{"id":"54ef4d4c-4c1c-4f7d-b025-f53133798c2b","arxiv_id":"2502.07134","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Vietoris-Rips complexes of n-by-n torus grids are shown to be tori, spheres, or wedges of spheres for several infinite families of grid sizes and scales.","lead":"This paper proves exact shapes (homotopy types) for Vietoris-Rips complexes of square grids on the torus at several scales, ranging from a torus to spheres and bouquets of spheres. The results give a precise reference point for topological data analysis on torus-shaped point clouds.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 5.12's proof misdescribes K ∩ L^1_a: it omits 3-simplex faces from M_{3k-1,k}, so the claimed S^1 intersections and wedge counts are unsupported.","rationale":"The reader's weakest_assumption was Lemma 5.5, the facet classification, and the reader flagged the missing proof of Lemma 5.11 and unshipped computational data. Those are real issues, but the most load-bearing problem I find is in the proof of Theorem 5.12 itself: the description of K ∩ L^1_a for a ≠ 0 omits 3-dimensional faces coming from M_{3k-1,k}. This is not merely an omitted proof; it is a false assertion in the argument for a main theorem. The theorem's conclusion is corroborated by the homology computations in Table 1 for small parameters, so I do not claim the theorem is false. However, the paper's current proof does not establish it. Since the existing CONDITIONAL verdict already requires revision, I keep that verdict unchanged, but now for a more specific and concrete reason: the intersection analysis in Theorem 5.12 must be corrected or replaced. The proposed check of H_3(K ∩ L^1_1) for k=3 would immediately confirm whether the stated S^1 claim fails and would guide a repair.","tokens_in":29054,"tokens_out":18339,"duration_ms":154881,"concrete_test":"For k=3 (n=8), build K with facets M_{8,3} ∪ {σ1, σ2}. Verify that the 3-simplex {([1],[0]),([1],[1]),([1],[2]),([1],[3])} is a face of the lift facet π_8(B[(1,1.5),1.5]∩Z^2), hence belongs to K ∩ L^1_1. Then compute the homology of K ∩ L^1_1. If H_3(K ∩ L^1_1) is nonzero, the proof's assertion that this intersection is homotopy equivalent to S^1 fails, and the inductive argument for Theorem 5.12 collapses. Reporting the largest omitted facets of K ∩ L^1_a for a=1 would settle whether the proof can be repaired by a corrected intersection description.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Theorem 5.12 asserts VR(T_{3k-1,3k-1};k) ≃ ⋁_{6k-3} S^2 ∨ ⋁_{6k-2} S^3. In its proof, K is the subcomplex with facets M_{3k-1,k} ∪ {σ1, σ2}, and for a ≠ 0 the proof claims K ∩ L^1_a has facets {([a],[b]),([a],[b+2k-1]),([a],[b+2k])}, {([a],[b+k]),([a],[b+2k-1]),([a],[b+2k])}, and {([a],[b]),([a],[b+k])}, hence is homotopy equivalent to S^1. This list is incomplete. For k=3, the arc {([1],[0]),([1],[1]),([1],[2]),([1],[3])} is a face of the lift facet π_8(B[(1,1.5),1.5] ∩ Z^2) ∈ M_{8,3}, so this 3-simplex lies in K ∩ L^1_1. The same construction gives every arc of VR(C_{3k-1};k) as a face of an M_{3k-1,k} facet, so for a ≠ 0 the intersection contains 3-simplices, contradicting the proof's facet list. Consequently the claimed S^1 homotopy type of K ∩ L^1_a, and the induction that produces the wedge sum of 6k-3 copies of S^2 and 6k-2 copies of S^3, are not established. The theorem may be true, and Table 1 is consistent with it, but the proof as written has a concrete false claim.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies Vietoris–Rips complexes of n×n torus grids with the l1 metric. It proves: for n > 3k, VR(T_{n,n};k) is homotopy equivalent to the torus; for k ≥ 2, VR(T_{3k,3k};k) is homotopy equivalent to a wedge of 6k^2−1 copies of S^2; for k ≥ 3, VR(T_{3k−1,3k−1};k) is homotopy equivalent to a wedge of 6k−3 copies of S^2 and 6k−2 copies of S^3; and for even n, VR(T_{n,n};n−1) is homeomorphic to a sphere. The paper also uses integral homology computations and Hurewicz–Whitehead arguments to prove VR(T_{5,5};3) ≃ ⋁_9 S^4 and VR(T_{7,7};4) ≃ S^3, and it classifies the facets of VR(Z^2;k). Several conjectures based on homology computations are stated, including a countable family of (n,k) for which VR(T_{n,n};k) is conjectured to be a 3-sphere.","tokens_in":29423,"tokens_out":36623,"duration_ms":299070,"significance":"If the main theorems are correct, these are among the few exact homotopy-type results for Vietoris–Rips complexes of finite metric spaces at all scales, and they provide a clean picture of the transition from torus at small scales to wedges of spheres at intermediate scales. The facet classification of VR(Z^2;k) and the nerve-lemma torus argument are elegant and likely reusable. The paper is transparent about its computational evidence and poses well-motivated open questions. However, the proof of the family VR(T_{3k−1,3k−1};k) currently contains a false claim about the intersection K ∩ L^1_a and an omitted proof of a load-bearing lemma, so the significance is conditional on repair.","major_comments":[{"comment":"Lemma 5.11 is load-bearing for Theorem 5.12 but its proof is omitted with only the note that it is analogous to Lemma 5.9. The analogy is not automatic: in Lemma 5.9 the added cells are 2-simplices attached along circles, whereas in Lemma 5.11 the added cells are 3-simplices, and one must verify which faces of τ1 and τ2 already lie in the subcomplex generated by M_{3k−1,k}. For k=3, the 2-face {([0],[0]),([0],[3]),([0],[5])} of τ1 is not a face of any M_{8,3} facet, whereas {([0],[0]),([0],[5]),([0],[6])} is; hence the union is not obtained by attaching a 3-ball along its full boundary, and the homotopy type does not follow from the proof of Lemma 5.9. This gap must be filled for Theorem 5.12 to be established.","section":"§5.2, Lemma 5.11"},{"comment":"The proof's description of K ∩ L^1_a for a ≠ 0 is false. For k=3 (n=8), the 3-simplex {([1],[0]),([1],[1]),([1],[2]),([1],[3])} is a face of the lift facet π_8(B_{R^2}[(1,1.5),1.5] ∩ Z^2) ∈ M_{8,3}, so it lies in K ∩ L^1_1. In general, every arc {([a],[b]),([a],[b+1]),...,([a],[b+k])} is a face of some M_{3k−1,k} facet, and these arcs are maximal simplices of the intersection; the simplices listed in the proof are proper faces of these arcs. Thus the asserted homotopy equivalence K ∩ L^1_a ≃ S^1 is not established by the given facet list, and the induction producing ⋁_{6k−3} S^2 ∨ ⋁_{6k−2} S^3 is unsupported as written. The theorem may be true and Table 1 is consistent with it, but the proof needs a correct analysis of the intersection, for example a nerve-lemma argument showing that the subcomplex generated by the arcs is homotopy equivalent to S^1.","section":"§5.2, Theorem 5.12"}],"minor_comments":[{"comment":"In the first direction of the proof, the sentence 'by Proposition 4.2 and the claim above we have σ = B[c,k/2] ∩ Z^2' is incorrect: Proposition 4.2 classifies facets, and a non-maximal simplex is only contained in such a ball, not equal to it. The corollary is true, but the proof should argue by extending σ to a facet.","section":"§4, Corollary 4.3"},{"comment":"In the proof of the second statement, 'isomorphic to the clique complex VR(C_{3k};k)' should read VR(C_{3k−1};k).","section":"§5.1, Lemma 5.5"},{"comment":"Lemma 5.6 is stated without proof. It follows from the discussion after Lemma 5.4 when n > 3k, but the implication should be made explicit.","section":"§5.1, Lemma 5.6"},{"comment":"The treatment of VR(T_{5,5};2) relies on unproved assertions labeled 'similar as Lemma 5.5' and 'similar approach as Lemma 5.9'. Given the false intersection description in the same theorem, this case should be reworked with a complete proof.","section":"§5.2, Theorem 5.12, T_{5,5} case"},{"comment":"The proofs of VR(T_{5,5};3) ≃ ⋁_9 S^4 and VR(T_{7,7};4) ≃ S^3 depend on Polymake homology computations. The authors should provide the code, input data, or a verification script so that the computational step is reproducible.","section":"§7, Theorem 7.1"},{"comment":"There are several typos: 'toplogically' in Proposition 6.2, 'Papaer No.' in reference [30], 'dimesnional' in Proposition 6.2, and the use of 'W' for wedge sums in the abstract and theorems is inconsistent with the standard '∨' notation.","section":"Various"}],"recommendation":"major_revision","confidential_remarks":"The paper is a strong candidate for publication if the proof of Theorem 5.12 is repaired. The two main obstacles are the omitted proof of Lemma 5.11 and the incorrect description of K ∩ L^1_a; both are localized to the proof of the 3k−1 family and appear fixable. The computational results in Section 7 should also be made reproducible. I do not see a fundamental flaw in the central claims, but the manuscript in its current form does not provide a valid proof of one of its headline theorems."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Background on arXiv:2502.07134. This paper is a genuine advance in the exact homotopy-type literature for Vietoris-Rips complexes of finite torus grids, but the proof of Theorem 5.12 has a concrete false claim that needs repair. I would not desk-reject it; I would send it out with the expectation of a major revision.\n\nWhat is new and what works: Section 4 gives a complete facet classification of VR(Z^2;k), which is useful in itself. The transfer lemmas (5.3–5.5) to the torus grid are careful, and the nerve-lemma proof of Theorem 5.8 (VR(T_{n,n};k) ≃ T^2 for n>3k) is clean. Theorem 5.10, for VR(T_{3k,3k};k) ≃ ∨^{6k^2−1} S^2, is convincing: the gluing argument uses Lemma 5.7 with explicit intersections and the induction is straightforward. The cross-polytope sphere result (Corollary 6.3) is standard but nicely applied. Citations look fine; the T_{4,4} ≅ Q_4 coincidence is correctly attributed as known.\n\nThe main soft spot: Theorem 5.12's induction is not established as written. The proof claims that for a≠0, K∩L^1_a has exactly three triangle facets and is homotopy equivalent to S^1. That is false. For k=3, the arc {([1],[0]),([1],[1]),([1],[2]),([1],[3])} is a face of the M_{3k−1,k} facet π_8(B[(1,1.5),1.5]∩Z^2), so it lies in K∩L^1_1. The same construction gives every arc of VR(C_{3k−1};k) as a face of an M_{3k−1,k} facet, so the intersection contains 3-simplices that are not in the listed facet set. Consequently, the claimed S^1 homotopy type of the intersection, and the wedge counts that follow, are unsupported. The theorem may well be true—Table 1 is consistent with it—but the proof has a hole.\n\nTwo smaller issues. Lemma 5.11 is used in Theorem 5.12 but its proof is omitted as “analogous”; given the error in the theorem it supports, that omission is no longer harmless. And Theorem 7.1 depends on Polymake homology computations that are not shipped; the authors should provide code or data so the Hurewicz–Whitehead argument is independently checkable.\n\nWho this is for: anyone working on VR complexes of finite metric spaces or clique complexes of graph powers. The torus and 3k families are likely correct and will be cited. The 3k−1 family is plausible but needs a fixed proof.\n\nRecommendation: send to peer review, but require the authors to repair Theorem 5.12 and release the computation scripts before acceptance.","headline":"Real new results for Vietoris-Rips complexes of torus grids, but Theorem 5.12's proof has a concrete false claim about K∩L^1_a and needs repair before the wedge-count family is established.","tokens_in":30003,"tokens_out":6683,"would_cite":true,"duration_ms":51893,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["55N31","05E45","05C69"],"pacs":[],"model":"deepseek-v4-flash","headline":"Torus grid complexes collapse to exact wedges of spheres","keywords":["Vietoris–Rips complexes","torus grid graphs","homotopy type","wedge sums of spheres","l1 metric","clique complexes","nerve lemma","facets"],"falsifier":"Enumerate all maximal cliques of the Vietoris–Rips graph of $T_{9,9}$ at scale 3 with an independent implementation, or compute $H_2(\\mathrm{VR}(T_{9,9};3);\\mathbb{Z})$. If the number of exceptional triangles is not 54, or if the second Betti number is not 53, Theorem 5.10 fails; the identical check for $\\mathrm{VR}(T_{8,8};3)$ should give Betti numbers 15 and 16 in degrees 2 and 3.","tokens_in":28870,"feed_emoji":"🧩","tokens_out":6534,"duration_ms":55698,"temperature":0.7,"pith_summary":"The paper proves exact homotopy-type formulas for Vietoris–Rips complexes of $n\\times n$ grids on the flat $l^1$ torus, for several families of grid size and scale. It shows that at small scales the complex is homotopy equivalent to the torus itself, that at the special sizes $n=3k$ and $n=3k-1$ the complex is a finite wedge sum of spheres, and that at diameter-minus-one scales the complex is a sphere. These are exact topological identities, not homology approximations, and they confirm a pattern visible in computer computations. The result matters because such grids approximate the continuous $l^1$ torus, so the formulas describe how coarse sampling of the torus retains or changes its topology under the standard Vietoris–Rips construction.","feed_headline":"Torus grid complexes collapse to exact wedges of spheres","feed_subtitle":"New proofs give exact homotopy equivalences for n by n l1-torus grids at scales k, including torus, sphere, and wedge-sum identities.","key_machinery":"The classification of maximal simplices, or facets, of $\\mathrm{VR}(\\mathbb{Z}^2;k)$: every facet of the infinite square lattice with the $l^1$ metric is the intersection of $\\mathbb{Z}^2$ with a closed $l^1$ ball of radius $k/2$ centered at a point whose coordinates are integers or half-integers. Quotienting by $n\\mathbb{Z}\\times n\\mathbb{Z}$ gives the facets of the torus-grid complex, except that for $n=3k$ and $n=3k-1$ there are extra non-liftable facets: triples equally spaced by $k$ along a row or column for $n=3k$, and quadruples spanned by $\\{v_i, v_{i+k}, v_{i+2k-1}, v_{i+2k}\\}$ for $n=3k-1$. The proof shows via the nerve lemma that the subcomplex generated by the liftable facets is homotopy equivalent to the torus, and each extra facet attaches a disk along a circle wrapping once around a meridian or longitude, converting the torus into the stated wedge sums of spheres.","core_discovery":"On the $n\\times n$ torus grid $T_{n,n}$ with the $l^1$ metric, the paper claims: $\\mathrm{VR}(T_{n,n};k)$ is homotopy equivalent to $T^2$ whenever $k\\geq 2$ and $n>3k$; for $k\\geq 2$, $\\mathrm{VR}(T_{3k,3k};k)\\simeq \\bigvee^{6k^2-1}S^2$; for $k\\geq 3$, $\\mathrm{VR}(T_{3k-1,3k-1};k)\\simeq \\left(\\bigvee_{6k-3}S^2\\right)\\vee \\left(\\bigvee_{6k-2}S^3\\right)$; and for $n\\geq 2$, $\\mathrm{VR}(T_{2n,2n};2n-1)$ is homeomorphic to $S^{2n^2-1}$. In addition, integral homology computations combined with the Hurewicz and Whitehead theorems give $\\mathrm{VR}(T_{5,5};3)\\simeq \\bigvee^{9}S^4$ and $\\mathrm{VR}(T_{7,7};4)\\simeq S^3$. Together these determine the full homotopy type, not merely Betti numbers, for several entire diagonals of the paper's table of computed complexes.","pith_inferences":["Beyond the paper's claims, the facet-lifting technique looks applicable to other quotient lattices, such as grids on $m$-dimensional tori or on other flat orbifolds, and would yield exact homotopy types rather than homology-only information.","If the conjectured countable family of $(n,k)$ pairs with $S^3$ homotopy type is correct, the paper's Heegaard-decomposition picture predicts a middle-scale regime where a single three-dimensional hole appears and then fills in as $k$ grows; this is testable by persistent homology in degree 3.","The exact wedge-sum formulas provide a natural sanity check for computational topology software: any implementation that computes persistence with integral coefficients should reproduce the stated Betti numbers, and any discrepancy would pinpoint a bug or a missed facet.","The paper's division into liftable facets and exceptional facets suggests a broader pattern: topology changes at scales where lattice points snap into periodic alignment, so the same dichotomy may organize the intermediate-scale topology of other periodic metric spaces."],"forward_implications":["If the scale $k$ satisfies $n>3k$, the Vietoris–Rips complex of the grid has the same homotopy type as the continuous torus, so no spurious topology arises in that regime.","At $n=3k$, all topology is two-dimensional: the complex is a wedge of $6k^2-1$ copies of $S^2$, so all higher Betti numbers vanish.","At $n=3k-1$, homology is nonzero only in degrees 2 and 3, with ranks $6k-3$ and $6k-2$, respectively.","For even grid sizes, at scale one below the diameter the complex is the boundary of an $n^2/2$-dimensional cross-polytope, hence a single sphere of dimension $n^2/2-1$.","The paper verifies one member of the conjectured 3-sphere family: $\\mathrm{VR}(T_{7,7};4)\\simeq S^3$.","If the paper's facet counts are correct, the complexes $\\mathrm{VR}(T_{9,9};3)$ and $\\mathrm{VR}(T_{8,8};3)$ provide concrete benchmarks: the former should have second Betti number $53$, and the latter should have second and third Betti numbers $15$ and $16$."],"supporting_citations":[{"why":"Supplies the nerve theorem and the Hurewicz–Whitehead consequence used to turn homology information into homotopy equivalences.","marker":"[32]"},{"why":"Provides the homotopy classification of Vietoris–Rips complexes of cycles, used to identify the cross-sectional subcomplexes in Theorem 5.12 as $S^3$ or contractible.","marker":"[1]"},{"why":"Provides the splitting lemma used to assemble the complex as iterated wedge sums with suspensions of intersection spaces.","marker":"[30]"},{"why":"Supplies the $ℤ/2$ homology computations that guided the conjectures and populate the table.","marker":"[6]"},{"why":"Supplies the integral homology computations for $\\mathrm{VR}(T_{5,5};3)$ and $\\mathrm{VR}(T_{7,7};4)$, which Theorem 7.1 turns into homotopy equivalences.","marker":"[29]"},{"why":"Provides the conic-complex connectivity criterion used to prove the simple connectedness needed in Theorem 7.1.","marker":"[22]"}],"fun_headline_variants":["Torus grid Rips complexes collapse to exact sphere wedges","Exact homotopy types: torus grid Rips become sphere wedges","Sphere wedges from torus grid Rips complexes","Exact sphere wedge homotopy types for torus grid Rips","Torus grids: Rips complexes become exact sphere wedges"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof counts facets exhaustively: for $n=3k$ or $n=3k-1$, every maximal simplex in the grid complex is either the image of a maximal simplex in the infinite square lattice or one of the explicitly listed special simplices, and if a single exceptional facet is missed, the wedge-sum formulas would acquire an extra sphere summand and fail.","fun_headline_variants_meta":{"raw":{"variants":["Torus grid Rips complexes collapse to exact sphere wedges","Exact homotopy types: torus grid Rips become sphere wedges","Sphere wedges from torus grid Rips complexes","Exact sphere wedge homotopy types for torus grid Rips","Torus grids: Rips complexes become exact sphere wedges"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001077,"raw_usage":{"total_tokens":4642,"prompt_tokens":1217,"completion_tokens":3425,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":833,"completion_tokens_details":{"reasoning_tokens":3335}},"tokens_in":833,"tokens_out":3425,"duration_ms":22723,"temperature":1.0,"reasoning_tokens":3335,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-08T13:42:17.071099+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Enumerate all maximal cliques of the Vietoris–Rips graph of $T_{9,9}$ at scale 3 with an independent implementation, or compute $H_2(\\mathrm{VR}(T_{9,9};3);\\mathbb{Z})$. If the number of exceptional triangles is not 54, or if the second Betti number is not 53, Theorem 5.10 fails; the identical check for $\\mathrm{VR}(T_{8,8};3)$ should give Betti numbers 15 and 16 in degrees 2 and 3.","supporting_citations":[{"cited_title":"Clique complexes and graph powers.Israel Journal of Mathematics, 196(1):295– 319, 2013","cited_arxiv_id":null,"evidence_quote":"Provides the homotopy classification of Vietoris–Rips complexes of cycles, used to identify the cross-sectional subcomplexes in Theorem 5.12 as $S^3$ or contractible."},{"cited_title":"Goyal, S","cited_arxiv_id":null,"evidence_quote":"Provides the splitting lemma used to assemble the complex as iterated wedge sums with suspensions of intersection spaces."},{"cited_title":"Ripser: efficient computation of Vietoris–Rips persistence barcodes.Journal of Applied and Computational Topology, pages 391–423, 2021","cited_arxiv_id":null,"evidence_quote":"Supplies the $ℤ/2$ homology computations that guided the conjectures and populate the table."},{"cited_title":"In Polytopes—combinatorics and computation (Oberwolfach, 1997), volume 29 ofDMV Sem., pages 43–73","cited_arxiv_id":null,"evidence_quote":"Supplies the integral homology computations for $\\mathrm{VR}(T_{5,5};3)$ and $\\mathrm{VR}(T_{7,7};4)$, which Theorem 7.1 turns into homotopy equivalences."},{"cited_title":"Large simplicial complexes: universality, randomness, and ampleness.Journal of Applied and Computational Topology, 8:1551–1574, 2023","cited_arxiv_id":null,"evidence_quote":"Provides the conic-complex connectivity criterion used to prove the simple connectedness needed in Theorem 7.1."}],"review_version":1}