{"id":"535b85fe-56f3-4321-b2f1-315740a1a354","arxiv_id":"2607.27535","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The expected persistence ratio measure for Čech/Rips complexes of a Poisson sample on any admissible compact C^2-triangulable space converges to the same universal limit as in Euclidean space, independent of geometry and density.","lead":"This paper proves that the distribution of persistence ratios from random point clouds on a wide class of curved and singular spaces converges to a universal limit, independent of the space and the sampling density. The result gives topological-data-analysis practitioners a single null distribution for data near manifolds, algebraic varieties, and Whitney stratified sets.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Lemma 4.6's metric stability proof assumes the Euclidean universal limit measure Π* is atomless at α; without a proof or citation of this continuity, Theorem 3.3 is unproven at atomic thresholds.","rationale":"The paper is a serious and largely coherent extension of Euclidean persistence-ratio universality to admissible triangulable spaces. The proof architecture—strata-wise decomposition, FK-subdivision flattening, interference control—is explicit and plausible. The reader's weakest assumption correctly identifies a real gap: Lemma 4.6 needs the Euclidean limiting measure to have no atom at the threshold α, but the paper neither proves nor cites this regularity. This is load-bearing because Theorem 3.4's metric stability condition is exactly what lets the proof pass from a curved simplex to a Euclidean one; without continuity, the error E(t) does not vanish and the transfer argument fails at those α. The concern is addressable, and it is reasonable to expect Π* is absolutely continuous, but the manuscript should state and prove or cite this. The secondary issue of f_max in Lemma 4.3 is also real, but the continuity gap is the more fundamental missing hypothesis for the central transfer theorem. Since the reader's verdict is already CONDITIONAL and my critique does not overturn the core claim, the verdict should remain UNCHANGED.","tokens_in":32286,"tokens_out":27046,"duration_ms":288677,"concrete_test":"Verify the Euclidean universal measure Π*_{k,F,d} from [8] is atomless. Concretely: derive the expected persistence intensity for transient cycles in R^d and check it is absolutely continuous with a locally integrable density in (birth, death) coordinates (equivalently, in the ratio coordinate). If the intensity has a density, then Π*({α})=0 for all α and Lemma 4.6 goes through. If no such result exists in [8], add atomlessness of Π* as an explicit hypothesis in Theorem 3.3, or restrict the statement to α outside a countable exceptional set.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The proof of metric stability (Lemma 4.6) defines E(t)=2·Π*_k([(1+t)^{-2}α,(1+t)^2α]) and asserts lim_{t↓0}E(t)=0 by continuity of the Euclidean universal limit Π*_k at α. The paper neither proves nor cites this atomlessness. If Π*_k has an atom at α, then lim_{t↓0}E(t)=2Π*_k({α})>0, so no δ>0 can satisfy E(3δ)≤ε/3 in the proof of Theorem 3.4. Consequently the metric stability condition (3.5) is not established for the persistence functional, and Theorem 3.3—stated for every α—has a proof gap exactly at atomic thresholds. This is not a challenge to the central universality claim, but a missing regularity hypothesis on the input measure from [8]; without it the transfer method of Section 4.1 cannot be applied at such α. The same proof also uses an unstated global upper bound on f in Lemma 4.3, but the continuity point is the more fundamental missing input because it affects all densities and all triangulated spaces.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper extends the Euclidean universality theorem for persistence ratios [8] to compact, purely d-dimensional, C^2-triangulable subspaces of R^D with a positive strata-separation angle (admissible triangulable spaces). Theorem 3.3 states that for a good density f on such an M, the expected persistence ratio measure for either the Vietoris–Rips or Čech filtration of a Poisson point process with intensity nf converges, after division by n, to the same universal limit Π*_{k,F,d} that appears in the Euclidean theorem, independently of M and f. Theorem 3.4 states the analogous result for general scale-invariant functionals satisfying conditions (1)–(7), including simplex additivity and metric stability. The proof proceeds by transferring the Euclidean result through a sequence of Freudenthal–Kuhn subdivisions and rectified triangulations, with control on interference across singular strata. The paper also provides detailed geometric estimates, an appendix on FK subdivisions, and a discussion of the necessity of vertex ordering and simplex quality.","tokens_in":32567,"tokens_out":13641,"duration_ms":130994,"significance":"If correct, the result is a substantial extension of a striking universality phenomenon from Euclidean space to manifolds and singular spaces, including semialgebraic and Whitney stratified sets. The geometric transfer framework — FK subdivision, rectification, skeleton interference control — is a useful contribution in its own right and appears to be carefully developed, with explicit constants and a self-contained appendix. The paper is not circular: it relies on the independently established Euclidean universality theorem [8]. However, the proof as written depends on an unstated regularity property of the Euclidean limit measure, and one step in the general-functional proof needs an additional normalization argument; both are local and repairable, but they affect the central claims.","major_comments":[{"comment":"The proof of metric stability for the persistence functional assumes that the Euclidean universal limit measure Π*_k is continuous at α. Equation (4.11) defines E(t)=2Π*_k([(1+t)^{-2}α,(1+t)^2α]) and asserts lim_{t↓0}E(t)=0 because Π*_k is continuous at α. The paper neither proves nor cites this atomlessness. If Π*_k({α})>0, then lim E(t)=2Π*_k({α})>0, and no δ>0 can satisfy E(3δ)≤ε/3, which is the condition used in the proof of Theorem 3.4. Since Lemma 4.6 is precisely the verification of condition (7) for the persistence functional, Theorem 3.3 as stated — for every α∈(1,∞) — is not established at atomic thresholds. This is a load-bearing gap. The fix is either to prove the atomlessness of Π*_k (or cite a proof if it is known), or to add it as an explicit hypothesis and restrict Theorem 3.3 to continuity points of Π*_k, adjusting the measure-convergence statement accordingly.","section":"Lemma 4.6, Eq. (4.11)"},{"comment":"The step labeled \"By the universality theorem for the Euclidean space case (Theorem 3.2)\" is not a direct application as printed. The density \\hat f_σ has total mass |σ|, not 1, so H_n(n\\hat f_σ) is an expectation under intensity n|σ| times the probability density g_σ=\\hat f_σ/|σ|. To justify (4.5), one must apply Theorem 3.2 to g_σ at control parameter m=n|σ|, obtaining (1/(n|σ|))H_{n|σ|}(n|σ|g_σ)→H*, and then use condition (2) with m=n|σ| and c=1/|σ| to replace H_{n|σ|}(n|σ|g_σ) by H_n(n|σ|g_σ). This gives (1/n)H_n(n\\hat f_σ)→|σ|H*, and summing yields (4.5). The authors should spell out this normalization/homogeneity argument; as written, the proof skips a load-bearing step.","section":"Proof of Theorem 3.4, Eq. (4.5)"}],"minor_comments":[{"comment":"The good-density condition is defined in the Euclidean setting; for f on a triangulable space M it should be made explicit that the condition is applied to the pullback of f under an admissible triangulation, or defined with respect to the d-dimensional Hausdorff measure on M.","section":"Section 3.2, Theorem 3.3"},{"comment":"The proof uses f_max, but the quantity is not defined in the statement of the lemma. It follows from the good-density condition that f is bounded, but this should be stated and f_max should be introduced explicitly.","section":"Lemma 4.3"},{"comment":"The bound (1+δ)^2 ≤ 1+3δ uses δ≤1. This assumption is present earlier in the proof, but it would help to restate it immediately before (4.2) to avoid ambiguity.","section":"Proof of Theorem 3.4, around (4.2)"},{"comment":"After Lemma C.2, it would be helpful to state explicitly that the global vertex order chosen in Definition B.8 avoids the incompatibility exhibited in Lemma C.2, since the induced order on a shared face is then the same from both adjacent simplices.","section":"Appendix C"},{"comment":"There are a few typographical issues, for example in Lemma 5.2 footnotes 8–10 are informal proof aids; these could be integrated into the proof. The phrase \"in [8, 29] it is shown\" in the remark after Theorem 3.3 is slightly vague about which claim is established where.","section":"General"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is technically rich and the transfer strategy is credible, but the two major comments above concern load-bearing steps. The continuity/atomlessness issue is the more serious one because it affects the statement of Theorem 3.3 at every α; I would ask the authors to either prove the atomlessness of Π*_k from [8] or state the theorem with this regularity hypothesis. The normalization issue in (4.5) is easier to repair but should be corrected. I do not see grounds for rejection: the framework is sound, the geometric estimates are detailed, and the missing pieces appear to be within the manuscript's scope."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Worth a careful read. The paper extends Bobrowski–Skraba universality from Euclidean domains to compact C^2-triangulable spaces, including singular ones, and the extension is real. The transfer machinery—FK subdivisions, rectification, skeleton-interference estimates—is the actual contribution, and it is built with care. The angle condition excluding cusps is a sensible quality assumption. The proof is not circular: the Euclidean theorem is an independent external benchmark, and the new conditions (6)–(7) carry the work. The paper also covers both Rips and Čech filtrations, and the universal limit is genuinely the same as in the Euclidean case.\n\nWhere it gets soft: the main theorem is not fully proven as written. Lemma 4.6 conditions metric stability on continuity of the Euclidean universal limit Π*_k at α, and Theorem 3.3 does not state that as a hypothesis. The proof needs lim_{t↓0} Π*_k([(1+t)^{-2}α,(1+t)^2α])=0. If Π*_k has an atom at α, the error E(t) does not vanish and the transfer fails at that threshold. This is a missing regularity input from [8], not an internal contradiction, but it is load-bearing. Either the authors cite or prove atomlessness, or Theorem 3.3 must be formulated for α off the atoms. Minor but real: Lemma 4.3 uses f_max, yet the “good density” definition does not guarantee boundedness—in the inf>0 case you can have compactly supported densities that are positive but unbounded. That is an easy fix: add boundedness or state the relevant upper bound. The stricter radius cutoff nρ^d = o(n^{1/(d^2+d+1)}) is called technical, but I would like a sentence on why the extra power does not change the limiting measure.\n\nOverall the central argument is plausible, the geometric estimates are substantial, and the missing pieces are addressable. This paper deserves a serious referee, not a desk rejection. I would tell the editor to send it, with the request that the referee verify the atomless issue and the density bound. I would cite it once the continuity question is settled.","headline":"A serious, original extension of persistence-ratio universality to curved and singular triangulable spaces, but Theorem 3.3 as stated is missing an atomlessness condition on the Euclidean limit and an explicit boundedness assumption on the density.","tokens_in":33036,"tokens_out":3361,"would_cite":true,"duration_ms":38404,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["60D05","55N31","60G55"],"pacs":[],"model":"deepseek-v4-flash","headline":"Random persistence statistics on curved and singular spaces converge to the same universal limit as in Euclidean space.","keywords":["persistent homology","universality","triangulable spaces","Poisson point process","Vietoris–Rips filtration","Čech filtration","singular spaces","Freudenthal–Kuhn subdivision"],"falsifier":"Compute or bound the Euclidean limit measure Π*_{k,F,d}: if it has an atom at any α > 1, then Lemma 4.6's error term E(t) fails to vanish at that α, and the transfer argument in Theorem 3.3 does not cover that threshold. Alternatively, simulate Poisson processes with matched intensity on a curved surface (e.g., a 2-sphere) and on a flat rectangle and compare (1/n)Π_{k,F}(nf)([α,∞)) for fixed α: a non-vanishing difference would contradict the claimed universality.","tokens_in":32154,"feed_emoji":"🔺","tokens_out":6011,"duration_ms":58477,"temperature":0.7,"pith_summary":"This paper proves that the statistical behavior of persistent homology on random point clouds is universal across a broad class of spaces. For any compact triangulable space with no cusp singularities—covering smooth manifolds, algebraic varieties, semialgebraic sets, and Whitney stratified spaces—the expected distribution of persistence ratios (death/birth) converges, after rescaling, to a single limit that depends only on the dimension, the filtration, and the homology degree, not on the shape of the space or the sampling density. The known Euclidean universality is transferred to curved and singular spaces by approximating them with recursively subdivided, rectified triangulations while controlling errors near singular strata. If correct, this means the same null distribution used for flat Euclidean data is valid for data sampled near curved or singular objects.","feed_headline":"Curved and singular spaces share one persistence-ratio law","feed_subtitle":"On curved and singular spaces, persistence ratios converge to the same universal limit as flat space.","key_machinery":"The load-bearing device is the geometric transfer method. The space M is triangulated, then repeatedly subdivided by the Freudenthal–Kuhn (edgewise) subdivision, which—unlike barycentric subdivision—keeps a positive lower bound on simplex quality. Each subdivision is rectified via its secant map, yielding a piecewise-linear metric on the complex that distorts the intrinsic metric of M by at most a factor of 1 + γ/ℓ (Proposition 3.7). Meanwhile the strata separation angle Θ_φ controls how close points on two different simplices can be: if they are close, both must lie near the skeleton, whose tube volume is only O(r) in the d-dimensional Hausdorff measure (Propositions 3.5 and Lemma 3.6). Tog","core_discovery":"The central claim is Theorem 3.3: for an admissibly triangulable space M—compact, C^2-triangulable, purely d-dimensional, with a positive angle between normal cones of adjacent simplices—and a good density f, the normalized expected persistence-ratio measure satisfies lim_{n→∞} (1/n)Π_{k,F}(nf) = Π*_{k,F,d}, the same universal limit as in Euclidean space, for both Rips and Čech filtrations. A broader theorem (3.4) extends this to any scale-invariant functional satisfying rigidity, homogeneity, continuity, and simplex additivity. The proof divides M into simplices, shows that points lying on two distinct simplices must stay near the singular skeleton, and approximates each curved simplex by a","pith_inferences":["If the universal limit Π* is in fact atom-free, a corollary not spelled out in the paper is that one can build bootstrap or subsampling inference for topological statistics on arbitrary such spaces using a single tabulated null distribution.","The transfer idea suggests testable extensions to non-compact or unbounded-curvature settings: whether the admissibility condition (positive strata separation angle) can be relaxed to allow cusp singularities, or whether universality survives at all, is a natural numerical experiment.","Since the proof only requires metric stability and simplex additivity, other functionals—such as intrinsic-dimension estimators or stabilization-based geometric statistics—should inherit the same universality; verifying this would generalize the framework.","The paper leaves open the regularity of Π*; settling whether the limiting measure is continuous would turn the metric-stability assumption into a theorem and likely sharpen the radius-cutoff conditions."],"forward_implications":["Data sampled from compact smooth manifolds, algebraic varieties, semialgebraic sets, and Whitney stratified spaces can be analyzed with the same universal null distribution for persistence ratios already used in Euclidean topological data analysis.","The universality extends to a general class of scale-invariant functionals, so other stabilizing geometric statistics on such spaces will also have space-independent limits.","The geometric transfer framework—FK subdivision plus rectification plus singular-strata control—provides explicit metric-distortion bounds, making it a reusable tool for transferring Euclidean stochastic-geometry results to triangulable spaces.","Under the (assumed) continuity of the universal limit, small metric perturbations change expected persistence ratios by an amount controlled by the metric distortion and by the mass the universal limit assigns to the intervening interval.","The results are stated for Poisson point processes; the paper notes that an analogue for binomial processes is a technical step away."],"fun_headline_variants":["Persistence ratios on singular spaces match flat-space limit","Universal topological law proven for curved and singular shapes","One persistence-ratio law governs curved and singular spaces","Triangulated spaces share universal persistence statistics","Singular spaces obey same persistence-ratio limit as flat"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The proof of the metric-stability step assumes that the already-proven Euclidean universal limit distribution Π*_k has no atom at the threshold α where the comparison is made; the paper states full convergence without proving this continuity, and if Π* jumps there, the transfer from curved to flat simplices fails at that scale.","fun_headline_variants_meta":{"raw":{"variants":["Persistence ratios on singular spaces match flat-space limit","Universal topological law proven for curved and singular shapes","One persistence-ratio law governs curved and singular spaces","Triangulated spaces share universal persistence statistics","Singular spaces obey same persistence-ratio limit as flat"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":9.5e-05,"raw_usage":{"total_tokens":821,"prompt_tokens":716,"completion_tokens":105,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":460,"completion_tokens_details":{"reasoning_tokens":31}},"tokens_in":460,"tokens_out":105,"duration_ms":2172,"temperature":1.0,"reasoning_tokens":31,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T06:06:56.400173+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute or bound the Euclidean limit measure Π*_{k,F,d}: if it has an atom at any α > 1, then Lemma 4.6's error term E(t) fails to vanish at that α, and the transfer argument in Theorem 3.3 does not cover that threshold. Alternatively, simulate Poisson processes with matched intensity on a curved surface (e.g., a 2-sphere) and on a flat rectangle and compare (1/n)Π_{k,F}(nf)([α,∞)) for fixed α: a non-vanishing difference would contradict the claimed universality.","supporting_citations":[],"review_version":1}