Electronic Journal of Probability

Minimax rate of convergence and the performance of empirical risk minimization in phase recovery

Guillaume Lecué and Shahar Mendelson

Full-text: Open access

Abstract

We study the performance of Empirical Risk Minimization in both noisy and noiseless phase retrieval problems, indexed by subsets of $\mathbb{R}^n$ and relative to subgaussian sampling; that is, when the given data is $y_i=\left a_i,x_0\right^2+w_i$ for a subgaussian random vector $a$, independent noise $w$ and a fixed but unknown $x_0$ that belongs to a given subset of $\mathbb{R}^n$.

We show that ERM produces $\hat{x}$ whose Euclidean distance to either $x_0$ or $-x_0$ depends on the gaussian mean-width of the indexing set and on the signal-to-noise ratio of the problem. The bound coincides with the one for linear regression when $\|x_0\|_2$ is of the order of a constant. In addition, we obtain a minimax lower bound for the problem and identify sets for which ERM is a minimax procedure. As examples, we study the class of $d$-sparse vectors in $\mathbb{R}^n$ and the unit ball in $\ell_1^n$.

Article information

Source
Electron. J. Probab., Volume 20 (2015), paper no. 57, 29 pp.

Dates
Accepted: 29 May 2015
First available in Project Euclid: 4 June 2016

Permanent link to this document
https://projecteuclid.org/euclid.ejp/1465067163

Digital Object Identifier
doi:10.1214/EJP.v20-3525

Mathematical Reviews number (MathSciNet)
MR3354617

Zentralblatt MATH identifier
1318.62178

Subjects
Primary: 62G99: None of the above, but in this section

Keywords
empirical process phase recovery minimax theory

Rights
This work is licensed under a Creative Commons Attribution 3.0 License.

Citation

Lecué, Guillaume; Mendelson, Shahar. Minimax rate of convergence and the performance of empirical risk minimization in phase recovery. Electron. J. Probab. 20 (2015), paper no. 57, 29 pp. doi:10.1214/EJP.v20-3525. https://projecteuclid.org/euclid.ejp/1465067163


Export citation

References

  • Birgé, Lucien. Approximation dans les espaces métriques et théorie de l'estimation. (French) [Approximation in metric spaces and the theory of estimation] Z. Wahrsch. Verw. Gebiete 65 (1983), no. 2, 181–237.
  • Candès, Emmanuel J.; Eldar, Yonina C.; Strohmer, Thomas; Voroninski, Vladislav. Phase retrieval via matrix completion. SIAM J. Imaging Sci. 6 (2013), no. 1, 199–225.
  • Candès, Emmanuel J.; Strohmer, Thomas; Voroninski, Vladislav. PhaseLift: exact and stable signal recovery from magnitude measurements via convex programming. Comm. Pure Appl. Math. 66 (2013), no. 8, 1241–1274.
  • Candès, Emmanuel J.; Li, Xiaodong; Soltanolkotabi, Mahdi. Phase retrieval from coded diffraction patterns. Appl. Comput. Harmon. Anal. 39 (2015), no. 2, 277–299.
  • Candès, Emmanuel J.; Li, Xiaodong; Soltanolkotabi, Mahdi. Phase retrieval via Wirtinger flow: theory and algorithms. IEEE Trans. Inform. Theory 61 (2015), no. 4, 1985–2007.
  • Candès, Emmanuel J.; Li, Xiaodong. Solving quadratic equations via PhaseLift when there are about as many equations as unknowns. Found. Comput. Math. 14 (2014), no. 5, 1017–1026.
  • Chafai, Djalil; Guédon, Olivier; Lecué, Guillaume; Pajor, Alain. Interactions between compressed sensing random matrices and high dimensional geometry. Panoramas et Synthèses [Panoramas and Syntheses], 37. Société Mathématique de France, Paris, 2012. 181 pp. ISBN: 978-2-85629-370-6
  • Eldar, Yonina C.; Mendelson, Shahar. Phase retrieval: stability and recovery guarantees. Appl. Comput. Harmon. Anal. 36 (2014), no. 3, 473–494.
  • J.R. Fienup. Phase retrieval algorithms: a comparison. Applied Optics, 21(15):2758-2769, 1982.
  • F. Fogel and I. Waldspurger and A. D'Aspremont. Phase Retrieval for Imaging Problems. arXiv:1304.7735.
  • Garnaev, A. Yu.; Gluskin, E. D. The widths of a Euclidean ball. (Russian) Dokl. Akad. Nauk SSSR 277 (1984), no. 5, 1048–1052.
  • de la Pena, Victor H.; Giné, Evarist. Decoupling. From dependence to independence. Randomly stopped processes. $U$-statistics and processes. Martingales and beyond. Probability and its Applications (New York). Springer-Verlag, New York, 1999. xvi+392 pp. ISBN: 0-387-98616-2
  • Gordon, Y.; Litvak, A. E.; Mendelson, S.; Pajor, A. Gaussian averages of interpolated bodies and applications to approximate reconstruction. J. Approx. Theory 149 (2007), no. 1, 59–73.
  • R. Gerchberg and W. Saxton. A practical algorithm for the determination of phase from image and diffraction plane pictures. Optik, 35:237-246, 1972.
  • V. Koltchinskii and S. Mendelson. Bounding the smallest singular value of a random matrix without concentration. To appear in International Mathematics Research Notices. arXiv:1312.3580.
  • G. Lecué, S. Mendelson, Learning subgaussian classes : Upper and minimax bounds arXiv:1305.4825.
  • Lecué, Guillaume; Mendelson, Shahar. On the optimality of the empirical risk minimization procedure for the convex aggregation problem. Ann. Inst. Henri Poincaré Probab. Stat. 49 (2013), no. 1, 288–306.
  • G. Lecué, S. Mendelson. Sparse recovery under weak moment assumption. To appear in Journal of the European Mathematical Society. arXiv:1401.2188.
  • Ledoux, Michel; Talagrand, Michel. Probability in Banach spaces. Isoperimetry and processes. Ergebnisse der Mathematik und ihrer Grenzgebiete (3) [Results in Mathematics and Related Areas (3)], 23. Springer-Verlag, Berlin, 1991. xii+480 pp. ISBN: 3-540-52013-9
  • S. Mendelson. On the geometry of subgaussian coordinate projections. Preprint.
  • S. Mendelson. Learning without concentration. To appear in Journal of the ACM. arXiv:1410.3192.
  • S. Mendelson. Learning without concentration for general losses. arXiv:1410.3192.
  • S. Mendelson. A remark on the diameter of random sections of convex bodies. Geometric Aspects of Functional Analysis (GAFA Seminar Notes). Lecture notes in Mathematics 2116, pages 395–404, 2014.
  • S. Mendelson. Upper bounds on product and multiplier empirical processes. arXiv:1410.8003.
  • Milman, Vitali D.; Schechtman, Gideon. Asymptotic theory of finite-dimensional normed spaces. With an appendix by M. Gromov. Lecture Notes in Mathematics, 1200. Springer-Verlag, Berlin, 1986. viii+156 pp. ISBN: 3-540-16769-2
  • Y. Shechtman, Y.C. Eldar, A. Szameit, and M. Segev. Sparsity based sub-wavelength imaging with partially incoherent light via quadratic compressed sensing. Optics Express, 19(16):14807-14822, 2011.
  • Y. Shechtman and Y.C. Eldar and O. Cohen and H.N. Chapman and J. Miao and M. Segev. Phase Retrieval with Application to Optical Imaging, IEEE signal processing society, 32(3), pp. 87 – 109, May 2015.
  • Tsybakov, Alexandre B. Introduction to nonparametric estimation. Revised and extended from the 2004 French original. Translated by Vladimir Zaiats. Springer Series in Statistics. Springer, New York, 2009. xii+214 pp. ISBN: 978-0-387-79051-0
  • J. Tropp. Convex recovery of a structured signal from independent random linear measurements. To appear in Sampling Theory, a Renaissance. arXiv:1405.1102.
  • van der Vaart, Aad W.; Wellner, Jon A. Weak convergence and empirical processes. With applications to statistics. Springer Series in Statistics. Springer-Verlag, New York, 1996. xvi+508 pp. ISBN: 0-387-94640-3
  • V. Voroninski and Z. Xu. A strong restricted isometry property, with an application to phaseless compressed sensing. arXiv:1404.3811.
  • Waldspurger, Irène; d'Aspremont, Alexandre; Mallat, Stéphane. Phase recovery, MaxCut and complex semidefinite programming. Math. Program. 149 (2015), no. 1-2, Ser. A, 47–81.
  • J.D. Watson and F.H. Crick. A structure for Deoxyribose Nucleic Acid. Nature, 171, 1953.
  • Yang, Yuhong; Barron, Andrew. Information-theoretic determination of minimax rates of convergence. Ann. Statist. 27 (1999), no. 5, 1564–1599.