Optimization-Based AMP for Phase Retrieval: The Impact of Initialization and \ell_{2} Regularization

DOI10.1109/TIT.2019.2893254zbMATH Open1432.94032arXiv1801.01170OpenAlexW2963519307MaRDI QIDQ5224019FDOQ5224019

Authors: Junjie Ma, Ji Xu, Arian Maleki 

Publication date: 19 July 2019

Published in: IEEE Transactions on Information Theory (Search for Journal in Brave)

Abstract: We consider an $el{l}_2$-regularized non-convex optimization problem for recovering signals from their noisy phaseless observations. We design and study the performance of a message passing algorithm that aims to solve this optimization problem. We consider the asymptotic setting $m,nightarrowinfty$, $m/nightarrowdelta$ and obtain sharp performance bounds, where $m$ is the number of measurements and $n$ is the signal dimension. We show that for complex signals the algorithm can perform accurate recovery with only $m=left(frac64p{i}^2-4ight)napprox2.5n$ measurements. Also, we provide sharp analysis on the sensitivity of the algorithm to noise. We highlight the following facts about our message passing algorithm: (i) Adding $el{l}_2$ regularization to the non-convex loss function can be beneficial. (ii) Spectral initialization has marginal impact on the performance of the algorithm. The sharp analyses in this paper, not only enable us to compare the performance of our method with other phase recovery schemes, but also shed light on designing better iterative algorithms for other non-convex optimization problems.

Full work available at URL: https://arxiv.org/abs/1801.01170

Recommendations

• Robust amplitude method with \(L_{1/2}\)-regularization for compressive phase retrieval
• Phase Retrieval by Alternating Minimization With Random Initialization
• Optimal Spectral Initialization for Signal Recovery With Applications to Phase Retrieval
• Compressive Phase Retrieval via Reweighted Amplitude Flow
• Phase Retrieval via Reweighted Amplitude Flow
• The global landscape of phase retrieval. I: Perturbed amplitude models
• Phase retrieval using alternating minimization in a batch setting
• PhaseEqual: Convex Phase Retrieval via Alternating Direction Method of Multipliers
• Phase Retrieval Using Alternating Minimization
• scientific article

zbMATH Keywords

message passing algorithm\(l_2\)-regularized nonconvex optimization problemrecovering of signals from noisy phaseless observationssharp performance bounds

Mathematics Subject Classification ID

Signal theory (characterization, reconstruction, filtering, etc.) (94A12) Nonconvex programming, global optimization (90C26)

Cited In (12)

• Approximate message passing with spectral initialization for generalized linear models*
• Gradient descent with random initialization: fast global convergence for nonconvex phase retrieval
• A Unifying Tutorial on Approximate Message Passing
• Asymptotic Properties of Stationary Solutions of Coupled Nonconvex Nonsmooth Empirical Risk Minimization
• Approximate message passing with rigorous guarantees for pooled data and quantitative group testing
• The distribution of the Lasso: uniform control over sparse balls and adaptive parameter tuning
• Marginals of a spherical spin Glass model with correlated disorder
• Immersed body motion: near-bottom added mass effects
• Convex and Nonconvex Optimization Are Both Minimax-Optimal for Noisy Blind Deconvolution Under Random Designs
• Phase transition and higher order analysis of \(L_q\) regularization under dependence
• Fundamental barriers to high-dimensional regression with convex penalties
• Generalized approximate survey propagation for high-dimensional estimation ^*

This page was built for publication: Optimization-Based AMP for Phase Retrieval: The Impact of Initialization and $\ell_{2}$ Regularization

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5224019)