{"id":"76268ae3-067d-4711-ad3a-3eb9e4c54bc7","arxiv_id":"2607.07528","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":8.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":0,"one_line_summary":"A higher-dimensional Kummer construction associates K3^[3]-type hyper-Kähler sixfolds to Kum^3-type sixfolds, with applications to motives, derived categories, and algebraic cycle conjectures.","lead":"The paper constructs a map from any 6-dimensional hyper-Kähler manifold of generalized Kummer type to one of K3^[3]-type, analogous to the classical Kummer construction of K3 surfaces from abelian surfaces. It proves lattice-theoretic characterizations, McKay correspondences for derived categories and motives, and applications including Beauville's weak splitting conjecture and Kimura-O'Sullivan finite-dimensionality for new K3 surfaces.","discovery_kind":"unclear","skeptic_critique":{"model":"glm-5.2","headline":"The deformation-to-K_3(A) strategy is sound, but the crepant resolution being hyper-Kähler for non-projective K is the least-verified link.","rationale":"The reader correctly identified the deformation argument as the weakest link. The concern is real but, on closer inspection, likely does not land as a genuine gap: the local singularity type of K/G is controlled by the symplectic linear algebra of the G-action on tangent spaces, which is deformation-invariant because G is defined via its trivial action on H^2 and H^4 (making the representation on cohomology, and hence on tangent spaces via the period map, constant in families). The paper's method of deforming to K_3(A) and using the topological nature of the statements is standard and well-established in the hyper-Kähler literature (cf. similar arguments in Huybrechts, Markman, Mongardi). The foundational construction from [Flo24] is used as a black box, but the present paper provides substantial independent verification through the lattice computations of Section 3, the Torelli-type arguments of Section 4, and the explicit examples of Section 8. The cohomological computations (Propositions 3.2–3.4) are verified at K_3(A) with complete explicit calculations (Lemmas 3.11–3.25) and transported via deformation, which is appropriate since these are statements about integral Hodge structures. The downstream results (Theorems B, D, J, L) follow logically from the construction and the cohomological analysis. The paper makes genuine new contributions: the birational characterization (Theorem B), the McKay correspondence (Theorem D), the reconstruction theorem (G), the comparison with MRS double covers (Theorems H, I), and the applications to Beauville's weak splitting conjecture (Theorem J) and Kimima-O'Sullivan finite-dimensionality (Theorem L) are all substantial. The construction recipe of Section 7 is acknowledged as incomplete at Step 2, but this is clearly stated as a limitation, not presented as a proven result. No circularity issues are present. The verdict of ACCEPT with HIGH confidence is appropriate.","tokens_in":74582,"tokens_out":1235,"duration_ms":577565,"concrete_test":"Verify that the local analytic type of the singularities of K/G is determined by the G-representation on the tangent space T_pK at fixed points, and that this representation is deformation-invariant. Concretely: at a fixed point p ∈ W^σ ⊂ K, the tangent space T_pK decomposes as T_pW^σ ⊕ N_p, where N_p is the normal 2-plane. Check that the stabilizer subgroup ⟨σ⟩ acts on N_p as -1 (a symplectic reflection of order 2 in dimension 2), making the local singularity of K/G at p analytically isomorphic to (C^2/±1) × (C^2)^2 — a product of an A_1 surface singularity with a smooth factor. If this local model holds for all K of Kum^3-type (not just K_3(A)), then the blow-up is crepant and smooth everywhere, and Theorem 2.3 is secure. This can be checked by computing the character of the G-representation on H^2(K,Z) and using the fact that the representation on T_pK is determined by the Verbitsky/B","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's strategy for Propositions 2.4–2.7 and the cohomological computations of Section 3 is: (1) establish the result at the generalized Kummer point K_3(A) by explicit computation, (2) use deformation invariance of Aut_0 (HT13, Theorem 2.1) to transport the group G and its fixed locus structure to arbitrary Kum^3-type K, (3) verify the statement is topological/Hodge-theoretic so it deforms. This is a well-established method in hyper-Kähler geometry. The key load-bearing step is that the crepant resolution Y_K → K/G is hyper-Kähler for ALL K of Kum^3-type, including non-projective K. Theorem 2.3 cites [Flo24] for this. The proof that the blow-up of K/G along its singular locus yields a smooth hyper-Kähler manifold requires that the singularities of K/G are locally products of A_1-surface singularities (so the blow-up is smooth) and that the resulting manifold carries a symplectic form extending the one on the smooth locus. At K_3(A), this is verified explicitly (Proposition 2.15, Lemma 2.17). For general K, the paper argues via deformation: the whole configuration deforms with K (Remark 2.8), and since the statement is topological, it suffices to check at one point. The potential gap is whether the local analytic structure of the singularities of K/G is deformation-invariant in the way required. The fixed loci W^σ are trianalytic (Remark 2.8), so they deform. But the local transverse structure to W^σ inside K — which determines the singularity type of K/G — is only verified at K_3(A). If the transverse singularity type were to jump in some non-algebraic deformation, the blow-up might fail to be crepant or smooth. However, the group G acts by symplectic automorphisms with fixed locus components of codimension 2 (the W^σ), and the local structure of symplectic automorphisms of hyper-Kähler manifolds is quite rigid: by the symplectic linear algebra of the G-action on the tangent space, the local model is determined by the representation on T_pK, which is constant.","agreement_with_reader":"partial"},"referee_report":{"model":"glm-5.2","summary":"This paper investigates the hyper-Kummer construction, which associates a hyper-Kähler manifold Y_K of K3^[3]-type to any hyper-Kähler sixfold K of Kum^3-type via crepant resolution of K/G, where G ≅ (Z/2Z)^5 is the group of automorphisms acting trivially on H^2 and H^4. The construction was established in [Flo24] (first author), and the present paper develops its analogies with the classical Kummer construction of K3 surfaces from abelian surfaces. The main results include: (1) a lattice-theoretic birational characterization of hyper-Kummer K3^[3]-manifolds (Theorem B), paralleling Nikulin's characterization of Kummer surfaces; (2) a McKay correspondence giving derived equivalences and motivic isomorphisms between Y_K and the quotient stack [K/G] (Theorem D); (3) a rich configuration of companion hyper-Kähler manifolds (K3^[2]-type fourfolds and K3 surfaces) canonically associated to any Kum^3-type manifold; (4) a comparison with the Mongardi-Rapagnetta-Saccà double covers of O'Grady-6 type manifolds (Theorems H, I); (5) a recipe for constructing locally complete families of Kum^3-type varieties (Section 7); (6) Beauville's weak splitting conjecture for all Kum^3-type varieties (Theorem J); and (7) Kimura-O'Sullivan finite-dimensionality for infinitely many new families of K3 surfaces of Picard rank 16 (Theorem L). The proofs proceed by deformation to the generalized Kummer variety K_3(A) associated to an abelian surface A, where explicit computations are performed, and then","tokens_in":74701,"tokens_out":3725,"duration_ms":163668,"significance":"This is a substantial and ambitious paper that establishes a comprehensive higher-dimensional analog of the classical Kummer construction. The lattice-theoretic characterizations (Theorems B, F, H) are clean and well-motivated, and the McKay correspondence (Theorem D) is a significant structural result. The proof of Beauville's weak splitting conjecture for an entire deformation type of hyper-Kähler manifolds of dimension > 2 (Theorem J) is a major contribution, as is the verification of Kimura-O'Sullivan finite-dimensionality for new K3 surfaces (Theorem L). The paper provides explicit, falsifiable lattice-theoretic criteria throughout (e.g., Theorem 4.4, Theorem 6.10), and the examples in Section 8 (Heisenberg-invariant quartics, genus-125 K3 surfaces, diagonal complete intersections) make the theory concrete. The comparison between hyper-Kummer sixfolds and MRS double covers (Section 6) is illuminating and ties together all three known deformation types of 6-dimensional hyper-Kähler manifolds. The construction of locally complete families of Kum^3-type varieties (Section 7) addresses a genuine gap in the literature.","major_comments":[],"minor_comments":[],"recommendation":"accept","confidential_remarks":"The stress-test concern about whether the crepant resolution Y_K → K/G is hyper-Kähler for non-projective K is, upon reading the paper, a concern about [Flo24] rather than about this manuscript. Theorem 2.3 explicitly cites [Flo24] for this result, and the present paper builds on it. The deformation arguments in Propositions 2.4–2.7 and Section 3 are standard for hyper-Kähler geometry: the fixed loci W^σ are trianalytic (hence deform), the group G is deformation-invariant by [HT13], and the local singularity type is determined by the G-representation on the tangent space, which is preserved under deformation since G acts trivially on cohomology. The Schur's lemma arguments (Lemmas 3.13, 3.22) are correctly applied. I did not find a load-bearing gap in the manuscript's own arguments. The paper is long (~100 pages) but the length is justified by the breadth of results. The dependence on [Flo24] for the foundational construction is clearly stated and is standard practice when one author's prior work establishes the existence of a construction that the present paper develops."},"author_rebuttal":{"model":"glm-5.2","summary":"We thank the referee for the careful and generous report. We are gratified that the referee recognizes the scope and significance of the paper, and we appreciate the detailed summary of our results. As the referee's report contains no major comments, specific criticisms, or requested revisions, there are no points requiring a detailed point-by-point response. We will of course continue to check the manuscript for typographical and minor expository issues before the final version.","responses":[],"tokens_in":73963,"tokens_out":170,"duration_ms":17498,"standing_objections":[]},"desk_editor":{"model":"glm-5.2","letter":"The main thing to know: this paper takes the hyper-Kummer construction from [Flo24] (which associates a K3^[3]-type manifold to any Kum^3-type sixfold via crepant resolution of K/G) and builds out a full program around it — birational characterization via Hodge lattices (Theorem B), McKay correspondence for derived categories and motives (Theorem D), comparison with MRS double covers (Theorems H, I), a reconstruction theorem (Theorem G), and applications to Beauville's weak splitting conjecture for all Kum^3-type varieties (Theorem J) and Kimura-O'Sullivan finite-dimensionality for new K3 families (Theorem L). The Beauville result is the first proof of weak splitting for an entire deformation type in dimension > 2. That alone makes the paper significant if it holds up.","headline":"Substantial paper extending the hyper-Kummer construction to Kum^3-type sixfolds; strong results, one foundational dependency worth checking.","tokens_in":75514,"tokens_out":247,"would_cite":true,"duration_ms":107732,"reading_group":"no","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"glm-5.2","headline":"Hyper-Kummer construction bridges two worlds of six-dimensional geometry","keywords":[],"falsifier":"If the group G fails to act with the prescribed fixed locus structure on some Kum^3-type manifold outside the algebraic setting, or if the crepant resolution of K/G fails to be hyper-Kähler for non-projective K, the companion manifold configuration and all downstream cohomological computations would be invalid.","tokens_in":74531,"feed_emoji":"🔗","tokens_out":1388,"duration_ms":236482,"temperature":0.7,"pith_summary":"The paper establishes that a construction discovered by the first author — which associates a hyper-Kähler manifold of K3^[3]-type to any hyper-Kähler sixfold of generalized Kummer type (Kum^3-type) — is a faithful six-dimensional analog of the classical Kummer construction that produces K3 surfaces from abelian surfaces. The classical Kummer construction quotients an abelian surface by the involution x ↦ -x, producing 16 singularities whose resolution yields a K3 surface. The hyper-Kummer construction quotients a Kum^3-type manifold K by a group G ≅ (Z/2Z)^5 of symplectic automorphisms acting trivially on H^2 and H^4, producing a singular variety whose crepant resolution Y_K is a hyper-Kähler manifold of K3^[3]-type. The paper proves that this construction satisfies direct parallels of every major theorem known for the classical Kummer construction: a lattice-theoretic characterization of the resulting manifolds up to birational equivalence (the Kummer lattice embeds primitively in the Néron–Severi group with complement U(2)^⊕3 ⊕ ⟨-4⟩), a McKay correspondence identifying the derived category and Chow motive of the resolution with those of the quotient stack, and a reconstruction theorem allowing one to reverse the construction. Beyond these analogs, the construction produces a rich configuration of companion manifolds — 16 fourfolds of K3^[2]-type and 120 K3 surfaces — canonically associated to any Kum^3-type manifold. In the projective case, these K3 surfaces (called hyper-Kummer K3 surfaces) are all isomorphic to each other and form countably many 4-dimensional families of generic Picard rank 16, generalizing classical Kummer surfaces. The paper also relates the hyper-Kummer manifolds to the Mongardi–Rapagnetta–Saccà double covers of O'Grady's six-dimensional hyper-Kähler manifolds, showing that in the projective case the two constructions produce overlapping but distinct families of K3^[3]-type manifolds, characterized by the Kummer lattice versus the Barnes–Wall lattice. As applications, the construction is used to prove Beauville's weak splitting conjecture for all varieties of Kum^3-type, to establish the Hodge and Tate conjectures for all powers of the varieties involved, and to prove that infinitely many families of hyper-Kummer K3 surfaces have abelian Chow motives, confirming the Kimura–O'Sullivan finite-dimensionality conjecture for new K3 Surfac","feed_headline":"Classical Kummer construction gets a six-dimensional successor","feed_subtitle":"A new bridge between the two most studied families of 6-dimensional hyper-Kähler manifolds yields lattice characterizations, McKay-type dual","key_machinery":"Crepant resolution of symplectic quotients, lattice embeddings into the Beauville–Bogomolov lattice, deformation to the generalized Kummer variety K_3(A) on an abelian surface, equivariant Hilbert schemes, the global Torelli theorem for K3^[3]-type manifolds, and the Barnes–Wall lattice BW_16 for comparison with O'Grady-6 type manifolds.","core_discovery":"The central object is the hyper-Kummer construction: for any hyper-Kähler manifold K of Kum^3-type, the group G ≅ (Z/2Z)^5 of automorphisms acting trivially on H^2(K,Z) and H^4(K,Z) acts on K, and the crepant resolution of K/G is a hyper-Kähler manifold Y_K of K3^[3]-type. The paper proves that Y_K is characterized up to birational equivalence by a primitive embedding of the Kummer lattice L_Km in the Néron–Severi group with orthogonal complement isometric to U(2)^⊕3 ⊕ ⟨-4⟩, exactly paralleling Nikulin's classical characterization of Kummer surfaces. The construction also canonically produces 120 K3 surfaces S_{σ,σ'} from the fixed loci of pairs of involutions in G, and in the projective set","pith_inferences":[],"forward_implications":["The hyper-Kummer construction provides the first general recipe for constructing locally complete families of projective Kum^3-type varieties, by reversing the construction: starting from K3^[3]-type manifolds with 16 divisors in a Kummer lattice configuration, one takes a (Z/2Z)^5-cover and contracts to obtain Kum^3-type varieties.","Beauville's weak splitting conjecture is proved for all Kum^3-type manifolds — the first time it has been established for an entire deformation type of hyper-Kähler manifolds in dimension greater than 2.","The Hodge and Tate conjectures are established for all powers of any variety involved in the hyper-Kummer construction, by reducing to the Kuga–Satake abelian variety and using known results for abelian fourfolds of Weil type.","Infinitely many 4-dimensional families of K3 surfaces of generic Picard rank 16 are shown to have abelian Chow motives, confirming the Kimura–O'Sullivan finite-dimensionality conjecture for these surfaces.","The construction reveals that all three known deformation types of 6-dimensional hyper-Kähler manifolds (Kum^3, K3^[3], and OG6) are linked by explicit geometric maps: the hyper-Kummer construction connects Kum^3 to K3^[3], and the MRS double cover connects K3^[3] to OG6."],"fun_headline_variants":["From Kum^3-type to K3^[3]-type: the hyper-Kummer construction","Six-dimensional Kummer analog mirrors classical surface results","Higher Kummer construction proves Hodge and Tate conjectures","Kummer lattice characterizes hyper-Kummer manifolds up to birational equivalence","McKay correspondence extends to hyper-Kummer derived categories"],"cache_read_input_tokens":0,"weakest_assumption_plain":"The entire construction depends on the group G ≅ (Z/2Z)^5 being deformation-invariant and acting with the prescribed fixed locus structure on every Kum^3-type manifold. This is established by deforming to the generalized Kummer variety K_3(A) on an abelian surface, computing everything explicitly there, and transporting the results back via the deformation invariance of the automorphism group acting trivially on H^2. If this deformation argument fails for non-algebraic or non","fun_headline_variants_meta":{"raw":{"variants":["From Kum^3-type to K3^[3]-type: the hyper-Kummer construction","Six-dimensional Kummer analog mirrors classical surface results","Higher Kummer construction proves Hodge and Tate conjectures","Kummer lattice characterizes hyper-Kummer manifolds up to birational equivalence","McKay correspondence extends to hyper-Kummer derived categories","Hyper-Kummer: lattice characterization, McKay correspondence, conjecture proofs","Hyper-Kummer construction proves Beauville weak splitting for Kum^3-type","Crepant resolution of Kum^3-type quotients yields K3^[3]-type manifolds"]},"model":"glm-5.2","effort":"low","cost_usd":0.0,"raw_usage":{"total_tokens":8703,"prompt_tokens":791,"completion_tokens":7912,"prompt_tokens_details":null},"tokens_in":791,"tokens_out":7912,"duration_ms":403704,"temperature":1.0,"reasoning_tokens":5914,"cache_read_input_tokens":0,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-09T07:57:51.190761+00:00","model_set":{"reader":"glm-5.2"},"falsifier":"If the group G fails to act with the prescribed fixed locus structure on some Kum^3-type manifold outside the algebraic setting, or if the crepant resolution of K/G fails to be hyper-Kähler for non-projective K, the companion manifold configuration and all downstream cohomological computations would be invalid.","supporting_citations":[],"review_version":1}