{"id":"02ee88f3-20c2-4a66-9d5d-d1f8667adc4d","arxiv_id":"2505.10334","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The Roller boundary of a countable CAT(0) cube complex of dimension n has Borel asymptotic dimension at most n, and its component equivalence relation is smooth.","lead":"This paper proves that the Roller boundary of any finite-dimensional CAT(0) cube complex, viewed as a Borel graph, has Borel asymptotic dimension at most the dimension of the complex. Generalists might care because it shows that potentially wild geometric boundaries are measurably simple, and it supplies new tools for Borel combinatorics.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The componentwise application of Wright's theorem in Proposition 7.10 needs uniform cobornologous constants; the paper does not state or prove this uniformity.","rationale":"The paper's central claim depends on a Borel version of Wright's construction. The Borel adaptation appears careful: the hyperplane spaces, controlled coloring, interpolation map, and projection are all shown Borel through standard descriptive set-theoretic arguments. I found no internal inconsistency in the componentwise Borel transfer itself. The soft spot is the passage from componentwise geometric estimates to a globally cobornologous map in Proposition 7.10. The proof states that each restriction is cobornologous, but Lemma 7.9 needs a single uniform bound across all components. This is not just a cosmetic issue: non-uniform R(r) would make the uniform boundedness conclusion fail, as in disjoint unions of pairs of points at growing distances. The cited Wright results may well provide uniform constants, since the controlled coloring parameter depends only on D, but the paper neither states this nor proves it. The reader's weakest assumption flagged the same region of the proof; I sharpen it to the uniformity of the cobornologous bounds. If the uniformity is confirmed, the theorem should stand; if not, the argument has a real gap. Hence a conditional recommendation is appropriate.","tokens_in":33453,"tokens_out":42904,"duration_ms":418984,"concrete_test":"Inspect the statements of [Wri12, Lemma 4.7, Lemma 4.8, Theorem 4.9]. Determine whether the cobornologous function R(r) is given by a formula depending only on D and r, e.g. R(r)=(3^{D-1}D+1)r, or whether it is allowed to depend on the complex X. If it is uniform in X, add an explicit uniformity statement to Proposition 7.10 and the proof is complete. If not, construct a countable Borel median graph with components witnessing R→∞, such as intervals of growing length for D=1 or suitable D-dimensional examples, and check whether asdim_B>D, which would contradict Proposition 7.10.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"Proposition 7.10 obtains a global ε-Lipschitz cobornologous Borel map f: X→eY by applying [Wri12, Lemma 4.8 and Theorem 4.9] componentwise and iterating. The proof only asserts that the restriction to every G-component is a (3^{D-1}D)/(3^{D-1}D+1)-Lipschitz cobornologous map to its quotient. Cobornologousness is a global property: it requires a single R(r) valid for all pairs. If the R supplied by Wright's theorem depends on the individual component, not just on the dimension bound D, then for a Borel graph with infinitely many components the constants can be unbounded, the quantity sup_v diam_Y(f^{-1}(St_{T2}(v))) may be infinite, and Lemma 7.9's uniform boundedness of F_r(U_ℓ) fails. Theorem 2.24 and [Wri12, Lemma 4.7] are imported as black boxes without defining 'controlled coloring' or stating whether the control function and cobornologous constants are uniform in X. Since Lemma 7.9 requires exactly this uniformity, the proof of Proposition 7.10, and hence Theorem 1.1, depends on an unstated uniformity assumption.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves Theorem 1.1: for any countable CAT(0) cube complex X of dimension n, the Borel median graph on the Roller boundary R(X) has Borel asymptotic dimension at most n, and the associated component equivalence relation is smooth. The proof passes through a more general statement, Proposition 1.2, for standard Borel spaces equipped with a countable Borel median graph whose components are CAT(0) cube complexes of dimension at most D and whose component equivalence relation is smooth. The author develops a Borel version of Wright's construction: he defines a standard Borel structure on the union of the CAT(0) cube complexes obtained by gluing cubes to the components, introduces Borel spaces of hyperplanes, proves a Borel quotient theorem, establishes Borel measurability of the Wright interpolation and projection maps, and finally uses a Borel triangulation argument (Lemma 7.9) to deduce the asymptotic-dimension bound. The paper is careful and detailed in its descriptive-set-theoretic arguments, and it does not assume local finiteness of the Borel graphs.","tokens_in":33684,"tokens_out":14103,"duration_ms":143798,"significance":"If the proof is correct, the result is a genuine strengthening of Wright's asymptotic-dimension theorem in the Borel setting: it gives a uniform upper bound on the Borel asymptotic dimension of Roller boundaries in terms of the dimension of the cube complex, without local finiteness or finiteness of hyperplanes. The paper also introduces reusable tools: a canonical Borel structure on the union of cubes over a Borel median graph, a Borel analogue of the Sageev-Roller duality, and a Borel version of Wright's projection. The Borel measurability arguments are mostly carried out with standard tools (Arsenin-Kunugui, smooth CBER selectors), and the paper is transparent about the black-box use of [Wri12]. The main conceptual contribution—that a naturally occurring family of non-locally-finite Borel median graphs has finite Borel asymptotic dimension—is interesting and timely.","major_comments":[{"comment":"The proof concludes that the iterated map f : X → eY is ε-Lipschitz and cobornologous, and Lemma 7.9 then uses a single R(r) uniformly over all f^{-1}(St_{T2}(v)). However, the proof only states that the restriction of each intermediate map to every G-component is cobornologous; if the R supplied by [Wri12, Lemma 4.7] can depend on the individual component, then sup_v diam_Y(f^{-1}(St_{T2}(v))) need not be finite and the argument for Lemma 7.9 fails. Please add an explicit statement that the cobornologous control in [Wri12, Lemma 4.7] depends only on D and on the controlled-coloring constant 3^{D-1}D, hence is uniform across the Borel family, and verify that this uniformity is preserved under the N-fold iteration. If this uniformity is not present in [Wri12], then the proof of Proposition 7.10 is incomplete as written.","section":"Section 7.10, proof of Proposition 7.10 (with Definition 7.8 and Lemma 7.9)"}],"minor_comments":[{"comment":"Since 'controlled coloring' is deliberately used as a black box, the paper should either include the definition from [Wri12, Definition 2.3] or state explicitly the uniformity property needed in Proposition 7.10, so that the reader can verify the componentwise application and the global cobornologousness.","section":"Section 2.2.1, Theorem 2.24"},{"comment":"There is a typo: 'countble union' should be 'countable union'.","section":"Section 3, proof of Lemma 3.9"},{"comment":"The proof sketch introduces constants δ1 and δ2 by assertion. It would be easier for the reader to check if the definitions of δ1 and δ2 were made explicit, for example by bounding the relevant segment lengths in terms of the fixed triangulation T2 of [0,1]^D.","section":"Section 7, Lemma 7.7(2)"},{"comment":"In the sentence 'By uniqueness of ξ_x, the map f satisfies x E^{R(X)}_{G_{R(X)}} y ⇐⇒ f(x) = f(y)', a brief justification that ξ_x depends only on the E-class of x would improve readability.","section":"Section 7, Theorem 7.11"},{"comment":"For D ≤ 1 the rank vector is constant 0; Remark 2.18 explains why, but adding a cross-reference to Remark 2.18 at Definition 2.17 would help the reader.","section":"Definition 2.17"}],"recommendation":"major_revision","confidential_remarks":"This is a solid and interesting paper. The referee's main concern is the uniformity of the cobornologous constants in Proposition 7.10; if the author confirms that [Wri12, Lemma 4.7] provides bounds depending only on D and the controlled-coloring constant, the gap is easily fixed and the paper should be acceptable. The Borel arguments are detailed, and the technique of gluing cubes to Borel median graphs is a useful contribution."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Nick, this one is worth reading. Oyakawa proves that for any countable finite-dimensional CAT(0) cube complex, the Borel median graph on the Roller boundary has Borel asymptotic dimension at most the dimension, and the component equivalence relation is smooth. That's a natural and previously open question in Borel combinatorics, and it sharpens Wright's classical asymptotic dimension result in the Borel setting. The proof is a serious Borel adaptation, not a black-box translation: the paper constructs a Borel extended metric space by gluing cubes to a Borel median graph, develops a Borel version of Sageev-Roller duality including a Borel space of hyperplanes, and proves a Borel projection theorem. The measurability arguments are detailed and use standard tools correctly. It also carefully avoids assuming local finiteness, which matters because Roller boundaries can have infinite degree.\n\nWhere are the soft spots? The main one is Proposition 7.10. The proof applies Wright's theorem componentwise and asserts that the resulting map is uniformly epsilon-Lipschitz and cobornologous. The Lipschitz constant is explicitly uniform (3^{D-1}D/(3^{D-1}D+1)), but cobornologousness is a global property: you need a single R(r) for all components. Wright's construction is quantitative, and I believe the control function depends only on D, so the uniformity is almost certainly true. But the paper never states this, and since it imports Wright's results as black boxes without defining 'controlled coloring', a strict referee cannot verify that claim from the text. That's a genuine expository gap, though not a fatal one. The author should either prove a lemma stating the uniform cobornologous bound or at least pin down exactly which statement in [Wri12] gives it.\n\nOtherwise, the paper checks out. No circularity, no fitted constants, and the citations look right. The proof is long, so moderate confidence is fair, but I didn't find a load-bearing flaw. The audience is Borel combinatorics and descriptive set theory people, and also geometric group theorists who care about Roller boundaries.\n\nMy recommendation: send it to peer review. The theorem deserves referee time. The only required change is clarifying the uniformity in Proposition 7.10; that can be done with a short argument or a precise reference.","headline":"A genuine Borel adaptation of Wright's construction that proves a sharp bound for Roller boundaries; the only real gap is an unstated uniformity assumption in Proposition 7.10.","tokens_in":34194,"tokens_out":4453,"would_cite":true,"duration_ms":43652,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["03E15","20F65"],"pacs":[],"model":"deepseek-v4-flash","headline":"For any countable CAT(0) cube complex of dimension n, the Borel median graph on its Roller boundary has Borel asymptotic dimension at most n, and its component equivalence relation is smooth.","keywords":["Borel asymptotic dimension","Roller boundary","CAT(0) cube complex","median graph","smooth equivalence relation","controlled coloring","Wright construction","hyperplane"],"falsifier":"A concrete search is to exhibit a countable two-dimensional CAT(0) cube complex whose Roller boundary has Borel asymptotic dimension at least 3, or contains a non-hyperfinite Borel subgraph; since finite Borel asymptotic dimension implies hyperfiniteness, such an example would contradict the theorem. The natural first test case is the locally finite 2-dimensional complex constructed in Remark 2.9, whose Roller boundary has a vertex of infinite valency.","tokens_in":33260,"feed_emoji":"📐","tokens_out":7289,"duration_ms":71684,"temperature":0.7,"pith_summary":"The paper establishes a uniform upper bound on the Borel asymptotic dimension of the Roller boundary of any countable finite-dimensional CAT(0) cube complex: if the complex has dimension n, the Borel median graph on the boundary has Borel asymptotic dimension at most n, and the equivalence relation that identifies points in the same boundary component is smooth. This is the Borel counterpart of Wright's finite-asymptotic-dimension theorem for CAT(0) cube complexes, but it is not a formal consequence of it, because Borel asymptotic dimension can be strictly larger than the usual asymptotic dimension of the components. The interest is that the bound needs no local finiteness of the boundary nor any finiteness of hyperplanes: even when the Roller boundary has vertices of infinite valency, its Borel large-scale geometry is still controlled by the dimension of the original complex.","feed_headline":"Roller boundary of n-dimensional cube complexes has Borel dim at most n","feed_subtitle":"The bound needs no local finiteness of the boundary, so even infinite-valency compactifications stay controlled by dimension.","key_machinery":"The engine is a Borelized version of Wright's construction. Given a smooth Borel median graph, the paper glues cubes to every component to form the Borel extended metric space eX, builds a standard Borel space of hyperplanes H_s(G) with halfspace relations, and then runs Wright's controlled-coloring argument componentwise: a Borel coloring c of hyperplanes, with K_c = $c^{{-1}}$(0), a quotient X to X_{K_c}, an interpolated contraction Psi_w into a larger cube complex, and a Borel contractive projection P back to the embedded complex. Iterating this contraction N times produces, for any epsilon > 0, an epsilon-Lipschitz cobornologous Borel map from the original Borel graph into another smooth Borel median graph of dimension at most D. Lemma 7.9 converts such maps into the Borel asymptotic dimension bound by a Borel triangulation of eX with a dimension-dependent separation constant delta.","core_discovery":"The central claim is Theorem 1.1: for any countable CAT(0) cube complex X of dimension n in N union {0}, the Borel median graph on the Roller boundary R(X) has Borel asymptotic dimension at most n, and the Borel equivalence relation of connected components of this graph is smooth. The proof runs through a more general proposition: any smooth countable Borel median graph whose connected components are CAT(0) cube complexes of dimension at most D has Borel asymptotic dimension at most D. The smoothness condition is essential and best possible: weakening it to hyperfiniteness is impossible, since hyperfinite Borel trees with infinite Borel asymptotic dimension exist. A corollary noted by the author is that Theorem 1.1 strengthens Wright's result, because the original complex embeds as a connected component of its Roller compactification and the classical asymptotic dimension of each component is bounded by the Borel asymptotic dimension of the whole Borel graph.","pith_inferences":["The same cube-gluing technique may apply to other natural compactifications with median-graph structure, giving Borel asymptotic dimension bounds wherever a smooth component relation is available.","Because smoothness is proven constructively via a Borel selector, the paper suggests that boundary components of a countable cube complex can be assigned basepoints in a definable way; this may have consequences for definable choices in cubical groups and actions.","A natural testable extension is whether the bound is sharp for every n: for the n-dimensional grid Z^n, which embeds as a component of its own Roller boundary, the Borel asymptotic dimension should be exactly n; verifying this would show Theorem 1.1 is optimal."],"forward_implications":["The Borel asymptotic dimension of the Roller boundary of any countable CAT(0) cube complex is bounded by the dimension of the complex, with no local-finiteness or hyperplane-finiteness assumptions.","The connected-component equivalence relation of the Roller boundary graph is smooth, so it admits a Borel selector and cannot realize Borel complexity above smoothness.","Wright's classical result that finite-dimensional CAT(0) cube complexes have finite asymptotic dimension is strengthened to the Borel setting for Roller boundaries.","Proposition 1.2 gives a general transfer: any smooth Borel median graph whose components are CAT(0) cube complexes of dimension at most D has Borel asymptotic dimension at most D.","Since finite Borel asymptotic dimension implies hyperfiniteness, the boundary component relations of countable finite-dimensional CAT(0) cube complexes are hyperfinite."],"supporting_citations":[{"why":"Supplies the controlled-coloring theorem and the Lipschitz/cobornologous estimates that the Borel construction imports componentwise.","marker":"[Wri12]"},{"why":"Defines Borel asymptotic dimension and supplies the criterion used in the definition and in the consequence that finite Borel asymptotic dimension implies hyperfiniteness.","marker":"[CJM+23]"},{"why":"Establishes the equivalence between median graphs and 1-skeleta of CAT(0) cube complexes, so each component is a cube complex.","marker":"[Che00]"},{"why":"Provides the unique boundary point selection used to prove smoothness of the component relation on the Roller boundary.","marker":"[Gen21]"},{"why":"Supplies Arsenin-Kunugui and related projection theorems used throughout for Borel measurability of countable sections.","marker":"[Kec95]"},{"why":"Supplies the descriptive-set-theoretic facts about countable Borel equivalence relations, selectors, and transversals used in the smoothness arguments.","marker":"[Tse22]"},{"why":"Provides the triangulation-based characterization of asymptotic dimension that Lemma 7.9 adapts to the Borel setting.","marker":"[Roe03]"}],"fun_headline_variants":["Borel asymptotic dim of Roller boundary ≤ dimension","Roller boundary Borel dim at most n for n-dim cube complexes","Smooth Borel median graph: asymptotic dim ≤ dimension","Countable CAT(0) cube complexes: boundary Borel dim ≤ n"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing assumption is that Wright's controlled-coloring theorem, a geometric estimate about coloring the hyperplanes of a finite-dimensional CAT(0) cube complex so that each color class is well separated, remains valid when applied componentwise to a Borel median graph; if that estimate or the Borel transfer of its Lipschitz and cobornologous bounds fails, the iterative contraction that produces the epsilon-Lipschitz cobornologous Borel map collapses and the dimension bound does not follow.","fun_headline_variants_meta":{"raw":{"variants":["Borel asymptotic dim of Roller boundary ≤ dimension","Roller boundary Borel dim at most n for n-dim cube complexes","Smooth Borel median graph: asymptotic dim ≤ dimension","Countable CAT(0) cube complexes: boundary Borel dim ≤ n"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001187,"raw_usage":{"total_tokens":4807,"prompt_tokens":757,"completion_tokens":4050,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":373,"completion_tokens_details":{"reasoning_tokens":3977}},"tokens_in":373,"tokens_out":4050,"duration_ms":30537,"temperature":1.0,"reasoning_tokens":3977,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T21:10:49.965066+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A concrete search is to exhibit a countable two-dimensional CAT(0) cube complex whose Roller boundary has Borel asymptotic dimension at least 3, or contains a non-hyperfinite Borel subgraph; since finite Borel asymptotic dimension implies hyperfiniteness, such an example would contradict the theorem. The natural first test case is the locally finite 2-dimensional complex constructed in Remark 2.9, whose Roller boundary has a vertex of infinite valency.","supporting_citations":[],"review_version":1}