Pith. sign in

REVIEW 3 major objections 3 minor 1 cited by

Eigenvector fluctuations and limit results for random graphs with infinite rank kernels

T0 review · 3 major / 3 minor · reviewed 2026-08-10 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read For latent position graphs with infinite-rank, indefinite link functions, the leading sample eigenvectors are shown to decompose into a Gaussian noise term plus a controlled residual, yielding entrywise graphon bounds and a rank-adaptive…

desk verdict Strong two-to-infinity perturbation theory for infinite-rank graphons, but the rank-adaptive test's N(0,2) claim rests on an unproved selector and fails as stated for exponentially decaying kernels. read the letter →

arxiv 2501.15725 v1 pith:S23CRAT2 submitted 2025-01-27 math.ST stat.TH

Minh Tang , Joshua R. Cape This is my paper · ORCID
classification math.STstat.TH MSC 05C8060B2062H1262H15
keywords latentpositiongraphsinfinite-rankkernelsindefinitetwo-to-infinitynormspectralembeddinggraphonestimationhypothesistestingrandom
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper extends spectral perturbation theory for random graphs beyond low-rank, positive-semidefinite models to the general latent position model, where the link function may have infinitely many nonzero eigenvalues and may take both signs. The main theorem packages the alignment residual as $\widehat U |\widehat\Lambda|^{1/2} W - U |\Lambda|^{1/2} = E U |\Lambda|^{-1/2} + Q$, with high-probability bounds in the maximum row norm, so the leading eigenvectors are not merely close to the population ones but have a Gaussian-dominated first-order behavior. On top of this expansion the paper derives entrywise estimation bounds for the edge probability matrix and a test for equality of latent positions whose null distribution is a weighted sum of independent chi-square variables; when the kernel has infinite rank the test statistic is asymptotically standard normal after centering. A sympathetic reader would say the upshot is that spectral embeddings remain statistically usable with growing dimension even when the kernel is indefinite and full rank.

What carries the argument

The central object is the expansion identity $\widehat U |\widehat\Lambda|^{1/2} W^{(n)} - U |\Lambda|^{1/2} = E U |\Lambda|^{-1/2} + Q$ measured in the two-to-infinity norm $\|M\|_{2\to\infty} = \max_i \|M_{i\cdot}\|$, i.e., the maximum Euclidean row length. The identity separates a linear noise term, whose rows are sums of independent mean-zero Bernoulli deviations, from a residual $Q$ controlled by the eigenvalue gap $\delta_r$ through sin-$\theta$ subspace alignment and matrix concentration inequalities. A leave-one-out analysis on the adjacency matrix, where one row and column are replaced by the population values, is what controls the delicate term $E(I-UU^\top)\widehat U$. The same decomposition is repeated for unweighted and eigenvalue-weighted embeddings, and a deterministic perturbation theorem abstracts the expansion so that only quantities linear in the noise matrix $E$ need to be bounded for a given application.

What would settle it

A decisive check is to simulate a latent position graph with an indefinite infinite-rank kernel such as $\kappa(x,y)=\cos(2\pi(x-y))$ on $[0,1]^2$, set $X_n=X_1$, choose $r$ by the data-driven rule, and compare the empirical null distribution of $T(\widehat X_1,\widehat X_n)$ over many replicates with $N(0,2)$ and with the weighted chi-square approximation using plug-in weights; systematic departures beyond Monte Carlo error would falsify Corollary 4.

Watch

Extended reading notes

Core claim

The central claim is that for independent-edge latent position graphs with kernels that may be indefinite and of infinite rank, the leading scaled sample eigenvectors can be aligned to the population eigenvectors by an orthogonal matrix $W$, after which the difference obeys $\widehat U |\widehat\Lambda|^{1/2} W - U |\Lambda|^{1/2} = E U |\Lambda|^{-1/2} + Q$ with explicit high-probability bounds on both the main term and the residual. The same program is carried out for $\widehat U W - U$ and $\widehat U \widehat\Lambda W - U \Lambda$. The results allow for repeated population eigenvalues, for the embedding dimension $r$ to grow with $n$, and for positive semidefinite as well as indefinite link functions; the indefinite case pays a heavier factor involving $|\lambda_r|^{-1/2}$ because $|P|$ has no entrywise closed form. These fine-grained bounds are then used to obtain entrywise high-probability errors for estimating the edge probability matrix and a plug-in, rank-adaptive test statistic whose null limit is a weighted sum of independent chi-square variables, reducing to $N(0,2)$ when the kernel has infinite rank.

Load-bearing premise

The load-bearing premise is that, for indefinite infinite-rank kernels, high-probability eigenvalue concentration sharp enough to let the embedding dimension $r$ grow with $n$ actually holds; the paper notes in Section 3.1 that such a bound is unavailable and only $O_p(n^{-1/2})$ rates are known under conditions that are difficult to verify.

Editorial extensions

If this is right

  • The expansion makes the leading embedding row-wise close to a linearized noise term, so entrywise statements about rows of spectral embeddings become feasible rather than only subspace-level or Frobenius-level statements.
  • For positive semidefinite kernels, the embedding dimension $r$ may grow with $n$ under explicit eigengap conditions, and entrywise estimation of the edge probability matrix achieves a rate involving $\lambda_r^{-1/2}(r^{1/2}+\log^{1/2} n)$ plus a bias term from truncation.
  • The test statistic $T(\widehat X_i, \widehat X_j)$ is computable from the adjacency matrix alone, adapts to the unknown kernel rank, and under the null hypothesis converges to a weighted sum of independent chi-square variables; for infinite-rank link functions it is asymptotically $N(0,2)$.
  • A row-wise central limit theorem for the scaled eigenvectors holds even for indefinite kernels, with covariance determined by the Bernoulli noise, as long as the eigenvalue gap satisfies the stated growth condition.
  • The data-driven rank selector based on the sample eigengap converges to the true finite rank when the kernel has finite rank and diverges when the kernel has infinite rank, removing the need for a user-specified embedding dimension in the testing problem.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the expansion is as sharp as claimed, the same main-term-plus-residual decomposition should carry over to general signal-plus-noise matrix models beyond Bernoulli graphs, because the deterministic perturbation result is linear in the noise matrix and does not use the graph structure itself.
  • The $N(0,2)$ limit for infinite-rank kernels is attractive but rests on spectral concentration that the paper does not fully establish for indefinite kernels; a safer practical route may be to use the weighted chi-square approximation with plug-in weights even when the data-driven rank does not grow.
  • Since repeated eigenvalues are explicitly allowed, the framework is plausibly applicable to kernels with symmetry, such as rotationally invariant kernels on spheres, where standard eigengap assumptions are known to fail; the paper's simulations with the Laplace kernel show near-zero gaps yet a small data-driven rank.
  • The numerical results suggest that, for smooth kernels, the effective dimension needed for hypothesis testing is far smaller than $n$, so the practical payoff of the infinite-rank theory may be to justify a low-dimensional approximation rather than to use a large number of spectral coordinates.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 3 minor

Summary. The paper studies spectral embeddings of latent position random graphs whose link function may have infinite rank and may be indefinite. Theorems 1–3 give two-to-infinity norm expansions of the form bU|bΛ|^{1/2}W(n) − U|Λ|^{1/2} = EU|Λ|^{-1/2} + Q, with explicit high-probability bounds on the main term and residual, for positive semidefinite and indefinite kernels; Theorem 5 is a deterministic perturbation counterpart. Corollary 2 provides row-wise normal approximations, Corollary 3 gives entrywise bounds for edge-probability estimation, and Theorem 4 with Lemma 1 and Corollary 4 propose a rank-adaptive test for equality of latent positions, claimed to converge to a weighted chi-square distribution, and to N(0,2) for infinite-rank kernels.

Significance. If the main theorems are correct, the paper is a substantial advance: it extends refined eigenvector fluctuation results from low-rank or positive semidefinite models to infinite-rank and indefinite kernels, supplies explicit constants in the main order term, and avoids population-level coherence assumptions. The proof machinery—leave-one-out analysis, matrix Bernstein bounds, and the deterministic perturbation framework of Theorem 5—is detailed and appears credible for Theorems 1–3, Corollaries 1–3, and the conditional statement in Theorem 4. No circularity or parameter fitting is present. However, the headline rank-adaptive inference claim in Corollary 4 is not proved, and as stated it is not compatible with the paper’s own assumptions for general infinite-rank indefinite kernels.

major comments (3)
  1. [§4.2, Corollary 4] The assertion that the data-driven rank b_r defined in Eq. (4.12) satisfies the hypotheses of Theorem 4 and Lemma 1 is unproved, and the thresholds in (4.12) do not match condition (4.11). Lemma 1 requires δ_r = ω(max{log^{3/2} n, sqrt(r n ρ_n) log n, (n ρ_n)^{3/4}/(r^{1/4}+log^{1/4} n)}), while the selector b_r only enforces sample eigengaps at least max{log^{7/4} n, sqrt(j dave(A)) log^{3/4} n, (dave(A))^{3/4}}. By Weyl's inequality this yields at best a population gap of order sqrt(j n ρ_n) log^{3/4} n, lacking the log n factor required by (4.11). Consequently b_r can overshoot the range in which the residual ε_n in (4.13) is controlled, and no argument is given that ε_n → 0 for the random, data-dependent b_r.
  2. [§4.2, Theorem 4 and Corollary 4] The claim that for every infinite-rank kernel 'r(n) → ∞ in probability' and that (4.15) holds is not established and is false without an explicit lower bound on eigenvalue gaps. A diverging sequence r(n) must satisfy δ_r = ω(sqrt(r n ρ_n) log n), equivalently μ_r − μ_{r+1} = ω(sqrt(r) log n). For kernels with exponentially decaying eigenvalues, as considered in Remark 6, every diverging r(n) fails this condition because μ_r − μ_{r+1} decays exponentially while sqrt(r) log n grows polynomially. Remark 7 acknowledges that gaps can be arbitrarily small but does not connect that observation to the rank-adaptive N(0,2) claim.
  3. [§3.1 and §4.2] For indefinite kernels the paper explicitly notes, after Eq. (3.21), that no high-probability spectral concentration inequality comparable to Eq. (2.5) is established; only O_p(n^{-1/2}) rates under conditions (3.22) or (3.23) are cited, and those conditions are acknowledged to be difficult to verify. Since Corollary 4 applies to general, possibly indefinite kernels and selects b_r from the eigenvalues of A, its proof would require a non-asymptotic lower bound on |λ_j| − |λ_{j+1}| in terms of the population spectrum. No such bound is provided, so the generality of the rank-adaptive test is not supported.
minor comments (3)
  1. [§4.2, Corollary 4] In item 2, 'probablity' should read 'probability'; in addition, since r(n) is a deterministic sequence in Theorem 4, the phrase 'r(n) → ∞ in probability' should be rephrased, for example as 'one may choose a diverging sequence r(n)'.
  2. [§4.1, Corollary 3] The displayed bound in Eq. (4.1) has an unmatched parenthesis, and the text 'we can reformulated these conditions' should read 'we can reformulate these conditions'.
  3. [§4.2, Corollary 4] The definition of b_r in Eq. (4.12) is an arg max over the set of j satisfying a threshold; if no such j exists the set is empty, and the convention for b_r should be stated explicitly.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the eigenvector expansions and limit theorems are derived from deterministic perturbation bounds and concentration inequalities, with no fitted parameter renamed as a prediction and no load-bearing self-citation chain.

full rationale

The paper's central claims are self-contained mathematical derivations. Theorems 1 and 2 decompose bU |bLambda|^{1/2}W - U |Lambda|^{1/2} into an explicit main term EU |Lambda|^{-1/2} and a residual Q, with the residual controlled by deterministic row-wise perturbation bounds (Theorem 5) and concentration lemmas for E = A - P (Lemmas 4-9). No unknown parameter is fitted to data and then reported as a prediction; the expansion terms are functions of P and E defined by the model. Corollary 1 converts population eigengap conditions into sample eigengap conditions via Weyl's inequality, which is a standard transfer argument and not circular. Theorem 4's null distribution for the latent-position test is derived from the entrywise expansion (A.79), a quadratic-form comparison theorem [48], and the Lindeberg-Feller CLT, with no assumption that the target limit holds beforehand. Lemma 1 proves consistency of the plug-in centering and scaling terms btheta and bsigma under explicit spectral-gap conditions, so the later corollary does not reduce to assuming its own conclusion. The paper's self-citations, e.g., [15], [16], [25], [49], and [53], are used as technical tools (two-to-infinity norm inequalities, signal-plus-noise blueprints, earlier low-rank RDPG testing, and entrywise eigenvector analysis) rather than as the sole justification of the new results, and no uniqueness theorem from the authors' own prior work is invoked to force the present modeling choice. The open issues flagged in the text, such as the lack of an Eq. (2.5)-type concentration bound for indefinite kernels in Section 3.1 and the unproven rank-adaptive behavior of br in Corollary 4, are correctness and assumption-verification gaps rather than circular reasoning: they concern whether stated conditions hold, not whether a claimed output is definitionally or statistically identical to its inputs. Accordingly, no specific circular step can be exhibited with quote and reduction.

Assumptions & free parameters 0 free parameters · 6 assumptions · 0 invented entities

No free parameters are fitted. The assumptions are explicit model, spectral gap, and concentration conditions, plus standard operator theory and probability tools. The invented-entities list is empty.

assumptions (6)
  • standard math Mercer's theorem and uniform convergence of kappa(x,y) = sum_r mu_r phi_r(x) phi_r(y) for continuous positive semidefinite kernels.
    Used in Proposition 1 and Remark 10 to relate the edge probability matrix P to the integral operator K; requires continuity and positive semidefiniteness.
  • standard math Spectral concentration inequality for kernel matrices, Eq. (2.5), citing [47, Theorem 7], holding with probability at least 1 - 2 n^{-c}.
    Used to derive data-driven rank selection and eigenvalue gap conditions in Remarks 4 through 6 and Corollary 1. The paper notes this inequality is not available for indefinite kernels.
  • standard math Matrix Bernstein inequalities, Davis-Kahan theorems, and leave-one-out independence arguments.
    Used throughout Appendix A to control ||E||, ||U^T E U||, and ||E(I - U U^T) bU||_{2 to infinity}.
  • domain assumption Latent position graph model with i.i.d. latent positions, independent Bernoulli edges, and sparsity parameter rho_n.
    Definition 1 is the data-generating model; all theoretical results are stated for this model.
  • domain assumption Population eigengap and eigenvalue magnitude conditions such as Eqs. (3.1), (3.2), (3.15), (3.16), and rate conditions in Eqs. (3.8), (3.20), and (4.11).
    These conditions on lambda_r and delta_r define the allowed embedding dimensions. The paper shows they can be satisfied for infinite-rank positive semidefinite kernels under eigenvalue decay assumptions, but they are assumptions on the unknown population matrix.
  • domain assumption For indefinite kernels, existence of a high-probability spectral convergence rate such as Eq. (3.22) or Eq. (3.23) when r grows with n.
    Section 3.1 explicitly states that only O_p(n^{-1/2}) results are known for indefinite kernels, and that the conditions needed are difficult to verify. Corollary 4's fully general rank-adaptive infinite-rank claim relies on this type of concentration being available.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Eigenvector fluctuations and limit results for random graphs with infinite rank kernels." pith.science (2026). https://pith.science/paper/S23CRAT2

@misc{pith2026250115725,
  author       = {Pith},
  title        = {Pith review of: Eigenvector fluctuations and limit results for random graphs with infinite rank kernels},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/S23CRAT2}},
  note         = {Machine review of arXiv:2501.15725}
}
read the original abstract

This paper systematically studies the behavior of the leading eigenvectors for independent edge undirected random graphs generated from a general latent position model whose link function is possibly infinite rank and also possibly indefinite. We first derive uniform error bounds in the two-to-infinity norm as well as row-wise normal approximations for the leading sample eigenvectors. We then build on these results to tackle two graph inference problems, namely (i) entrywise bounds for graphon estimation and (ii) testing for the equality of latent positions, the latter of which is achieved by proposing a rank-adaptive test statistic that converges in distribution to a weighted sum of independent chi-square random variables under the null hypothesis. Our fine-grained theoretical guarantees and applications differ from the existing literature which primarily considers first order upper bounds and more restrictive low rank or positive semidefinite model assumptions. Further, our results collectively quantify the statistical properties of eigenvector-based spectral embeddings with growing dimensionality for large graphs.

Figures

Figures reproduced from arXiv: 2501.15725 by the authors.

Figure 1
Figure 1. Plots of the forty largest eigenvalues (left panel) and gap between consecutive eigenvalues (right [PITH_FULL_IMAGE:figures/full_fig_p008_1.png] view at source ↗
Figure 2
Figure 2. Empirical histograms, based on 500 Monte Carlo replicates, for [PITH_FULL_IMAGE:figures/full_fig_p021_2.png] view at source ↗
Figure 3
Figure 3. Empirical histograms, based on 500 Monte Carlo replicates, for [PITH_FULL_IMAGE:figures/full_fig_p022_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. OpenAlex reports about 2 citations worldwide. Full citation record

  1. Predictive Subsampling for Scalable Inference in Networks

    stat.ME 2026-02 conditional novelty 6.0 of 10

    Predictive Subsampling estimates a GRDPG network by spectral embedding a random subgraph plus out-of-sample prediction, with O(n m d) cost and consistency in 2-to-infinity and Frobenius norms.

Reference graph

Works this paper leans on

68 extracted references · 63 canonical work pages · cited by 1 Pith paper

  1. [1]

    E. Abbe, J. Fan, K. Wang, and Y. Zhong. Entrywise eigenvector analysis of random matrices with low expected rank. Annals of Statistics , 48:1452–1474, 2020

    work page 2020

  2. [2]

    Agterberg, Z

    J. Agterberg, Z. Lubberts, and C. E. Priebe. Entrywise estimation of singular vectors of low-rank matrices with heteroskedasticity and dependence. IEEE Transactions on Information Theory , 68:4618– 4650, 2022

    work page 2022

  3. [3]

    E. M. Airoldi, T. B. Costa, and S. H. Chan. Stochastic blockmodel approximation of a graphon: Theory and consistent estimation. Advances in Neural Information Processing Systems , 26:692–700, 2013

    work page 2013

  4. [4]

    Athreya, V

    A. Athreya, V. Lyzinski, D. J. Marchette, C. E. Priebe, D. L. Sussman, and M. Tang. A limit theorem for scaled eigenvectors of random dot product graphs. Sankhya A , 78:1–18, 2016

    work page 2016

  5. [5]

    A. S. Bandeira and R. Van Handel. Sharp nonasymptotic bounds on the norm of random matrices with independent entries. Annals of Probability, 44:2479–2506, 2016

    work page 2016

  6. [6]

    M. Belkin. Approximation beats concentration? An approximation view on inference with smooth radial kernels. In Proceedings of the 31st Conference on Learning Theory , pages 1348–1361, 2018

    work page 2018

  7. [7]

    R. Bhatia. Matrix Analysis. Springer, 1997

    work page 1997

  8. [8]

    Billingsley

    P. Billingsley. Probability and Measure. John Wiley & Sons, Inc., 3 edition, 1995

    work page 1995

Show all 68 references
  1. [9]

    Bollob´ as, S

    B. Bollob´ as, S. Janson, and O. Riordan. The phase transition in inhomogeneous random graphs.Random Structures & Algorithms , 31:3–122, 2007

    2007

  2. [10]

    Borgs, J

    C. Borgs, J. Chayes, H. Cohn, and N. Holden. Sparse exchangeable graphs and their limits via graphon processes. Journal of Machine Learning Research , 18:7740–7810, 2018

    2018

  3. [11]

    Boucheron, G

    S. Boucheron, G. Lugosi, and P. Massart. Concentration Inequalities: A nonasymptotic theory of independence. Oxford University Press, 2013

    2013

  4. [12]

    C. Cai, G. Li, Y. Chi, H. V. Poor, and Y. Chen. Subspace estimation from unbalanced and incomplete data matrices: ℓ2,∞ statistical guarantees. Annals of Statistics , 49:944–967, 2021. 23

    2021

  5. [13]

    Cai and A

    T. Cai and A. Zhang. Rate-optimal perturbation bounds for singular subspaces with applications to high-dimensional statistics. Annals of Statistics , 46:60–89, 2018

    2018

  6. [14]

    Candes and B

    E. Candes and B. Recht. Exact matrix completion via convex optimization. Communications of the ACM, 55:111–119, 2012

    2012

  7. [15]

    J. Cape, M. Tang, and C. E. Priebe. The two-to-infinity norm and singular subspace geometry with applications to high-dimensional statistics. Annals of Statistics , 47:2405–2439, 2019

    2019

  8. [16]

    J. Cape, M. Tang, and C. E. Priebe. Signal-plus-noise matrix models: eigenvector deviations and fluctuations. Biometrika, 106:243–250, 2019

    2019

  9. [17]

    Chatterjee

    S. Chatterjee. Matrix estimation by universal singular value thresholding. Annals of Statistics , 43: 177–214, 2015

    2015

  10. [18]

    Chen and J

    Y. Chen and J. Lei. Minimax optimal probability matrix estimation for graphon with spectral decay. arXiv preprint #2410.01073, 2024

    2024 arXiv

  11. [19]

    Y. Chen, J. Fan, C. Ma, and Y. Yan. Inference and uncertainty quantification for noisy matrix comple- tion. Proceedings of the National Academy of Sciences , 116:22931–22937, 2019

    2019

  12. [20]

    Y. Chen, Y. Chi, J. Fan, and C. Ma. Spectral methods for data science: a statistical perspective. Foundations and Trends® in Machine Learning , 14:566–806, 2021

    2021

  13. [21]

    Cheng, Y

    C. Cheng, Y. Wei, and Y. Chen. Tackling small eigen-gaps: fine-grained eigenvector estimation and inference under heteroscedastic noise. IEEE Transactions on Information Theory , 67:7380–7419, 2021

    2021

  14. [22]

    Damle and Y.Sun

    A. Damle and Y.Sun. Uniform bounds for invariant subspace perturbations. SIAM Journal on Matrix Analysis and Its Applications , 41:1208–1236, 2020

    2020

  15. [23]

    Davis and W

    C. Davis and W. Kahan. The rotation of eigenvectors by a pertubation. III. Siam Journal on Numerical Analysis, 7:1–46, 1970

    1970

  16. [24]

    Drineas and I

    P. Drineas and I. C. F. Ipsen. Low-rank matrix approximations do not need a singular value gap. SIAM Journal on Matrix Analysis and Applications , 40:299–319, 2019

    2019

  17. [25]

    Du and M

    X. Du and M. Tang. Hypothesis testing for equality of latent positions in random graphs. Bernoulli, 29:3221–3254, 2023

    2023

  18. [26]

    J. Fan, W. Wang, and Y. Zhong. An ℓ∞ eigenvector perturbation bound and its application to robust covariance estimation. Journal of Machine Learning Research , 18:7608–7649, 2018

    2018

  19. [27]

    J. Fan, Y. Fan, X. Han, and J. Lv. Simple: Statistical inference on membership profiles in large networks. Journal of the Royal Statistical Society, Series B , 84:630–653, 2022

    2022

  20. [28]

    R. W. Farebrother. Algorithm AS204: The distribution of a positive linear combination of χ2 random variables. Journal of the Royal Statistical Society, Series C. , 33:332–339, 1984

    1984

  21. [29]

    C. Gao, Y. Lu, Z. Ma, and H. H. Zhou. Rate-optimal graphon estimation. Annals of Statistics , 43: 2624–2652, 2015

    2015

  22. [30]

    P. D. Hoff, A. E. Raftery, and M. S. Handcock. Latent space approaches to social network analysis. Journal of the American Statistical Association , 97(460):1090–1098, 2002

    2002

  23. [31]

    Horn and C

    R. Horn and C. Johnson. Topics in Matrix Analysis . Cambridge University Press, 1991

    1991

  24. [32]

    R. A. Horn. Norm bounds for Hadamard products and an arithmetic-geometric mean inequality for unitarily invariant norms. Linear Algebra and its Applications , 223:355–361, 1995

    1995

  25. [33]

    Hsing and R

    T. Hsing and R. Eubank. Theoretical foundations of functional data analysis with an introduction to linear operators. John Wiley and Sons, 2015. 24

    2015

  26. [34]

    Janson and S

    S. Janson and S. Ohlede. Can smooth graphons in several dimensions be represented by smooth graphons on [0, 1]? Examples and Counterexamples , 1:100011, 2021

    2021

  27. [35]

    Javanmard and A

    A. Javanmard and A. Montanari. Debiasing the Lasso: optimal sample size for Gaussian designs. Annals of Statistics , 46(6A):2593–2622, 2018

    2018

  28. [36]

    Klopp, A

    O. Klopp, A. Tsybakov, and N. Verzelen. Oracle inequalities for network models and sparse graphon estimation. Annals of Statistics , 45:316–354, 2017

    2017

  29. [37]

    Koltchinskii and E

    V. Koltchinskii and E. Gin´ e. Random matrix approximation of spectra of integral operators. Bernoulli, 6:113–167, 2000

    2000

  30. [38]

    J. Lei. Network representation using graph root distributions. Annals of Statistics , 49:745–768, 2021

    2021

  31. [39]

    L. Lei. Unified ℓ2→∞ eigenspace perturbation theory for symmetric random matrices. arXiv preprint #1909.04798, 2019

    1909 arXiv

  32. [40]

    Lov´ asz.Large networks and graph limits

    L. Lov´ asz.Large networks and graph limits . American Mathematical Society, 2012

    2012

  33. [41]

    Luo and C

    Y. Luo and C. Gao. Computational lower bounds for graphon estimation via low-degree polynomials. Annals of Statistics , 52:2318–2348, 2024

    2024

  34. [42]

    X. Mao, P. Sarkar, and D. Chakrabarti. Estimating mixed memberships with sharp eigenvector devia- tions. Journal of the American Statistical Association , 116:1928–1940, 2021

    1928

  35. [43]

    A. Modell. Entrywise error bounds for low-rank approximations of kernel matrices. In Advances in Neural Information Processing Systems 37 , 2024

    2024

  36. [44]

    M. Pensky. Dynamic network models and graphon estimation. Annals of Statistics, 47:2378–2403, 2019

    2019

  37. [45]

    M. Pensky. Davis-Kahan theorem in the two-to-infinity norm and its application to perfect clustering. arXiv preprint #2411.11728, 2024

    2024 arXiv

  38. [46]

    Y. Qin, L. Yu, and Y. Li. Iterative connecting probability estimation for networks. Advances in Neural Information Processing Systems, 34:1155–1166, 2021

    2021

  39. [47]

    Rosasco, M

    L. Rosasco, M. Belkin, and E. D. Vito. On learning with integral operators. Journal of Machine Learning Research, 11:905–934, 2010

    2010

  40. [48]

    V. I. Rotar. On the distribution of a quadratic form in many random variables. Theory of Probability and its Applications , 20:880–882, 1976

    1976

  41. [49]

    Rubin-Delanchy, J

    P. Rubin-Delanchy, J. Cape, M. Tang, and C. E. Priebe. A statistical interpretation of spectral embed- ding: the generalised random dot product graph. Journal of the Royal Statistical Society: Series B , 84: 1446–1473, 2022

    2022

  42. [50]

    Scetbon and Z

    M. Scetbon and Z. Harchaoui. A spectral analysis of dot-product kernels. In Proceedings of the 24th International Conference on Artificial Intelligence and Statistics , pages 3394–3402, 2021

    2021

  43. [51]

    Steinwart and A

    I. Steinwart and A. Christmann. Support Vector Machines. Springer, 2008

    2008

  44. [52]

    Takhanov

    R. Takhanov. On the speed of uniform convergence in Mercer’s theorem. Journal of Mathematical Analysis and Applications , 518:126718, 2023

    2023

  45. [53]

    M. Tang, D. L. Sussman, and C. E. Priebe. Universally consistent vertex classification for latent position graphs. Annals of Statistics , 41:1406 – 1430, 2013

    2013

  46. [54]

    J. A. Tropp. User-friendly tail bounds for sums of random matrices. Foundations of Computational Mathematics, 12:389–434, 2012

    2012

  47. [55]

    Udell and A

    M. Udell and A. Townsend. Why are big data matrices approximately low rank. SIAM Journal on Mathematics of Data Science , 1:144–160, 2019. 25

    2019

  48. [56]

    E. A. Valdivia. Relative concentration bounds for the spectrum of kernel matrices. arXiv preprint #1812.02108, 2018

    2018 arXiv

  49. [57]

    A. W. Van der Vaart. Asymptotic statistics . Cambridge University Press, 2000

    2000

  50. [58]

    Vershynin

    R. Vershynin. High-dimensional probability: an introduction with applications in data science, volume 47. Cambridge University Press, 2018

    2018

  51. [59]

    P. J. Wolfe and S. C. Olhede. Nonparametric graphon estimation. arXiv preprint #1309/5936, 2013

    2013

  52. [60]

    F. Xie. Entrywise limit theorems for eigenvectors of signal-plus-noise matrix models with weak signals. Bernoulli, 30:388–418, 2024

    2024

  53. [61]

    J. Xu. Rates of convergence of spectral methods for graphon estimation. In Proceedings of the 35th International Conference on Machine Learning , pages 5433–5442, 2018

    2018

  54. [62]

    Y. Yan, Y. Chen, and J. Fan. Inference for heteroskedastic pca with missing data. Annals of Statistics , 52:729–756, 2024

    2024

  55. [63]

    J. J. Yang, Q. Han, and E. M. Airoldi. Nonparametric estimation and testing of exchangeable graph models. In Proceedings of the Seventeenth International Conference on Artificial Intelligence and Statis- tics, pages 1060–1067, 2014

    2014

  56. [64]

    Zhang, E

    Y. Zhang, E. Levina, and J. Zhu. Estimating network edge probabilities by neighbourhood smoothing. Biometrika, 104:771–783, 2017

    2017

  57. [65]

    U ⊤ + bU+ − W (+) 0 0 U ⊤ − bU− − W (−) # +

    Y. Zhong and N. Boumal. Near-optimal bounds for phase synchronization. SIAM Journal on Optimiza- tion, 28:989–1016, 2018. 26 A Proofs of stated results A.1 Proof of Theorem 1 (positive semidefinite kernel κ) Recall that by convention, we index the eigenvalues of A and P in dec...

    2018

  58. [66]

    + 32C1(ν)(rρn log n)1/2∥E∥ δrλr + 16(∥E∥ · ∥U ∥2→∞ + ∥EU ∥2→∞)(Cν log n + 2∥E∥) δrλr , ∥Y1∥2→∞ ≤ 24C2(ν) logn + 64δ−1 r ∥E∥2 λr ∥T∗∥2→∞ + ∥R1∥2→∞ + ∥R2∥2→∞ + ∥Y0∥2→∞ . Once again, substituting the bounds for ψ0, ψ1, ψ2, ψ(k) 3 and ∥E∥ in the proof of Theorem 1 together with so...

  59. [67]

    For any A → ∞, sup 1≤k≤n Z |x|≥A x2dGk(x) → 0. (A.82)

  60. [68]

    (A.84) 53 Then, Q(W (n), F) → Q(W (n), G) in distribution as n → ∞

    For any fixed but arbitrary ϵ >0, nX k=1 nX ℓ=k+1 m2 kℓ Z Z |xy|≥|ϵ/mkℓ| |xyΨk(x)Ψℓ(y)| dx dy → 0, (A.83) nX k=1 s2 k(n) Z |x|≥ϵ/sk(n) |xΨk(x)| dx → 0. (A.84) 53 Then, Q(W (n), F) → Q(W (n), G) in distribution as n → ∞. Recall that we denote d2 k = 2pik(1 − pik). We now verify...

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.