quot scheme, the authors find a closed-form formula for the virtual number of maps from a fixed curve to a hypersurface inside a space of planes, with prescribed point conditions. The result extends a classical counting formula for Grassmannians, and under a dimension hypothesis it becomes a true count of maps for large degree.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The central virtual count (12) depends on the imported virtual-class compatibility (9)/Proposition 2, which is not re-proven in the paper; an unchecked sign or excess term in that pushforward would invalidate the main formula.","rationale":"The reader's weakest_assumption lists Conjecture 1 first and the virtual-class compatibility (9) second. I elevate (9) because it is the load-bearing input for the paper's central unconditional claim, Theorem 1. If (9) is wrong, the virtual count itself is wrong; Conjecture 1 only affects the conditional enumerative interpretation, which the paper explicitly states as open. I checked the internal computation in Lemma 1 and the use of Proposition 3 from [MO]; the signs in the Chern character computation are consistent with the convention δ_iδ_{i+g}=η and with the anticommutation of the odd b-classes, and the summation yielding (n−ℓ)^g/n^g is correct. The Vafa–Intriligator formula (11) also reproduces known genus-zero and genus-one checks. Thus no flaw was found in the algebraic core of the proof. The remaining source of uncertainty is the imported compatibility (9), which is cited to [CKM] without re-proof. Given that the paper is otherwise careful and the cited result is standard in the field, this concern warrants keeping the reviewer's CONDITIONAL verdict rather than rejecting. The proposed localization test is a concrete way to settle whether the sign and excess-term structure of (9) are correct in a non-trivial example.","tokens_in":13753,"tokens_out":53979,"duration_ms":594838,"concrete_test":"Verify (9) in a small non-smooth case by torus localization: take C = P^1 (g = 0), G(2,4), ℓ = 1, d = 2, and compute both ι_*[Quot_d(C,X_1)]^vir and c_top(E_1) ∩ [Quot_d(C,G(2,4))]^vir at the torus-fixed points of the Quot scheme. If the two classes agree as Chow classes, Proposition 2 passes this check; if they differ by a sign or an extra factor, (9) is false and Theorem 1's formula (12) fails.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Theorem 1's formula (12) is obtained by pushing forward the virtual class of Quot_d(C,X_l) via the compatibility (9): ι_*[Quot_d(C,X_l)]^vir = c_top(E_l) ∩ [Quot_d(C,G(r,n))]^vir. This identity is imported from [CKM, Prop. 6.2.2] and is not re-derived. The paper's Proposition 1 only sketches the existence of a perfect obstruction theory for Quot_d(C,X_l) using the GIT/quasimap formalism; it does not prove (9). Since Quot_d(C,G(r,n)) is singular in general, the zero-locus formula for a section of E_l is not automatic: one must show that the obstruction theory of Quot_d(C,X_l) is exactly the cone of the section morphism E_l^∨[-1] → L_{Quot_d(C,G)}, and that the section is 'perfectly obstructed' relative to the ambient virtual class. A sign error in the exponential term of Lemma 1 or in the triangle defining the obstruction theory would change the factor (n−ℓ)^g/n^g in (12) to something like (n+ℓ)^g/n^g, altering all the paper's virtual counts. The assumption dℓ > 2g−2 ensures E_l is locally free and R^1π_*(det E∨)^⊗ℓ = 0, but it does not by itself guarantee the virtual pullback formula when the ambient Quot scheme is singular. Because the central claim of the paper is the virtual count (Theorem 1), this imported compatibility is the most load-bearing unverified input.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper computes virtual counts of degree-d maps from a fixed smooth projective curve C of genus g to a degree-ℓ hypersurface X_ℓ in a Grassmannian G(r,n), with pointwise incidence conditions to special Schubert subvarieties, using the Quot scheme compactification. The main result, Theorem 1, expresses the virtual intersection numbers on Quot_d(C,X_ℓ) as a Vafa-Intriligator integral on the ambient Quot_d(C,G(r,n)) times an explicit factor (n−ℓ)^g ℓ^{dℓ−g+1}/n^g. The proof combines a virtual-class compatibility (Proposition 2, imported from [CKM]) with an explicit computation of the top Chern class of the pushforward bundle E_ℓ (Lemma 1) and known b-class integrals over the Quot scheme (Proposition 3). The paper then discusses asymptotic enumerativity: Theorem 2, conditional on the expected-dimension hypothesis for Mor_d(C,X_ℓ), shows that the virtual counts are actual counts for large d and for incidence conditions of codimension < n−ℓ. A complete-intersection generalization (Theorem 3) and a projective-space specialization (Corollaries 1 and 2) are also given.","tokens_in":13971,"tokens_out":19739,"duration_ms":226215,"significance":"If the technical hypotheses are satisfied, the paper gives a short, explicit, and parameter-free formula for virtual counts of maps to hypersurfaces in a broad class of targets, extending the Vafa-Intriligator calculus from Grassmannians to their hypersurfaces. The derivation of Theorem 1 is transparent and checkable: Lemma 1 computes c_top(E_ℓ) via Grothendieck-Riemann-Roch, Proposition 3 supplies the required b-class integrals, and the final binomial sum is closed form. The conditional enumerativity theorem connects the virtual counts to classical enumerative geometry and to recent work on Tevelev degrees and weak convexity, with a clearly stated open conjecture. The manuscript is honest about the conditional nature of the enumerative half and about the imported virtual-class compatibility (9).","major_comments":[{"comment":"The main formula (12) depends entirely on the virtual-class compatibility ι_*[Quot_d(C,X_ℓ)]^{vir} = c_top(E_ℓ) ∩ [Quot_d(C,G(r,n))]^{vir}. This identity is imported from [CKM, Prop. 6.2.2] and is not re-proven in the paper. Since Quot_d(C,G(r,n)) is singular for d>0, the zero-locus formula is not automatic: one must know that the perfect obstruction theory of Quot_d(C,X_ℓ) is exactly the cone of the section E_ℓ^∨[−1] → L_{Quot_d(C,G)}, with no excess or sign corrections. The paper should either state precisely the hypotheses of [CKM, Prop. 6.2.2] under which it applies to this fixed-domain Quot-scheme setting, or give an outline of the verification (for instance via Manolache's virtual pullback, which is only mentioned in Remark 7). As it stands, Proposition 2 is a load-bearing unproven input; if it fails, all formulas in Theorems 1 and 3 and in the corollaries would change.","section":"Section 2, Proposition 2 and Eq. (9)"},{"comment":"The computation of c_top(E_ℓ) in Lemma 1 is central to Theorem 1, and it depends on a sign convention in the Künneth decomposition (14). The paper says only that {1, δ_1, …, δ_{2g}, η} is a symplectic basis, but does not specify whether δ_jδ_{j+g} = η or δ_jδ_{j+g} = −η. With the first convention the Grothendieck-Riemann-Roch calculation in the proof of Lemma 1 appears to produce (ℓa_1)^{dℓ−g+1} e^{+ℓφ/a_1} rather than e^{−ℓφ/a_1}, which would change the final factor in (12) from (n−ℓ)^g/n^g to (n+ℓ)^g/n^g. The sign convention should be fixed and the expansion leading to Eq. (15) checked against it. This is not a cosmetic issue, because the exponent of the exponential determines the genus-dependent factor in all the paper's counts.","section":"Lemma 1, Eq. (15) and the symplectic basis convention"}],"minor_comments":[{"comment":"The step writing ∫_{[Quot_d(C,X_ℓ)]^{vir}} P = ∫_{[Quot_d(C,G)]^{vir}} c_top(E_ℓ) · P implicitly uses the projection formula for the inclusion ι; this should be stated, since P is a pullback of a class on the ambient Quot scheme.","section":"Proof of Theorem 1, after Eq. (17)"},{"comment":"In the GRR calculation, the phrase 'A careful calculation further yields the total Chern class' omits the intermediate Newton-identity step; adding one line showing that c_t(E_ℓ) = (1+tℓa_1)^{dℓ−g+1} exp(−tℓ^2φ/(1+tℓa_1)) from the displayed Chern character would make the proof fully self-contained.","section":"Lemma 1, proof"},{"comment":"The inclusion (24) and the subsequent dimension estimate are stated succinctly; a few more words explaining why the dimension of the first alternative in (23) (the case where the torsion support contains some p_k) is negligible would help the reader follow the boundary analysis.","section":"Section 4, proof of Theorem 2"},{"comment":"The comparison with Tevelev degrees is interesting but the notation Tev_{g,n,β} is introduced without a precise definition of n; the authors should clarify the role of the integer n in that discussion.","section":"Remark 4"},{"comment":"In the algebraic proof of Proposition 4, the identity e_i(ζ_1,…,ζ_r) = h_i(−ζ_{r+1},…,−ζ_n) is used for i ≤ r; it would be helpful to indicate which elementary identity for the n-th roots justifies it, since the reader must otherwise reconstruct the relation from ∏(1+ζ_k t) = 1+t^n.","section":"Section 3.2, Proposition 4"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is well written and the main virtual-count formula is likely correct, but the referee report requests that the authors either prove or precisely verify the imported virtual-class compatibility (9) in this singular Quot-scheme setting, and that they fix the sign convention in Lemma 1. These are fixable within the manuscript's scope, so the paper should not be rejected, but they are load-bearing. I would be comfortable with acceptance after a revision that addresses these two points."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's the short version: the paper's central formula (12) is a genuine new result, the derivation is explicit and checkable, and the paper is honest that the enumerative interpretation depends on an open conjecture. The soft spots are real but not fatal: the virtual-class compatibility is imported, and Section 4 is a sketch. I'd send it to review.\n\nWhat's new: nobody had computed virtual counts for maps from a fixed curve to a hypersurface in a Grassmannian with Schubert incidence conditions. The formula interpolates between the Vafa-Intriligator counts for the Grassmannian and the Tevelev-degree counts for projective hypersurfaces. The proof is short and concrete: Lemma 1 computes the top Chern class of the pushed-forward line bundle via GRR; Proposition 3 from Marian-Oprea handles the b-class integrals; and the rest is bookkeeping. The special cases check out—projective space, LG(2,4), and the isotropic Quot scheme of Sinha—which is strong evidence the signs and constants are right.\n\nThe main soft spot is the imported compatibility (9), from [CKM, Prop. 6.2.2]. It's not re-proven, and a sign error there would shift the factor. But it's a cited published result, the paper states the hypotheses (dℓ > 2g−2) under which it holds, and the consistency checks make a systematic sign error unlikely. Not a reason to reject.\n\nThe bigger limitation is that the enumerative half, Theorem 2, is conditional on Conjecture 1 (expected dimension for general hypersurfaces), which is open. The paper says this clearly. The proof of Theorem 2 is also compressed—the large-m boundary case is asserted rather than derived. I'd like more detail there, but it's a minor issue for the virtual count.\n\nWho this is for: anyone working on Quot scheme invariants, Vafa-Intriligator formulas, or fixed-domain curve counts. It's a short note that makes a clean point. Worth reading, worth citing for the virtual formula. Yes, send it to a serious referee.","headline":"A clean, parameter-free virtual count for maps to hypersurfaces in Grassmannians, with a conditional enumerative half that is honestly flagged.","tokens_in":14696,"tokens_out":2306,"would_cite":true,"duration_ms":24086,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14N35","14C17","14H60","14M15"],"pacs":[],"model":"deepseek-v4-flash","headline":"For a fixed curve and a hypersurface in a Grassmannian, the virtual number of degree-d maps with Schubert incidence is a universal prefactor times a Vafa-Intriligator integral over the ambient Quot scheme.","keywords":["Quot scheme compactification","hypersurfaces in Grassmannians","Vafa-Intriligator formula","virtual fundamental class","special Schubert subvarieties","enumerative geometry","maps from curves","complete intersections"],"falsifier":"Evaluate formula (12) in a case where both sides are accessible independently, for instance $C=\\mathbb{P}^1$, $G(2,4)$, $\\ell=1$ with $X_1=LG(2,4)$: the right side is a finite Vafa-Intriligator root-of-unity sum, and a direct count of degree-$d$ rational curves with the prescribed Schubert incidences must match it, or the virtual identity fails. On the enumerative side, a single hypersurface for which $\\operatorname{Mor}_d(C,X_\\ell)$ has dimension larger than expected for infinitely many $d$ would refute Theorem 2's conclusion.","tokens_in":13388,"feed_emoji":"📐","tokens_out":17029,"duration_ms":146707,"temperature":0.7,"pith_summary":"This paper gives a closed formula for counting degree-$d$ maps from a fixed genus-$g$ curve to a hypersurface $X_\\ell$ in a Grassmannian $G(r,n)$, with prescribed points of the curve sent to special Schubert subvarieties of the hypersurface. Using a Quot scheme compactification of the map space, the paper shows that the virtual count equals $((n-\\ell)^g \\ell^{d\\ell-g+1}/n^g)$ times a Vafa-Intriligator integral over the ambient Grassmannian Quot scheme, so the count is a finite sum over $r$-tuples of $n$-th roots of unity. The same mechanism handles complete intersections and gives explicit numbers such as $\\ell^{d\\ell-g+1}(r+1-\\ell)^g$ for hypersurfaces in projective space. The paper then studies when this virtual count is an honest count: if the space of maps has expected dimension for all sufficiently large $d$, the virtual number counts actual maps with incidence to Schubert cycles of codimension less than $n-\\ell$. A reader should care because the enumerative interpretation of the formula rests entirely on one expected-dimension conjecture, which the paper isolates explicitly.","feed_headline":"Map counts to Grassmannian hypersurfaces collapse to one integral","feed_subtitle":"A closed formula gives the virtual count; a dimension conjecture upgrades it to an honest count.","key_machinery":"The central object is the Quot scheme $\\operatorname{Quot}_d(C,G(r,n))$, which parametrizes rank-$r$ degree-$(-d)$ subsheaves of the trivial rank-$n$ bundle on $C$; the hypersurface Quot scheme $\\operatorname{Quot}_d(C,X_\\ell)$ is its zero locus under a section of the vector bundle $E_\\ell = \\pi_*(\\det E^{\\vee})^{\\otimes \\ell}$, which has rank $d\\ell-g+1$ when $d\\ell>2g-2$. The load-bearing identity is the virtual class compatibility (9), imported from [CKM, Proposition 6.2.2]: $\\iota_*[\\operatorname{Quot}_d(C,X_\\ell)]^{\\mathrm{vir}} = c_{\\mathrm{top}}(E_\\ell) \\cap [\\operatorname{Quot}_d(C,G(r,n))]^{\\mathrm{vir}}$. Lemma 1 evaluates $c_{\\mathrm{top}}(E_\\ell)$ explicitly in terms of the first Chern class $a_1$ and the pairing $\\phi = \\sum_{j=1}^g b^j_1 b^{j+g}_1$, and Proposition 3 evaluates pairings involving the $b$-classes as multiples of $a_1^s/n^s$. The Vafa-Intriligator formula, a finite sum over $r$-tuples of distinct $n$-th roots of unity, closes the computation by evaluating the remaining $a$-class intersections on the ambient Quot scheme.","core_discovery":"The paper's central claim is Theorem 1: for $d\\ell > 2g-2$ and any monomial $P=\\prod_{k=1}^t a_{i_k}$ of weighted degree $e_\\ell=e-(d\\ell-g+1)$, the virtual intersection on the hypersurface Quot scheme equals $$\\int_{[\\operatorname{Quot}_d(C,X_\\ell)]^{\\mathrm{vir}}} P = \\frac{(n-\\ell)^g\\,\\$ell^{{d\\ell-g+1}}$}{n^g} \\int_{[\\operatorname{Quot}_d(C,G(r,n))]^{\\mathrm{vir}}} $a_1^{{d\\ell-g+1}}$ P.$$ The right-hand integral is evaluated in closed form by the Vafa-Intriligator formula. The mechanism is the virtual class compatibility $\\iota_*[\\operatorname{Quot}_d(C,X_\\ell)]^{\\mathrm{vir}} = c_{\\mathrm{top}}(E_\\ell) \\cap [\\operatorname{Quot}_d(C,G(r,n))]^{\\mathrm{vir}}$, together with an explicit computation of $c_{\\mathrm{top}}(E_\\ell) = (\\ell a_1)^{d\\ell-g+1} e^{-\\ell\\phi/a_1}$ and a known rule that replaces powers of $\\phi$ by powers of $a_1/n$. Theorem 2 then states that, under the expected-dimension conjecture for $\\operatorname{Mor}_d(C,X_\\ell)$, the virtual count is enumerative for all sufficiently large $d$ whenever each incidence class has codimension $i_k < n-\\ell$; a complete-intersection analogue is given in Theorem 3.","pith_inferences":["The prefactor $(n-\\ell)^g/n^g$ looks like a universal correction for replacing the ambient Grassmannian by a single hypersurface in a determinant representation; a natural test is whether the same prefactor shape governs hypersurfaces in other homogeneous spaces.","The codimension bound $i_k < n-\\ell$ in Theorem 2 is probably not technical: the paper's own Remark 8 exhibits a comparison with Tevelev degrees where larger incidences pick up an extra $(\\ell^\\ell/\\ell!)^t$ factor from boundary contributions, so enumerativity should fail outside that range.","Any proof of Conjecture 1 would immediately promote the closed virtual formula to honest counts for every genus, since the virtual class compatibility already holds in the entire range $d\\ell>2g-2$."],"forward_implications":["For any hypersurface of degree $\\ell$ and any fixed curve of genus $g$, the virtual count is computable in closed form once $d\\ell>2g-2$; no further intersection theory is needed.","For a hypersurface $X_\\ell \\subset \\mathbb{P}^r$, the formula gives $\\ell^{d\\ell-g+1}(r+1-\\ell)^g$ virtual maps meeting general codimension-$i_k$ planes, and Corollary 1 makes this an actual count under the expected-dimension hypothesis.","For the Lagrangian Grassmannian $LG(2,4)\\subset G(2,4)$, the virtual intersections reduce to $2^{2d-m_2-g+1}\\,3^g$ with $m_1+2m_2=3(d-g+1)$.","If Conjecture 1 holds, then for all sufficiently large $d$ the virtual count of Theorem 1 counts actual maps with incidence to general Schubert varieties of type $c_{i_k}(S^\\vee)$ with $i_k