pith. sign in

arxiv: 2308.02480 · v3 · submitted 2023-08-04 · 🧮 math.ST · stat.TH

Statistical Inference for Linear Functions of Eigenvectors with Small Eigengaps

Joshua Agterberg This is my paper

Pith reviewed 2026-05-24 06:50 UTC · model grok-4.3

classification 🧮 math.ST stat.TH
keywords eigenvector inferencematrix denoisingspiked PCAsmall eigengapsconfidence intervalsGaussian approximationminimax optimalitystatistical inference
0
0 comments X

The pith

Debiased linear forms of eigenvectors admit approximate Gaussian limits under Gaussian noise even with small eigengaps, yielding valid confidence intervals.

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

The paper shows that in matrix denoising and spiked principal component analysis, linear functions of eigenvectors become approximately Gaussian after a debiasing step when the additive noise is Gaussian. This limiting distribution holds in the challenging regime where eigenvalue gaps are much smaller than the spiked eigenvalues themselves. From the Gaussian limit the authors derive plug-in estimators for bias and variance that produce confidence intervals directly from the observed data matrix. They prove these intervals achieve widths that are minimax optimal up to constant factors. A reader would care because the result supplies a practical route to inference on spectral quantities in high-dimensional settings where classical eigenvector perturbation bounds are too coarse.

Core claim

We prove the approximate Gaussianity for debiased linear forms in the matrix denoising model and the spiked principal component analysis model, both under Gaussian noise. Based on this limiting behavior, we propose estimators for the appropriate bias and variance quantities resulting in approximately valid confidence intervals. We then investigate the optimality of these confidence intervals and show that their widths are minimax optimal up to constant factors. Of note, our proposed confidence intervals can be computed directly from data without the need for any sample-splitting.

What carries the argument

Debiased linear forms whose approximate Gaussianity is established under Gaussian noise for matrix denoising and spiked PCA models with arbitrarily small eigengaps.

If this is right

  • Approximately valid confidence intervals for linear functionals of eigenvectors can be constructed without sample splitting.
  • The same procedure applies uniformly to both the matrix denoising model and the spiked PCA model.
  • The widths of the resulting intervals are minimax optimal up to constant factors.
  • The intervals remain valid when eigenvalue gaps are arbitrarily small relative to the spiked eigenvalues.

Where Pith is reading between the lines

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

  • The Gaussian approximation may extend to inference on other smooth functionals of the same eigenvectors beyond linear forms.
  • The method could be applied in network community detection or factor analysis where small eigengaps arise naturally from the data-generating process.
  • Relaxing the Gaussian noise assumption while preserving the limiting distribution would be a natural next theoretical step.

Load-bearing premise

The noise in the matrix denoising and spiked PCA models is Gaussian.

What would settle it

Repeated Monte Carlo trials under Gaussian noise with small eigengaps in which the empirical coverage of the constructed intervals falls substantially below the nominal 1-alpha level would falsify the approximate validity result.

Figures

Figures reproduced from arXiv: 2308.02480 by Joshua Agterberg.

Figure 1
Figure 1. Figure 1: Empirical (dotted) and theoretical (solid) ellipses for the quantity [PITH_FULL_IMAGE:figures/full_fig_p024_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Empirical (dotted) and theoretical (solid) ellipses for the quantity [PITH_FULL_IMAGE:figures/full_fig_p026_2.png] view at source ↗
read the original abstract

Spectral methods have myriad applications in high-dimensional statistics and data science, and while previous works have primarily focused on $\ell_2$ or $\ell_{2,\infty}$ eigenvector and singular vector perturbation theory, in many settings these analyses fall short of providing the fine-grained guarantees required for various inferential tasks. In this paper we study statistical inference for linear functions of eigenvectors and principal components with a particular emphasis on the setting where gaps between eigenvalues may be extremely small relative to the corresponding spiked eigenvalue, a regime which has been oft-neglected in the literature. First, we prove the approximate Gaussianity for debiased linear forms in the matrix denoising model and the spiked principal component analysis model, both under Gaussian noise. Based on this limiting behavior, we propose estimators for the appropriate bias and variance quantities resulting in approximately valid confidence intervals. We then investigate the optimality of these confidence intervals and show that their widths are minimax optimal up to constant factors. Of note, our proposed confidence intervals can be computed directly from data without the need for any sample-splitting.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

0 major / 2 minor

Summary. The paper proves approximate Gaussianity for debiased linear forms of eigenvectors in the matrix denoising model and the spiked principal component analysis model under Gaussian noise, even when eigengaps are small relative to the spike. It develops data-driven estimators for the associated bias and variance terms to produce approximately valid confidence intervals for linear functionals, shows that the widths of these intervals are minimax optimal up to constant factors, and emphasizes that the intervals are computable directly from the data without sample splitting.

Significance. If the derivations hold, the work fills an important gap in high-dimensional spectral inference by delivering valid, optimal-width confidence intervals in the small-eigengap regime that prior perturbation analyses often exclude. The combination of limiting distributional results, explicit bias/variance estimators, and minimax optimality (without sample splitting) constitutes a substantive theoretical and practical advance for applications of eigenvector-based methods.

minor comments (2)
  1. [Main theorems (e.g., Theorems 3.1 and 4.2)] In the statements of the main theorems, explicitly restate the precise conditions on the eigengap relative to the spike and the noise level so that the small-gap regime is immediately visible without cross-reference to the model definitions.
  2. [Sections 2 and 3] The notation distinguishing the matrix denoising model from the spiked PCA model could be made more uniform across Sections 2 and 3 to reduce the chance of reader confusion when comparing the two settings.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for their positive assessment of our work and the recommendation for minor revision. The referee's summary accurately reflects the paper's contributions on approximate Gaussianity, data-driven bias/variance estimation, and minimax-optimal confidence intervals without sample splitting in the small-eigengap regime.

Circularity Check

0 steps flagged

No significant circularity; derivation is self-contained proof

full rationale

The paper's central results consist of a mathematical proof establishing approximate Gaussianity for debiased linear forms of eigenvectors in the matrix denoising and spiked PCA models under explicit Gaussian noise assumptions, followed by explicit construction of bias/variance estimators from the data and a minimax optimality argument for the resulting CI widths. No load-bearing step reduces by construction to a fitted parameter, self-defined quantity, or self-citation chain; the limiting distributions and optimality bounds are derived directly from the model and stated assumptions without renaming empirical patterns or smuggling ansatzes. The absence of sample-splitting is presented as a feature of the direct estimators, not a circularity. This is the normal case for a self-contained theoretical statistics paper.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

Abstract-only review yields no explicit free parameters, invented entities, or additional axioms beyond the stated Gaussian-noise modeling assumption.

axioms (1)
  • domain assumption Gaussian noise in matrix denoising and spiked PCA models
    Abstract states the limiting Gaussianity holds 'both under Gaussian noise'.

pith-pipeline@v0.9.0 · 5707 in / 1245 out tokens · 22817 ms · 2026-05-24T06:50:57.657349+00:00 · methodology

discussion (0)

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

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

84 extracted references · 84 canonical work pages

  1. [1]

    Entrywise eigenvector analysis of random matrices with low expected rank

    Emmanuel Abbe, Jianqing Fan, Kaizheng Wang, and Yiqiao Zhong. Entrywise eigenvector analysis of random matrices with low expected rank. The Annals of Statistics, 48 0 (3): 0 1452--1474, June 2020. ISSN 0090-5364, 2168-8966. doi:10.1214/19-AOS1854

    work page doi:10.1214/19-aos1854 2020

  2. [2]

    An \ ell\_p\ theory of PCA and spectral clustering

    Emmanuel Abbe, Jianqing Fan, and Kaizheng Wang. An \ ell\_p\ theory of PCA and spectral clustering. The Annals of Statistics, 50 0 (4): 0 2359--2385, August 2022. ISSN 0090-5364, 2168-8966. doi:10.1214/22-AOS2196

    work page doi:10.1214/22-aos2196 2022

  3. [3]

    An Overview of Asymptotic Normality in Stochastic Blockmodels : Cluster Analysis and Inference , May 2023

    Joshua Agterberg and Joshua Cape. An Overview of Asymptotic Normality in Stochastic Blockmodels : Cluster Analysis and Inference , May 2023. arXiv:2305.06353 [math, stat]

    work page arXiv 2023

  4. [4]

    Entrywise Recovery Guarantees for Sparse PCA via Sparsistent Algorithms

    Joshua Agterberg and Jeremias Sulam. Entrywise Recovery Guarantees for Sparse PCA via Sparsistent Algorithms . In Proceedings of The 25th International Conference on Artificial Intelligence and Statistics , pages 6591--6629. PMLR, May 2022. ISSN: 2640-3498

    work page 2022

  5. [5]

    Estimating Higher - Order Mixed Memberships via the \ ell\_\ 2, infty\ \ Tensor Perturbation Bound , December 2022

    Joshua Agterberg and Anru Zhang. Estimating Higher - Order Mixed Memberships via the \ ell\_\ 2, infty\ \ Tensor Perturbation Bound , December 2022. arXiv:2212.08642 [math, stat]

    work page arXiv 2022

  6. [6]

    Joshua Agterberg, Zachary Lubberts, and Carey E. Priebe. Entrywise Estimation of Singular Vectors of Low - Rank Matrices With Heteroskedasticity and Dependence . IEEE Transactions on Information Theory, 68 0 (7): 0 4618--4650, July 2022. ISSN 1557-9654. doi:10.1109/TIT.2022.3159085

    work page doi:10.1109/tit.2022.3159085 2022

  7. [7]

    Phase transition of the largest eigenvalue for nonnull complex sample covariance matrices

    Jinho Baik, Gérard Ben Arous, and Sandrine Péché. Phase transition of the largest eigenvalue for nonnull complex sample covariance matrices. The Annals of Probability, 33 0 (5): 0 1643--1697, September 2005. ISSN 0091-1798, 2168-894X. doi:10.1214/009117905000000233

    work page doi:10.1214/009117905000000233 2005

  8. [8]

    Singular vector and singular subspace distribution for the matrix denoising model

    Zhigang Bao, Xiucai Ding, and and Ke Wang. Singular vector and singular subspace distribution for the matrix denoising model. The Annals of Statistics, 49 0 (1), February 2021. ISSN 0090-5364. doi:10.1214/20-AOS1960

    work page doi:10.1214/20-aos1960 2021

  9. [9]

    Statistical inference for principal components of spiked covariance matrices

    Zhigang Bao, Xiucai Ding, Jingming Wang, and Ke Wang. Statistical inference for principal components of spiked covariance matrices. The Annals of Statistics, 50 0 (2): 0 1144--1169, April 2022. ISSN 0090-5364, 2168-8966. doi:10.1214/21-AOS2143

    work page doi:10.1214/21-aos2143 2022

  10. [10]

    The eigenvalues and eigenvectors of finite, low rank perturbations of large random matrices

    Florent Benaych-Georges and Raj Rao Nadakuditi. The eigenvalues and eigenvectors of finite, low rank perturbations of large random matrices. Advances in Mathematics, 227 0 (1): 0 494--521, May 2011. ISSN 0001-8708. doi:10.1016/j.aim.2011.02.007

    work page doi:10.1016/j.aim.2011.02.007 2011

  11. [11]

    The singular values and vectors of low rank perturbations of large rectangular random matrices

    Florent Benaych-Georges and Raj Rao Nadakuditi. The singular values and vectors of low rank perturbations of large rectangular random matrices. Journal of Multivariate Analysis, 111: 0 120--135, October 2012. ISSN 0047-259X. doi:10.1016/j.jmva.2012.04.019

    work page doi:10.1016/j.jmva.2012.04.019 2012

  12. [12]

    Entry-wise dissipation for singular vector perturbation bounds, April 2023

    Abhinav Bhardwaj and Van Vu. Entry-wise dissipation for singular vector perturbation bounds, April 2023. arXiv:2304.00328 [cs, math, stat]

    work page arXiv 2023

  13. [13]

    On the principal components of sample covariance matrices

    Alex Bloemendal, Antti Knowles, Horng-Tzer Yau, and Jun Yin. On the principal components of sample covariance matrices. Probability Theory and Related Fields, 164 0 (1): 0 459--552, February 2016. ISSN 1432-2064. doi:10.1007/s00440-015-0616-x

    work page doi:10.1007/s00440-015-0616-x 2016

  14. [14]

    Vincent Poor, and Yuxin Chen

    Changxiao Cai, Gen Li, Yuejie Chi, H. Vincent Poor, and Yuxin Chen. Subspace estimation from unbalanced and incomplete data matrices: \ ell\_\ 2, infty\ \ statistical guarantees. The Annals of Statistics, 49 0 (2): 0 944--967, April 2021 a . ISSN 0090-5364, 2168-8966. doi:10.1214/20-AOS1986

    work page doi:10.1214/20-aos1986 2021

  15. [15]

    Vincent Poor, and Yuxin Chen

    Changxiao Cai, H. Vincent Poor, and Yuxin Chen. Uncertainty Quantification for Nonconvex Tensor Completion : Confidence Intervals , Heteroscedasticity and Optimality . IEEE Transactions on Information Theory, 69 0 (1): 0 407--452, January 2023. ISSN 1557-9654. doi:10.1109/TIT.2022.3205781

    work page doi:10.1109/tit.2022.3205781 2023

  16. [16]

    Tony Cai and Zijian Guo

    T. Tony Cai and Zijian Guo. Confidence intervals for high-dimensional linear regression: Minimax rates and adaptivity. The Annals of Statistics, 45 0 (2): 0 615--646, April 2017. ISSN 0090-5364, 2168-8966. doi:10.1214/16-AOS1461

    work page doi:10.1214/16-aos1461 2017

  17. [17]

    Tony Cai and Anru Zhang

    T. Tony Cai and Anru Zhang. Rate-optimal perturbation bounds for singular subspaces with applications to high-dimensional statistics. The Annals of Statistics, 46 0 (1): 0 60--89, February 2018. ISSN 0090-5364, 2168-8966. doi:10.1214/17-AOS1541

    work page doi:10.1214/17-aos1541 2018

  18. [18]

    Tony Cai, Zongming Ma, and Yihong Wu

    T. Tony Cai, Zongming Ma, and Yihong Wu. Sparse PCA : Optimal rates and adaptive estimation. The Annals of Statistics, 41 0 (6): 0 3074--3110, December 2013. ISSN 0090-5364, 2168-8966. doi:10.1214/13-AOS1178

    work page doi:10.1214/13-aos1178 2013

  19. [19]

    Optimal Structured Principal Subspace Estimation : Metric Entropy and Minimax Rates

    Tony Cai, Hongzhe Li, and Rong Ma. Optimal Structured Principal Subspace Estimation : Metric Entropy and Minimax Rates . Journal of Machine Learning Research, 22 0 (46): 0 1--45, 2021 b . ISSN 1533-7928

    work page 2021

  20. [20]

    Signal-plus-noise matrix models: eigenvector deviations and fluctuations

    J Cape, M Tang, and C E Priebe. Signal-plus-noise matrix models: eigenvector deviations and fluctuations. Biometrika, 106 0 (1): 0 243--250, March 2019 a . ISSN 0006-3444. doi:10.1093/biomet/asy070

    work page doi:10.1093/biomet/asy070 2019

  21. [21]

    Joshua Cape, Minh Tang, and Carey E. Priebe. The two-to-infinity norm and singular subspace geometry with applications to high-dimensional statistics. Annals of Statistics, 47 0 (5): 0 2405--2439, October 2019 b . ISSN 0090-5364, 2168-8966. doi:10.1214/18-AOS1752

    work page doi:10.1214/18-aos1752 2019

  22. [22]

    Capitaine, C

    M. Capitaine, C. Donati-Martin, and D. Féral. Central limit theorems for eigenvalues of deformations of Wigner matrices. Annales de l'I.H.P. Probabilités et statistiques, 48 0 (1): 0 107--133, 2012. ISSN 1778-7017. doi:10.1214/10-AIHP410

    work page doi:10.1214/10-aihp410 2012

  23. [23]

    Limiting Eigenvectors of Outliers for Spiked Information - Plus - Noise Type Matrices

    Mireille Capitaine. Limiting Eigenvectors of Outliers for Spiked Information - Plus - Noise Type Matrices . In Catherine Donati-Martin, Antoine Lejay, and Alain Rouault, editors, Séminaire de Probabilités XLIX , Lecture Notes in Mathematics , pages 119--164. Springer International Publishing, Cham, 2018. ISBN 978-3-319-92420-5. doi:10.1007/978-3-319-92420-5_4

    work page doi:10.1007/978-3-319-92420-5_4 2018

  24. [24]

    The largest eigenvalues of finite rank deformation of large Wigner matrices: Convergence and nonuniversality of the fluctuations

    Mireille Capitaine, Catherine Donati-Martin, and Delphine Féral. The largest eigenvalues of finite rank deformation of large Wigner matrices: Convergence and nonuniversality of the fluctuations. The Annals of Probability, 37 0 (1): 0 1--47, January 2009. ISSN 0091-1798, 2168-894X. doi:10.1214/08-AOP394

    work page doi:10.1214/08-aop394 2009

  25. [25]

    Carpentier, O

    A. Carpentier, O. Klopp, and M. Löffler. Constructing Confidence Sets for the Matrix Completion Problem . In Patrice Bertail, Delphine Blanke, Pierre-André Cornillon, and Eric Matzner-Løber, editors, Nonparametric Statistics , Springer Proceedings in Mathematics & Statistics , pages 103--118, Cham, 2018. Springer International Publishing. ISBN 978-3-319-9...

    work page doi:10.1007/978-3-319-96941-1_7 2018

  26. [26]

    On signal detection and confidence sets for low rank inference problems

    Alexandra Carpentier and Richard Nickl. On signal detection and confidence sets for low rank inference problems. Electronic Journal of Statistics, 9 0 (2): 0 2675--2688, January 2015. ISSN 1935-7524, 1935-7524. doi:10.1214/15-EJS1087

    work page doi:10.1214/15-ejs1087 2015

  27. [27]

    Uncertainty Quantification for Matrix Compressed Sensing and Quantum Tomography Problems

    Alexandra Carpentier, Jens Eisert, David Gross, and Richard Nickl. Uncertainty Quantification for Matrix Compressed Sensing and Quantum Tomography Problems . In Nathael Gozlan, Rafał Latała, Karim Lounici, and Mokshay Madiman, editors, High Dimensional Probability VIII , Progress in Probability , pages 385--430, Cham, 2019. Springer International Publishi...

    work page doi:10.1007/978-3-030-26391-1_18 2019

  28. [28]

    Pinhan Chen, Chao Gao, and Anderson Y. Zhang. Partial recovery for top-k ranking: Optimality of MLE and SubOptimality of the spectral method. The Annals of Statistics, 50 0 (3): 0 1618--1652, June 2022. ISSN 0090-5364, 2168-8966. doi:10.1214/21-AOS2166

    work page doi:10.1214/21-aos2166 2022

  29. [29]

    Spectral method and regularized MLE are both optimal for top-\ K \ ranking

    Yuxin Chen, Jianqing Fan, Cong Ma, and Kaizheng Wang. Spectral method and regularized MLE are both optimal for top-\ K \ ranking. The Annals of Statistics, 47 0 (4): 0 2204--2235, August 2019. ISSN 0090-5364, 2168-8966. doi:10.1214/18-AOS1745

    work page doi:10.1214/18-aos1745 2019

  30. [30]

    Asymmetry helps: Eigenvalue and eigenvector analyses of asymmetrically perturbed low-rank matrices

    Yuxin Chen, Chen Cheng, and Jianqing Fan. Asymmetry helps: Eigenvalue and eigenvector analyses of asymmetrically perturbed low-rank matrices. The Annals of Statistics, 49 0 (1): 0 435--458, February 2021 a . ISSN 0090-5364, 2168-8966. doi:10.1214/20-AOS1963

    work page doi:10.1214/20-aos1963 2021

  31. [31]

    Spectral Methods for Data Science : A Statistical Perspective

    Yuxin Chen, Yuejie Chi, Jianqing Fan, and Cong Ma. Spectral Methods for Data Science : A Statistical Perspective . Foundations and Trends® in Machine Learning, 14 0 (5): 0 566--806, 2021 b . ISSN 1935-8237, 1935-8245. doi:10.1561/2200000079

    work page doi:10.1561/2200000079 2021

  32. [32]

    Tackling Small Eigen - Gaps : Fine - Grained Eigenvector Estimation and Inference Under Heteroscedastic Noise

    Chen Cheng, Yuting Wei, and Yuxin Chen. Tackling Small Eigen - Gaps : Fine - Grained Eigenvector Estimation and Inference Under Heteroscedastic Noise . IEEE Transactions on Information Theory, 67 0 (11): 0 7380--7419, November 2021. ISSN 1557-9654. doi:10.1109/TIT.2021.3111828

    work page doi:10.1109/tit.2021.3111828 2021

  33. [33]

    Inference for Low - Rank Models , January 2023

    Victor Chernozhukov, Christian Hansen, Yuan Liao, and Yinchu Zhu. Inference for Low - Rank Models , January 2023. arXiv:2107.02602 [econ, math, stat]

    work page arXiv 2023

  34. [34]

    Lu, and Yuxin Chen

    Yuejie Chi, Yue M. Lu, and Yuxin Chen. Nonconvex Optimization Meets Low - Rank Matrix Factorization : An Overview . IEEE Transactions on Signal Processing, 67 0 (20): 0 5239--5269, October 2019. ISSN 1941-0476. doi:10.1109/TSP.2019.2937282

    work page doi:10.1109/tsp.2019.2937282 2019

  35. [35]

    Uniform Bounds for Invariant Subspace Perturbations

    Anil Damle and Yuekai Sun. Uniform Bounds for Invariant Subspace Perturbations . SIAM Journal on Matrix Analysis and Applications, 41 0 (3): 0 1208--1236, January 2020. ISSN 0895-4798. doi:10.1137/19M1262760

    work page doi:10.1137/19m1262760 2020

  36. [36]

    Chandler Davis and W. M. Kahan. The Rotation of Eigenvectors by a Perturbation . III . SIAM Journal on Numerical Analysis, 7 0 (1): 0 1--46, 1970. ISSN 0036-1429

    work page 1970

  37. [37]

    High dimensional deformed rectangular matrices with applications in matrix denoising

    Xiucai Ding. High dimensional deformed rectangular matrices with applications in matrix denoising. Bernoulli, 26 0 (1): 0 387--417, February 2020. ISSN 1350-7265. doi:10.3150/19-BEJ1129

    work page doi:10.3150/19-bej1129 2020

  38. [38]

    Spiked sample covariance matrices with possibly multiple bulk components

    Xiucai Ding. Spiked sample covariance matrices with possibly multiple bulk components. Random Matrices: Theory and Applications, 10 0 (01): 0 2150014, January 2021. ISSN 2010-3263. doi:10.1142/S2010326321500143

    work page doi:10.1142/s2010326321500143 2021

  39. [39]

    Spiked separable covariance matrices and principal components

    Xiucai Ding and Fan Yang. Spiked separable covariance matrices and principal components. The Annals of Statistics, 49 0 (2): 0 1113--1138, April 2021. ISSN 0090-5364, 2168-8966. doi:10.1214/20-AOS1995

    work page doi:10.1214/20-aos1995 2021

  40. [40]

    Unperturbed: spectral analysis beyond Davis - Kahan

    Justin Eldridge, Mikhail Belkin, and Yusu Wang. Unperturbed: spectral analysis beyond Davis - Kahan . In Algorithmic Learning Theory , pages 321--358, April 2018

    work page 2018

  41. [41]

    An \ ell\_\ infty\ \ Eigenvector Perturbation Bound and Its Application

    Jianqing Fan, Weichen Wang, and Yiqiao Zhong. An \ ell\_\ infty\ \ Eigenvector Perturbation Bound and Its Application . Journal of Machine Learning Research, 18 0 (207): 0 1--42, 2018. ISSN 1533-7928

    work page 2018

  42. [42]

    Estimating Mixed Memberships With Sharp Eigenvector Deviations

    Jianqing Fan, Yingying Fan, Xiao Han, and Jinchi Lv. Asymptotic Theory of Eigenvectors for Random Matrices With Diverging Spikes . Journal of the American Statistical Association, 0 0 (0): 0 1--14, October 2020. ISSN 0162-1459. doi:10.1080/01621459.2020.1840990

    work page doi:10.1080/01621459.2020.1840990 2020

  43. [43]

    Uncertainty quantification in the Bradley – Terry – Luce model

    Chao Gao, Yandi Shen, and Anderson Y Zhang. Uncertainty quantification in the Bradley – Terry – Luce model. Information and Inference: A Journal of the IMA, 12 0 (2): 0 1073--1140, June 2023. ISSN 2049-8772. doi:10.1093/imaiai/iaac032

    work page doi:10.1093/imaiai/iaac032 2023

  44. [44]

    Horn and C.R

    R.A. Horn and C.R. Johnson. Matrix Analysis . Cambridge University Press, 2012. ISBN 978-1-139-78888-5

    work page 2012

  45. [45]

    Mixed membership estimation for social networks

    Jiashun Jin, Zheng Tracy Ke, and Shengming Luo. Mixed membership estimation for social networks. Journal of Econometrics, February 2023. ISSN 0304-4076. doi:10.1016/j.jeconom.2022.12.003

    work page doi:10.1016/j.jeconom.2022.12.003 2023

  46. [46]

    Relative perturbation bounds with applications to empirical covariance operators

    Moritz Jirak and Martin Wahl. Relative perturbation bounds with applications to empirical covariance operators. Advances in Mathematics, 412: 0 108808, January 2023. ISSN 00018708. doi:10.1016/j.aim.2022.108808

    work page doi:10.1016/j.aim.2022.108808 2023

  47. [47]

    Johnstone and Arthur Yu Lu

    Iain M. Johnstone and Arthur Yu Lu. On Consistency and Sparsity for Principal Components Analysis in High Dimensions . Journal of the American Statistical Association, 104 0 (486): 0 682--693, June 2009. ISSN 0162-1459. doi:10.1198/jasa.2009.0121

    work page doi:10.1198/jasa.2009.0121 2009

  48. [48]

    Perturbation Theory for Linear Operators

    Tosio Kato. Perturbation Theory for Linear Operators . Classics in Mathematics . Springer-Verlag, Berlin Heidelberg, 2 edition, 1995. ISBN 978-3-540-58661-6. doi:10.1007/978-3-642-66282-9

    work page doi:10.1007/978-3-642-66282-9 1995

  49. [49]

    The Outliers of a Deformed Wigner Matrix

    Antti Knowles and Jun Yin. The Outliers of a Deformed Wigner Matrix . The Annals of Probability, 42 0 (5): 0 1980--2031, 2014. ISSN 0091-1798

    work page 1980

  50. [50]

    Asymptotics and concentration bounds for bilinear forms of spectral projectors of sample covariance

    Vladimir Koltchinskii and Karim Lounici. Asymptotics and concentration bounds for bilinear forms of spectral projectors of sample covariance. Annales de l'Institut Henri Poincaré, Probabilités et Statistiques, 52 0 (4): 0 1976--2013, November 2016. ISSN 0246-0203. doi:10.1214/15-AIHP705

    work page doi:10.1214/15-aihp705 1976

  51. [51]

    Concentration inequalities and moment bounds for sample covariance operators

    Vladimir Koltchinskii and Karim Lounici. Concentration inequalities and moment bounds for sample covariance operators. Bernoulli, 23 0 (1): 0 110--133, February 2017 a . ISSN 1350-7265. doi:10.3150/15-BEJ730

    work page doi:10.3150/15-bej730 2017

  52. [52]

    Normal approximation and concentration of spectral projectors of sample covariance

    Vladimir Koltchinskii and Karim Lounici. Normal approximation and concentration of spectral projectors of sample covariance. The Annals of Statistics, 45 0 (1): 0 121--157, February 2017 b . ISSN 0090-5364, 2168-8966. doi:10.1214/16-AOS1437

    work page doi:10.1214/16-aos1437 2017

  53. [53]

    Perturbation of Linear Forms of Singular Vectors Under Gaussian Noise

    Vladimir Koltchinskii and Dong Xia. Perturbation of Linear Forms of Singular Vectors Under Gaussian Noise . In Christian Houdré, David M. Mason, Patricia Reynaud-Bouret, and Jan Rosiński, editors, High Dimensional Probability VII , Progress in Probability , pages 397--423, Cham, 2016. Springer International Publishing. ISBN 978-3-319-40519-3. doi:10.1007/...

    work page doi:10.1007/978-3-319-40519-3_18 2016

  54. [54]

    Efficient estimation of linear functionals of principal components

    Vladimir Koltchinskii, Matthias Löffler, and Richard Nickl. Efficient estimation of linear functionals of principal components. Annals of Statistics, 48 0 (1): 0 464--490, February 2020. ISSN 0090-5364, 2168-8966. doi:10.1214/19-AOS1816

    work page doi:10.1214/19-aos1816 2020

  55. [55]

    Laurent and P

    B. Laurent and P. Massart. Adaptive estimation of a quadratic functional by model selection. The Annals of Statistics, 28 0 (5): 0 1302--1338, October 2000. ISSN 0090-5364, 2168-8966. doi:10.1214/aos/1015957395

    work page doi:10.1214/aos/1015957395 2000

  56. [56]

    Unified \ ell\_\ 2 rightarrow infty\ \ Eigenspace Perturbation Theory for Symmetric Random Matrices

    Lihua Lei. Unified \ ell\_\ 2 rightarrow infty\ \ Eigenspace Perturbation Theory for Symmetric Random Matrices . arXiv:1909.04798 [math, stat], September 2019

    work page arXiv 1909

  57. [57]

    Vincent Poor, and Yuxin Chen

    Gen Li, Changxiao Cai, H. Vincent Poor, and Yuxin Chen. Minimax Estimation of Linear Functions of Eigenvectors in the Face of Small Eigen - Gaps , July 2022. arXiv:2104.03298 [cs, math, stat]

    work page arXiv 2022

  58. [58]

    Near-optimal performance bounds for orthogonal and permutation group synchronization via spectral methods

    Shuyang Ling. Near-optimal performance bounds for orthogonal and permutation group synchronization via spectral methods. Applied and Computational Harmonic Analysis, 60: 0 20--52, September 2022. ISSN 1063-5203. doi:10.1016/j.acha.2022.02.003

    work page doi:10.1016/j.acha.2022.02.003 2022

  59. [59]

    Yuetian Luo, Rungang Han, and Anru R. Zhang. A Schatten -q low-rank matrix perturbation analysis via perturbation projection error bound. Linear Algebra and its Applications, 630: 0 225--240, December 2021. ISSN 0024-3795. doi:10.1016/j.laa.2021.08.005

    work page doi:10.1016/j.laa.2021.08.005 2021

  60. [60]

    Zhang, and Harrison H

    Matthias Löffler, Anderson Y. Zhang, and Harrison H. Zhou. Optimality of spectral clustering in the Gaussian mixture model. The Annals of Statistics, 49 0 (5): 0 2506--2530, October 2021. ISSN 0090-5364, 2168-8966. doi:10.1214/20-AOS2044

    work page doi:10.1214/20-aos2044 2021

  61. [61]

    Estimating Mixed Memberships With Sharp Eigenvector Deviations

    Xueyu Mao, Purnamrita Sarkar, and Deepayan Chakrabarti. Estimating Mixed Memberships With Sharp Eigenvector Deviations . Journal of the American Statistical Association, 0 0 (0): 0 1--13, April 2020. ISSN 0162-1459. doi:10.1080/01621459.2020.1751645

    work page doi:10.1080/01621459.2020.1751645 2020

  62. [62]

    Finite sample approximation results for principal component analysis: A matrix perturbation approach

    Boaz Nadler. Finite sample approximation results for principal component analysis: A matrix perturbation approach. The Annals of Statistics, 36 0 (6), December 2008. ISSN 0090-5364. doi:10.1214/08-AOS618

    work page doi:10.1214/08-aos618 2008

  63. [63]

    A Novel and Optimal Spectral Method for Permutation Synchronization , March 2023

    Duc Nguyen and Anderson Ye Zhang. A Novel and Optimal Spectral Method for Permutation Synchronization , March 2023. arXiv:2303.12051 [cs, math, stat]

    work page arXiv 2023

  64. [64]

    Asymptotics of the principal components estimator of large factor models with weakly influential factors

    Alexei Onatski. Asymptotics of the principal components estimator of large factor models with weakly influential factors. Journal of Econometrics, 168 0 (2): 0 244--258, June 2012. ISSN 0304-4076. doi:10.1016/j.jeconom.2012.01.034

    work page doi:10.1016/j.jeconom.2012.01.034 2012

  65. [65]

    Random perturbation of low rank matrices: Improving classical bounds

    Sean O'Rourke, Van Vu, and Ke Wang. Random perturbation of low rank matrices: Improving classical bounds. Linear Algebra and its Applications, 540: 0 26--59, March 2018. ISSN 0024-3795. doi:10.1016/j.laa.2017.11.014

    work page doi:10.1016/j.laa.2017.11.014 2018

  66. [66]

    Matrices with Gaussian noise: optimal estimates for singular subspace perturbation, January 2023

    Sean O'Rourke, Van Vu, and Ke Wang. Matrices with Gaussian noise: optimal estimates for singular subspace perturbation, January 2023. arXiv:1803.00679 [cs, math, stat]

    work page arXiv 2023

  67. [67]

    Asymptotics of Sample Eigenstructure for a Large Dimensional Spiked Covariance Model

    Debashis Paul. Asymptotics of Sample Eigenstructure for a Large Dimensional Spiked Covariance Model . Statistica Sinica, 17 0 (4): 0 1617--1642, 2007. ISSN 1017-0405

    work page 2007

  68. [68]

    Hypothesis testing for eigenspaces of covariance matrix, February 2020

    Igor Silin and Jianqing Fan. Hypothesis testing for eigenspaces of covariance matrix, February 2020. arXiv:2002.09810 [math, stat]

    work page arXiv 2020

  69. [69]

    G. W. Stewart and Ji-guang Sun. Matrix Perturbation Theory . Elsevier Science, June 1990. ISBN 978-0-12-670230-9. Google-Books-ID: l78PAQAAMAAJ

    work page 1990

  70. [70]

    High- Dimensional Probability : An Introduction with Applications in Data Science

    Roman Vershynin. High- Dimensional Probability : An Introduction with Applications in Data Science . Cambridge Series in Statistical and Probabilistic Mathematics . Cambridge University Press, 2018. doi:10.1017/9781108231596

    work page doi:10.1017/9781108231596 2018

  71. [71]

    Wainwright

    Martin J. Wainwright. High- Dimensional Statistics : A Non - Asymptotic Viewpoint . Cambridge Series in Statistical and Probabilistic Mathematics . Cambridge University Press, Cambridge, 2019. ISBN 978-1-108-49802-9. doi:10.1017/9781108627771

    work page doi:10.1017/9781108627771 2019

  72. [72]

    Perturbation bounds in connection with singular value decomposition

    Per-Åke Wedin. Perturbation bounds in connection with singular value decomposition. BIT Numerical Mathematics, 12 0 (1): 0 99--111, March 1972. ISSN 1572-9125. doi:10.1007/BF01932678

    work page doi:10.1007/bf01932678 1972

  73. [73]

    Confidence Region of Singular Subspaces for Low - Rank Matrix Regression

    Dong Xia. Confidence Region of Singular Subspaces for Low - Rank Matrix Regression . IEEE Transactions on Information Theory, 65 0 (11): 0 7437--7459, November 2019. ISSN 1557-9654. doi:10.1109/TIT.2019.2924900

    work page doi:10.1109/tit.2019.2924900 2019

  74. [74]

    Normal approximation and confidence region of singular subspaces

    Dong Xia. Normal approximation and confidence region of singular subspaces. Electronic Journal of Statistics, 15 0 (2): 0 3798--3851, January 2021. ISSN 1935-7524, 1935-7524. doi:10.1214/21-EJS1876

    work page doi:10.1214/21-ejs1876 2021

  75. [75]

    Statistical Inferences of Linear Forms for Noisy Matrix Completion

    Dong Xia and Ming Yuan. Statistical Inferences of Linear Forms for Noisy Matrix Completion . Journal of the Royal Statistical Society Series B: Statistical Methodology, 83 0 (1): 0 58--77, February 2021. ISSN 1369-7412. doi:10.1111/rssb.12400

    work page doi:10.1111/rssb.12400 2021

  76. [76]

    The Sup -norm Perturbation of HOSVD and Low Rank Tensor Denoising

    Dong Xia and Fan Zhou. The Sup -norm Perturbation of HOSVD and Low Rank Tensor Denoising . Journal of Machine Learning Research, 20 0 (61): 0 1--42, 2019. ISSN 1533-7928

    work page 2019

  77. [77]

    Zhang, and Yuchen Zhou

    Dong Xia, Anru R. Zhang, and Yuchen Zhou. Inference for low-rank tensors—no need to debias. The Annals of Statistics, 50 0 (2): 0 1220--1245, April 2022. ISSN 0090-5364, 2168-8966. doi:10.1214/21-AOS2146

    work page doi:10.1214/21-aos2146 2022

  78. [78]

    Inference for Heteroskedastic PCA with Missing Data

    Yuling Yan, Yuxin Chen, and Jianqing Fan. Inference for Heteroskedastic PCA with Missing Data . arXiv:2107.12365 [cs, math, stat], July 2021

    work page arXiv 2021

  79. [79]

    Y. Yu, T. Wang, and R. J. Samworth. A useful variant of the Davis — Kahan theorem for statisticians. Biometrika, 102 0 (2): 0 315--323, 2015. ISSN 0006-3444

    work page 2015

  80. [80]

    Exact Minimax Optimality of Spectral Methods in Phase Synchronization and Orthogonal Group Synchronization , September 2022

    Anderson Ye Zhang. Exact Minimax Optimality of Spectral Methods in Phase Synchronization and Orthogonal Group Synchronization , September 2022. arXiv:2209.04962 [cs, math, stat]

    work page arXiv 2022

Showing first 80 references.