Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
134 tokens/sec
GPT-4o
9 tokens/sec
Gemini 2.5 Pro Pro
47 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Topological complexity of spiked random polynomials and finite-rank spherical integrals (2312.12323v1)

Published 19 Dec 2023 in math.PR, math.ST, stat.ML, and stat.TH

Abstract: We study the annealed complexity of a random Gaussian homogeneous polynomial on the $N$-dimensional unit sphere in the presence of deterministic polynomials that depend on fixed unit vectors and external parameters. In particular, we establish variational formulas for the exponential asymptotics of the average number of total critical points and of local maxima. This is obtained through the Kac-Rice formula and the determinant asymptotics of a finite-rank perturbation of a Gaussian Wigner matrix. More precisely, the determinant analysis is based on recent advances on finite-rank spherical integrals by [Guionnet, Husson 2022] to study the large deviations of multi-rank spiked Gaussian Wigner matrices. The analysis of the variational problem identifies a topological phase transition. There is an exact threshold for the external parameters such that, once exceeded, the complexity function vanishes into new regions in which the critical points are close to the given vectors. Interestingly, these regions also include those where critical points are close to multiple vectors.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (40)
  1. Robert J. Adler and Jonathan E. Taylor “Random fields and geometry”, Springer Monographs in Mathematics Springer, New York, 2007, pp. xviii+448
  2. “Matrix regularizing effects of Gaussian perturbations” In Commun. Contemp. Math. 19.3, 2017, pp. 1750028\bibrangessep22 DOI: 10.1142/S0219199717500286
  3. Greg W. Anderson, Alice Guionnet and Ofer Zeitouni “An introduction to random matrices” 118, Cambridge Studies in Advanced Mathematics Cambridge University Press, Cambridge, 2010, pp. xiv+492 URL: https://www.cambridge.org/core/books/an-introduction-to-random-matrices/8992DA8EB0386651E8DA8214A1FC7241
  4. Antonio Auffinger and Gerard Ben Arous “Complexity of random smooth functions on the high-dimensional sphere” In Ann. Probab. 41.6, 2013, pp. 4214–4247 DOI: 10.1214/13-AOP862
  5. Antonio Auffinger, Gerard Ben Arous and Zhehua Li “Sharp complexity asymptotics and topological trivialization for the (p,k)𝑝𝑘(p,k)( italic_p , italic_k ) spiked tensor model” In J. Math. Phys. 63.4, 2022, pp. Paper No. 043303\bibrangessep21 DOI: 10.1063/5.0070300
  6. Antonio Auffinger, Gérard Ben Arous and Jiřı́ Černý “Random matrices and complexity of spin glasses” In Comm. Pure Appl. Math. 66.2, 2013, pp. 165–201 DOI: 10.1002/cpa.21422
  7. “Level sets and extrema of random processes and fields” John Wiley & Sons, Inc., Hoboken, NJ, 2009, pp. xii+393 DOI: 10.1002/9780470434642
  8. Jinho Baik, Gérard Ben Arous and Sandrine Péché “Phase transition of the largest eigenvalue for nonnull complex sample covariance matrices” In Ann. Probab. 33.5, 2005, pp. 1643–1697 DOI: 10.1214/009117905000000233
  9. “Triviality of the geometry of mixed p𝑝pitalic_p-spin spherical Hamiltonians with external field” In J. Stat. Phys. 186.1, 2022, pp. Paper No. 12\bibrangessep34 DOI: 10.1007/s10955-021-02855-6
  10. G. Ben Arous, A. Dembo and A. Guionnet “Aging of spherical spin glasses” In Probab. Theory Related Fields 120.1, 2001, pp. 1–67 DOI: 10.1007/PL00008774
  11. Gérard Ben Arous, Paul Bourgade and Benjamin McKenna “Exponential growth of random determinants beyond invariance” In Probab. Math. Phys. 3.4, 2022, pp. 731–789 DOI: 10.2140/pmp.2022.3.731
  12. Gérard Ben Arous, Paul Bourgade and Benjamin McKenna “Landscape complexity beyond invariance and the elastic manifold” In Comm. Pure Appl. Math. 77.2, 2024, pp. 1302–1352 DOI: 10.1002/cpa.22146
  13. Gérard Ben Arous, Reza Gheissari and Aukosh Jagannath “Algorithmic thresholds for tensor PCA” In Ann. Probab. 48.4, 2020, pp. 2052–2087 DOI: 10.1214/19-AOP1415
  14. Gérard Ben Arous, Daniel Zhengyu Huang and Jiaoyang Huang “Long random matrices and tensor unfolding” In Ann. Appl. Probab. 33.6B, 2023, pp. 5753–5780 DOI: 10.1214/23-aap1958
  15. “The landscape of the spiked tensor model” In Comm. Pure Appl. Math. 72.11, 2019, pp. 2282–2330 DOI: 10.1002/cpa.21861
  16. A J Bray and M A Moore “Metastable states in spin glasses” In Journal of Physics C: Solid State Physics 13.19, 1980, pp. L469 DOI: 10.1088/0022-3719/13/19/002
  17. Andrea Cavagna, Irene Giardina and Giorgio Parisi “Stationary points of the Thouless-Anderson-Palmer free energy” In Phys. Rev. B 57 American Physical Society, 1998, pp. 11251–11257 DOI: 10.1103/PhysRevB.57.11251
  18. “Thouless-Anderson-Palmer approach to the spherical p-spin glass model” In Journal de Physique I 5.7, 1995, pp. 805–813 DOI: 10.1051/jp1:1995164
  19. S. Dallaporta “Eigenvalue variance bounds for Wigner and covariance random matrices” In Random Matrices Theory Appl. 1.3, 2012, pp. 1250007\bibrangessep28 DOI: 10.1142/S2010326312500074
  20. “Large deviations techniques and applications” Corrected reprint of the second (1998) edition 38, Stochastic Modelling and Applied Probability Springer-Verlag, Berlin, 2010, pp. xvi+396 DOI: 10.1007/978-3-642-03311-7
  21. Yan V. Fyodorov “Complexity of random energy landscapes, glass transition, and absolute value of the spectral determinant of random matrices” In Phys. Rev. Lett. 92.24, 2004, pp. 240601\bibrangessep4 DOI: 10.1103/PhysRevLett.92.240601
  22. Yan V. Fyodorov “High-dimensional random fields and random matrix theory” In Markov Process. Related Fields 21.3, 2015, pp. 483–518 URL: http://math-mprf.org/journal/articles/id1385/
  23. “Concentration of the spectral measure for large matrices” In Electron. Comm. Probab. 5, 2000, pp. 119–136 DOI: 10.1214/ECP.v5-1026
  24. “Asymptotics of k𝑘kitalic_k dimensional spherical integrals and applications” In ALEA Lat. Am. J. Probab. Math. Stat. 19.1, 2022, pp. 769–797 DOI: 10.30757/alea.v19-30
  25. Samuel B. Hopkins, Jonathan Shi and David Steurer “Tensor principal component analysis via sum-of-square proofs” In Proceedings of The 28th Conference on Learning Theory 40, Proceedings of Machine Learning Research Paris, France: PMLR, 2015, pp. 956–1006 URL: https://proceedings.mlr.press/v40/Hopkins15.html
  26. “Power Iteration for Tensor PCA” In Journal of Machine Learning Research 23.128, 2022, pp. 1–47 URL: http://jmlr.org/papers/v23/21-1290.html
  27. “Spherical Integrals of Sublinear Rank”, 2023 arXiv:2208.03642 [math.PR]
  28. Iain M. Johnstone “On the distribution of the largest eigenvalue in principal components analysis” In Ann. Statist. 29.2, 2001, pp. 295–327 DOI: 10.1214/aos/1009210544
  29. Tosio Kato “Perturbation theory for linear operators” Reprint of the 1980 edition, Classics in Mathematics Springer-Verlag, Berlin, 1995, pp. xxii+619 DOI: 10.1007/978-3-642-66282-9
  30. J. Kurchan “Replica trick to calculate means of absolute values: applications to stochastic equations” In J. Phys. A 24.21, 1991, pp. 4969–4979 URL: http://stacks.iop.org/0305-4470/24/4969
  31. Mylène Maïda “Large deviations for the largest eigenvalue of rank one deformations of Gaussian ensembles” In Electron. J. Probab. 12.11, 2007, pp. 1131–1150 DOI: 10.1214/EJP.v12-438
  32. Antoine Maillard, Gérard Ben Arous and Giulio Biroli “Landscape Complexity for the Empirical Risk of Generalized Linear Models” In Proceedings of The First Mathematical and Scientific Machine Learning Conference 107, Proceedings of Machine Learning Research PMLR, 2020, pp. 287–327 URL: https://proceedings.mlr.press/v107/maillard20a.html
  33. Elizabeth S. Meckes and Mark W. Meckes “Concentration and convergence rates for spectral measures of random matrices” In Probab. Theory Related Fields 156.1-2, 2013, pp. 145–164 DOI: 10.1007/s00440-012-0423-6
  34. “A statistical model for tensor PCA” In Advances in Neural Information Processing Systems 27 Curran Associates, Inc., 2014, pp. 2897–2905 URL: https://proceedings.neurips.cc/paper/2014/file/b5488aeff42889188d03c9895255cecc-Paper.pdf
  35. “Complex energy landscapes in spiked-tensor and simple glassy models: Ruggedness, arrangements of local minima, and phase transitions” In Physical Review X 9.1 APS, 2019, pp. 011003 URL: https://journals.aps.org/prx/pdf/10.1103/PhysRevX.9.011003
  36. Valentina Ros and Yan V. Fyodorov “The High-dimensional Landscape Paradigm: Spin-Glasses, and Beyond” In Spin Glass Theory and Far Beyond, 2023, pp. 95–114 DOI: 10.1142/9789811273926_0006
  37. “Who is Afraid of Big Bad Minima? Analysis of gradient-flow in spiked matrix-tensor models” In Advances in Neural Information Processing Systems 32 Curran Associates, Inc., 2019 URL: https://proceedings.neurips.cc/paper/2019/file/fbad540b2f3b5638a9be9aa6a4d8e450-Paper.pdf
  38. “Passed & Spurious: Descent Algorithms and Local Minima in Spiked Matrix-Tensor Models” In Proceedings of the 36th International Conference on Machine Learning 97, Proceedings of Machine Learning Research PMLR, 2019, pp. 4333–4342 URL: https://proceedings.mlr.press/v97/mannelli19a.html
  39. Eliran Subag “The complexity of spherical p𝑝pitalic_p-spin models—a second moment approach” In Ann. Probab. 45.5, 2017, pp. 3385–3450 DOI: 10.1214/16-AOP1139
  40. “Concentration of the complexity of spherical pure p𝑝pitalic_p-spin models at arbitrary energies” In J. Math. Phys. 62.12, 2021, pp. Paper No. 123301\bibrangessep15 DOI: 10.1063/5.0070582
Citations (1)
View on Semantic Scholar

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now