{"id":"6f20c7b3-1c5b-49eb-a1a6-8adfcc6aef9d","arxiv_id":"2506.11918","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For the new Random Connection Model of simplicial complexes, a generalized Euler characteristic is shown to satisfy quantitative central limit theorems in high-intensity and large-window limits.","lead":"This paper introduces a random simplicial complex model that extends the Random Connection Model from graphs to higher-dimensional shapes and proves quantitative central limit theorems for a generalized Euler characteristic. It unifies many existing random complex models, including Čech and Vietoris-Rips complexes, within one framework.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Lemma 6.5's linear-in-|W| bounds for (25)–(29) are asserted by analogy rather than proved, and Theorem 6.6's rates collapse if any of these bounds contains an extra |W| factor.","rationale":"The reader's conditional verdict is appropriate. The CLT mechanism is clear, and the detailed estimates in Lemma 6.1 and parts of Lemma 6.5 are plausible. The single load-bearing gap is the completeness of Lemma 6.5: the increasing-window CLT and the multivariate CLT both rest on the linear-in-|W| bounds (23)–(29), but the proof only fully treats (23) and (24) and asserts the rest by analogy. I do not see an actual counterexample, and the sketch contains the right ingredients, but because the entire asymptotics depend on these bounds, a full verification of the omitted cases is needed before full acceptance. If the verification succeeds, the paper's claims should stand.","tokens_in":30555,"tokens_out":28467,"duration_ms":380985,"concrete_test":"Independently re-derive the omitted cases (25)–(29) of Lemma 6.5 for α=2. In the Mecke expansion of the left side of (28), show explicitly that the simplicial complex associated with every term in E[(Λ^2_{x1,x2}F)^4] contains the edge {x1,x2}, so the term is bounded by a constant times φ1(x1,x2) (or a product containing it). Then verify that ∫_{W^2} φ1(x1,x2)^{1/2} d(x1,x2) ≤ ν|W| under (22). If any expanded term lacks this φ1-factor, the bound becomes |W|^2 and the rate γ_6=O(|W|^{-1/2}) in Theorem 6.6 is invalid.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Theorem 6.6 and hence Theorem 7.3 depend on Lemma 6.5: each integral θ_i from (23)–(28) must be ≤ C_i|W| so that γ_i = O(|W|^{-1/2}) after division by Var = Θ(|W|). The paper proves (24) in detail and (23) in detail, but (25)–(29) are disposed of with 'in principle, this approach can be applied.' The non-obvious part is exactly where a mistake would matter: after expanding products of difference operators via the multivariate Mecke equation, every term must be a connected integral representation with the distinguished vertices (x1,x2,x3, or x) supplying the only |W| factor. For (28), for instance, one must check that each summand from E[(Λ^2_{x1,x2}F)^4] contains at least one factor φ1(x1,x2), so that after the outer exponent 1/2 the x2-integral is bounded by ν from (22); a naive bound treating this expectation as O(1) would produce |W|^2 instead of |W|. The same careful cancellation of a |W| factor is needed in (23) and (26). The estimates are probably correct, but as written the decisive step is a sketch, not a proof.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces a higher-dimensional generalization of the random connection model: vertices are drawn from a Poisson process on a Borel space, and simplices of each dimension are added independently with probabilities given by symmetric connection functions. For the resulting random simplicial complex, the paper studies the generalized Euler characteristic with an arbitrary coefficient vector. It derives first- and second-moment formulas, a Fock-space-type variance representation, and quantitative normal approximation bounds based on a transfer theorem for Poisson functionals. The main results are a central limit theorem as the intensity β tends to infinity (Theorem 6.2), a central limit theorem as the observation window grows under the global integrability condition (22) and a lower-growth condition (36) (Theorem 6.6), and a multivariate central limit theorem for simplex counts in the marked stationary case (Theorem 7.3).","tokens_in":30857,"tokens_out":26145,"duration_ms":309908,"significance":"If the missing details are supplied, this is a substantial contribution. The model unifies several existing random simplicial complex constructions, the setting of an arbitrary Borel space is genuinely general, and the results come with explicit Wasserstein and Kolmogorov distance rates that are not tied to fitted parameters. The moment formulas in Section 4 and the operator transfer in Theorem 5.1 are carefully developed, and the proof of the β→∞ CLT is convincing. The paper also gives a transparent Fock-space lower bound for the variance and a clean combinatorial mechanism via connected integral representations in Lemma 6.4. The main unresolved point is the completeness of Lemma 6.5, on which the increasing-window CLT and the multivariate CLT rest.","major_comments":[{"comment":"Lemma 6.5 is the load-bearing estimate for Theorem 6.6 and Theorem 7.3, but its proof is not complete as written. The proof details (24) and partially (23), then states that (25)–(29) can be established by the same method. The delicate point is that a naive bound such as f_i^{x_1,...,x_k} ≤ τ^{i+1-k} would give powers of |W| larger than 1; the required linear bound depends on expanding via the multivariate Mecke equation and using connectivity of the associated complexes together with condition (22). For (28), for example, one must show that the term E[(Λ²_{x_1,x_2}F)^4] produces a factor φ_1(x_1,x_2) that, after the outer exponent 1/2, lets the x_2-integral be controlled by ν from (22); without this cancellation one would obtain |W|² instead of |W|. Since the rates in Theorem 6.6 and the multivariate CLT in Theorem 7.3 depend directly on all six bounds, the proof should be completed for each of (25)–(29), or at least the reductions should be written out precisely enough that this cancellation is verifiable.","section":"§6.2, Lemma 6.5"},{"comment":"The proof of (23) is also written in a compressed way that makes it difficult to verify. The two expectations in (33) come from difference operators with different numbers of auxiliary points (l=2 and l=3), so the underlying complexes have different vertex sets; the intermediate notation f_L(x_1,…,x_{s+3}) is not defined for a complex L on s+2 vertices. Please rewrite the reduction with explicit vertex sets and relabelling, so that the final display (35) and the formation of the complex M in (34) are unambiguous.","section":"§6.2, proof of Lemma 6.5, display (23)"}],"minor_comments":[{"comment":"In the displayed formula for κ_m, the product is over ∅≠I⊆{0,…,m+1}; it should be over ∅≠I⊆{0,…,m}, since κ_m is a function on X^{m+1}. The example for κ_2 in the following display confirms this.","section":"Definition 4.1"},{"comment":"The proof of Lemma 6.1 explicitly treats only the first displayed estimate; the remaining estimates are dismissed with 'this approach can easily be extended.' Since Theorem 6.2 uses all of them, please provide the short calculation or state the combinatorial bound that yields the claimed powers of β.","section":"§6.1, Lemma 6.1"},{"comment":"The final remark that the outside exponents of the expectations in (23)–(29) lie in [1/2,1] is not literally correct for (25), (27), and the second summand of (28), where the outside exponent is 1; for exponent 1 the direct expansion used for (24) applies, so the remark should be rephrased.","section":"§6.2, proof of Lemma 6.5"},{"comment":"When Theorem 6.6 is applied to a coefficient vector whose top nonzero entry is b_s with s<α, the proof should explicitly say that the model is reduced to dimension s and that the higher-dimensional connection functions do not affect the functional; as written, the invocation of Theorem 6.6 is not literally covered by its statement, which assumes a_α≠0.","section":"§7, proof of Theorem 7.3"},{"comment":"There are several typographical slips, including 'allmost' and 'for allmost all n' in the proof of Theorem 6.6, an extraneous 'an' after 'F_n:=χ_a(Δ_{W_n})', 'connections functions' in the captions of Figures 1 and 2, and a missing parenthesis in the statement of Proposition 5.4.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The paper is well within the scope of a probability journal and the central ideas appear sound. My recommendation is driven by the incomplete proof of Lemma 6.5; if the author supplies the missing estimates, which I expect are correct, the manuscript should be publishable. I do not see issues with novelty or citation practice: the reliance on [12] and [13] is standard, and the self-citation to [15] is transparent. The referee should ask for a full proof of Lemma 6.5 before acceptance."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"One thing you should know about this paper: it proves quantitative CLTs for a generalized Euler characteristic in a new random simplicial complex model (the simplicial RCM), and the main results are probably correct, but the proof of Lemma 6.5—the O(|W|) bounds that drive the large-window CLT—is partly a sketch rather than a full proof. That is the spot to focus on in review.\n\nWhat's genuinely new: the model itself, which extends Penrose's RCM to higher-dimensional complexes via symmetric connection functions on a Borel space, with a careful marking construction. The modified difference operators in Section 5.1 and the multivariate CLT for simplex counts in the marked stationary case (Theorem 7.3) are also new. The paper recovers Čech, Vietoris-Rips and soft random simplicial complexes as special cases. The first and second moment formulas in Section 4 are cleanly derived using Mecke's equation, and the Fock space representation for the variance (Proposition 5.4) is a nice addition.\n\nThe soft spot is Lemma 6.5. The proof gives (24) in full detail and (23) in considerable detail—good, because (23) is the most complicated. But (25)-(29) are dispatched with 'in principle, this approach can be applied.' The stress-test note is on point: for (28), you need to verify that each summand acquires a factor φ1(x1,x2) so that after the outer exponent 1/2 the x2-integral is bounded by ν; a naive bound would produce |W|^2. This is the kind of step where a subtle mistake could change the scaling in Theorem 6.6 and hence Theorem 7.3. I don't think the estimates are wrong—the method shown for (23) is convincing and the author explicitly relies on the same connectedness argument—but a referee should demand explicit derivations for all the bounds, perhaps in an appendix. The paper also leans on the author's PhD thesis for some auxiliary statements; that is acceptable if the results are independently checkable.\n\nThe integrability condition (22) is strong—it excludes the highly non-integrable case—but it is a natural analogue of the bounded-degree condition in the RCM literature, and the paper is transparent about the technical role of the exponent 1/2.\n\nThis paper is for researchers in stochastic geometry and random simplicial complexes. It deserves a serious referee: the results are new, the strategy is sound, and the required fix is localized. I'd accept it for review and ask for a completed proof of Lemma 6.5 before publication.","headline":"Solid new model and likely-correct CLTs; the large-window proof hinges on Lemma 6.5, whose most delicate bounds are still a sketch.","tokens_in":31348,"tokens_out":3215,"would_cite":true,"duration_ms":35051,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["60F05","60D05","60C05"],"pacs":[],"model":"deepseek-v4-flash","headline":"A generalized Euler characteristic of higher-dimensional random connection models is asymptotically normal in both high-intensity and large-window regimes, with explicit Wasserstein and Kolmogorov error rates.","keywords":["random connection model","simplicial complexes","Euler characteristic","central limit theorem","Poisson process","normal approximation","marked stationary model","integral representations"],"falsifier":"Take $X=\\mathbb{R}$ with Lebesgue intensity and $\\varphi_1(x,y)=\\min\\{1,|x-y|^{-p}\\}$ with $p<1/2$, so that $\\int \\varphi_1^{1/2}\\,d\\lambda=\\infty$; then compute $\\operatorname{Var}(\\chi_a(\\Delta_{[-n,n]}))$ as $n\\to\\infty$. Linear growth would mean condition (22) is stronger than needed; superlinear growth, or a non-normal limit under $\\sqrt{n}$ scaling, would show the theorem has reached its natural boundary.","tokens_in":30323,"feed_emoji":"🔺","tokens_out":9716,"duration_ms":112964,"temperature":0.7,"pith_summary":"The paper extends the Random Connection Model---a random graph in which a Poisson process supplies vertices and each pair connects with probability $\\varphi_1(x,y)$---to higher-dimensional simplicial complexes: for each $j$, every potential $j$-simplex whose boundary simplices are already present is included independently with probability $\\varphi_j(x_0,\\dots,x_j)$. It then proves quantitative central limit theorems for the generalized Euler characteristic $\\chi_a(\\Delta_W)=\\sum_{i=0}^{\\alpha} a_i f_i(\\Delta_W)$ with an arbitrary coefficient vector $a=(a_0,\\dots,a_\\alpha)$ and $a_\\alpha\\neq 0$. For a fixed observation window the standardized statistic converges to the standard normal as the intensity $\\beta\\to\\infty$, with Wasserstein and Kolmogorov distance bounds of order $\\beta^{-1/2}$, one term of order $\\beta^{-3/2}$, and one of order $\\beta^{-1}$. For growing windows the same conclusion holds under a global integrability condition on $\\varphi_1$ and a positivity condition on the top-dimensional integral representation, with all six error bounds of order $|W_n|^{-1/2}$. A multivariate version for the vector of simplex counts follows in the translation-invariant marked stationary case. The upshot is that one argument covers a broad family of random complexes, from Čech and Vietoris–Rips complexes to soft random simplicial complexes, with explicit Gaussian fluctuation bounds.","feed_headline":"Gaussian limits for the Euler characteristic in random complexes","feed_subtitle":"Higher-dimensional random connection models converge to the normal law as intensity or window grows.","key_machinery":"The argument runs on integral representations $I_K(W)=\\int_{W^r} \\prod_{\\sigma\\in K} \\varphi_{|\\sigma|-1}(x_\\sigma)\\,\\lambda^r(d(x_1,\\dots,x_r))$, which encode the probability weight of a fixed combinatorial simplicial complex $K$ appearing inside $\\Delta_W$. The difference operators $\\Lambda^k_{(x_1,U_1),\\dots,(x_l,U_l)}$ are constructed so that, applied to the Euler characteristic, they simply count simplices containing all $k$ added points and none of the others; this makes products of such operators tractable. A Fock-space representation expresses the variance as a sum of nonnegative terms and yields the lower bound $\\operatorname{Var}(\\chi_a(\\Delta_W))\\ge a_\\alpha^2 \\beta^{\\alpha+1} I^{\\alpha+1}_{\\alpha+1,\\alpha+1}(W)/(\\alpha+1)!$. The normal approximation comes from Poisson-space normal approximation bounds adapted to these modified operators, which replace the usual add-one operators. Condition (22) is the step that forces every connected integral representation to be at most $C|W|$, which is what turns all six error terms into order $|W|^{-1/2}$ after standardization.","core_discovery":"The central claim is that the generalized Euler characteristic of the random connection model is asymptotically Gaussian in the regimes where the amount of randomness goes to infinity, either through the intensity or through the observation window. Concretely, Theorem 6.2 states that if $I^{\\alpha+1}_{\\alpha+1,\\alpha+1}(W)>0$, then $(\\chi_a(\\Delta_W)-\\mathbb{E}[\\chi_a(\\Delta_W)])/\\sqrt{\\operatorname{Var}(\\chi_a(\\Delta_W))}$ converges in distribution to $N(0,1)$ as $\\beta\\to\\infty$, with the six error quantities of Theorem 5.1 satisfying $\\gamma_1,\\gamma_3,\\gamma_4,\\gamma_5=O(\\beta^{-1/2})$, $\\gamma_2=O(\\beta^{-3/2})$, and $\\gamma_6=O(\\beta^{-1})$. Theorem 6.6 establishes the analogous result for sequences of windows with $|W_n|\\to\\infty$ under condition (22), $\\nu=\\sup_x \\int_X \\varphi_1(x,y)^{1/2}\\lambda(dy)<\\infty$, together with a linear lower bound on the expected number of top-dimensional simplices; in that regime all six error terms are $O(|W_n|^{-1/2})$. Theorem 7.3 derives a multivariate CLT for the simplex-count vector in the marked stationary case, with a positive definite limiting covariance matrix $\\Sigma$. The underlying discovery is that, despite the complex dependence created by the simplex-formation rules, the generalized Euler characteristic has variance of order the expected number of top-dimensional simplices, and its fluctuations are normal with explicit rates.","pith_inferences":["The exponent $1/2$ in condition (22) is forced by the smallest outside exponent appearing in the six error integrals, so relaxing the condition would likely require a different decomposition rather than a small tweak; this suggests the theorem is close to the natural limit of its method.","A natural testable extension is to replace the supremum in (22) with an $L^p$-type average over $x$; the paper's method may still work when the averaged quantity is small, but this is not stated and would need a separate argument.","Because the classical Euler characteristic is a special case, the CLTs imply that the signed simplex counts of Čech and Vietoris–Rips complexes fluctuate normally whenever the intensity or window growth makes the top-dimensional simplex count dominate the variance."],"forward_implications":["For a fixed window, the standardized generalized Euler characteristic is asymptotically normal as the intensity grows, with explicit Wasserstein and Kolmogorov error rates of order $\\beta^{-1/2}$ and faster.","For growing windows, condition (22) together with a linear lower bound on the expected top-simplex count yields a CLT with all six error bounds of order $|W_n|^{-1/2}$, so the $\\sqrt{|W_n|}$ normalization is the right scaling.","In the marked stationary case, the vector of simplex counts $(f_0(\\Delta_W),\\dots,f_\\alpha(\\Delta_W))$ satisfies a multivariate CLT with a positive definite limiting covariance matrix $\\Sigma$.","The classical Euler characteristic ($a_i=(-1)^i$) is included, and so are the Čech and Vietoris–Rips complexes as special connection-function choices.","The variance lower bound is proportional to the expected number of top-dimensional simplices, so positivity of $I^{\\alpha+1}_{\\alpha+1,\\alpha+1}$ is the non-degeneracy condition controlling the whole theorem."],"supporting_citations":[{"why":"Supplies the general random simplicial complex construction whose framework the paper extends to the Random Connection Model.","marker":"[4]"},{"why":"Introduced the Random Connection Model as a random graph, the object generalized to higher-dimensional complexes here.","marker":"[16]"},{"why":"Developed the edge-marked Poisson process representation and the difference-operator estimates for the classical RCM that the paper adapts to higher dimensions.","marker":"[12]"},{"why":"Provides the Poisson-space normal approximation theorems, Theorems 1.1 and 1.2, on which Theorem 5.1 is built.","marker":"[13]"},{"why":"Supplies the Poisson process tools---Mecke's formula, markings, thinning, and Fock-space representation---used throughout the proofs.","marker":"[14]"},{"why":"Contains the PhD thesis version of the fourth-moment calculations and the technical estimates that the paper draws on.","marker":"[15]"},{"why":"Defines soft random simplicial complexes, which the paper identifies as special cases of the new model.","marker":"[3]"}],"fun_headline_variants":["Gaussian limits for Euler characteristic in random connection models","Asymptotic normality for Euler characteristic in random simplicial complexes","Higher-dimensional random complexes yield Gaussian Euler characteristic","Euler characteristic of random simplicial complexes is Gaussian","Central limit theorems for Euler characteristic in random simplicial complexes"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The entire growing-window theorem depends on the global bound $\\nu=\\sup_{x\\in X}\\int_X \\varphi_1(x,y)^{1/2}\\lambda(dy)<\\infty$; if that constant is infinite, the estimates that force the variance to grow only linearly in $|W|$ break down, and the $\\sqrt{|W|}$ normalization is no longer justified.","fun_headline_variants_meta":{"raw":{"variants":["Gaussian limits for Euler characteristic in random connection models","Asymptotic normality for Euler characteristic in random simplicial complexes","Higher-dimensional random complexes yield Gaussian Euler characteristic","Euler characteristic of random simplicial complexes is Gaussian","Central limit theorems for Euler characteristic in random simplicial complexes"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000604,"raw_usage":{"total_tokens":2835,"prompt_tokens":977,"completion_tokens":1858,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":593,"completion_tokens_details":{"reasoning_tokens":1778}},"tokens_in":593,"tokens_out":1858,"duration_ms":14796,"temperature":1.0,"reasoning_tokens":1778,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T01:01:40.928958+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take $X=\\mathbb{R}$ with Lebesgue intensity and $\\varphi_1(x,y)=\\min\\{1,|x-y|^{-p}\\}$ with $p<1/2$, so that $\\int \\varphi_1^{1/2}\\,d\\lambda=\\infty$; then compute $\\operatorname{Var}(\\chi_a(\\Delta_{[-n,n]}))$ as $n\\to\\infty$. Linear growth would mean condition (22) is stronger than needed; superlinear growth, or a non-normal limit under $\\sqrt{n}$ scaling, would show the theorem has reached its natural boundary.","supporting_citations":[{"cited_title":"Costa and M","cited_arxiv_id":null,"evidence_quote":"Supplies the general random simplicial complex construction whose framework the paper extends to the Random Connection Model."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduced the Random Connection Model as a random graph, the object generalized to higher-dimensional complexes here."},{"cited_title":"Therandomconnectionmodelandfunctionsofedge-markedPois- son processes: Second order properties and normal approximation.The Annals of Applied Probability, 31(1):128–168, 2021","cited_arxiv_id":null,"evidence_quote":"Developed the edge-marked Poisson process representation and the difference-operator estimates for the classical RCM that the paper adapts to higher dimensions."},{"cited_title":"NormalapproximationonPoissonspaces: Mehler’sformula,second order Poincaré inequalities and stabilization.Probability Theory and Related Fields, 165:667–723,","cited_arxiv_id":null,"evidence_quote":"Provides the Poisson-space normal approximation theorems, Theorems 1.1 and 1.2, on which Theorem 5.1 is built."},{"cited_title":"Last and M","cited_arxiv_id":null,"evidence_quote":"Supplies the Poisson process tools---Mecke's formula, markings, thinning, and Fock-space representation---used throughout the proofs."},{"cited_title":"Pabst.Das Random Connection Model für höherdimensionale Simplizialkomplexe","cited_arxiv_id":null,"evidence_quote":"Contains the PhD thesis version of the fourth-moment calculations and the technical estimates that the paper draws on."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defines soft random simplicial complexes, which the paper identifies as special cases of the new model."}],"review_version":1}