{"id":"f4b2d68d-36a9-45bb-b20d-8a6c2ddd58b1","arxiv_id":"1908.07814","paper_version":3,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":8.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"The paper defines asymptotic expanders, characterizes quasi-locality of averaging projections via them, and proves the uniform quasi-local algebra is nuclear iff the space has Property A.","lead":"Mathematicians introduce a new family of sparse graphs, called asymptotic expanders, and show they control when an averaging projection is quasi-local. They also prove the associated operator algebra is nuclear exactly when the underlying space has Property A.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified; Lemma 5.3 is unproved but its statement is a standard truncation/renormalization equivalence, so the main theorem stands.","rationale":"The paper's central claim is the equivalence in Theorem 5.5, and the only unproved ingredient I could identify is Lemma 5.3, which the reader also flagged. I agree that this is where a critic should look, but on inspection the lemma is correct: the strong summability condition (c) is exactly what allows one to pass from asymptotically supported maps to uniformly supported ones by truncation to B(x,S) and renormalization; the error is controlled by the uniform tail mass, which tends to zero. Hence the converse direction (3) => (1) in Theorem 5.5 is supported. I checked the other potentially delicate steps: in Theorem 3.11, the induction sets N_{n-kR0}(A_n) remain subsets of N_n(A_n) and therefore stay below half of X_n, so the expander condition remains applicable; in Theorem 3.15, the negation of asymptotic expanders is handled correctly by choosing c = 1/m and passing to subsequences; in Theorem 4.4, the contradiction with Proposition 3.9 is quantitative and valid. The remaining minor omissions are the missing proof of Lemma 5.3 and the lightly justified adaptation of [16, Lemma 1] in Corollary 3.16, neither of which affects the main theorem. No significant objection identified.","tokens_in":21394,"tokens_out":15143,"duration_ms":154641,"concrete_test":"Supply the missing proof of Lemma 5.3: for fixed R, eps, choose delta and S with sup_x sum_{z notin B(x,S)} |eta_x(z)|^2 < delta; define xi_x = chi_{B(x,S)} eta_x / ||chi_{B(x,S)} eta_x|| and verify ||xi_x - xi_y|| <= (eps + 2 sqrt(delta))/(1-delta) + 2 delta/(1-delta)^2. If this cannot be made < eps by choosing delta small, the converse of Theorem 5.5 collapses; otherwise the theorem is confirmed.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is Theorem 5.5: C*_uq(X) is nuclear iff X has Property A. The converse direction (3) => (1) relies on Lemma 5.3, which is stated without proof. This is indeed the weakest point in the paper, as the reader noted. However, the omitted proof is routine and the lemma is true: given eta satisfying (a)-(c), truncate to B(x,S) and renormalize. Condition (c) makes the omitted mass uniformly small, so the normalization error is o(1), and condition (b) passes through the truncation with arbitrarily small loss. Thus the unproved lemma is not a load-bearing weakness. The other potentially delicate steps — the iteration in Theorem 3.11, the negation argument in Theorem 3.15, and the quantitative contradiction in Theorem 4.4 — are internally consistent. Corollary 3.16 relies on a lightly justified adaptation of [16, Lemma 1], but this is not essential to the main nuclearity theorem.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces a weakening of expanders called asymptotic expanders, shows that the averaging projection on a coarse disjoint union of finite metric spaces is quasi-local if and only if the sequence is a sequence of asymptotic expanders (Theorem 3.11), proves that this property is preserved under piece-preserving coarse equivalences and under coarse equivalence of D-connected pieces (Theorem 3.15 and Corollary 3.16), and shows that a coarse disjoint union of asymptotic expanders is not uniformly locally amenable (Theorem 4.4). The main result, Theorem 5.5, states that for a metric space X of bounded geometry the following are equivalent: X has Property A; the uniform quasi-local algebra C*_uq(X) is nuclear; the inclusion C*_u(X) → C*_uq(X) is nuclear; all ghost operators in C*_uq(X) are compact; and ℓ∞(X) separates ideals of C*_uq(X). The paper also gives a Cartan subalgebra characterization of the equality C*_u(X)=C*_uq(X) (Proposition 6.1) and discusses open questions.","tokens_in":21590,"tokens_out":18244,"duration_ms":152908,"significance":"If the results hold, they provide a new and sharp geometric test for the difference between uniform Roe algebras and uniform quasi-local algebras, and a new characterization of Property A. The notion of asymptotic expanders is natural and likely to be useful. The proofs of the main theorems are detailed and internally consistent, and the paper is careful with hypotheses. The central theorem is shown to reduce to an elementary (but omitted) reformulation of Property A in Lemma 5.3, which the authors state is standard; this is the only point requiring attention. The paper also gives credit to related work and clearly states remaining open questions.","major_comments":[],"minor_comments":[{"comment":"Lemma 5.3 is stated without proof. Since it is used in the proof of Theorem 5.5(3)⇒(1), the authors should include the short proof (truncation and renormalization) or provide a reference.","section":"Section 5, Lemma 5.3"},{"comment":"In the proof of Corollary 3.16, the adaptation of [16, Lemma 1] from connected graphs to D-connected metric spaces is only sketched; please add a brief justification.","section":"Section 3.3, Corollary 3.16"},{"comment":"There are minor typos: in Lemma 5.4, 'Cauchy-Schwartz' should be 'Cauchy-Schwarz'; in the statement of Lemma 3.18, the displayed inequality appears to have a missing fraction bar and should be formatted as 1 - N_Y(D)·|∂_D(B)|/|B| ≥ 1/2.","section":"Section 5, Lemma 5.4 and Section 3.2, Lemma 3.18"}],"recommendation":"accept","confidential_remarks":"The manuscript is within the scope of the journal and the results are likely to be of broad interest to the coarse geometry and C*-algebra communities. The reliance on previous work of the same authors (e.g., [34]) is appropriate. I recommend acceptance."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe headline is that Theorem 5.5 delivers: for bounded geometry, C*_uq(X) is nuclear iff X has Property A. That closes nuclearity as a way to distinguish the uniform quasi-local algebra from the uniform Roe algebra, which is exactly the motivating question. The paper's own new tool, asymptotic expanders, is the more interesting contribution. Definition 3.12 is a genuine weakening of expanders, and Example 3.7 plus Corollary 3.13 show the notion is not vacuous. Theorem 3.11, identifying quasi-locality of the averaging projection with the asymptotic expander condition, is a clean and original characterization; the expansion argument in the proof is sound.\n\nThe coarse invariance results (Theorem 3.15, Corollary 3.16) and the non-ULA Theorem 4.4 are solid. Theorem 4.4's contradiction argument with disjoint neighborhoods is careful and works. The nuclearity proof follows the Brown-Ozawa strategy but replaces uniform support with strong summability, and the construction of the Property A vectors is carried out in detail. The Cartan subalgebra proposition is a nice extra. The citation pattern is fine: the paper leans on earlier work by the same group [34], but that result is independent and not being smuggled in; no circularity.\n\nSoft spots are minor. Lemma 5.3 is stated without proof and is load-bearing for the converse of Theorem 5.5. I agree with the stress-test note: the lemma is a routine truncation/renormalization characterization of Property A, and condition (c) makes the normalization error vanish uniformly, so I do not see a real gap. Still, the authors should put the proof in the final version; leaving a load-bearing equivalence to the reader is not ideal. The adaptation of Khukhro--Valette's lemma in Corollary 3.16 is lightly justified, but the argument is plausible and not central to the main theorem.\n\nThe paper is for researchers working on coarse geometry, Roe algebras, and Property A. It deserves a serious referee. I would accept it after the authors supply the proof of Lemma 5.3 and tighten the Corollary 3.16 justification. No reason to desk reject.","headline":"Solid paper: Theorem 5.5 closes nuclearity as a separator between uniform quasi-local and uniform Roe algebras, and the asymptotic expander notion is a genuinely useful new tool; the only real gap is an unproved but standard lemma.","tokens_in":22119,"tokens_out":2183,"would_cite":true,"duration_ms":20823,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["46H35","46L05","20F65","05C99"],"pacs":[],"model":"deepseek-v4-flash","headline":"A uniform quasi-local algebra is nuclear precisely when its metric space has Property A, and the averaging projection detects a new class of 'asymptotic expanders'.","keywords":["quasi-local algebras","uniform Roe algebras","Property A","asymptotic expanders","averaging projection","nuclearity","coarse geometry","expander graphs"],"falsifier":"Find a bounded-geometry metric space $X$ that satisfies the strong-summability condition of Lemma 5.3(2) but fails Property A; then the unproved lemma is false and the proof of Theorem 5.5 loses its converse. Alternatively, compute the limit $\\sup_{A,B \\subseteq X_n,\\ d(A,B) \\ge R} |A||B|/|X_n|^2$ for a concrete candidate sequence: if Theorem 3.11 is right, this limit vanishes exactly when every $\\alpha$-sized subset has $|\\partial_R A| > c|A|$ for some $c,R$ independent of $n$.","tokens_in":21188,"feed_emoji":"📐","tokens_out":5726,"duration_ms":136086,"temperature":0.7,"pith_summary":"This paper tries to establish that the uniform quasi-local algebra $C^*_{uq}(X)$ of a bounded-geometry metric space is nuclear if and only if $X$ has Property A. It also introduces a weakening of expander graphs, called asymptotic expanders, and proves that the averaging projection over a coarse disjoint union is quasi-local precisely when the pieces form such a sequence. The point of the comparison is that quasi-locality is a proposed intrinsic test for membership in the uniform Roe algebra, and the paper shows that nuclearity cannot tell the two algebras apart. A sympathetic reader should care because the results convert a $C^*$-algebraic property, nuclearity, into a geometric one, Property A, and expose a concrete operator whose quasi-locality is a sharp geometric test.","feed_headline":"Quasi-local algebra is nuclear only on Property A spaces","feed_subtitle":"New 'asymptotic expanders' make the averaging projection quasi-local and sharpen where the two algebras diverge.","key_machinery":"The load-bearing objects are the uniform quasi-local algebra $C^*_{uq}(X)$, the $C^*$-algebra of all operators on $\\ell^2(X)$ whose matrix entries decay uniformly with distance between their indices, and the averaging projection $P_X$ onto the span of the constant functions on each piece $X_n$. The mechanism connecting them is the norm identity $\\|\\chi_A P_F \\chi_B\\| = \\sqrt{|A||B|}/\\sqrt{|F|}$, which turns quasi-locality of $P_X$ into a statement about products of sizes of far-apart subsets. The proof of the nuclearity converse relies on an auxiliary characterization of Property A, Lemma 5.3, in which the usual uniformly bounded supports of Property A vectors are replaced by strong summability at infinity; quasi-locality of the approximating operators supplies exactly that summability.","core_discovery":"The central discovery is a pair of equivalences. On the geometric side, for a coarse disjoint union $X = \\bigsqcup_n X_n$ of finite metric spaces, the averaging projection $P_X$ is quasi-local if and only if $\\{X_n\\}$ is a sequence of asymptotic expanders, meaning that for every $\\alpha > 0$ there are $c \\in (0,1)$ and $R > 0$ such that every subset $A$ of size between $\\alpha |X_n|$ and $|X_n|/2$ satisfies $|\\partial_R A| > c|A|$. On the $C^*$-algebraic side, for a bounded-geometry metric space $X$, the uniform quasi-local algebra $C^*_{uq}(X)$ is nuclear if and only if $X$ has Property A; the same equivalence holds for nuclearity of the inclusion $C^*_u(X) \\hookrightarrow C^*_{uq}(X)$, for compactness of all ghost operators in $C^*_{uq}(X)$, and for $\\ell^\\infty(X)$ separating ideals. Consequently nuclearity cannot distinguish $C^*_{uq}(X)$ from $C^*_u(X)$. Being a sequence of asymptotic expanders is also a coarse property when the pieces are sufficiently connected, and any coarse disjoint union of asymptotic expanders fails uniform local amenability and hence Property A.","pith_inferences":["If Lemma 5.3 turns out to fail, the present proof of the converse in Theorem 5.5 would need a different route, although the theorem itself could still be true.","The quasi-locality criterion for $P_X$ is a finite, checkable condition on boundary growth; it could be used computationally to test random graph sequences for this weak form of expansion.","One might expect a coarse embedding obstruction: if asymptotic expanders are eventually shown to be non-embeddable in Hilbert space, then Proposition 7.4 would force $C^*_u(X) = C^*_{uq}(X)$ for all such spaces, aligning them with ordinary expanders.","The strong-summability form of Property A in Lemma 5.3 suggests a possible bridge to metric sparsification and operator norm localization, where similar decay conditions replace uniform support bounds."],"forward_implications":["If $X$ has Property A, then $C^*_u(X) = C^*_{uq}(X)$, so the two algebras are identical and nuclearity is the same property for both.","A coarse disjoint union of asymptotic expanders is never uniformly locally amenable, and in particular never has Property A.","If an asymptotic expander sequence coarsely embeds into Hilbert space, its averaging projection would be quasi-local but outside $C^*_u(X)$, giving the first strict inclusion $C^*_u(X) \\subsetneq C^*_{uq}(X)$ and answering the corresponding open question from the quasi-locality literature.","Every expander sequence is an asymptotic expander sequence, but the converse is false: the deformed expander sequence of Example 3.7 is asymptotic yet not an expander sequence.","Under a piece-respecting coarse equivalence, which is automatic for $D$-connected pieces such as connected graphs, the asymptotic expander property is preserved."],"supporting_citations":[{"why":"Supplies the equality $C^*_u(X) = C^*_{uq}(X)$ under Property A, the baseline inclusion that the paper extends and tests.","marker":"[34]"},{"why":"Provides the theorem that $C^*_u(X)$ is nuclear if and only if $X$ has Property A, the anchor for the nuclearity equivalence.","marker":"[30]"},{"why":"Defines uniform local amenability and proves Property A implies it, the background used in Theorem 4.4.","marker":"[3]"},{"why":"Shows coarse embeddability into Hilbert space excludes non-compact ghost projections, used in Proposition 7.4.","marker":"[11]"},{"why":"Supplies the ideal-structure result connecting ideal separation by $\\ell^\\infty(X)$ to compactness of ghost operators.","marker":"[2]"},{"why":"Provides the nuclearity and Hilbert-module machinery used in the proof of the converse direction of Theorem 5.5.","marker":"[5]"}],"fun_headline_variants":["New asymptotic expanders bridge quasi-local algebras and Property A","Nuclearity of quasi-local algebra: equivalent to Property A","Averaging projection quasi-local iff asymptotic expanders","Asymptotic expanders: a weaker expander for coarse geometry","Quasi-local algebra nuclearity: exactly Property A spaces"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The converse direction of the nuclearity theorem rests on an unproved characterization of Property A, Lemma 5.3, which replaces uniformly bounded supports with a strong-summability condition at infinity; if that characterization is false, the argument collapses.","fun_headline_variants_meta":{"raw":{"variants":["New asymptotic expanders bridge quasi-local algebras and Property A","Nuclearity of quasi-local algebra: equivalent to Property A","Averaging projection quasi-local iff asymptotic expanders","Asymptotic expanders: a weaker expander for coarse geometry","Quasi-local algebra nuclearity: exactly Property A spaces"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000933,"raw_usage":{"total_tokens":3980,"prompt_tokens":921,"completion_tokens":3059,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":537,"completion_tokens_details":{"reasoning_tokens":2977}},"tokens_in":537,"tokens_out":3059,"duration_ms":453363,"temperature":1.0,"reasoning_tokens":2977,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:55:00.038454+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Find a bounded-geometry metric space $X$ that satisfies the strong-summability condition of Lemma 5.3(2) but fails Property A; then the unproved lemma is false and the proof of Theorem 5.5 loses its converse. Alternatively, compute the limit $\\sup_{A,B \\subseteq X_n,\\ d(A,B) \\ge R} |A||B|/|X_n|^2$ for a concrete candidate sequence: if Theorem 3.11 is right, this limit vanishes exactly when every $\\alpha$-sized subset has $|\\partial_R A| > c|A|$ for some $c,R$ independent of $n$.","supporting_citations":[{"cited_title":"Quasi-locality and Property A","cited_arxiv_id":null,"evidence_quote":"Supplies the equality $C^*_u(X) = C^*_{uq}(X)$ under Property A, the baseline inclusion that the paper extends and tests."},{"cited_title":"The coarse Baum-Connes conjecture and groupoids","cited_arxiv_id":null,"evidence_quote":"Provides the theorem that $C^*_u(X)$ is nuclear if and only if $X$ has Property A, the anchor for the nuclearity equivalence."},{"cited_title":"Niblo, J´ an ˇSpakula, Rufus Willett, and Nick Wright","cited_arxiv_id":null,"evidence_quote":"Defines uniform local amenability and proves Property A implies it, the background used in Theorem 4.4."},{"cited_title":"Fibred coarse embeddings, a-T-menability and t he coarse analogue of the Novikov conjecture","cited_arxiv_id":null,"evidence_quote":"Shows coarse embeddability into Hilbert space excludes non-compact ghost projections, used in Proposition 7.4."},{"cited_title":"Ideal structure and pure infinite ness of ample groupoid C∗ - algebras","cited_arxiv_id":null,"evidence_quote":"Supplies the ideal-structure result connecting ideal separation by $\\ell^\\infty(X)$ to compactness of ghost operators."},{"cited_title":"Brown and Narutaka Ozawa","cited_arxiv_id":null,"evidence_quote":"Provides the nuclearity and Hilbert-module machinery used in the proof of the converse direction of Theorem 5.5."}],"review_version":1}