Phase retrieval via overparametrized nonconvex optimization: nonsmooth amplitude loss landscapes

We study nonconvex optimization for phase retrieval and the more general problem of semidefinite low-rank matrix sensing; in particular, we focus on the global nonconvex landscape of overparametrized versions of the nonsmooth amplitude least-squares loss as well as a smooth reformulation of this loss based on the PhaseCut approach. We first give a general, deterministic result on properties of second-order critical points for a general class of loss functions; we then specialize this result to the nonsmooth amplitude loss and, additionally, prove nearly identical results for a smooth reformulation (similar to PhaseCut) as a synchronization problem over spheres. Finally, we show the usefulness of these tools by proving high-probability landscape guarantees in two settings: (1) phase retrieval with isotropic sub-Gaussian measurements, and (2) phase retrieval in a general (possibly infinite-dimensional) Hilbert space with Gaussian measurements. In both cases, our results give state-of-the-art and statistically optimal guarantees with only a constant amount of overparametrization (in the well-studied case of isotropic sub-Gaussian measurements, such statistical guarantees had previously required greater degrees of overparametrization/relaxation); this demonstrates the potential of overparametrized nonconvex optimization as a principled and scalable algorithmic approach to phase retrieval.

• Publication year

  2025

• Venue

  Unknown venue

• Publication date

  2025-11-24

• Fields of study

  Mathematics, Computer Science, Engineering

• Identifiers
  arXiv 2511.19045
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• Phase retrieval and matrix sensing via benign and overparametrized nonconvex optimization

  2025influential reference
• Convergence analysis of Wirtinger Flow for Poisson phase retrieval

  2025influential reference
• Benign landscape for Burer-Monteiro factorizations of MaxCut-type semidefinite programs

  2024cited by this paper
• On the Hardness of Meaningful Local Guarantees in Nonsmooth Nonconvex Optimization

  2024cited by this paper
• Noisy phase retrieval from subgaussian measurements

  2024cited by this paper
• Truncated amplitude flow with coded diffraction patterns

  2024cited by this paper
• Local Geometry Determines Global Landscape in Low-rank Factorization for Synchronization

  2023cited by this paper
• Maximum-likelihood estimation in ptychography in the presence of Poisson-Gaussian noise statistics.

  2023influential reference
• The Local Landscape of Phase Retrieval Under Limited Samples

  2023cited by this paper
• The Riemannian Elevator for certifiable distance-based localization

  2022cited by this paper
• Phase Retrieval: From Computational Imaging to Machine Learning: A tutorial

  2022cited by this paper
• Nearly optimal bounds for the global geometric landscape of phase retrieval

  2022cited by this paper
• Smoothed amplitude flow-based phase retrieval algorithm

  2021cited by this paper
• Sharp Global Guarantees for Nonconvex Low-Rank Recovery in the Noisy Overparameterized Regime

  2021cited by this paper
• The global landscape of phase retrieval I: perturbed amplitude models

  2021cited by this paper
• Lipschitz Analysis of Generalized Phase Retrievable Matrix Frames

  2021cited by this paper
• Local and Global Linear Convergence of General Low-Rank Matrix Recovery Problems

  2021cited by this paper
• The numerics of phase retrieval

  2020cited by this paper
• Solving orthogonal group synchronization via convex and low-rank optimization: tightness and landscape analysis

  2020cited by this paper
• Complex Dynamics in Simple Neural Networks: Understanding Gradient Flow in Phase Retrieval

  2020cited by this paper
• Sample complexity and effective dimension for regression on manifolds

  2020cited by this paper
• Phase retrieval for sub-Gaussian measurements

  2019cited by this paper
• Phase retrieval via smoothed amplitude flow

  2019cited by this paper
• First-order methods almost always avoid strict saddle points

  2019cited by this paper
• Complex Phase Retrieval from Subgaussian Measurements

  2019cited by this paper
• Gradient descent with random initialization: fast global convergence for nonconvex phase retrieval

  2018cited by this paper
• Deterministic Guarantees for Burer‐Monteiro Factorizations of Smooth Semidefinite Programs

  2018cited by this paper
• Phase Retrieval via Reweighted Amplitude Flow

  2018cited by this paper
• Concise complexity analyses for trust region methods

  2018cited by this paper
• Toward the Optimal Construction of a Loss Function Without Spurious Local Minima for Solving Quadratic Equations

  2018cited by this paper
• Implicit Regularization in Nonconvex Statistical Estimation: Gradient Descent Converges Linearly for Phase Retrieval, Matrix Completion, and Blind Deconvolution

  2017cited by this paper
• A Nonconvex Approach for Phase Retrieval: Reshaped Wirtinger Flow and Incremental Algorithms

  2017cited by this paper
• The nonsmooth landscape of phase retrieval

  2017cited by this paper
• Solving Systems of Random Quadratic Equations via Truncated Amplitude Flow

  2016cited by this paper
• Reshaped Wirtinger Flow for Solving Quadratic System of Equations

  2016cited by this paper
• A Geometric Analysis of Phase Retrieval

  2016cited by this paper
• The non-convex geometry of low-rank matrix optimization

  2016cited by this paper
• Phase Retrieval Without Small-Ball Probability Assumptions

  2016cited by this paper
• On Lipschitz Analysis and Lipschitz Synthesis for the Phase Retrieval Problem

  2015cited by this paper
• Experimental robustness of Fourier Ptychography phase retrieval algorithms

  2015influential reference
• Phase Retrieval with Application to Optical Imaging: A contemporary overview

  2015cited by this paper
• Concentration inequalities and moment bounds for sample covariance operators

  2014cited by this paper
• Phase Retrieval: Stability and Recovery Guarantees

  2012cited by this paper
• Phase recovery, MaxCut and complex semidefinite programming

  2012influential reference

• No citing papers are available for this paper.