{"id":"d2126d0f-da48-439d-a125-a5353ca6183d","arxiv_id":"2607.06048","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":3,"one_line_summary":"Scattering network separation capacity on rectifiable sets is bounded by tangent-space rank and a second-moment matrix, yielding two filter design criteria: sufficient spectral coverage and well-conditioned frame-geometry coupling.","lead":"This paper derives mathematical bounds on how well scattering networks can classify data living on low-dimensional geometric structures, yielding concrete filter design criteria. A smart generalist reads it to understand how the geometry of data should drive neural network architecture choices.","discovery_kind":"unclear","skeptic_critique":{"model":"glm-5.2","headline":"Lower-bound design criteria for general rectifiable sets (Thm 3.8) require polynomial bi-Lipschitz parametrizations, which excludes smooth manifolds with transcendental parametrizations; the paper does not discuss this limitation.","rationale":"The reader's verdict of CONDITIONAL is appropriate. The mathematical content is sound: the proofs are detailed, the chain of lemmas from Lemma 2.2 through Theorem 3.8 is logical, and the design criteria follow from the bounds. The concern about the polynomial parametrization assumption is real and load-bearing for the lower-bound design criteria, which are the paper's main actionable contribution for general rectifiable sets. However, I note that the sparse-signal results (Theorem 3.5, Corollary 3.5.1) do not require the polynomial assumption — they use linear parametrizations ψ_{S,j}(x) = Ξ_S(x' + i x''), which are trivially polynomial of degree 1. So the concern applies specifically to Section 3.3's general rectifiable set results, not to the sparse-signal results. The reader's assessment of novelty at 7.0 is reasonable given the dependence on companion papers [7, 9] for the framework, though the particularization to scattering networks and the specific design criteria are new. The correctness risk remains 'unknown' in the sense that no machine-checked proofs or reproducible code are provided, but the proofs read carefully and I did not find errors. The CONDITIONAL verdict is justified: the paper should discuss the scope limitation imposed by the polynomial assumption and ideally address whether real-analytic or smooth parametrizations can be handled, perhaps through polynomial approximation arguments.","tokens_in":20150,"tokens_out":1045,"duration_ms":390260,"concrete_test":"Attempt to construct a concrete example of a countably H^s-rectifiable set E with a smooth (C^∞) bi-Lipschitz parametrization ψ that is not polynomial, and check whether the matrix T_{ψ̂} in Theorem 3.8 can be defined. Specifically, take E to be the unit circle in R^2 parametrized by ψ(θ) = (cos θ, sin θ), θ ∈ [0, 2π). Compute v_{M,d}(ψ̂(θ)) for d=2 and verify whether this lies in a finite-dimensional space of monomials in θ. If it does not, the factorization v_{M,d}(ψ̂(x)) = T_{ψ̂} w_{s,n d}(x) fails, confirming that Theorem 3.8 does not apply to this natural example. Then check whether the bound can be recovered by approximating ψ with polynomial parametrizations and taking a limit.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The reader correctly identifies the load-bearing concern. Theorem 3.8 and Corollary 3.8.8 — which provide the actionable lower-bound design criterion (well-conditioning of T_{ψ̂_j}^H A_{Λ,χ} T_{ψ̂_j}) for general rectifiable sets — depend critically on the assumption in Section 3.3 that each bi-Lipschitz parametrization ψ_j is a polynomial of degree n_j. This assumption enters the proof of Theorem 3.8 through the factorization v_{M,d}(ψ̂_j(x)) = T_{ψ̂_j} w_{s,n_j d}(x), which requires that the Veronese lift of the Fourier transform of ψ_j be expressible as a linear map applied to a monomial vector of bounded degree. If ψ_j is not polynomial, the image v_{M,d}(ψ̂_j(K_j)) is not contained in any finite-dimensional space of monomials, the matrix T_{ψ̂_j} is not well-defined as a finite-dimensional object, and the entire chain of inequalities in (3.23) and the subsequent trace bounds break down. The upper bounds (Theorem 3.7) also use the polynomial assumption to bound the span dimension via the degree n_j, but they are less critical since the upper bound is not the actionable design criterion. The concern is not about internal inconsistency — the proofs are correct under the stated assumptions — but about the scope of applicability. Many natural low-dimensional data models (smooth submanifolds with non-algebraic parametrizations, certain fractal sets) do not admit polynomial bi-Lipschitz parametrizations, and the paper provides no discussion of which natural data models satisfy this condition or whether the assumption can be relaxed (e.g., to real-analytic parametrizations, which would suffice for Corollary 2.9.1 but not for Theorem 3.8's factorization). This is a genuine gap between the paper's framing ('for data modeled as rectifiable sets') and the actual scope of the lower-bound results.","agreement_with_reader":"agree"},"referee_report":{"model":"glm-5.2","summary":"This paper studies the separation capacity of pooling-free scattering networks with fixed monomial nonlinearities on data modeled as countably H^s-rectifiable sets. The authors first establish general bounds on the s-separation capacity of feature extractors: a lower bound in terms of the rank of the differential restricted to approximate tangent spaces (Lemma 2.7) and an upper bound via a second-moment matrix (Lemma 2.9), with an exact formula under real-analyticity (Corollary 2.9.1). These are then particularized to scattering networks for two data models: sparse signals and polynomially parametrized rectifiable sets. The resulting design criteria are: (i) filter spectral supports should not be contained in a coset of a proper subgroup (Theorem 3.3), and (ii) matrices coupling the filter frame to the data geometry should be well-conditioned (Theorems 3.5, 3.8 and Corollaries 3.5.1, 3.8.1). The proofs proceed through a clean chain from the subspace characterization (Lemma 2.2) through geometric measure theory tools to concrete matrix conditions.","tokens_in":20417,"tokens_out":1681,"duration_ms":307098,"significance":"The paper provides a mathematically rigorous bridge between geometric measure theory and the separation capacity of scattering networks, yielding actionable filter-design criteria expressed in terms of well-conditioning of data-dependent matrices. The derivation chain is carefully executed: the area formula application in Corollary 2.9.1, the Veronese map factorization in Lemma 3.2, and the restricted isometry constant formulation in Theorem 3.5 are notable technical contributions. The reduction of the sparse-signal lower bound to an RIP condition on (T^H_{bXi} A_{Lambda,chi} T_{bXi})^{1/2} is a clean, falsifiable criterion. The upper-bound design criterion (spectral supports not in a coset of a proper subgroup) is parameter-free in the sense that it depends only on the group structure of Z/MZ and the support sets. The paper is self-contained in its proof structure, though it relies on two companion papers [7, 9] for background results on Cover-type separation capacity.","major_comments":[{"comment":"Section 3.3, Theorem 3.8: The lower bound for general rectifiable sets depends on the assumption that each bi-Lipschitz parametrization psi_j is a polynomial of degree n_j. This assumption is load-bearing: the factorization v_{M,d}(hat{psi}_j(x)) = T_{hat{psi}_j} w_{s,n_j d}(x) in the proof of Theorem 3.8 requires the Veronese lift of hat{psi}_j to land in a finite-dimensional monomial space. If psi_j is not polynomial, T_{hat{psi}_j} is not a finite-dimensional object and the trace bounds in (3.23) break down. The paper does not discuss which natural data models satisfy this polynomial assumption, nor does it acknowledge that smooth manifolds with non-algebraic parametrizations (a primary motivating example per the introduction's reference to [1, 2]) are excluded. This is a scope limitation on the central actionable design criterion (Corollary 3.8.1). The authors should either (a) add a","section":null},{"comment":"Section 3.3, Theorem 3.7: The upper bound also relies on the polynomial assumption to bound dim_C(span_C(v_{M,d}(F_M(E_j)))) <= C(s + n_j d, n_j d) via the degree of x -> v_{M,d}(F_M(psi_j(x))). While the upper bound is less directly actionable, the same scope concern applies. The paper should clarify whether the polynomial assumption is essential for the upper bound as well, or whether a more general bound (e.g., in terms of the Hausdorff dimension of the Fourier support of psi_j(K_j)) could replace it.","section":null},{"comment":"Theorem 3.5 and Corollary 3.5.1: The lower bound for sparse signals involves the term inf_{j in J_S} 2(Tr(C_{S,j}))^2 / Tr((C_{S,j})^2), which depends on the parametrization domains K_{S,j} but not on the frame Xi. The paper does not discuss whether this term can be bounded below in terms of s and d alone, or whether it can be arbitrarily small for adversarial choices of K_{S,j}. If the latter, the practical utility of the design criterion (minimizing delta_{s,d}(Xi; Lambda, chi)) is diminished, as the overall lower bound could still be small. A brief remark on the behavior of this term would strengthen the result.","section":null}],"minor_comments":[{"comment":"The notation T_{hat{psi}_j} in Theorem 3.8 uses a hat on psi_j, but the text in Section 3.3 defines psi_j without a hat. The hat presumably denotes the Fourier transform, but this should be stated explicitly.","section":null},{"comment":"In the proof of Lemma 3.2, the matrix A is defined with the condition supp(alpha) subseteq H_{lambda,S}, but in the subsequent definition of A_lambda (footnote 2 on page 11), this condition is dropped. The relationship between A and A_lambda should be stated more explicitly to avoid confusion.","section":null},{"comment":"Theorem 3.3: The upper bound includes a term 4 min_{S} [...] rather than 2 min_{S} [...]. The factor of 2 relative to the real-valued case is presumably due to the complex-to-real identification, but this should be noted explicitly, perhaps with a reference to equation (3.3).","section":null},{"comment":"In the proof of Theorem 3.8(b), the chain of inequalities bounding Tr(G_j A_Lambda G_j A_Lambda) uses the Loewner ordering notation (preceq, succeq) without explicitly defining it. While standard, a brief note would aid readability.","section":null},{"comment":"Reference [9] (arXiv:2607.01010) is cited for the identity SC_s(Phi) = min_{j in J} SC_s(Phi|_{psi_j(K_j)}) used in Corollary 2.9.1, and for the identity in (2.2). Since these are central to the derivation, the reader would benefit from a brief statement of the relevant results from [9] rather than relying entirely on the companion paper.","section":null},{"comment":"The abstract states 'no pooling' but Section 3.1 defines the general scattering network with pooling operators P_n. The restriction to P_n = Id is stated in Section 3.1 but could be mentioned in the abstract for precision.","section":null}],"recommendation":"major_revision","confidential_remarks":"The paper depends heavily on two companion papers by the same authors ([7] and [9], both arXiv 2026). The central identity SC_s(Phi) = 2 min dim_R(V) (Lemma 2.2) and the factorization SC_s(Phi) = min_j SC_s(Phi|_{psi_j(K_j)}) (Corollary 2.9.1) both cite [9]. If [9] has not yet been refereed, the correctness of the present paper's main results cannot be fully verified independently. The editor may wish to consider the refereeing of these papers in coordination. The polynomial parametrization assumption in Section 3.3 is the main substantive concern; it is correctly flagged by the reader and is a genuine scope limitation that should be addressed before publication."},"author_rebuttal":{"model":"glm-5.2","summary":"We thank the referee for a careful reading and for identifying a genuine scope limitation in Section 3.3 that we will address in revision. The referee's three major comments are interconnected: the first two concern the polynomial parametrization assumption in Theorems 3.7 and 3.8, and the third concerns the parametrization-domain-dependent term in the sparse-signal lower bound. We agree that the polynomial assumption is load-bearing and that the manuscript does not adequately discuss its scope or which data models satisfy it. We will add a dedicated remark addressing this. On the third comment, we will add discussion of the behavior of the trace-ratio term, including both the lower bound it admits and the adversarial cases the referee identifies.","responses":[{"response":"The referee is correct that the polynomial assumption is load-bearing for Theorem 3.8. Specifically, the factorization v_{M,d}(hat{psi}_j(x)) = T_{hat{psi}_j} w_{s,n_j d}(x) in the proof requires the Veronese lift of hat{psi}_j to land in a finite-dimensional monomial space, which fails for non-polynomial psi_j. We acknowledge that this excludes smooth manifolds with non-algebraic parametrizations, which the introduction's reference to [1, 2] might suggest are covered. We will add a dedicated remark in Section 3.3 explicitly stating this scope limitation, clarifying that the results of Theorem 3.8 and Corollary 3.8.1 apply to polynomially parametrized rectifiable sets (which include sparse-signal models as a special case, as well as algebraic varieties and their finite unions). Regarding option (b), extending to non-polynomial parametrizations would require replacing the finite-dimensional matrix T_{hat{psi}_j} with an operator acting on an infinite-dimensional function space, and the trace bounds in (3.23) would no longer apply in their current form. We believe this extension is beyond the scope of the present paper, but we will note it as a direction for future work. We will also adjust the introduction to avoid implying that general smooth submanifolds are covered by the polynomial results.","revision_made":"yes","referee_comment":"Section 3.3, Theorem 3.8: The lower bound depends on the polynomial parametrization assumption. The paper does not discuss which natural data models satisfy this, nor that smooth manifolds with non-algebraic parametrizations are excluded. The authors should either (a) add a discussion of scope, or (b) extend to non-polynomial parametrizations."},{"response":"The referee correctly identifies that the bound dim_C(span_C(v_{M,d}(F_M(E_j)))) <= C(s + n_j d, n_j d) in Lemma 3.6 relies on x -> v_{M,d}(F_M(psi_j(x))) being a multivariate polynomial of degree at most n_j d, which requires psi_j to be polynomial. For the upper bound, however, the situation is somewhat different from the lower bound. The term |H_{d,lambda,psi_j} cap supp(hat{chi})| in (3.21) does not depend on the polynomial assumption and remains valid for general bi-Lipschitz parametrizations. Only the combinatorial term C(s + n_j d, n_j d) requires polynomial structure. For a general (non-polynomial) bi-Lipschitz parametrization, the Veronese lift v_{M,d}(hat{psi}_j(x)) need not lie in any finite-dimensional subspace, so the combinatorial bound breaks down. A replacement bound in terms of the Hausdorff dimension of the Fourier support of psi_j(K_j) is an interesting possibility, but it would require a fundamentally different proof technique, as the current argument proceeds through finite-dimensional polynomial degree counting. We will add a remark clarifying that the polynomial assumption is essential for the combinatorial term in the upper bound, while the spectral-support term is not, and that extending the upper bound to non-polynomial parametrizations is left open.","revision_made":"yes","referee_comment":"Section 3.3, Theorem 3.7: The upper bound also relies on the polynomial assumption to bound dim_C(span_C(v_{M,d}(F_M(E_j)))) <= C(s + n_j d, n_j d). The paper should clarify whether the polynomial assumption is essential for the upper bound, or whether a more general bound could replace it."},{"response":"The referee raises a valid point. The term 2(Tr(C_{S,j}))^2 / Tr((C_{S,j})^2) equals 2 times the effective rank of C_{S,j}, which is at least 2 and at most 2*C(s-1+d, d). For well-behaved parametrization domains (e.g., K_{S,j} containing an open ball in R^{2s}), C_{S,j} is full-rank and the term equals 2*C(s-1+d, d), the maximum possible value. However, the referee is correct that for adversarial choices of K_{S,j} — for instance, domains concentrated near a lower-dimensional subset — the matrix C_{S,j} can become rank-deficient and the term can be arbitrarily small. We will add a remark noting both the upper bound 2*C(s-1+d, d) (achieved for full-dimensional domains) and the fact that the term can degenerate for adversarial domains. We would also note that in the typical setting where the parametrization domains are fixed by the data model (not chosen adversarially), this term is a geometric constant of the dataset, and the design criterion of minimizing delta_{s,d}(Xi; Lambda, chi) remains the actionable lever for the practitioner, as the frame Xi is the only design variable.","revision_made":"yes","referee_comment":"Theorem 3.5 and Corollary 3.5.1: The term inf_{j in J_S} 2(Tr(C_{S,j}))^2 / Tr((C_{S,j})^2) depends on the parametrization domains K_{S,j} but not on the frame Xi. The paper does not discuss whether this term can be bounded below in terms of s and d alone, or whether it can be arbitrarily small for adversarial choices of K_{S,j}."}],"tokens_in":20130,"tokens_out":1352,"duration_ms":275268,"standing_objections":[]},"desk_editor":{"model":"glm-5.2","letter":"The paper gives two actionable filter-design criteria for scattering networks on low-dimensional data: (i) filter spectral supports should not sit inside a coset of a proper subgroup of Z/MZ, and (ii) the matrices T_{ψ̂_j}^H A_{Λ,χ} T_{ψ̂_j} should be well-conditioned. The first comes from an upper bound on separation capacity, the second from a lower bound. Both are new results. The reduction to RIP minimization for sparse signals is clean and connects this work to compressed sensing in a natural way. The proof chain is solid: Lemma 2.2 (subspace characterization) → Lemma 2.7 (tangent-space lower bound) → Lemma 2.9 (second-moment upper bound) → Corollary 2.9.1 (exact formula under real-analyticity) → Theorems 3.3–3.8 (scattering network instantiation). The use of GMT machinery — rectifiable sets, approximate tangent spaces, area formula — to get Cover-type separation bounds is the right tool for the job and is applied correctly. The sparse-signal results (Theorem 3.5, Corollary 3.5.1) are the strongest part: the RIP-based lower bound is concrete and the connection to the Veronese lift is elegant. The upper bound in Theorem 3.3 and its interpretation via sumset growth in Z/MZ is also clean. The soft spot is real but bounded. The lower-bound design criterion for general rectifiable sets (Theorem 3.8, Corollary 3.8.1) requires each bi-Lipschitz parametrization ψ_j to be a polynomial of degree n_j. This enters through the factorization v_{M,d}(ψ̂_j(x)) = T_{ψ̂_j} w_{s,n_j d}(x), which needs the Veronese lift of ψ̂_j to land in a finite-dimensional monomial space. If ψ_j is not polynomial, T_{ψ̂_j} is not finite-dimensional and the trace bounds in (3.23) break down. The upper bounds (Theorem 3.7) also use this assumption but are less sensitive — they only need the degree to bound the span dimension. The paper does not discuss which natural data models satisfy the polynomial assumption, or whether it can be relaxed (e.g., to real-analytic parametrizations, which would suffice for Corollary 2.9.1 but not for Theorem 3.8's factorization). This is a gap between the framing ('for data modeled as rectifiable sets') and the actual scope of the lower-bound results. It does not invalidate anything — the proofs are correct under the stated assumptions — but it limits the practical reach of the actionable lower-bound criterion. The reader's CONDITIONAL verdict is slightly too cautious. The polynomial assumption is a genuine scope limitation, but the upper-bound criterion (Theorem 3.3) and the sparse-signal lower bound (Theorem 3.5) do not depend on it and are already useful results. The paper is for researchers in mathematical deep learning theory who care about principled filter design. It deserves a serious referee who can check the GMT details and push the authors to address the scope question.","headline":"Scattering network filter design criteria derived from geometric measure theory; lower bounds need polynomial parametrization assumption","tokens_in":21262,"tokens_out":751,"would_cite":true,"duration_ms":133014,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"glm-5.2","headline":"Filters must meet data on enough frequencies to maximize separation","keywords":[],"falsifier":"If one could exhibit a rectifiable dataset E with a non-polynomial bi-Lipschitz parametrization where the conditioning-based design criterion fails to correlate with actual separation capacity, the practical applicability of the lower-bound design recommendations would be called into question.","tokens_in":20307,"feed_emoji":"📐","tokens_out":1115,"duration_ms":175796,"temperature":0.7,"pith_summary":"This paper asks a precise question: given a scattering network with a fixed nonlinearity and no pooling, how should you choose the convolutional filters to maximize the network's ability to separate (i.e., linearly classify) data that lives on a low-dimensional structure inside a high-dimensional space? The authors model low-dimensional data as rectifiable sets from geometric measure theory, which generalizes smooth manifolds and unions of linear subspaces (including sparse-signal models). Using Cover's separation capacity, they prove two-sided bounds that connect the network's classification power to the geometry of the data. The upper bound says the separation capacity is limited by the rank of a second-moment matrix that couples the feature extractor to the global geometry of the dataset. The lower bound says the separation capacity is at least twice the essential infimum of the rank of the differential of the feature map restricted to approximate tangent spaces of the data. When particularized to scattering networks, these bounds yield two concrete filter-design criteria. First, the spectral supports of the filters, when intersected with the data's spectrum, should not be contained in a coset of a proper subgroup of the frequency group. Second, certain filter-dependent matrices that couple the frame to the data geometry should be as well-conditioned as possible. For sparse signals, the second criterion reduces to minimizing a restricted isometry constant.","feed_headline":"Filters must meet data on enough frequencies to maximize separation","feed_subtitle":"Two geometric bounds yield concrete design rules for scattering network filters on low-dimensional data","key_machinery":"The key objects are: (1) countably H^s-rectifiable sets as a model for low-dimensional data; (2) Cover's s-separation capacity as the measure of classification power; (3) approximate tangent spaces T_f E of rectifiable sets; (4) the second-moment matrix C_Phi = integral of y y^T d Phi_sharp(H^s restricted to E); (5) the vector Veronese map v_{M,d}, which encodes the effect of the monomial nonlinearity of degree d; (6) the matrix A_Lambda constructed from filter Fourier coefficients, which couples the filters to the Veronese lift; and (7) the matrices T_{hat Psi_j}^H A_{Lambda,chi} T_{hat Psi_j}, whose conditioning controls the lower bound.","core_discovery":"The central mechanism is a pair of bounds on the s-separation capacity SC_s of a feature extractor on a countably H^s-rectifiable set. The lower bound is governed by the rank of the differential of the feature map on approximate tangent spaces (local geometry), while the upper bound is governed by the rank of a second-moment matrix formed by pushing forward the Hausdorff measure through the feature map (global geometry). For scattering networks specifically, the upper bound translates into a requirement on the spectral overlap between filters and data, and the lower bound translates into a conditioning requirement on matrices that couple the filter frame to the parametrization of the data. A","pith_inferences":["The polynomial parametrization assumption required for the lower bounds may exclude natural data manifolds with non-algebraic structure (e.g., smooth manifolds parametrized by transcendental maps). Extending the lower-bound argument to general bi-Lipschitz parametrizations would broaden the applicability of the conditioning-based design criterion.","The framework treats the nonlinearity degree d and network depth n_d as fixed; the interaction between depth and the conditioning of the coupling matrices is not fully explored and could reveal depth-dependent tradeoffs in separation capacity.","The separation capacity as defined measures linear separability of the network's output features. A natural extension would investigate whether the geometric bounds transfer to nonlinear classifiers applied on top of the scattering features."],"forward_implications":["Filter design for scattering networks on low-dimensional data can be guided by two checkable conditions: spectral coverage of the data spectrum and numerical conditioning of data-dependent matrices, rather than by empirical tuning alone.","For sparse-signal models, the design criterion reduces to minimizing a restricted isometry constant, directly connecting scattering network filter design to the compressed sensing literature.","The upper bound's group-theoretic condition (spectral supports not in a coset of a proper subgroup) provides a concrete, checkable necessary condition for a filter bank to achieve maximum separation capacity.","The framework applies to any Lipschitz feature extractor on rectifiable sets, so the geometric bounds could in principle be applied to feature extractors beyond scattering networks."],"fun_headline_variants":["Separation capacity of scattering networks tied to two geometric bounds","Filter spectral overlap and conditioning bound scattering network separation","Scattering filters need enough frequency overlap to separate low-dimensional data","Local and global geometry bound separation capacity of feature extractors","Two geometric criteria govern scattering network filter design for separation"],"cache_read_input_tokens":0,"weakest_assumption_plain":"The lower bounds, which yield the actionable conditioning-based design criteria, require the bi-Lipschitz parametrizations of the rectifiable data set to be polynomial maps. Many natural low-dimensional data structures (smooth manifolds with non-algebraic parametrizations, fractal sets) may not admit such polynomial bi-Lipschitz parametrizations, limiting the scope of the design recommendations.","fun_headline_variants_meta":{"raw":{"variants":["Separation capacity of scattering networks tied to two geometric bounds","Filter spectral overlap and conditioning bound scattering network separation","Scattering filters need enough frequency overlap to separate low-dimensional data","Local and global geometry bound separation capacity of feature extractors","Two geometric criteria govern scattering network filter design for separation"]},"model":"glm-5.2","effort":"high","cost_usd":0.0,"raw_usage":{"total_tokens":564,"prompt_tokens":502,"completion_tokens":62,"prompt_tokens_details":null},"tokens_in":502,"tokens_out":62,"duration_ms":33565,"temperature":1.0,"reasoning_tokens":null,"cache_read_input_tokens":0,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-08T18:02:58.995594+00:00","model_set":{"reader":"glm-5.2"},"falsifier":"If one could exhibit a rectifiable dataset E with a non-polynomial bi-Lipschitz parametrization where the conditioning-based design criterion fails to correlate with actual separation capacity, the practical applicability of the lower-bound design recommendations would be called into question.","supporting_citations":[],"review_version":1}