{"id":"bda9c042-abc7-4d27-b676-ef4b0907a6d8","arxiv_id":"2607.24146","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"Space-filling families in Assouad-Nagata dimension n have intersection-graph asymptotic dimension at most n+1, and this is tight; boundaries preserve the dimension up to max with 1.","lead":"Intersection graphs of nice families of sets in n-dimensional space have asymptotic dimension at most n+1. The paper also shows that replacing sets by their boundaries barely changes that dimension, settling bounds for balls and spheres.","discovery_kind":"extension","skeptic_critique":{"model":"moonshotai/kimi-k3","headline":"No significant objection identified. The proof of the central bound (Theorem 1.3/6.4) is internally consistent at every point I probed; the only residual exposure is black-box reliance on published external lemmas, which are invoked with matching hypotheses.","rationale":"I read the proof of the strongest claim (Theorem 1.3/6.4 and the matching lower bound Theorem 1.4) looking for the least secure load-bearing step. My candidate concerns all dissolved under direct verification: the space-filling count in Lemma 7.3 (the apparent gap about witnessing points lying outside the ball is closed because S(p) ⊆ Q_i ⊆ Q_{i_0}); the constant-matching between Lemma 6.3's choice b = 3^n r and Lemma 5.11's 3^{s−1}r-compliance requirement; the constrained-map constants in Lemmas 5.6/5.9; the KKM application in Lemma 7.4 including the apex-identification subtlety; and the uniformity required for graph-class asdim in the final assembly of Theorem 6.4. The lower-bound construction is explicit, parameter-free, and its counting lemmas (7.2) have correct volume arguments.\n\nWhat remains is dependence on four external results: [7, Lemmas 7–8] (the structural keystone turning Assouad–Nagata control into laminar webs and centered tree-decompositions), [3, Theorem A.2] (finite-to-infinite reduction), [11, Lemma 5.7] and [4, Theorem 4.1] (weak-diameter colorings for centered tree-decompositions and bounded treewidth). These are published in refereed venues, the quoted statements match their applications, and one ([7]) shares an author with this paper, which lowers but does not eliminate citation risk. This is the correct target for a verification pass, but it does not constitute an identified flaw.\n\nThe reader's verdict (ACCEPT, high confidence, low correctness risk) and weakest-assumption analysis (necessity of the space-filling hypothesis via the Dvořák–Norin expander obstruction; the non-covering condition in the invariance theorem) align with my own read. The acknowledged n vs. n+1 gap for balls/spheres is an open problem, not a correctness issue, and the authors state it honestly. I recommend no verdict change.","tokens_in":28081,"tokens_out":8944,"duration_ms":147143,"concrete_test":"Cross-check the one non-self-contained step in the proof of Theorem 1.3: open [7, Lemma 7] and [7, Lemma 8] as published (European J. Combin. 123 (2025), 103631) and verify that (a) [7, Lemma 7] yields an (n+1)-laminar C-web for any metric space admitting a dilation g as n-dimensional control function, with C depending only on g; (b) [7, Lemma 8] yields an ((f(C),1))-centered tree-decomposition of I(S_i) for an f-space-filling family caught by one laminar family, with no additional hypotheses (finiteness, uniform diameter bounds, integer-valued f) beyond those held in Lemma 6.2. If both match as invoked, the central theorem's keystone is secure. As an independent sanity check, for n=2, k small (e.g., k=4), brute-force verify the claim of Lemma 7.4 on the explicit grid family in R^2: every 3-coloring of I(∪_{i≤4} S_i) has a monochromatic component of weak diameter ≥ min{k/2, 2^{k/2}n^{-3/","verdict_should_be":"UNCHANGED","load_bearing_attack":"I stress-tested the central claim at its most exposed joints and each probe resolved:\n\n(1) Lemma 7.3's space-filling count initially looked vulnerable: a set S of diameter ≥ s meeting B(p,r) has its witnessing Q_{i_0}-point only within diam(S) of the ball, which at coarse levels is ~4√n ≫ r. But S(p) ⊆ P_i ∪ P_{i−1} ⊆ Q_i ⊆ Q_{i_0}, so the intersection point with B(p,r) is itself a Q_{i_0} point, and disjointness makes witnessing points distinct; the count |B(p,r) ∩ Q_{i_0}| is correct.\n\n(2) The scale-matching in Lemma 6.3: b = 3^n r is exactly 3^{s−1}r for s = n+1, so the b-compliant (k, 4b+2)-centered tree-decomposition from Claim 1 matches Lemma 5.11's 3^{s−1}r-compliance hypothesis; the pullback diameter bound 2(2b+1) = 4b+2 is correct.\n\n(3) The (r/(2R),0,r)-constrained identity map in Lemma 5.6: ℓ ≥ (r/2R)·d pointwise on both J-edges and non-edges, so d' ≥ (r/2R)d, and d' ≤ d gives the second condition; constants check.\n\n(4) Theorem 7.1 (KKM variant) as applied in Lemma 7.4: X_C are finite unions of closed sets, cover ∆′, no point lies in n+2 of them (n+1 colors, same-color components disjoint, and the apex identification is safe because all level-0 sets are pairwise adjacent). The diameter-≥1/n estimate for a set meeting all facets of ∆ follows from Σ(p_0)_j = 1.\n\n(5) The finite-to-infinite passage via Theorem 2.2 and the weak-diameter characterization (Lemma 2.4) are applied with the uniformity (control function independent of G) the class-asdim definition requires.\n\nThe genuinely least self-contained point is Lemma 6.2's invocation of [7, Lemmas 7–8] (laminar webs → centered tree-decompositions), plus Lemmas 5.5 and 5.7 quoted from [11] and [4]. All four are from refereed publications and are quoted with hypotheses that visibly match their use here; this is standard modular citation, not circularity. The reader's identified weakest assumptions (necessity of the space-filling hypothesis, the non-covering condition in Theorem 1.7) are correctly stated and the ","agreement_with_reader":"agree"},"referee_report":{"model":"moonshotai/kimi-k3","summary":"The paper proves two main results about asymptotic dimension of intersection graphs. First (Theorem 1.3/6.4): for any f-space-filling family F of subsets of a metric space of Assouad–Nagata dimension at most n, the class of intersection graphs of subfamilies of F has asymptotic dimension at most n+1, improving the 2n+1 bound of Dvořák–Norin [7]; the bound is shown quantitatively tight (Theorem 1.4) by an explicit construction of space-filling families in the n-simplex whose intersection graphs force monochromatic components of arbitrarily large weak diameter under any (n+1)-coloring, via a discrete application of the KKM theorem. Qualitative optimality (necessity of the space-filling and Assouad–Nagata hypotheses, and that n+1 cannot be replaced by bounded Assouad–Nagata dimension) is also argued. Corollaries include asdim of ball intersection graphs in R^n lying in {n, n+1} (Theorem 1.2) and a bound for compact convex sets of bounded aspect ratio (Corollary 1.6). Second (Theorem 1.7/1.9): under mild non-covering and connectivity hypotheses, passing from a family of closed connected sets to the family of their boundaries does not change the intersection asymptotic dimension (up to the max{·,1} correction), yielding n ≤ asdim(sphere intersection graphs in R^n) ≤ n+1 for n ≥ 2 (Corollary 1.8). The proofs proceed through a reduction to finite graphs (Theorem 2.2), weak-diameter coloring characterizations (Lemma 2.4), an inductive extension argument for augmentation families (§3–","tokens_in":28548,"tokens_out":7712,"duration_ms":557960,"significance":"If correct, this is a strong and definitive contribution: it determines the optimal asymptotic-dimension bound (n+1) for intersection graphs of space-filling families, settling quantitatively a line of work initiated by Dvořák–Norin [7] and improving their 2n+1 to the exact n+1 with a matching lower bound. Theorem 1.4 is genuinely parameter-free in the relevant sense — no fitted constants, an explicit self-contained construction in the simplex, and a classical topological fixed-point ingredient — and the qualitative optimality discussion (thin long boxes realizing expanders; ambient asdim 0 in the lower-bound construction; infinite Assouad–Nagata dimension of ball intersection graphs) demonstrates the hypotheses are necessary, not an artifact of the proof. Corollary 1.8 improves the concurrent 2n+2 bound of Davies–Georgakopoulos–Hatzel–McCarty [5] on Georgakopoulos's sphere question, and the invariance theorem (1.7/1.9) is a clean, natural statement of independent interest with a sharp max{·,1} correction term illustrated by an explicit example. The technical development in §5 (controlled colorings pulled back along (α,β,r)-constrained maps, extension lemmas 5.8–5.10, and the induc","major_comments":[{"comment":"Theorem 7.1 is the sole external ingredient in the proof of the tightness theorem (Theorem 1.4, via Lemma 7.4), but it is stated in a 'dual' form that does not match the classical KKM theorem proved in the cited reference [9] (Knaster–Kuratowski–Mazurkiewicz 1929). The classical statement asserts non-empty total intersection for covers respecting the face structure; the statement used here — a closed cover of the (n+1)-simplex with no (n+2)-fold point has some member meeting every facet — is a known equivalent (essentially the Lebesgue covering dimension of the simplex, or the KKM theorem applied to a suitable barycentric refinement / nerve map), but the implication is not immediate and is load-bearing for the optimality claim of the whole paper. Please supply a short derivation of Theorem 7.1 from the standard KKM theorem (or Sperner's lemma), or a precise citation where this exact form","section":"§7, Theorem 7.1"}],"minor_comments":[{"comment":"Reference [6] (Dranishnikov–Smith, On asymptotic Assouad–Nagata dimension) appears in the bibliography but is never cited in the text; either cite it where relevant (e.g., the discussion of Assouad–Nagata vs asymptotic dimension in §1) or remove it.","section":"References"},{"comment":"Footnote 2 states 'All graphs are finite and simple in this paper unless otherwise specified', but infinite intersection graphs are central to the paper (int-asdim is defined via possibly infinite subfamilies, and Theorem 2.2/Lemma 2.3 exist precisely to handle them). Consider rewording to avoid confusion on first reading.","section":"§1, footnote 2"},{"comment":"Hyphenation of 'f-space-filling' is inconsistent: 'f-space filling' (without the second hyphen) appears in the statements of Theorem 1.4, Lemma 6.3, and Theorem 6.4.","section":"§1, §6"},{"comment":"Lemma 2.4 is quoted with weak diameter measured 'in G^ℓ', while the cited [3, Proposition 1.17] and the applications later in the paper (e.g., the proof of Theorem 6.4, 'weak diameter in I(S) (and hence in (I(S))^r)') move between weak diameter in G and in G^ℓ. A sentence stating the convention and the inequality dist_{G^ℓ} ≤ dist_G that justifies the parenthetical 'and hence' would make §2 and §6 easier to audit.","section":"§2, Lemma 2.4; §6, proof of Theorem 6.4"},{"comment":"In Lemma 7.2(1), the n=1 case is dismissed as 'easy to verify'; since the inductive midpoint argument given for n>1 is the only hint, one line for n=1 (or a uniform argument) would be helpful. Similarly, in Lemma 7.3 the implication 'maximality of i0 gives 2^{i0+1} ≤ 8√n/s' is the key arithmetic step and could be made explicit.","section":"§7, Lemmas 7.2–7.3"},{"comment":"Lemma 6.2's statement promises a (k,2)-centered tree-decomposition while the proof establishes the stronger (k,1)-centered property via [7, Lemma 8]. Since Claim 1 of Lemma 6.3 only uses the (k,2) version (inflated to (k,4b+2) after pullback), this is harmless, but stating the stronger conclusion would slightly simplify the audit of the constants in §6.","section":"§6, Lemma 6.2"},{"comment":"In the discussion after Theorem 1.7, the example showing max{int-asdim(F),1} ≠ int-asdim(F) uses the family of tangent circles C_i; a one-line verification that the corresponding disks pairwise intersect (e.g., they are nested/concentric in pairs) would save the reader a computation.","section":"§1.2"},{"comment":"It may be worth one remark in §7 or the introduction on whether the present framework offers any route toward closing the residual gap between n and n+1 for balls (Question 1.1) and toward the conjecture asdim = n of [5] for spheres — e.g., whether the obstruction is in the n+2-coloring step of Lemma 5.11 or is intrinsic. This is purely a suggestion for context, not a requirement.","section":"§1.1, §7"}],"recommendation":"minor_revision","confidential_remarks":"The manuscript is technically careful and self-contained beyond clearly identified black boxes; the relation to the concurrent Davies–Georgakopoulos–Hatzel–McCarty paper [5] is disclosed transparently, with proper attribution of the motivating question and their conjecture. The only item I would ask the authors to address before acceptance is the documentation of the KKM variant (Theorem 7.1); I verified the rest of the load-bearing arguments (scale matching in §5–6, the space-filling count in Lemma 7.3, the cover properties in Lemma 7.4) to my satisfaction. Recommend minor revision; no re-review of substance should be needed."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"The punchline is simple: under the natural packing condition they call f-space-filling, intersection graphs of subsets of an Assouad-Nagata-n space have asymptotic dimension at most n+1, and the bound is tight. That improves the earlier 2n+1 result and immediately gives the ball and bounded-aspect-ratio convex cases. The second theorem says that, under mild connectivity and non-covering hypotheses, you can pass to boundaries (or closed augmentations) without changing the intersection asymptotic dimension by more than the max-with-1 adjustment; spheres in R^n therefore sit at n or n+1.\n\nWhat is actually new is the quantitative jump, the matching lower-bound construction via a discrete KKM argument on a simplex grid, and the invariance statement. The proofs are modular and written out: weak-diameter colorings, reduction to finite graphs, the tree-decomposition machinery for metric spaces in Section 5, preservation of space-filling under shallow unions, and an inductive coloring argument. The lower-bound packing count and the scale-matching in the coloring lemmas check out on a close read. Reliance on prior centered-tree-decomposition and coloring lemmas (their own and others’) is ordinary modular citation, not circularity.\n\nSoft spots are modest and already flagged by the authors. The space-filling hypothesis is essential; without it you recover expanders from thin boxes. The invariance needs the global non-covering condition. The remaining n-versus-n+1 gap for balls and spheres is left open (concurrent work only reached 2n+2). No code or formal verification, but this is pure combinatorial metric geometry and does not need them.\n\nThis is for people working on asymptotic dimension of geometric graph classes, separators, and coarse geometry. It deserves a serious referee. I would bring it to reading group and expect to cite the n+1 bound and the sphere corollary when those come up.","headline":"Clean improvement of the Dvořák–Norin bound from 2n+1 to the optimal n+1 for space-filling families, plus a useful boundary-invariance theorem that pins spheres to n or n+1.","tokens_in":29741,"tokens_out":502,"would_cite":true,"duration_ms":15224,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["05C12","05C62","51F30","54F45"],"pacs":[],"model":"grok-4.5","headline":"Intersection graphs of space-filling families in spaces of Assouad-Nagata dimension n have asymptotic dimension at most n+1.","keywords":["asymptotic dimension","Assouad-Nagata dimension","intersection graphs","space-filling families","nerve theorem","sphere graphs","weak diameter coloring","tree-decomposition"],"falsifier":"Produce an f-space-filling family inside a space of Assouad-Nagata dimension n whose finite intersection graphs have unbounded weak-diameter monochromatic components under every (n+2)-coloring of every power; the paper’s own discrete-simplex construction already realizes the matching lower bound of n+1.","tokens_in":29297,"feed_emoji":"📐","tokens_out":1032,"duration_ms":37305,"temperature":0.7,"pith_summary":"The paper proves a nerve-type theorem for asymptotic dimension: if a family of sets packs reasonably inside a metric space of Assouad-Nagata dimension n, then the graph that records pairwise intersections has asymptotic dimension at most n+1. The packing rule is simple—every ball of radius r meets at most f(r/s) pairwise-disjoint members of diameter at least s—and is both necessary and sufficient for the bound. The result is sharp: a discrete construction inside the Euclidean simplex forces asymptotic dimension at least n+1, and it immediately caps the asymptotic dimension of ball graphs and of intersection graphs of compact convex sets of bounded aspect ratio in R^n by n+1. A second invariance theorem shows that, under mild connectivity and non-covering hypotheses, replacing each set by its boundary (or by a closed connected augmentation) changes the intersection asymptotic dimension by at most one; consequently sphere graphs in R^n also sit between n and n+1. Together the theorems let one read the large-scale dimension of geometric intersection graphs directly from the ambient geometry.","feed_headline":"Nice set families in n-space yield graphs of asdim ≤ n+1","feed_subtitle":"Packing plus Assouad-Nagata dimension caps intersection-graph dimension; balls and spheres included.","key_machinery":"The f-space-filling packing condition, which limits how many large pairwise-disjoint members can meet any ball. Combined with Assouad-Nagata control functions it produces (k,R)-centered tree-decompositions of the metric space of sets; those decompositions yield controlled weak-diameter colorings of graph powers and therefore bound asymptotic dimension.","core_discovery":"If F is an f-space-filling family of subsets of a metric space of Assouad-Nagata dimension at most n, then every intersection graph of a subfamily of F has asymptotic dimension at most n+1. The bound is quantitatively tight by an explicit space-filling construction in R^n that forces asymptotic dimension at least n+1. Under mild connectivity assumptions, the intersection asymptotic dimension of a family of closed connected sets equals that of their boundaries up to a possible additive 1.","pith_inferences":["The same packing-plus-Assouad-Nagata template should bound asymptotic dimension for other roundish families (bounded-eccentricity ellipsoids, geodesic balls in manifolds of bounded geometry).","The invariance theorem implies that many hollow geometric classes—sphere graphs, annulus graphs—inherit their large-scale dimension from the solid bodies they bound.","A natural sharpening left open is whether the upper bound drops from n+1 to n for balls or spheres, matching a conjecture mentioned for sphere graphs.","The tree-decomposition and weak-diameter-coloring machinery developed for the metric of sets can be reused for other nerve-type questions in coarse geometry."],"forward_implications":["Intersection graphs of closed balls in R^n have asymptotic dimension between n and n+1.","The same upper bound holds for any family of compact convex sets of bounded aspect ratio in R^n.","Intersection graphs of spheres (or of connected sets obtained by removing interior points from balls) in R^n have asymptotic dimension n or n+1 for n≥2.","Assouad-Nagata dimension, not ordinary asymptotic dimension, is the scale-invariant ambient invariant that controls the intersection graphs.","Replacing sets by closed connected augmentations or by their boundaries changes intersection asymptotic dimension by at most 1."],"fun_headline_variants":["Assouad-Nagata dim n plus packing bounds intersection graphs by asdim n+1","Space-filling families in AN-dim n force intersection asdim at most n+1","Compact convex sets of bounded aspect in R^n give graphs of asdim ≤ n+1","Intersection asdim of connected sets equals that of their boundaries","Spheres in R^n yield intersection graphs of asdim n or n+1"],"cache_read_input_tokens":16512,"weakest_assumption_plain":"The family must obey a uniform packing bound: no ball may meet too many large pairwise-disjoint members; without it, thin long boxes already realize intersection graphs of infinite asymptotic dimension.","fun_headline_variants_meta":{"raw":{"variants":["Assouad-Nagata dim n plus packing bounds intersection graphs by asdim n+1","Space-filling families in AN-dim n force intersection asdim at most n+1","Compact convex sets of bounded aspect in R^n give graphs of asdim ≤ n+1","Intersection asdim of connected sets equals that of their boundaries","Spheres in R^n yield intersection graphs of asdim n or n+1"]},"model":"grok-4.5","effort":"low","cost_usd":0.003723,"raw_usage":{"total_tokens":1297,"prompt_tokens":906,"num_sources_used":0,"completion_tokens":112,"cost_in_usd_ticks":37228000,"prompt_tokens_details":{"text_tokens":906,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":279,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":906,"tokens_out":112,"duration_ms":5781,"temperature":1.0,"reasoning_tokens":279,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-31T22:33:21.314449+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"Produce an f-space-filling family inside a space of Assouad-Nagata dimension n whose finite intersection graphs have unbounded weak-diameter monochromatic components under every (n+2)-coloring of every power; the paper’s own discrete-simplex construction already realizes the matching lower bound of n+1.","supporting_citations":[],"review_version":1}