{"id":"63af9651-a1fe-4aad-91b7-4b739bacd2fa","arxiv_id":"2412.11256","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The KSBA compactifications of four maximal order-3 and one order-4 K3 moduli spaces are explicitly described as semitoroidal compactifications via ADE root lattices.","lead":"This paper computes Baily-Borel, toroidal, and KSBA stable-pair compactifications of moduli spaces of K3 surfaces with order-3 and order-4 nonsymplectic automorphisms. The main results identify the KSBA compactifications as semitoroidal compactifications governed by explicit ADE root lattices, including all four maximal order-3 families and one order-4 family.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The order-4 theorem rests on an unproved smoothness/K-triviality assertion in Proposition 4.8; without it, the KSBA semifan A1^2 is not established.","rationale":"The reader's weakest_assumption focused on Proposition 6.14 for the order-3 cases. That concern is real but the proof of Proposition 6.14 is supported by Proposition 6.11 and a boundary-dominance argument, and I did not find a concrete flaw there. The order-3 arguments appear structurally sound: dominance on the boundary implies dominance on the interior when the source is irreducible and the boundary maps dominantly to a boundary divisor, and the Torelli arguments for anti-canonical pairs justify the contraction calculations. The clearest concrete gap in the manuscript is Proposition 4.8, where the central geometric input for Theorem 4.10 is explicitly delegated to the reader. This is load-bearing for the order-4 theorem, but it is a verification gap rather than a demonstrated error, so the appropriate verdict remains conditional rather than rejection.","tokens_in":49127,"tokens_out":33983,"duration_ms":325843,"concrete_test":"Independently verify Proposition 4.8 by a local computation: write the cyclic degree-4 cover in local coordinates over the double curve and over the A1 singularities of the base, and compute the canonical class via the Hurwitz formula for X→~Y. Confirm that X is nonsingular and K_X∼C 0. If this fails, check whether a modified branch divisor smoothing yields a Kulikov model; if not, the order-4 theorem is not proven.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The most concrete load-bearing gap is Proposition 4.8, on which Theorem 4.10 (the order-4 case of Theorem 1.2) entirely rests. The proof constructs a purported Kulikov model X→Δ as a degree-4 cyclic cover of a degenerate threefold and then states: 'We leave it to the reader to verify that X is indeed smooth and K-trivial.' This is not a routine clarification: smoothness of a cyclic cover fails in general when the base has a non-normal central fiber or when the branch divisor is not transverse to the singular locus, and K-triviality is essential for the central fiber to be a genuine Type II Kulikov surface. Without these facts, the period point ψ∈AJ used in Theorem 4.10 is not the period of a Kulikov degeneration, and the conclusion that the KSBA semifan is exactly the A1^2 summand does not follow. The paper also asserts without proof that B⊂~Y can be smoothed; this is part of the same verification.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies moduli spaces of K3 surfaces carrying a purely non-symplectic automorphism of order 3 or 4 whose fixed locus contains a curve of genus at least 2. For the four maximal order-3 families with (n,k)=(0,2), (0,1), (1,1), (2,1) and for cyclic quadruple covers of P2 branched along a quartic, the authors identify the KSBA stable-pair compactification with an explicit semitoroidal compactification: the semifans are listed in Table 1 and Theorem 1.2 (FJ=A1^2 inside J-perp/J=D4^2+A1^2). The proof uses the general framework of [AEH24] to reduce KSBA compactifications to semitoroidal data, then computes the data via a new 'triple Tschirnhausen construction' relating degenerations of admissible triple covers to Type II Kulikov models, lattice computations, and the Torelli theorem for anti-canonical del Pezzo pairs.","tokens_in":49297,"tokens_out":10254,"duration_ms":83009,"significance":"If correct, this gives a complete and remarkably simple lattice-theoretic description of the boundary of these compactifications, and demonstrates that the KSBA boundary can be computed explicitly in nontrivial cases. The paper's main contribution is the identification of the exact semifan, not merely a lower bound, and the triple Tschirnhausen construction is an original and plausible computational tool. The lattice tables are detailed, and the logical architecture (reduction to boundary divisors, dominance via anti-canonical Torelli) is coherent. However, two load-bearing verifications are presented sketchily: the smoothness/K-triviality of the constructed order-4 Kulikov model in Proposition 4.8, and the generic finiteness/dominance statement for the order-3 case in Propositions 6.11 and 6.14. These gaps are localizable and appear fixable, but they should be completed before the theorems can be regarded as fully proved.","major_comments":[{"comment":"Proposition 4.8, on which Theorem 4.10 and hence Theorem 1.2 rest, constructs a Kulikov model as a cyclic degree-4 cover of a singular threefold and concludes 'We leave it to the reader to verify that X is indeed smooth and K-trivial.' This is a load-bearing assertion: if the total space is singular or has nontrivial canonical class, the central fiber is not a genuine Type II Kulikov degeneration, and the period point psi used in Theorem 4.10 is not a period of a Kulikov degeneration. The earlier sentence 'It is easy to see that B subset ~Y can be smoothed' is part of the same gap. Smoothness of cyclic covers can fail when the branch locus is not transverse to the singular locus of the base, and K-triviality is not automatic for such covers. Please provide a complete verification, for example by local coordinate analysis near the double locus and the blown-up fibers, or by a precise reference covering this specific degeneration.","section":"§4.1, Proposition 4.8"},{"comment":"The exactness of all four order-3 semifans in Table 1 depends on the dominance statement of Proposition 6.14: without it, the computations in Sections 7–10 can only show that the listed sublattice is contained in the KSBA semifan. The proof of Proposition 6.14 reduces to Proposition 6.11, whose generic-finiteness argument is compressed. In particular, the case d=9 is dismissed as 'automatic', and the finiteness for blow-ups in Z/3Z-orbits is justified by 'only finitely many embeddings A2 to Lambda'. Since a failure of generic finiteness would leave the equality FJ = ... unproved, I ask that this argument be expanded to a complete proof, or that the missing finiteness statement be stated and proved as a separate lemma.","section":"§6.3–6.4, Propositions 6.11 and 6.14"},{"comment":"Theorem 10.5 lists the fourth cusp as (J-perp/J)_root = E8 + A2 with FJ = <A2>^sat, but Section 10.1, the table in Theorem 5.5, and Table 1 in the introduction all give this cusp as E8 + A2^2, and the proof in §10.3.4 concludes that FJ is the saturation of the complementary A2^2 summand. Please correct the statement of Theorem 10.5 (and any parallel statement in Theorem 1.1 or Table 1 if needed) so that the theorem, the table, and the proof are mutually consistent.","section":"§10.3, Theorem 10.5"}],"minor_comments":[{"comment":"The first data row of Table 1 prints (n,k) as (0,0); it should be (0,2).","section":"Table 1"},{"comment":"There is a typo in the expression 'A1^2 tensor Z[i]]E'; the stray bracket should be removed.","section":"Proof of Theorem 4.10"},{"comment":"The notation R* for an index-3 overlattice is defined only after the theorem statement; it would be clearer to define it before or inside the statement.","section":"Section 5.3, Theorem 5.5"},{"comment":"The sentence 'The pairs that give a type II surface correspond to those with reducible Y' is imprecise because only generic points of the locus Z are discussed; please clarify the genericity assumption.","section":"Section 4.1, before Proposition 4.5"},{"comment":"In Theorem 8.4 the map Y_rho to the Baily-Borel compactification is asserted to be regular; the surrounding text would benefit from a sentence explaining why this map is regular rather than merely rational.","section":"Section 8.4"}],"recommendation":"major_revision","confidential_remarks":"The structural equivalence between KSBA and semitoroidal compactifications is imported from [AEH24], which is from the same group; the actual semifan computations in this paper are independent and substantial. An external check of Proposition 4.8 and of the dominance proof in Proposition 6.14 would materially strengthen the paper. The scope is a good fit for an algebraic geometry journal."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Read it. The order-3 part is the real payoff: four maximal families, explicit semifans in Table 1, and a clean mechanism (triple Tschirnhausen + MMP) explaining where each A2/E6/E8 summand goes. The lattice tables and cusp computations look careful, and Proposition 6.14's dominance argument is the right tool to turn lower bounds into exact KSBA semifans. I do not think the dependence on [AEH24] is a problem here: the framework comes from there, but the semifan computation is done in this paper.\n\nThe genuine soft spot is Proposition 4.8, exactly where the stress-test note points. The proof says \"We leave it to the reader to verify that X is indeed smooth and K-trivial.\" That is load-bearing for Theorem 4.10: without a genuine Type II Kulikov model, the period point in AJ is not known to come from a degeneration, and the claimed A1^2 semifan is not established. It is probably fixable—the construction is explicit and the components have the right dP2 geometry—but it has no business being left as an exercise. The final step of Theorem 4.10 (\"by a dimension count, FJ must have rank 2\") is also terse; you need the dominance statement to rule out a larger FJ, and that deserves one sentence with the right Proposition numbers.\n\nMinor things: the tables in Section 5 are dense and several embeddings into Niemeier lattices are only presented by their output, not by the intermediate discriminant-form checks; that is acceptable for a research paper but slow for a reader. The case-by-case MMP in Sections 7–10 is long, but the pattern is clear and the figures help.\n\nWho this is for: people working in moduli of K3 surfaces, KSBA, or ball quotients. It is a serious computation paper, not a speculative one. I would send it to a good referee. The required fix is small in spirit and big in verification: move the smoothness/K-triviality proof into the paper, or at least replace it with a precise lemma stated as Proposition 4.8.","headline":"Solid computation paper; order-3 part convincing, order-4 theorem needs the delegated smoothness/K-triviality check in Prop 4.8 actually done.","tokens_in":49820,"tokens_out":1895,"would_cite":true,"duration_ms":18213,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14J28","14J50","14D22"],"pacs":[],"model":"deepseek-v4-flash","headline":"For K3 moduli with order-3 and order-4 automorphisms, the compactifying boundary is explicitly identified as a semifan of ADE root lattices.","keywords":["K3 surfaces","nonsymplectic automorphism","KSBA compactification","semitoroidal compactification","Kulikov degenerations","ADE root lattices","triple Tschirnhausen construction","quadruple plane quartics"],"falsifier":"Choose the cusp of the $(0,1)$ family with root lattice $E_8\\oplus E_6\\oplus A_2$ and compute the KSBA stable limit directly for two one-parameter families whose periods differ only in the $E_6$-summand; the paper's argument predicts these limits are distinct, by the Torelli theorem for anti-canonical pairs, so if direct calculation produced isomorphic stable pairs the claimed semifan would be a strict lower bound and the dominance step in Proposition 6.14 would fail.","tokens_in":48918,"feed_emoji":"📐","tokens_out":12297,"duration_ms":104181,"temperature":0.7,"pith_summary":"This paper establishes what the boundary looks like when moduli spaces of K3 surfaces equipped with a nonsymplectic automorphism—one that multiplies the holomorphic 2-form by a nontrivial root of unity—are compactified by adding stable pairs. For the four maximal families with an order-3 automorphism whose fixed locus contains a curve of genus at least 2, and for cyclic degree-4 covers of the plane branched along a quartic, the KSBA compactification is identified as a semitoroidal compactification. In these settings the distinction among Baily–Borel, toroidal, and KSBA boundaries is controlled by a semifan: for each boundary cusp $J$, a primitive sublattice $F_J \\subset J^\\perp_{T_\\rho}/J$ whose translates are contracted when passing from the fine toroidal boundary to the coarse stable-pair boundary. The main theorems list these sublattices explicitly, in terms of $A_1$, $A_2$, $E_6$, and $E_8$ root lattices. The upshot is that the geometric question of which K3 degenerations are identified at the boundary becomes a concrete lattice computation.","feed_headline":"K3 moduli boundaries identified as semifans of ADE lattices","feed_subtitle":"Four maximal order-3 families and the degree-4 quartic-cover case now have fully explicit KSBA boundaries.","key_machinery":"The load-bearing mechanism is the triple Tschirnhausen construction: starting from an admissible triple cover $C\\to\\mathbb{P}^1$ with marked points, one blows up the threefold $\\mathbb{P}E$ at the marked fibers, takes a normalized cyclic triple cover branched over the proper transforms, and blows down the resulting $(-1)$-curves; the output is a Kulikov degeneration of K3 surfaces with an order-3 automorphism. This construction is shown to dominate the boundary of the period domain, so every boundary period is realized by such a degeneration. The other half of the machinery is the lattice of numerically Cartier divisors of a Type II Kulikov surface, which is identified with $J^\\perp/J$ and, in the $\\rho$-equivariant setting, with $J^\\perp_{T_\\rho}/J$; the period of the degeneration lives in $J^\\perp_{T_\\rho}/J \\otimes_{\\mathbb{Z}[\\zeta_3]} E$. Running the minimal model program on the divisor $R$, the fixed curve of the automorphism, produces the unique KSBA stable limit, and comparing which periods give the same stable limit yields exactly the sublattices $F_J$ in Table 1. For order 4 the construction is simpler: an explicit degeneration of $\\mathbb{P}^2$ to two quadric cones, pulled back by the cyclic quadruple cover, gives the Kulikov surface, and the same contraction argument gives $F_J = A_1^{\\oplus2}$.","core_discovery":"The central claim, stated as Theorem 1.1, is that for the four maximal order-3 families the KSBA compactification $F^{\\mathrm{KSBA}}_\\rho$ is the semitoroidal compactification of the ball quotient $D_\\rho/\\Gamma_\\rho$ with the semifan $F_J$ given in Table 1, and the paper proves this by splitting it into Theorems 7.1, 8.2, 9.3, and 10.5. The ambient lattice at a cusp is $J^\\perp_{T_\\rho}/J$, and the KSBA semifan is the saturation, or primitive closure, of the parenthesized $A_2$-summands, or zero where no summand is parenthesized. For $(n,k)=(0,2)$ this gives $F_J=0$, so the KSBA boundary is the toroidal boundary; for $(0,1)$, $(1,1)$, and $(2,1)$ the contracted directions are exactly the listed $A_2$ directions. Theorem 1.2 covers the order-4 case: the space of cyclic quadruple covers of $\\mathbb{P}^2$ branched at a quartic has a unique cusp with $J^\\perp_{T_\\rho}/J=D_4^{\\oplus2}\\oplus A_1^{\\oplus2}$, and the KSBA semifan is the $A_1^{\\oplus2}$ summand, so the boundary contracts precisely the translates of $A_1^{\\oplus2}\\otimes_{\\mathbb{Z}[i]}E$.","pith_inferences":["An extension of the paper's method would restrict Table 1 to the remaining, non-maximal order-3 families; since those families arise from the four maximal ones by specialization, one expects their KSBA semifans to be governed by the same $A_2$-coded contractions, though the paper does not carry out this restriction.","The pattern that the contracted directions are exactly the $A_2$-summands created by blowing up orbits of fixed points suggests a broader principle: for other cyclic automorphisms, the KSBA semifan may be determined by the root summands coming from such blow-ups, while the $E_6$ and $E_8$ summands inherited from del Pezzo surfaces remain uncontracted.","A testable extension is to run the same Tschirnhausen-based stable-model computation on an order-3 family with $g\\ge 2$ that is not maximal, and check whether the predicted restriction of Table 1 reproduces the actual contractions; a mismatch would pinpoint where the dominance of the boundary period map breaks down.","The order-4 computation is explicitly ad hoc, so a natural next step is to adapt the triple-cover-style construction to cyclic quadruple covers branched along curves other than quartics, to see whether the $A_1^{\\oplus2}$ contraction pattern persists or is special to this family."],"forward_implications":["For each of the four maximal order-3 families, the KSBA compactification and its boundary strata are now explicitly computable from Table 1; the boundary points are stable pairs whose isomorphism type is determined by the period restricted to the non-contracted lattice directions.","In the $(n,k)=(0,2)$ case the KSBA and toroidal compactifications coincide, so the geometric boundary retains all of the period information of the toroidal boundary.","In the order-4 case, the KSBA boundary over the unique cusp is a finite quotient of an abelian variety with the translates of $A_1^{\\oplus2}\\otimes_{\\mathbb{Z}[i]}E$ contracted; this gives a complete description of the KSBA compactification of the moduli space of cyclic quadruple covers of $\\mathbb{P}^2$ branched at a quartic.","For the $(0,1)$ family, the quotient by the automorphism identifies the KSBA compactification with the compactification of stable log quadrics, so the boundary strata of the K3 moduli space match the geometrically defined boundary strata of that log-surface moduli space."],"supporting_citations":[{"why":"Supplies the theorem that the KSBA compactification is a semitoroidal compactification and the extended period map; this is the framework in which the semifan $F_J$ is defined.","marker":"[AEH24]"},{"why":"Classifies the period lattices $T(n,k)$ and fixed lattices $S(n,k)$ of order-3 nonsymplectic automorphisms; Table 2 of the paper is built on this classification.","marker":"[AS08]"},{"why":"Provides the description of periods of Type II Kulikov surfaces and the isometry $\\Lambda\\simeq J^\\perp/J$ used to identify the boundary period domain with $J^\\perp_{T_\\rho}/J\\otimes_{\\mathbb{Z}[\\zeta_3]}E$.","marker":"[AE23]"},{"why":"Gives the original construction of periods for Type II K3 degenerations, on which the Kulikov-surface period computations rest.","marker":"[Kon85]"},{"why":"Supplies the Torelli theorem for rational surfaces with an anti-canonical cycle used in Proposition 6.12; this is what makes the dominance of the period map and the identification of the contracted sublattices work.","marker":"[Loo81]"},{"why":"Describes the KSBA compactification of plane quartics in the order-4 case; Proposition 4.5 and the identification of $Y_\\rho$ rest on it.","marker":"[Hac04]"},{"why":"Determines the lattices $S_\\rho$ and $T_\\rho$ for the cyclic quadruple cover of the plane, giving the ambient lattice $D_4^{\\oplus2}\\oplus A_1^{\\oplus2}$.","marker":"[Kon00]"},{"why":"Shows the Baily–Borel compactification for the order-4 family has a unique cusp, locating the single boundary point used in Theorem 4.10.","marker":"[Art09]"}],"fun_headline_variants":["KSBA boundaries as ADE semifans in K3 moduli","K3 moduli compactifications via root lattices","Order-3 and 4 K3 families get explicit KSBA","Semifan descriptions of K3 moduli boundaries","ADE semifans for K3 automorphism boundaries"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that every one-parameter degeneration near a cusp is produced by the triple Tschirnhausen construction for order 3, or by the stated degeneration model for order 4; if any degeneration is missed, the listed semifans would only be lower bounds on the true contractions.","fun_headline_variants_meta":{"raw":{"variants":["KSBA boundaries as ADE semifans in K3 moduli","K3 moduli compactifications via root lattices","Order-3 and 4 K3 families get explicit KSBA","Semifan descriptions of K3 moduli boundaries","ADE semifans for K3 automorphism boundaries"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000204,"raw_usage":{"total_tokens":1396,"prompt_tokens":959,"completion_tokens":437,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":575,"completion_tokens_details":{"reasoning_tokens":356}},"tokens_in":575,"tokens_out":437,"duration_ms":4536,"temperature":1.0,"reasoning_tokens":356,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T15:08:48.433200+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Choose the cusp of the $(0,1)$ family with root lattice $E_8\\oplus E_6\\oplus A_2$ and compute the KSBA stable limit directly for two one-parameter families whose periods differ only in the $E_6$-summand; the paper's argument predicts these limits are distinct, by the Torelli theorem for anti-canonical pairs, so if direct calculation produced isomorphic stable pairs the claimed semifan would be a strict lower bound and the dominance step in Proposition 6.14 would fail.","supporting_citations":[],"review_version":1}