Pith. sign in

REVIEW 3 major objections 4 minor 108 references

This paper proves that four low-cost tests—ridge/group-LASSO regression and two canonical-correlation variants—reliably detect whether node-level covariates are associated with the latent structure of a low-rank network, even when the covar

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

Four permutation tests (ridge, group LASSO, and two CCA variants) detect association between node covariates and random-dot-product-graph latent structure, with consistency theorems and cheaper computation than prior dependency tests.

T0 review reviewed 2026-08-05 challenge →

load-bearing objection Solid, useful network dependency testing toolkit with honest proofs; the abstract overclaims high-dimensional coverage for the CCA-based methods, which are fixed-p only. the 3 major comments →

arxiv 2509.03772 v1 pith:TUOKXXO7 submitted 2025-09-03 stat.ML cs.LGmath.STstat.TH

Testing for correlation between network structure and high-dimensional node covariates

Alexander Fuchs-Kreiss , Keith Levin This is my paper
classification stat.ML cs.LGmath.STstat.TH MSC 62H2062F0305C8062H12
keywords network dependency testingrandom dot product graphcanonical correlation analysisspectral embeddinghigh-dimensional covariatespermutation testsgroup LASSOridge regression
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

The reading

The paper shows how to test whether node-level covariates are correlated with network structure without observing the latent variables that drive the network. Its central claim is that plugging spectral estimates of latent positions into ridge or group-LASSO regression, or into canonical correlation analysis, yields statistics that converge to the oracle quantities one would get from the true latent positions. The CCA guarantees hold for both the estimated-latent-position version and a regularized version that works directly on the adjacency matrix, avoiding the need to choose the latent dimension. All four methods run on matrix-vector products and leading singular values, so they scale to large networks and high-dimensional feature vectors. Simulations and two real-data analyses indicate the tests control their level and detect association under linear, nonlinear, and assortative-mixing generative models.

Core claim

The central discovery is that network dependency testing can be carried out consistently on estimated latent positions rather than true ones. The paper proves (Theorem 3.12) that |rho(hat X, Z) - rho(X,Z)| = oP(1) when spectral embeddings recover the latent positions with (2,infinity)-norm error xi_n = o(1); the unknown orthogonal rotation that makes hat X an estimate of X cancels because the CCA coefficient is a subspace-alignment quantity. It proves further (Theorem 3.13) that canonical correlation between the adjacency matrix itself and Z, with a ridge-style regularization gamma I on the network sample covariance, also tracks the oracle coefficient at rate OP(eta_n/sqrt(gamma n) + |gamma|

What carries the argument

The central object is the CCA alignment score rho(X,Z), the leading singular value of the normalized cross-covariance matrix Sigma_X^{-1/2} Sigma_{XZ} Sigma_Z^{-1/2}, and its plug-in analogues CCA(hat X, Z) and CCA_gamma(A,Z). The workhorse is the two-to-infinity norm spectral embedding guarantee (Assumption 2): hat X recovers X up to an orthogonal rotation with row-wise error xi_n = o(1). Because the CCA coefficient is invariant to orthogonal rotation of either data block, the unknown rotation cancels; sin-theta subspace perturbation bounds then transfer row-wise embedding error into singular-value error. For the adjacency CCA, the paper proves a pseudo-inverse lemma showing that the nonzer

Load-bearing premise

All guarantees hang on the assumption that the observed network is truly a low-rank latent-position network whose edges are independent conditional on the latent positions and concentrate as required, with the latent dimension d known: if edges are dependent or heavy-tailed, or d is misspecified, the plug-in statistics need not converge to the oracle correlation.

What would settle it

Fix a stochastic blockmodel with d=2 true latent dimensions and Bernoulli edges, set the covariates independent of X, and run the spectral-embedding CCA permutation test with d incorrectly chosen as 3; if rejection rates at nominal 5% exceed binomial sampling error substantially as n grows, Assumption 2's correct-dimension premise fails and Theorem 3.12 no longer applies. Similarly, generate weighted RDPG edges with heavy-tailed noise and check whether the deviation of the regularized adjacency-CCA statistic from the oracle stays within the eta_n, zeta_n rates claimed in Theorem 3.13.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

If this is right

  • Network scientists can test whether node attributes track latent structure without strong parametric assumptions, using spectral embedding plus CCA and permutation p-values.
  • Ridge and group-LASSO methods remain consistent when p grows with n, with rates slowed only by the spectral embedding error xi_n, so covariate selection on network-linked data is feasible.
  • The regularized adjacency CCA avoids choosing the latent dimension d, replacing a model-selection problem with a regularization parameter gamma that has a wide safe range.
  • The methods are computationally cheaper than diffusion-map distance-correlation baselines, relying only on matrix-vector products and leading singular values.
  • The theorems give formal level control for the permutation tests under the null in the low-rank model, and power follows against linear association (and empirically under nonlinear misspecification).

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • A testable extension the paper leaves open is replacing the sample covariance of Z in CCA(hat X, Z) with sparse or regularized CCA when p >> n; simulations suggest power would then survive high-dimensional regimes where dense CCA currently loses to LASSO.
  • The proof of Theorem 3.13 relies only on concentration and subspace perturbation, not on Bernoulli edges, so the same consistency should transfer to generalized random dot product graphs and graph-root models, as the paper anticipates.
  • The group-LASSO covariance test derived in Appendix F is a standalone contribution to multivariate sparse regression: it yields a closed-form first-entry alpha and test statistic without cross-validation, independent of the network setting.
  • Since the plug-in CCA error is driven by xi_n, the practical bottleneck is spectral dimension selection (Remark 2.3); improving automated rank selection would strengthen all four tests more than improving covariate regularization.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. The paper introduces four (or, counting group LASSO, five) methods for testing association between observed node-level covariates and latent network structure under a low-rank latent-space model: ridge regression, (group) LASSO, CCA applied to estimated latent positions, and a regularized CCA applied directly to the adjacency matrix. The main theoretical results are Theorem 3.4 (ridge error bound), Theorem 3.8 (group-LASSO convergence rate with p allowed to grow), Theorem 3.12 (consistency of the plug-in CCA coefficient for fixed p), and Theorem 3.13 (consistency of the regularized adjacency-CCA coefficient for fixed p). The paper also contains simulations and real-data analyses using permutation tests based on these statistics.

Significance. The paper addresses a practically important problem and proposes computationally attractive methods. The proofs are based on standard and appropriate tools: the ridge and group-LASSO arguments follow the Buhlmann--van de Geer framework with an extra term for the estimated latent positions, and the CCA theorems reduce to singular-subspace perturbation bounds. The group-LASSO theorem, in particular, gives a concrete high-dimensional rate in terms of the embedding error xi_n and the compatibility constant. If the scope were stated accurately, the fixed-p CCA consistency results and the high-dimensional regression results would be a solid contribution. However, the advertised high-dimensional guarantee for the two CCA-based methods is not delivered, and the paper does not provide test-level guarantees for the permutation procedures that form the actual inferential proposal.

major comments (3)
  1. [Abstract and Section 2.2, Eq. (12)-(14); Remark 3.11] The abstract claims theoretical guarantees for all four methods when node covariates are high-dimensional. This is unsupported for the CCA-based methods. Section 2.2 explicitly assumes p is fixed, Remark 3.11 states that high-dimensional CCA is deferred to future work, and Theorems 3.12 and 3.13 both use eSigma_Z^{-1/2} under Assumption 4 requiring Sigma_Z invertible. When p_n > n, eSigma_Z is singular and the proofs' spectral-norm perturbation arguments cannot be controlled. The only p -> infinity theory in the paper is for ridge (Theorem 3.4/Corollary 3.5) and group LASSO (Theorem 3.8). The abstract and Section 5 should be revised to distinguish the high-dimensional guarantees for regression methods from the fixed-p consistency results for CCA methods, or new high-dimensional CCA theory must be supplied.
  2. [Section 4.1; Theorems 3.12 and 3.13] The paper's stated goal is testing, but Theorems 3.12 and 3.13 only establish that the test statistic converges in probability to the oracle population CCA coefficient. They do not establish that the permutation test controls level under H0 or has power. For the CCA-based methods, H0 is defined in Section 4.1 as zero population covariance, which is weaker than the row-exchangeability of Z given X needed for exact permutation inference. Unless the null model is taken to include independence of X and Z, or an asymptotic analysis of the permutation null distribution is supplied, the central testing claim is not proven. This is a load-bearing gap, not a presentation issue.
  3. [Section 4.2, Tables and Figures; Remark 3.11] The simulations apply CCA and netCCA with n=100, p=200 using the Fisher-Sun shrinkage estimator, a setting not covered by Theorem 3.12 or 3.13. The paper acknowledges this in Remark 3.11, but the abstract's high-dimensional promise and the simulation section should clearly separate these heuristic demonstrations from the proved guarantees. Otherwise the reader cannot tell which simulation results are covered by theory.
minor comments (4)
  1. [Abstract and Section 5] The paper alternates between four and five methods. The abstract says four novel methods, but Section 5 summarizes five methods (ridge, LASSO, group LASSO, CCA, netCCA). This should be reconciled, e.g., by treating group LASSO as a variant of the LASSO-based approach.
  2. [Section 1, paragraph 2] There is a typo: 'constitistutes' should be 'constitutes'.
  3. [Figure 3 caption] The bottom-left and bottom-right panels are described in the caption as 'sSNR' but appear to refer to rSNR; please check and standardize the notation.
  4. [Remark 3.11 and Section 4.2] It would be helpful to add a short table or remark listing which of the six simulated scenarios are covered by which theorem, and which are outside the theorem assumptions. This would prevent the reader from thinking the high-dimensional CCA simulations are theoretically guaranteed.

Circularity Check

0 steps flagged

No significant circularity; the core derivations are self-contained under the stated assumptions.

full rationale

The paper's central theorems (3.4, 3.8, 3.12, 3.13) are consistency and oracle-type bounds derived from explicit concentration and spectral-embedding assumptions (Assumptions 1-3) via external perturbation theory (Cai & Zhang 2018; Yu et al. 2014; Cape et al. 2019). The target quantities (B, rho_{X,Z}) are defined on latent quantities X and Z, not on the observed adjacency matrix, so replacing X by Xhat or A is not a self-definitional reduction: the theorems prove convergence to those oracle quantities rather than assuming it. The regularization parameters (ridge alpha, LASSO alpha, netCCA gamma) are not fitted to force the theorems; Theorem 3.13 and Corollary 3.14 hold for ranges of gamma, and Theorem 3.8 for alpha satisfying stated lower bounds. Self-citations such as Levin et al. (2022) provide concrete rates for weighted RDPGs, but they are not load-bearing because the main results are stated under generic Assumptions 1-3 and remain conditional on those assumptions. No fitted input is renamed as a prediction, and no cited uniqueness theorem is used to declare a choice forced. The abstract's claim of high-dimensional guarantees for all four methods is broader than the body's fixed-p CCA theorems (Section 2.2 explicitly says 'we assume that the dimension p of the features Z is fixed with respect to n'; Remark 3.11 defers high-dimensional CCA), but this is a scope/overclaim, not circularity. Overall, the derivation chain does not reduce to its own inputs.

Axiom & Free-Parameter Ledger

4 free parameters · 9 axioms · 0 invented entities

The central results are conditional on a package of model assumptions that the paper states explicitly (Assumptions 1-4 and the linear model in Equation (2) for the regression methods). None of these assumptions is proven from more primitive data conditions; they define the regime. The practical choices (d, alpha, gamma, sigma_hat^2) are tuning decisions that the theory requires or permits but does not pin down, which is why they are listed as free parameters rather than inputs. The paper postulates no new entities: latent positions, spectral embeddings, ridge/LASSO, and CCA are all imported from the cited literature.

free parameters (4)
  • Latent dimension d = heuristic: 14 (PPI), 55 (Wikipedia) via Li et al. (2020); unnamed in simulations, capped at 20 in Scenario (iv)
    All ASE-based theorems treat d as fixed and known; Remark 2.3 concedes correct dimension selection is a major open problem, so the guarantees inherit the user's choice.
  • Ridge/LASSO tuning alpha = cross-validated in experiments; theory allows any alpha >= 2 alpha_0,n (LASSO) or alpha_k > 0 (ridge)
    The ridge bound degrades as 1/min alpha_k and applies only for alpha_k > 0; Corollary 3.5's favorable alpha_k depends on unknown ||b(k)|| and sigma_k^2, so the stated rates are not fully adaptive.
  • netCCA shrinkage gamma = gamma = sqrt(n) in experiments
    Theorem 3.13 requires only eta_n^2/n << gamma << n; gamma = sqrt(n) is a heuristic inside the allowed range, so the experimental regime is covered for standard rates but the choice is not data-driven.
  • Covariance test variance sigma_hat^2 = empirical variance of embedded observations under H0 (Remark F.4)
    The covariance test statistic T_cov is inversely proportional to sigma_hat^2; under H0 the paper substitutes the empirical variance of Xhat, an estimator whose plug-in validity with estimated latent positions is not proven.
axioms (9)
  • domain assumption The observed network is a generalized RDPG: edges are conditionally independent given latent positions X, with E[A | X] = XX^T
    Equation (1) and surrounding text; all four methods and all theorems are conditional on this edge-generation structure.
  • domain assumption Assumption 1: ||A - XX^T|| = O_P(eta_n)
    Section 3; instantiated by Lei and Rinaldo (eta_n = sqrt(n)) and subgamma bounds (Levin et al. 2022).
  • domain assumption Assumption 2: an estimator Xhat exists with ||Xhat - XQ||_{2,infinity} = O_P(xi_n) for some orthogonal Q
    Section 3; load-bearing for Theorems 3.4, 3.8 and 3.12. Requires a spectral gap and correct d, so real networks without such structure fall outside.
  • domain assumption Assumption 3: W^T(A - XX^T)V = O_P(zeta_n) for arbitrary bounded-norm test matrices W, V
    Section 3; load-bearing for Theorem 3.13 (network CCA).
  • domain assumption Assumption 4: (Xi, Zi) i.i.d. with invertible Sigma_X and Sigma_Z; p fixed for the CCA theorems
    Section 3 and Remark 3.11; defines the target rho_X,Z and the fixed-p regime of the CCA results.
  • domain assumption Exact linear model X = ZB + E with independent rows (Equation 2)
    Section 2.1; required by Theorems 3.4 and 3.8; simulation scenarios (v) and (vi) deliberately violate it.
  • domain assumption Subgamma tail conditions on the entries of Z and E
    Theorems 3.4, 3.8 and Lemma B.3; standard but unverified for real covariates.
  • domain assumption Random multitask compatibility condition 1/phi_S = O_P(1)
    Definition 3.7 and Section 3.1.2; a random-covariate analogue of the classical LASSO compatibility condition.
  • ad hoc to paper Correct selection of the latent dimension d
    Remark 2.3; the theory is stated under the assumption that d has been selected correctly, while the paper concedes this is a major open problem.

reviewed 2026-08-05 · how reviews work

0 comments
Cite this review

Pith. "Pith review of Testing for correlation between network structure and high-dimensional node covariates." pith.science (2026). https://pith.science/paper/TUOKXXO7

@misc{pith2026250903772,
  author       = {Pith},
  title        = {Pith review of: Testing for correlation between network structure and high-dimensional node covariates},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TUOKXXO7}},
  note         = {Machine review of arXiv:2509.03772}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

In many application domains, networks are observed with node-level features. In such settings, a common problem is to assess whether or not nodal covariates are correlated with the network structure itself. Here, we present four novel methods for addressing this problem. Two of these are based on a linear model relating node-level covariates to latent node-level variables that drive network structure. The other two are based on applying canonical correlation analysis to the node features and network structure, avoiding the linear modeling assumptions. We provide theoretical guarantees for all four methods when the observed network is generated according to a low-rank latent space model endowed with node-level covariates, which we allow to be high-dimensional. Our methods are computationally cheaper and require fewer modeling assumptions than previous approaches to network dependency testing. We demonstrate and compare the performance of our novel methods on both simulated and real-world data.

Figures

Figures reproduced from arXiv: 2509.03772 by Alexander Fuchs-Kreiss, Keith Levin.

Figure 1
Figure 1. Figure 1: Graphs describing the different ways how [PITH_FULL_IMAGE:figures/full_fig_p011_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Rejection rates as a function of rSNR for the five proposed methods in the six [PITH_FULL_IMAGE:figures/full_fig_p028_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Rejection rates of the five different tests for different choices of [PITH_FULL_IMAGE:figures/full_fig_p029_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Plots show the rejection rates (left) and run-times (right) of the different tests for [PITH_FULL_IMAGE:figures/full_fig_p030_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Protein network analyzed in Section 4.3.1. Vertex sizes are proportional to their degree and the vertices are colored according to the four most relevant protein categories as identified by our LASSO-based method: [T] - Transcription, [F] - Protein fate, [P] - Translation, [U] - Uncharacterized 32 [PITH_FULL_IMAGE:figures/full_fig_p032_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Plots of the Wikipedia network with vertices colored according to four different [PITH_FULL_IMAGE:figures/full_fig_p034_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: QQ-plot of the empirical distributions of the p-values generated from the different [PITH_FULL_IMAGE:figures/full_fig_p077_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: QQ-plot of the empirical distributions of the p-values generated from the different [PITH_FULL_IMAGE:figures/full_fig_p078_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Eigenvalues (largest to smallest) of the adjacency matrices of the respective real [PITH_FULL_IMAGE:figures/full_fig_p079_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Protein network analyzed in Section 4.3.1. Vertex sizes are proportional to their degree. Vertex colors in plot k correspond to the values of the k-th column of ZBˆgLASSO. 80 [PITH_FULL_IMAGE:figures/full_fig_p080_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Protein network analyzed in Section 4.3.1. Vertex sizes are proportional to their degree. Vertex colors in plot k correspond to the values of the k-th column of ZBˆgLASSO. 81 [PITH_FULL_IMAGE:figures/full_fig_p081_11.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

108 extracted references · 71 canonical work pages · 7 internal anchors

  1. [1]

    Acemoglu

    D. Acemoglu. Systemic risk and stability in financial networks. American Economic Review, 105 0 (2): 0 564--608, 2015

    2015

  2. [2]

    Acharyya, J

    A. Acharyya, J. Agterberg, M. W. Trosset, Y. Park, and C. E. Priebe. Semisupervised regression in latent structure networks on unknown manifolds. Applied Network Science, 8 0 (75), 2023

    2023

  3. [3]

    E. M. Airoldi, D. M. Blei, S. E. Fienberg, and E. P. Xing. Mixed Membership Stochastic Blockmodels . Journal of Machine Learning Research, 9: 0 1981--2014, 2008

    1981

  4. [4]

    D. J. Aldous. Representations for partially exchangeable arrays of random variables. Journal of Multivariate Analysis, 11 0 (4): 0 581--598, 1981

    1981

  5. [5]

    U. Alon. Biological networks: The tinkerer as an engineer. Science, 301 0 (5641): 0 1866--1867, 2003

    2003

  6. [6]

    Athreya, D

    A. Athreya, D. E. Fishkind, K. Levin, V. Lyzinski, Y. Park, Y. Qin, D. L. Sussman, M. Tang, J. T. Vogelstein, and C. E. Priebe. Statistical inference on random dot product graphs: a survey. Journal of Machine Learning Research, 18 0 (226): 0 1--92, 2018

    2018

  7. [7]

    Barab \'a si, N

    A.-L. Barab \'a si, N. Gulbahce, and J. Loscalzo. Network medicine: a network-based approach to human disease. Nature reviews genetics, 12 0 (1): 0 56--68, 2011

    2011

  8. [8]

    R. Bhatia. Matrix Analysis. Springer, 1997

    1997

  9. [9]

    P. J. Bickel and E. Levina. Regularized estimation of large covariance matrices. The Annals of Statistics, 36: 0 199--227, 2008

    2008

  10. [10]

    Binkiewicz, J

    N. Binkiewicz, J. T. Vogelstein, and K. Rohe. Covariate-assisted spectral clustering. Biometrika, 104 0 (2): 0 361--377, 2017

    2017

  11. [11]

    Boucheron, G

    S. Boucheron, G. Lugosi, and P. Massart. Concentration Inequalities. Oxford University Press, 2013

    2013

  12. [12]

    B\" u hlmann and S

    P. B\" u hlmann and S. van de Geer. Statistics for High-Dimensional Data. Springer, 2011

    2011

  13. [13]

    T. T. Cai and A. Zhang. Rate-optimal perturbation bounds for singular subspaces with applications to high-dimensional statistics,. The Annals of Statistics, 46 0 (1): 0 60--89, 2018

    2018

  14. [14]

    J. Cape, M. Tang, and C. E. Priebe. The two-to-infinity norm and singular subspace geometry with applications to high-dimensional statistics. The Annals of Statistics, 47 0 (5): 0 2405--2439, 2019

    2019

  15. [15]

    Chandna, S

    S. Chandna, S. C. Olhede, and P. J. Wolfe. Local linear graphon estimation using covariates. Biometrika, 109 0 (3): 0 721--734, 2022

    2022

  16. [16]

    Profile least squares estimation in networks with covariates

    S. Chandna, B. Bagozzi, and S. Chatterjee. Profile least squares estimation in networks with covariates. arXiv:2412.16298, 2024

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  17. [17]

    T. Chaney. The network structure of international trade. American Economic Review, 104 0 (11): 0 3600--3634, 2014

    2014

  18. [18]

    J. H. Chang and S. Paul. Embedding network autoregression for time series analysis and causal peer effect inference. arXiv:2406.05944, 2024

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  19. [19]

    F. Chen, S. Roch, K. Rohe, and S. Yu. Estimating graph dimension with cross-validated eigenvalues. arXiv:2108.03336, 2021

    arXiv 2021

  20. [20]

    Coppersmith

    G. Coppersmith. Vertex nomination. WIRE s Computational Statistics , 6 0 (2): 0 144--153, 2014

    2014

  21. [21]

    Elliott, B

    M. Elliott, B. Golub, and M. O. Jackson. Financial networks and contagion. American Economic Review, 104 0 (10): 0 3115--3153, 2014

    2014

  22. [22]

    T. J. Fisher and X. Sun. Improved S tein-type shrinkage estimators for the high-dimensional multivariate normal covariance matrix. Computational Statistics & Data Analysis, 55 0 (5): 0 1909--1918, 2011

    1909

  23. [23]

    D. E. Fishkind, V. Lyzinski, H. Pao, L. Chen, and C. E. Priebe. Vertex nomination schemes for membership prediction. The Annals of Applied Statistics, 9 0 (3): 0 1510--1532, 2015

    2015

  24. [24]

    B. K. Fosdick and P. D. Hoff. Testing and modeling dependencies between a network and nodal attributes. Journal of the American Statistical Association, 110 0 (511): 0 1047--1056, 2015

    2015

  25. [25]

    C. Gao, Z. Ma, and H. H. Zhou. Sparse CCA : Adaptive estimation and computational barriers. The Annals of Statistics, 45 0 (5): 0 2074--2101, 2017

    2074

  26. [26]

    C. Gao, Z. Ma, A. Y. Zhang, and H. H. Zhou. Community detection in degree-corrected block models. The Annals of Statistics, 46 0 (5): 0 2153--2185, 2018

    2018

  27. [27]

    J. Golbeck. Analyzing the Social Web. Morgan Kaufmann, 2013

    2013

  28. [28]

    Hakes, J

    L. Hakes, J. W. Pinney, D. L. Robertson, and S. C. Lovell. Protein-protein interaction networks and biology---what's the connection? Nature Biotechnology, 26: 0 69--72, 2008

    2008

  29. [29]

    F. Han, Y. Gu, G. L. Brown, X. Zhang, and X. Liu. Neuroimaging contrast across the cortical hierarchy is the feature maximally linked to behavior and demographics. Neuroimage, 215, 2020

    2020

  30. [30]

    X. Han, Q. Yang, and Y. Fan. Universal rank inference via residual subsampling with application to large networks. The Annals of Statistics, 51 0 (3): 0 1109--1133, 2023

    2023

  31. [31]

    Hastie, R

    T. Hastie, R. Tibshirani, and M. Wainwright. Statistical Learning with Sparsity. CRC Press, 2015

    2015

  32. [32]

    M. J. Hawrylycz, E. S. Lein, A. L. Guillozet-Bongaarts, E. H. Shen, L. Ng, et al. An anatomically comprehensive atlas of the adult human brain transcriptome. Nature, 489: 0 391--399, 2012

    work page 2012

  33. [33]

    Hayes and K

    A. Hayes and K. Levin. Peer effects in the linear-in-means model may be inestimable even when identified. arXiv:2410.10772, 2024

    work page arXiv 2024

  34. [34]

    Hayes, M

    A. Hayes, M. M. Fredrickson, and K. Levin. Estimating network-mediated causal effects via principal components network regression. Journal of Machine Learning, 26 0 (13): 0 1--99, 2025

    work page 2025

  35. [35]

    P. D. Hoff. Bilinear mixed-effects models for dyadic data. Journal of the American Statistical Association, 100 0 (469): 0 286--295, 2005

    work page 2005

  36. [36]

    P. D. Hoff, A. E. Raftery, and M. S. Handcock. Latent space approaches to social network analysis. Journal of the American Statistical Association, 97 0 (460): 0 1090--1098, 2002

    work page 2002

  37. [37]

    P. W. Holland, K. B. Laskey, and S. Leinhardt. Stochastic blockmodels: First steps. Social Networks, 5 0 (2): 0 109--137, 1983

    work page 1983

  38. [38]

    Hotelling

    H. Hotelling. Relations between two sets of variates. Biometrika, 28 0 (3--4): 0 321--377, 1936

    work page 1936

  39. [39]

    James, D

    C. James, D. Yuan, I. Gaynanova, and J. Arroyo. Learning joint and individual structure in network data with covariates. arXiv:2406.08776, 2024

    Pith/arXiv arXiv 2024

  40. [40]

    P. Ji, J. Jin, Z. T. Ke, and W. Li. Co-citation and co-authorship networks of statisticians. Journal of Business & Economic Statistics, 40 0 (2): 0 469--485, 2022

    2022

  41. [41]

    J. Jin, Z. T. Ke, and S. Luo. Mixed membership estimation for social networks. Journal of Econometrics, 239 0 (2): 0 105369, 2024

    work page 2024

  42. [42]

    Karrer and M

    B. Karrer and M. E. J. Newman. Stochastic blockmodels and community structure in networks. Physical Review E, 83 0 (1): 0 016107, 2011

    work page 2011

  43. [43]

    Computational Inference for Directions in Canonical Correlation Analysis

    D. Kessler and E. Levina. Computational inference for directions in canonical correlation analysis. arXiv:2308.11218, 2023

    work page internal anchor Pith review Pith/arXiv arXiv 2023

  44. [44]

    Kreiss, E

    A. Kreiss, E. Mammen, and W. Polonik. Common drivers in sparsely interacting hawkes processes, 2025. URL https://arxiv.org/abs/2504.03916

    Pith/arXiv arXiv 2025

  45. [45]

    Latouche, S

    P. Latouche, S. Robin, and S. Ouadah. Goodness of fit of logistic regression models for random graphs. Journal of Computational and Graphical Statistics, 27 0 (1): 0 98--109, 2018

    work page 2018

  46. [46]

    C. M. Le, K. Levin, P. J. Bickel, and E. Levina. Comment: Ridge regression and regularization of large matrices. Technometrics, 62 0 (4): 0 443--446, 2020

    work page 2020

  47. [47]

    Ledoit and M

    O. Ledoit and M. Wolf. Nonlinear shrinkage estimation of large-dimensional covariance matrices. The Annals of Statistics, 40: 0 1024--1060, 2012

    work page 2012

  48. [48]

    Ledoit and M

    O. Ledoit and M. Wolf. Optimal estimation of a large-dimensional covariance matrix under S tein’s loss. Bernoulli, 24 0 (4B): 0 3791--3832, 2018

    work page 2018

  49. [49]

    Y. Lee, C. Shen, C. E. Priebe, and J. T. Vogelstein. Network dependence testing via diffusion maps and distance-based correlations. Biometrika, 106 0 (4): 0 857--873, 2019

    work page 2019

  50. [50]

    J. Lei. Network representation using graph root distributions. The Annals of Statistics, 49 0 (2): 0 745--768, 2021

    work page 2021

  51. [51]

    Lei and A

    J. Lei and A. Rinaldo. Consistency of spectral clustering in stochastic block models. The Annals of Statistics, 43 0 (1): 0 215--237, 2015

    work page 2015

  52. [52]

    Levin and V

    K. Levin and V. Lyzinski. Laplacian eigenmaps from sparse, noisy similarity measurements. IEEE Transactions on Signal Processing , 65 0 (8): 0 1988--2003, 2017

    work page 1988

  53. [53]

    Levin, A

    K. Levin, A. Athreya, M. Tang, V. Lyzinski, Y. Park, and C. E. Priebe. A central limit theorem for an omnibus embedding of random dot product graphs. arXiv:1705.09355v5, 2017

    Pith/arXiv arXiv 2017

  54. [54]

    Vertex Nomination in Richly Attributed Networks

    K. Levin, C. E. Priebe, and V. Lyzinski. Vertex nomination in richly attributed networks. arXiv:2005.02151, 2020

    work page internal anchor Pith review Pith/arXiv arXiv 2005

  55. [55]

    Levin, A

    K. Levin, A. Lodhia, and E. Levina. Recovering shared structure from multiple networks with unknown edge distributions. Journal of Machine Learning Research, 23 0 (3): 0 1--48, 2022. URL http://jmlr.org/papers/v23/19-1056.html

    work page 2022

  56. [56]

    J. Li, G. Xu, and J. Zhu. High-dimensional factor analysis for network-linked data. arXiv:2403.17121, 2024 a

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  57. [57]

    T. Li, E. Levina, and J. Zhu. Prediction models for network-linked data. The Annals of Applied Statistics, 13 0 (1): 0 132--164, 2019

    work page 2019

  58. [58]

    T. Li, E. Levina, and J. Zhu. Network cross-validation by edge sampling. Biometrika, 107 0 (2): 0 257--276, 2020

    2020

  59. [59]

    Z. Li, B. K. Johnson, D. L. Sussman, C. E. Priebe, and V. Lyzinski. Gotta match 'em all: Solution diversification in graph matching matched filters. IEEE Transactions on Signal and Information Processing over Networks , 10: 0 752--764, 2024 b

    work page 2024

  60. [60]

    Lockhart, J

    R. Lockhart, J. Taylor, R. J. Tibshirani, and R. Tibshirani. A significance test for the lasso . The Annals of Statistics, 42 0 (2): 0 413 -- 468, 2014. doi:https://doi.org/10.1214/13-AOS1175

    work page doi:10.1214/13-aos1175 2014

  61. [61]

    Loomba, J

    S. Loomba, J. Straehle, V. Gangadharan, N. Heike, A. Khalifa, A. Motta, N. Ju, M. Sievers, J. Gempt, H. S. Meyer, and M. Helmstaedter. Connectomic comparison of mouse and human cortex. Science, 377 0 (6602), 2022

    work page 2022

  62. [62]

    Lov\' a sz

    L. Lov\' a sz. Large Networks and Graph Limits. American Mathematical Society, 2012

    work page 2012

  63. [63]

    Lyzinski, M

    V. Lyzinski, M. Tang, A. Athreya, Y. Park, and C. E. Priebe. Community detection and classification in hierarchical stochastic blockmodels. IEEE Transactions on Network Science and Engineering , 4 0 (1): 0 13--26, 2017

    work page 2017

  64. [64]

    Marchette, C

    D. Marchette, C. E. Priebe, and G. Coppersmith. Vertex nomination via attributed random dot product graphs. In Proceedings of the 58th World Statistical Congress. International Statistical Institute, 2011

    work page 2011

  65. [65]

    McCulloh, H

    I. McCulloh, H. Armstrong, and A. Johnson. Social Network Analysis with Applications. Wiley, 1st edition, 2013

    work page 2013

  66. [66]

    McFowland III and C

    E. McFowland III and C. R. Shalizi. Estimating causal peer influence in homophilous social networks by inferring latent locations. Journal of the American Statistical Association, 118 0 (541): 0 707--718, 2023

    work page 2023

  67. [67]

    C. Mu, A. Mele, L. Hao, J. Cape, A. Athreya, and C. E. Priebe. On spectral algorithms for community detection in stochastic blockmodel graphs with vertex covariates. IEEE Transactions on Network Science and Engineering, 9 0 (5): 0 3373--3384, 2022. doi:10.1109/TNSE.2022.3177708

    work page arXiv 2022

  68. [68]

    R. J. Muirhead. Aspects of Multivariate Statistical Theory. Wiley Series in Probability and Statistics. John Wiley & Sons, 1982

    work page 1982

  69. [69]

    S. Nath, K. Warren, and S. Paul. Identifying peer influence in therapeutic communities adjusting for latent homophily. The Annals of Applied Statistics, 19 0 (1): 0 529--565, 2025

    work page 2025

  70. [70]

    M. E. J. Newman. Networks: an introduction. Oxford University Press, 2nd edition, 2018

    work page 2018

  71. [71]

    Orbanz and D

    P. Orbanz and D. M. Roy. Bayesian models of graphs, arrays and other exchangeable random structures. IEEE Transactions on Pattern Analysis and Machine Intelligent , 37 0 (2): 0 437--461, 2015

    work page 2015

  72. [72]

    P. O. Perry and P. J. Wolfe. Point process modelling for directed interaction networks. Journal of the Royal Statistical Society Series B: Statistical Methodology, 75 0 (5): 0 821--849, 2013

    work page 2013

  73. [73]

    K. Rohe, S. Chatterjee, and B. Yu. Spectral clustering and the high-dimensional stochastic blockmodel. The Annals of Statistics, 39 0 (4): 0 1878--1915, 2011

    work page 1915

  74. [74]

    Rozemberczki, C

    B. Rozemberczki, C. Allen, and R. Sarkar. Multi-scale attributed node embedding. Journal of Complex Networks, 9 0 (2): 0 cnab014, 2021. doi:https://doi.org/10.1093/comnet/cnab014

    doi:10.1093/comnet/cnab014 2021

  75. [75]

    Rubin-Delanchy, J

    P. Rubin-Delanchy, J. Cape, M. Tang, and C. E. Priebe. A statistical interpretation of spectral embedding: the generalised random dot product graph. Journal of the Royal Statistical Society: Series B (Statistical Methodology) , 84 0 (4): 0 1446--1473, 2022

    work page 2022

  76. [76]

    L. L. Scharf and C. T. Mullis. Canonical coordinates and the geometry of inference, rate, and capacity. IEEE Transactions on Signal Processing , 48 0 (3): 0 824--831, 2000

    work page 2000

  77. [77]

    Setayesh, Z

    A. Setayesh, Z. S. H. Zadeh, and B. Bahrak. Analysis of the global trade network using exponential random graph models. Applied Network Science, 7 0 (38), 2022

    work page 2022

  78. [78]

    C. R. Shalizi and A. C. Thomas. Homophily and contagion are generically confounded in observational social network studies. Sociological Methods Research, 40 0 (2): 0 211--239, 2011

    work page 2011

  79. [79]

    Y. Song, P. J. Schreier, D. Ram\' i rez, and T. Hasija. Canonical correlation analysis of high-dimensional data with very small sample support. Signal Processing, 128: 0 449--458, 2016

    work page 2016

  80. [80]

    Sosa and B

    J. Sosa and B. Betancourt. Bayesian inference of financial networks. Model Assisted Statistics and Applications, 18: 0 295--310, 2023

    work page 2023

Showing first 80 references.

This paper was first reviewed by deepseek-v4-flash on August 5, 2026.