{"id":"e2512615-31e1-4e15-9e2d-81411f0ab383","arxiv_id":"2411.16545","paper_version":3,"verdict":"REJECT","confidence":"HIGH","novelty_score":4.0,"correctness_risk":"high","formal_verification":"none","parameter_count":0,"one_line_summary":"The paper constructs infimum and supremum double complexes of differential forms for hypergraphs on manifolds and claims they are quasi-isomorphic under vertex deletion.","lead":"This paper builds double complexes of differential forms on configuration spaces and on hypergraphs whose vertices move on a manifold, and shows that a 'supremum' and an 'infimum' version are quasi-isomorphic under vertex deletion. The construction is meant as a foundation for manifold learning and motion planning, but it rests on an unproven and generally false assumption about extending forms from open submanifolds.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Lemma 3.2 is false for open embedded submanifolds, which the paper explicitly allows: dx/x on (0,1) has no smooth extension to R. This surjectivity is used in (3.7), Corollary 3.5, and Lemma 3.8, so the surjective maps in Theorems 1.2/1.3 are not proved for arbitrary hypergraphs.","rationale":"The paper's central claim is that every hypergraph on a manifold admits a canonical ladder of double complexes with surjective maps and a quasi-isomorphism in the middle. The proof of this claim funnels through Lemma 3.2, where restriction of global differential forms to an embedded submanifold is asserted to be surjective. The reader's diagnosis is correct: the lemma is false for open submanifolds, and the paper's own examples and corollaries make openness unavoidable by treating Conf_n(M,r) for r>0 as open hypergraph components. The concrete obstruction is elementary and local: dx/x on (0,1) is smooth yet admits no smooth extension to R. Because this surjectivity underpins the identifications of the infimum and supremum complexes with spaces of forms on the hypergraph, the surjective homomorphism part of Theorems 1.2 and 1.3 is not established for arbitrary hypergraphs. This is a correctness failure for the stated generality, not merely a missing hypothesis; a repaired version would need closed submanifolds or a sheaf-theoretic formulation. I also note that Definition 4's existence of the largest contained ∆-submanifold is asserted rather than proved, but the surjectivity issue is already sufficient. Since the reader already recommended REJECT and my analysis supports that verdict, no adjustment to the reader's verdict is needed.","tokens_in":30523,"tokens_out":10234,"duration_ms":113762,"concrete_test":"Take M=R, A=Conf_1(R)=R, and B=(0,1)⊂A. Verify that ω=dx/x is a smooth 1-form on B. Then check whether Lemma 3.2's claimed surjectivity holds: if there were α=g(x)dx∈Ω^1(R) with α|_B=ω, then g(x)x=1 for all x∈(0,1); taking x→0+ gives 0=1, a contradiction. Next trace this example through the proof of Theorem 10.2/1.2 to confirm that the asserted surjective homomorphism Ω(δH)←... or the identification of Inf/Sup complexes fails at the stage where Ω(B) is used. As a positive control, repeat the extension test with B=[0,1] instead; the surjectivity holds for this closed submanifold, isolating openness as the obstruction.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing step is Lemma 3.2, which asserts that for any embedded submanifold A'_• of a ∆-manifold A_•, the pull-back ι#: Ω(A_•)→Ω(A'_•) is surjective. This is true for closed embedded submanifolds via tubular neighborhoods, but the paper only assumes each H_n(M) is a differentiable manifold (Section 1.4 and Section 4) and explicitly allows open submanifolds: Example 4.4 treats Conf_n(M,r) for r>0 as open manifolds, and Corollary 9.3/9.6 apply Lemma 3.2 to these open submanifolds. For an open submanifold the restriction map is not surjective: take M=R, A_1=Conf_1(R)=R, and B_1=(0,1). The 1-form ω=dx/x on (0,1) is smooth, but no smooth α∈Ω^1(R) restricts to it: if α=g(x)dx, then g(x)=1/x on (0,1), which is impossible at x=0 since g is smooth and x g(x)→1 while x g(x)→0. Consequently, the identifications in (3.7) involving Ω(B_•) as a quotient of Ω(A_•) by Ker(ε#), and the quotient/surjection claims in Corollary 3.5 and Lemma 3.8 that rely on them, are unjustified for open hypergraph components. This is not a cosmetic gap: Theorems 1.2 and 1.3 state surjective homomorphisms for any hyperdigraph/hypergraph, so the central claim fails at the stated level of generality. A repair would require restricting to closed submanifolds or replacing Ω(-) by a sheaf of extendable forms, but that is a substantive hypothesis change. Definition 4 also asserts the existence of the largest ∆-submanifold δB without proof, but the surjectivity failure is already decisive.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies hypergraphs whose hyperedges are smooth submanifolds of ordered or unordered configuration spaces of a Riemannian manifold M. It constructs double complexes of differential forms on the ambient configuration-space ∆-manifolds and, for a hypergraph B, defines infimum and supremum double complexes as quotients of the forms on the ambient manifold. The main results (Theorems 1.2 and 1.3, proved via Lemmas 3.8, 10.2, and 10.5) assert a ladder of surjective double-complex homomorphisms from forms on the associated ∆-manifold ∆B to the supremum and infimum complexes and then to forms on the lower-associated ∆-manifold δB, with the middle map q a quasi-isomorphism with respect to ∂, and with all three maps identities exactly when the hypergraph is a ∆-submanifold. The paper also develops automorphism groups of hypergraphs on manifolds, estimates orders of associated vector bundles, and gives obstructions to k-regular embeddings.","tokens_in":30983,"tokens_out":7668,"duration_ms":76992,"significance":"If the main ladder were valid, it would attach to every sufficiently regular manifold-valued hypergraph a canonical double complex whose vertex-deletion homology is invariant under replacing the hypergraph by its infimum or supremum closure, thereby connecting the construction to the embedded homology of [13]. The paper deserves credit for giving a self-contained proof of the quotient quasi-isomorphism in Lemma 2.1 and for spelling out several explicit examples. However, the surjectivity assertions in the main theorems rest on Lemma 3.2, which is false for the open submanifolds that the paper explicitly allows, and the existence of the lower-associated ∆-manifold δB in Definition 4 is asserted without proof. The central claim is therefore not established at the stated level of generality; the failure is load-bearing rather than cosmetic, because the paper prominently features open hard-disk configuration spaces as motivating examples.","major_comments":[{"comment":"The claim that the pullback ι#: Ω(A•) → Ω(A'•) is surjective is false for embedded submanifolds that are not closed. The paper explicitly allows such submanifolds: Section 1.4 assumes only that each H_n(M) is a differentiable manifold, and Example 4.4 states that Conf_n(M,r) for r>0 is an open manifold when M is boundaryless. For instance, take A_1 = R and A'_1 = (0,1); the 1-form dx/x on (0,1) is smooth but has no smooth extension to R. This surjectivity is used to identify Ω(A'•) with the quotient in (3.2), to obtain (3.7), and in Corollary 3.5, Lemma 3.8, Corollaries 9.3 and 9.6, and therefore in Theorems 10.2, 10.5, 1.2, and 1.3. Since the hypotheses do not require closedness, the asserted surjective homomorphisms are not proved and the main theorem fails as stated.","section":"§3.1, Lemma 3.2"},{"comment":"Definition 4 asserts the existence of the smallest ∆-submanifold ∆B containing B and the largest ∆-submanifold δB contained in B, but no proof is given that the displayed unions over p ∈ B are smooth ∆-manifolds. Lemma 3.7 only proves independence from the ambient ∆-manifold A•, not existence or smoothness. Since δB is the target of the final map in Lemma 3.8 and in Theorems 1.2 and 1.3, the main theorems are unsupported for arbitrary graded submanifolds. In addition, the map Ω•(B•) → Ω•(δB•) in the proof of Lemma 3.8 again relies on the false surjectivity assertion from Lemma 3.2.","section":"§3.3, Definition 4"},{"comment":"There is a quantifier mismatch in the statements of the main theorems. Theorem 1.2 says 'for any hyperdigraph H(M) on M', while Section 1.4 assumes each H_n(M) is a differentiable manifold and Section 4 defines a hyperdigraph as any subspace of the configuration space. For an arbitrary subspace, the space of differential forms on the hypergraph is not defined; for arbitrary submanifolds, Lemma 3.2 fails. The main theorems therefore require an explicit restricted class of hypergraphs, and that class must be closed under the constructions ∆ and δ. This is not a purely presentational point, since the surjectivity of the ladder and the existence of δB are both used in the statements.","section":"§1.4 and Theorems 1.2/1.3"}],"minor_comments":[{"comment":"The quotient map q is called 'induced by (2.3)', but (2.3) is the inclusion Inf•(D) → Sup•(D); the map C/Sup•(D) → C/Inf•(D) is more naturally induced by the identity of C. Please clarify the direction and the inducing map.","section":"§2, around (2.4)"},{"comment":"The sentence 'Then (9.4) a sub-double complex ...' is missing the verb 'is'.","section":"§9, Theorem 9.5"},{"comment":"In the case M = S^m, the text refers to Stab(Conf_n(S^1,r)/Σ_n); this should presumably be Conf_n(S^m,r)/Σ_n.","section":"§5, Example 5.4(3)"},{"comment":"There are unmatched parentheses in the displayed statements of the sub-double complexes in (10.1) and (10.4); please correct the notation.","section":"§10, Theorems 10.1 and 10.4"}],"recommendation":"reject","confidential_remarks":"The rejection is based on a load-bearing mathematical error, not on lack of novelty. If the author restricts attention to closed submanifolds or replaces Ω by a sheaf of extendable forms, parts of the ladder may be recoverable, but that would exclude the hard-disk configuration spaces emphasized in Example 4.4 and Corollaries 9.3 and 9.6. The manuscript also mixes several largely independent topics—automorphism groups, vector-bundle orders, and regular-embedding obstructions—so a revision would need to reshape the hypotheses before the main claim can be assessed."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The genuinely new part is the construction of infimum and supremum double complexes of differential forms for hypergraphs on manifolds, together with the claim that the middle comparison map is a quasi-isomorphism with respect to the vertex-deletion boundary operator. That is a real idea, and the paper gives a clean algebraic engine for it in Lemma 2.1, where the quotient quasi-isomorphism is reproved directly rather than just quoted. The ambient double complex on configuration spaces is essentially the standard de Rham double complex of a simplicial/Delta manifold, but framing it this way is reasonable and the ordered/unordered configuration-space part is mostly fine. Sections 7 and 8 are honest citations of known vector-bundle results and apply them to hypergraph pullbacks; nothing fraudulent there.\n\nThe soft spot is load-bearing, and the stress-test note is right. Lemma 3.2 asserts that restriction of forms from a manifold to an embedded submanifold is surjective. That is true for closed embedded submanifolds via tubular neighborhoods, but the paper defines hypergraphs as arbitrary differentiable manifolds inside configuration spaces and even highlights open examples such as Conf_n(M,r) for r>0. For an open submanifold, surjectivity fails: on (0,1) inside R, the form dx/x is smooth, but no smooth form on R restricts to it. Since this lemma is used in (3.7), Corollary 3.5, and Lemma 3.8, the surjectivity of the maps in Theorems 1.2 and 1.3 is not established for general hypergraphs. The claim that every hypergraph on a manifold carries the full surjective ladder therefore fails at the stated level of generality. Definition 4 also postulates the existence of the largest Delta-submanifold delta B without proof; that is a lesser issue, but it is another spot where the paper asserts rather than proves.\n\nIt is fair to add that the failure is repairable in principle: restrict to closed submanifolds, or replace Omega(-) by a sheaf of extendable forms and work with the appropriate quotient. That is a substantive hypothesis change, not a cosmetic fix. As written, the central theorems are not supported.\n\nWho gets value from this? Someone working on algebraic models for hypergraphs on manifolds might find the inf/sup double-complex formalism worth adapting, provided they are alerted to the closedness issue. But as a paper, it needs major revision before its main claims can be trusted.\n\nMy recommendation: send it to a referee rather than desk-rejecting. The construction is serious enough that a careful referee could specify the missing hypothesis and decide whether the theorems survive in the closed case. I would not cite it as is.","headline":"The main theorems overstate their scope: Lemma 3.2's surjectivity of restriction is false for the open submanifolds the paper explicitly allows, so the surjective ladder in Theorems 1.2/1.3 is not proved as stated.","tokens_in":31485,"tokens_out":1748,"would_cite":false,"duration_ms":20712,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["57N65","57N75","55U05","55U15"],"pacs":[],"model":"deepseek-v4-flash","headline":"Every hypergraph on a manifold carries a canonical ladder of four double complexes, and the middle two are homologically identical under vertex deletion.","keywords":["configuration spaces","double complexes","hypergraphs","simplicial manifolds","differential forms","quasi-isomorphism","vertex deletion"],"falsifier":"Test the surjectivity lemma on the hypergraph $H$ on $M = \\mathbb{R}$ with $H_1(M) = (0,1)$, an open submanifold of $\\operatorname{Conf}_1(\\mathbb{R})$. The smooth $1$-form $\\omega = (1/x)\\,dx$ on $(0,1)$ does not extend to any smooth form on $\\mathbb{R}$, since it is unbounded near $0$; therefore the restriction map is not surjective, and the quotient identification underlying the main theorem collapses for this allowed example.","tokens_in":30276,"feed_emoji":"📐","tokens_out":12740,"duration_ms":98214,"temperature":0.7,"pith_summary":"The paper claims that any hypergraph on a manifold—whose vertices are distinct points moving smoothly, and whose hyperedges form submanifolds of the configuration space—carries a canonical four-term ladder of double complexes built from differential forms. The ladder runs from the forms on the smallest $\\Delta$-manifold enclosing the hypergraph, through the supremum and infimum double complexes, to the forms on the largest $\\Delta$-manifold contained in it. The middle two complexes are quasi-isomorphic with respect to the vertex-deletion boundary, so the homology that counts hyperedges is unchanged when the hypergraph is replaced by either of its two canonical closures. If established, this gives a smooth-geometry toolkit for hypergraphs, with applications to motion planning, manifold learning of network data, and obstructions to regular embeddings.","feed_headline":"Every manifold hypergraph fits a four-term double-complex ladder","feed_subtitle":"Vertex-deletion homology is the same whether you close the hypergraph up or down.","key_machinery":"The load-bearing object is the double complex $(\\Omega^\\bullet(A_\\bullet), d, \\partial)$ attached to a $\\Delta$-manifold $A_\\bullet$, where $d$ is the exterior derivative and $\\partial$ is the alternating sum of pullbacks along the face maps that delete vertices. For a graded submanifold $B$ with embedding $\\varepsilon$, the infimum complex $\\operatorname{Inf} = \\operatorname{Ker}\\varepsilon \\cap \\partial^{-1}(\\operatorname{Ker}\\varepsilon)$ and the supremum complex $\\operatorname{Sup} = \\operatorname{Ker}\\varepsilon + \\partial(\\operatorname{Ker}\\varepsilon)$ are subcomplexes of the ambient forms; quotienting by them yields the double complexes of forms on $B$. The associated $\\Delta$-manifolds $\\Delta B$ and $\\delta B$ are built pointwise from $\\Delta$-closures, making the whole sequence depend only on $B$, not on the ambient configuration space. The middle quasi-isomorphism is the established inclusion $\\operatorname{Inf} \\subset \\operatorname{Sup}$ inducing an isomorphism of $\\partial$-homology, transferred to the quotients.","core_discovery":"For a hyperdigraph $\\vec{H}(M)$ whose $n$-uniform parts are differentiable submanifolds of the ordered configuration spaces $\\operatorname{Conf}_n(M)$, the paper constructs the double complex $(\\Omega^\\bullet(\\operatorname{Conf}_\\bullet(M)), d, \\partial)$ and then shows that the infimum and supremum double complexes $\\operatorname{Inf}^\\bullet(\\vec{H})$ and $\\operatorname{Sup}^\\bullet(\\vec{H})$ fit into a sequence of surjective double-complex homomorphisms $$\\$\\Omega$(\\$\\Delta$\\vec{H}) \\to \\operatorname{Sup}^\\bullet(\\vec{H}) \\to \\operatorname{Inf}^\\bullet(\\vec{H}) \\to \\$\\Omega$(\\delta\\vec{H}).$$ The middle map is a quasi-isomorphism with respect to $\\partial$, and $\\vec{H}$ is a $\\Delta$-submanifold exactly when all three maps are the identity. The unordered analogue holds whenever $M$ admits a continuous total order. Consequently every hypergraph on a manifold has a vertex-deletion homology that is invariant under passing to the smallest enclosing or largest contained $\\Delta$-manifold.","pith_inferences":["A likely repair for the surjectivity gap is to require hypergraphs to be closed submanifolds; whether the quasi-isomorphism survives for open ones under a weaker extension condition is a question the paper leaves open.","The hard-disk filtrations already present in Section 9 suggest a persistence version of the ladder, producing double-complex persistence modules that would connect to persistent homology of hypergraphs on manifolds.","The maps assigning $\\Delta B$ and $\\delta B$ to a graded submanifold behave like adjoint closure and interior operations; formalizing that adjunction could explain the ambient-independence the paper proves."],"forward_implications":["If the main theorem is correct, the vertex-deletion homology of a hypergraph on a manifold is unchanged when the hypergraph is replaced by its infimum or supremum closure.","A hypergraph on a manifold is a $\\Delta$-submanifold exactly when the smallest enclosing and largest contained $\\Delta$-manifolds coincide with it, making all four double complexes identical.","For a discrete vertex set, the construction reduces to the known infimum and supremum chain complexes of hypergraphs, with trivial exterior derivative.","For manifolds of dimension at most one admitting a continuous total order, the double complexes contain only $0$-forms and $1$-forms, giving a concrete computable model."],"supporting_citations":[{"why":"Defines the infimum and supremum chain complexes and provides the quasi-isomorphism between them that the ladder's middle map transfers to forms.","marker":"[13]"},{"why":"Restates the infimum–supremum quasi-isomorphism, the other cited source for the key algebraic step.","marker":"[28]"},{"why":"Introduces the associated simplicial complex of a hypergraph, the combinatorial ancestor of the $\\Delta$-closure and lower closure used in the main theorem.","marker":"[37]"},{"why":"Supplies the directed hyperedge and hyperdigraph formalism on which the ordered version of the main theorem is built.","marker":"[15]"}],"fun_headline_variants":["Hypergraph double complexes: inf and sup quasi-isomorphic","Vertex-deletion homology invariant under hypergraph closure","Infimum and supremum complexes agree on vertex-deletion homology","Manifold hypergraphs share homology across double complex ladder"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole construction depends on the claim that every differential form on a submanifold of the configuration space extends smoothly to the whole configuration space, which is only guaranteed for closed submanifolds and not for the open ones the paper's hypotheses allow.","fun_headline_variants_meta":{"raw":{"variants":["Hypergraph double complexes: inf and sup quasi-isomorphic","Vertex-deletion homology invariant under hypergraph closure","Infimum and supremum complexes agree on vertex-deletion homology","Manifold hypergraphs share homology across double complex ladder"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000268,"raw_usage":{"total_tokens":1587,"prompt_tokens":880,"completion_tokens":707,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":496,"completion_tokens_details":{"reasoning_tokens":641}},"tokens_in":496,"tokens_out":707,"duration_ms":6965,"temperature":1.0,"reasoning_tokens":641,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T13:00:09.906042+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Test the surjectivity lemma on the hypergraph $H$ on $M = \\mathbb{R}$ with $H_1(M) = (0,1)$, an open submanifold of $\\operatorname{Conf}_1(\\mathbb{R})$. The smooth $1$-form $\\omega = (1/x)\\,dx$ on $(0,1)$ does not extend to any smooth form on $\\mathbb{R}$, since it is unbounded near $0$; therefore the restriction map is not surjective, and the quotient identification underlying the main theorem collapses for this allowed example.","supporting_citations":[{"cited_title":"Bressan, J","cited_arxiv_id":null,"evidence_quote":"Defines the infimum and supremum chain complexes and provides the quasi-isomorphism between them that the ladder's middle map transfers to forms."},{"cited_title":"Grbić, J","cited_arxiv_id":null,"evidence_quote":"Restates the infimum–supremum quasi-isomorphism, the other cited source for the key algebraic step."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces the associated simplicial complex of a hypergraph, the combinatorial ancestor of the $\\Delta$-closure and lower closure used in the main theorem."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the directed hyperedge and hyperdigraph formalism on which the ordered version of the main theorem is built."}],"review_version":1}