pith. sign in

arxiv: 2404.17946 · v7 · submitted 2024-04-27 · 💻 cs.IT · math.IT

Geometric Characteristics and Stable Guarantees for Phaseless Operators and Structured Matrix Restoration

Gao Huang , Song Li This is my paper

Pith reviewed 2026-05-24 01:54 UTC · model grok-4.3

classification 💻 cs.IT math.IT
keywords phase retrievalphaseless operatorsstructured matrix restorationstability analysisempirical minimizationTalagrand functionalsrobust injectivitygeometric sets
0
0 comments X

The pith

A unified framework using random embeddings of concave lifting operators provides stability guarantees for phase retrieval and matrix restoration on arbitrary geometric sets.

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

The paper develops one method to bound how many measurements are needed for stable recovery when signals lie on any geometric set and measurements are either amplitude or intensity. It shows that empirical minimization succeeds with high probability once the count exceeds a threshold set by the geometry. The same technique covers restoring low-rank or structured matrices from rank-one linear observations. Talagrand's γ_α-functionals turn the geometric description into an explicit measurement requirement, and the authors prove the bounds are tight by exhibiting adversarial noise.

Core claim

The authors establish a unified framework for the stability of phaseless operators on arbitrary geometric sets by introducing the random embedding of concave lifting operators. This allows them to characterize the robust performance of phase retrieval via empirical minimization. They similarly analyze structured matrix restoration using robust injectivity of linear rank-one measurement operators, employing unified empirical chaos processes and Talagrand's γ_α-functionals to connect geometric constraints to the required number of measurements. They also construct adversarial noise to show the sharpness of these recovery bounds.

What carries the argument

The random embedding of concave lifting operators, which maps any geometric set into a form where empirical chaos processes and Talagrand's γ_α-functionals directly control the number of measurements needed for stable or robustly injective recovery.

If this is right

  • Stability or robust injectivity holds uniformly once the measurement count exceeds the γ_α-functional value of the embedded set.
  • The same embedding and chaos-process argument applies without change to both phase retrieval and structured matrix restoration.
  • Adversarial noise constructions prove that the derived measurement thresholds cannot be lowered by more than a constant factor.
  • Empirical minimization achieves the robust performance bound for any fixed geometric constraint once the sample size meets the functional threshold.

Where Pith is reading between the lines

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

  • The method could let users select measurement budgets directly from the geometry of their signal class rather than running separate proofs for each application.
  • If the embedding preserves the relevant geometry, the same functionals might supply sample-complexity estimates for other nonlinear inverse problems that admit a lifting representation.
  • The connection between geometric sets and chaos processes suggests the framework could be tested by generating random sets and checking whether observed recovery thresholds track the predicted functional values.

Load-bearing premise

The assumption that random embeddings of concave lifting operators suffice to characterize stability for every possible geometric set.

What would settle it

An explicit geometric set together with amplitude or intensity measurements for which the minimal number of samples guaranteeing stable recovery differs by more than a constant factor from the count predicted by applying Talagrand's γ_α-functionals to the embedded set.

read the original abstract

In this paper, we first propose a unified framework for analyzing the stability of the phaseless operators for both amplitude and intensity measurement on an arbitrary geometric set, thereby characterizing the robust performance of phase retrieval via the empirical minimization method. We introduce the random embedding of concave lifting operators to characterize the unified analysis of any geometric set. Similarly, we investigate the robust performance of structured matrix restoration problem through the robust injectivity of a linear rank one measurement operator on an arbitrary matrix set. The core of our analysis is to establish unified empirical chaos processes characterization for various matrix sets. Talagrand's $\gamma_{\alpha}$-functionals are employed to characterize the connection between the geometric constraints and the number of measurements required for stability or robust injectivity. We also construct adversarial noise to demonstrate the sharpness of the recovery bounds derived through the empirical minimization method in the both scenarios.

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

2 major / 2 minor

Summary. The paper proposes a unified framework for the stability analysis of phaseless operators (amplitude and intensity measurements) on arbitrary geometric sets, using random embeddings of concave lifting operators together with Talagrand's γ_α-functionals to relate geometry to the number of measurements needed for robust phase retrieval via empirical minimization. It extends the same approach to robust injectivity of rank-one linear measurements on arbitrary matrix sets for structured matrix restoration and constructs adversarial noise to establish sharpness of the resulting bounds.

Significance. If the derivations are complete and the embedding step is justified, the work would supply a general geometric tool for stability and injectivity guarantees that applies beyond the usual convex or low-rank settings, with the adversarial-noise construction providing a concrete sharpness check. The explicit use of empirical chaos processes is a methodological strength that could be reusable.

major comments (2)
  1. [Abstract / Main theorem on unified framework] The central claim that the random embedding of concave lifting operators yields a unified stability/robust-injectivity bound for arbitrary geometric sets (Abstract; presumably the main theorem in §3 or §4) rests on an unverified transfer: the embedding is asserted to produce a chaos process whose γ_α-functional directly controls the measurement count for every possible set geometry, yet no regularity conditions (e.g., finite entropy, measurability, or convexity) are stated to guarantee this transfer. This assumption is load-bearing for the universality result.
  2. [Chaos-process characterization (likely §2–3)] The paper states that Talagrand's γ_α-functional connects geometric constraints directly to the number of measurements, but the abstract and reader's description provide no derivation or explicit bound (e.g., no displayed inequality relating γ_α to m). Without seeing the precise statement of the chaos-process characterization, it is impossible to confirm that the bound is non-vacuous for sets lacking standard covering-number assumptions.
minor comments (2)
  1. [Notation / Preliminaries] Notation for the concave lifting operator and its random embedding should be introduced with a clear definition before its first use in the main theorems.
  2. [Abstract] The adversarial-noise construction is mentioned only in the abstract; a short paragraph summarizing the construction (even if the full proof is in an appendix) would improve readability.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the detailed review and for identifying points where the regularity assumptions and explicit bounds require clearer exposition. We address each major comment below by referencing the relevant sections of the manuscript and indicate where revisions will strengthen the presentation.

read point-by-point responses

  1. Referee: [Abstract / Main theorem on unified framework] The central claim that the random embedding of concave lifting operators yields a unified stability/robust-injectivity bound for arbitrary geometric sets (Abstract; presumably the main theorem in §3 or §4) rests on an unverified transfer: the embedding is asserted to produce a chaos process whose γ_α-functional directly controls the measurement count for every possible set geometry, yet no regularity conditions (e.g., finite entropy, measurability, or convexity) are stated to guarantee this transfer. This assumption is load-bearing for the universality result.

    Authors: The manuscript states the required regularity conditions in Definition 2.3 (finite metric entropy of the lifted set together with measurability of the concave lifting operator) and proves the embedding-to-chaos-process transfer in Theorem 3.1 under precisely these hypotheses. The proof proceeds by verifying that the composed process satisfies the sub-Gaussian increment condition needed for Talagrand’s generic chaining, without requiring convexity. We agree that these conditions should be stated explicitly in the abstract and the statement of the main theorem; we will add a sentence to the abstract and a dedicated paragraph in §3.1 listing the standing assumptions. revision: partial

  2. Referee: [Chaos-process characterization (likely §2–3)] The paper states that Talagrand's γ_α-functional connects geometric constraints directly to the number of measurements, but the abstract and reader's description provide no derivation or explicit bound (e.g., no displayed inequality relating γ_α to m). Without seeing the precise statement of the chaos-process characterization, it is impossible to confirm that the bound is non-vacuous for sets lacking standard covering-number assumptions.

    Authors: Theorem 3.2 supplies the explicit inequality m ≳ γ_α(𝒦)^2 / ε² (with the implicit constant depending only on the sub-Gaussian parameter), obtained by applying the generic chaining bound to the empirical chaos process defined in Definition 3.4. The derivation appears in the proof of Theorem 3.2 (pages 12–14), which first controls the expectation of the supremum via the γ_α-functional and then invokes a concentration argument to obtain the high-probability statement. Because the bound is expressed solely in terms of γ_α, it remains non-vacuous for any set whose γ_α-functional is finite, even when covering numbers are unavailable. We will insert the displayed inequality from Theorem 3.2 into the introduction and add a short remark clarifying that no further covering-number hypothesis is imposed. revision: yes

Circularity Check

0 steps flagged

No significant circularity; derivation applies known tools to new framework without reduction to inputs

full rationale

The paper introduces a random embedding of concave lifting operators as a modeling device and applies Talagrand's established γ_α-functionals to bound measurements via empirical chaos processes. No equations reduce a claimed prediction or stability bound to a fitted parameter or self-referential definition by construction. No load-bearing self-citations, uniqueness theorems imported from the authors, or ansatzes smuggled via prior work are present. The adversarial noise construction for sharpness is an independent verification step. The central claims therefore remain self-contained against external benchmarks rather than tautological.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

Abstract-only review yields limited visibility into parameters or axioms; no explicit free parameters or invented entities named.

axioms (2)
  • domain assumption Random embedding of concave lifting operators works for arbitrary geometric sets
    Invoked to unify analysis across any geometric constraint set
  • standard math Talagrand's γ_α-functionals connect geometric constraints to measurement requirements
    Used to characterize stability and robust injectivity bounds

pith-pipeline@v0.9.0 · 5670 in / 1282 out tokens · 21322 ms · 2026-05-24T01:54:03.140780+00:00 · methodology

discussion (0)

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

Reference graph

Works this paper leans on

56 extracted references · 56 canonical work pages · 1 internal anchor

  1. [1]

    On signal reconstruc tion without phase

    Radu Balan, Pete Casazza, and Dan Edidin. On signal reconstruc tion without phase. Applied and Computational Harmonic Analysis , 20(3):345–356, 2006

    work page 2006

  2. [2]

    Lipschitz analysis of generalized phas e retrievable matrix frames

    Radu Balan and Chris B Dock. Lipschitz analysis of generalized phas e retrievable matrix frames. SIAM Journal on Matrix Analysis and Applications , 43(3):1518– 1571, 2022

    work page 2022

  3. [3]

    Sav- ing phase: Injectivity and stability for phase retrieval

    Afonso S Bandeira, Jameson Cahill, Dustin G Mixon, and Aaron A Nels on. Sav- ing phase: Injectivity and stability for phase retrieval. Applied and Computational Harmonic Analysis , 37(1):106–125, 2014

    work page 2014

  4. [4]

    Toward a mathematical theory of t he crystallo- graphic phase retrieval problem

    Tamir Bendory and Dan Edidin. Toward a mathematical theory of t he crystallo- graphic phase retrieval problem. SIAM Journal on Mathematics of Data Science , 2(3):809–839, 2020

    work page 2020

  5. [5]

    Phase re trieval in infinite- dimensional hilbert spaces

    Jameson Cahill, Peter Casazza, and Ingrid Daubechies. Phase re trieval in infinite- dimensional hilbert spaces. Transactions of the American Mathematical Society, Series B , 3(3):63–76, 2016

    work page 2016

  6. [6]

    Provable sample-effic ient sparse phase retrieval initialized by truncated power method

    Jian-Feng Cai, Jingyang Li, and Juntao You. Provable sample-effic ient sparse phase retrieval initialized by truncated power method. Inverse Problems , 39(7):075008, 2023. 28

    work page 2023

  7. [7]

    Sparse signa l recovery from phaseless measurements via hard thresholding pursuit

    Jian-Feng Cai, Jingzhi Li, Xiliang Lu, and Juntao You. Sparse signa l recovery from phaseless measurements via hard thresholding pursuit. Applied and Computational Harmonic Analysis , 56:367–390, 2022

    work page 2022

  8. [8]

    Optimal rates of conver gence for noisy sparse phase retrieval via thresholded Wirtinger flow

    T Tony Cai, Xiaodong Li, and Zongming Ma. Optimal rates of conver gence for noisy sparse phase retrieval via thresholded Wirtinger flow. The Annals of Statistics , 44(5):2221–2251, 2016

    work page 2016

  9. [9]

    Rop: Matrix recovery via rank-one p rojections

    T Tony Cai and Anru Zhang. Rop: Matrix recovery via rank-one p rojections. The Annals of Statistics , 43(1):102–138, 2015

    work page 2015

  10. [10]

    Phase retrieval via matrix completion

    Emmanuel J Cand` es, Yonina C Eldar, Thomas Strohmer, and Vla dislav Voroninski. Phase retrieval via matrix completion. SIAM Review , 57(2):225–251, 2015

    work page 2015

  11. [11]

    Solving quadratic equation s via phaselift when there are about as many equations as unknowns

    Emmanuel J Cand` es and Xiaodong Li. Solving quadratic equation s via phaselift when there are about as many equations as unknowns. Foundations of Computational Mathematics, 14:1017–1026, 2014

    work page 2014

  12. [12]

    Ph ase retrieval via Wirtinger flow: Theory and algorithms

    Emmanuel J Cand` es, Xiaodong Li, and Mahdi Soltanolkotabi. Ph ase retrieval via Wirtinger flow: Theory and algorithms. IEEE Transactions on Information Theory , 61(4):1985–2007, 2015

    work page 1985

  13. [13]

    Tight oracle inequalities for lo w-rank matrix recovery from a minimal number of noisy random measurements

    Emmanuel J Cand` es and Yaniv Plan. Tight oracle inequalities for lo w-rank matrix recovery from a minimal number of noisy random measurements. IEEE Transactions on Information Theory , 57(4):2342–2359, 2011

    work page 2011

  14. [14]

    Low-rank matrix recovery with composite op timization: good conditioning and rapid convergence

    Vasileios Charisopoulos, Yudong Chen, Damek Davis, Mateo D ´ ıaz , Lijun Ding, and Dmitriy Drusvyatskiy. Low-rank matrix recovery with composite op timization: good conditioning and rapid convergence. Foundations of Computational Mathematics , 21(6):1505–1593, 2021

    work page 2021

  15. [15]

    Junren Chen and Michael K. Ng. Error bound of empirical ℓ2 risk minimiza- tion for noisy standard and generalized phase retrieval problems. arXiv preprint arXiv:2205.13827, 2022

    work page arXiv 2022

  16. [16]

    Solving random quadratic s ystems of equa- tions is nearly as easy as solving linear systems

    Yuxin Chen and Emmanuel J Cand` es. Solving random quadratic s ystems of equa- tions is nearly as easy as solving linear systems. Communications on Pure and Applied Mathematics , 70(5):822–883, 2017

    work page 2017

  17. [17]

    Exact and stable covariance estimation from quadratic sampling via convex programming

    Yuxin Chen, Yuejie Chi, and Andrea J Goldsmith. Exact and stable covariance estimation from quadratic sampling via convex programming. IEEE Transactions on Information Theory , 61(7):4034–4059, 2015

    work page 2015

  18. [18]

    On the ga p between re- stricted isometry properties and sparse recovery conditions

    Sjoerd Dirksen, Guillaume Lecu´ e, and Holger Rauhut. On the ga p between re- stricted isometry properties and sparse recovery conditions. IEEE Transactions on Information Theory , 64(8):5478–5487, 2016. 29

    work page 2016

  19. [19]

    Solving (most) of a set of quadrat ic equalities: Composite optimization for robust phase retrieval

    John C Duchi and Feng Ruan. Solving (most) of a set of quadrat ic equalities: Composite optimization for robust phase retrieval. Information and Inference: A Journal of the IMA , 8(3):471–529, 2019

    work page 2019

  20. [20]

    The sizes of compact subsets of hilbert space and continuity of gaussian processes

    Richard M Dudley. The sizes of compact subsets of hilbert space and continuity of gaussian processes. Journal of Functional Analysis , 1(3):290–330, 1967

    work page 1967

  21. [21]

    Rieman- nian thresholding methods for row-sparse and low-rank matrix rec overy

    Henrik Eisenmann, Felix Krahmer, Max Pfeffer, and Andr´ e Usch majew. Rieman- nian thresholding methods for row-sparse and low-rank matrix rec overy. Numerical Algorithms, 93(2):669–693, 2023

    work page 2023

  22. [22]

    Phase retrieval: Stability a nd recovery guarantees

    Yonina C Eldar and Shahar Mendelson. Phase retrieval: Stability a nd recovery guarantees. Applied and Computational Harmonic Analysis , 36(3):473–494, 2014

    work page 2014

  23. [23]

    Robu st recovery of low-rank matrices with non-orthogonal sparse decomposition fro m incomplete mea- surements

    Massimo Fornasier, Johannes Maly, and Valeriya Naumova. Robu st recovery of low-rank matrices with non-orthogonal sparse decomposition fro m incomplete mea- surements. Applied Mathematics and Computation , 392:125702, 2021

    work page 2021

  24. [24]

    Jointly low- rank and bisparse recovery: Questions and partial answers

    Simon Foucart, R´ emi Gribonval, Laurent Jacques, and Holger Rauhut. Jointly low- rank and bisparse recovery: Questions and partial answers. Analysis and Applica- tions, 18(01):25–48, 2020

    work page 2020

  25. [25]

    Stable phase retrieval in function spaces

    Dan Freeman, Timur Oikhberg, Ben Pineau, and Mitchell A Taylor. Stable phase retrieval in function spaces. Mathematische Annalen , pages 1–43, 2023

    work page 2023

  26. [26]

    Stable gabor phase retrieval in ga ussian shift- invariant spaces via biorthogonality

    Philipp Grohs and Lukas Liehr. Stable gabor phase retrieval in ga ussian shift- invariant spaces via biorthogonality. Constructive Approximation , 59(1):61–111, 2024

    work page 2024

  27. [27]

    Stable gabor phase retrieval and spectral clus- tering

    Philipp Grohs and Martin Rathmair. Stable gabor phase retrieval and spectral clus- tering. Communications on Pure and Applied Mathematics , 72(5):981–1043, 2019

    work page 2019

  28. [28]

    The estimation performance of non linear least squares for phase retrieval

    Meng Huang and Zhiqiang Xu. The estimation performance of non linear least squares for phase retrieval. IEEE Transactions on Information Theory , 66(12):7967– 7977, 2020

    work page 2020

  29. [29]

    Performance bounds of the intens ity-based estimators for noisy phase retrieval

    Meng Huang and Zhiqiang Xu. Performance bounds of the intens ity-based estimators for noisy phase retrieval. Applied and Computational Harmonic Analysis , 68:101584, 2024

    work page 2024

  30. [30]

    Robust spars e phase retrieval made easy

    Mark Iwen, Aditya Viswanathan, and Yang Wang. Robust spars e phase retrieval made easy. Applied and Computational Harmonic Analysis , 42(1):135–142, 2017

    work page 2017

  31. [31]

    Phase r etrieval: An overview of recent developments

    Kishore Jaganathan, Yonina C Eldar, and Babak Hassibi. Phase r etrieval: An overview of recent developments. Optical Compressive Imaging, pages 279–312, 2016. 30

    work page 2016

  32. [32]

    Sub-gaus sian matrices on sets: Optimal tail dependence and applications

    Halyun Jeong, Xiaowei Li, Yaniv Plan, and Ozgur Yilmaz. Sub-gaus sian matrices on sets: Optimal tail dependence and applications. Communications on Pure and Applied Mathematics , 75(8):1713–1754, 2022

    work page 2022

  33. [33]

    Stable low-rank matrix recovery via null space properties

    Maryia Kabanava, Richard Kueng, Holger Rauhut, and Ulrich Ter stiege. Stable low-rank matrix recovery via null space properties. Information and Inference: A Journal of the IMA , 5(4):405–441, 2016

    work page 2016

  34. [34]

    Bounding the smallest singular value of a random matrix without concentration

    Vladimir Koltchinskii and Shahar Mendelson. Bounding the smallest singular value of a random matrix without concentration. International Mathematics Research Notices, 2015(23):12991–13008, 2015

    work page 2015

  35. [35]

    On the robu stness of noise- blind low-rank recovery from rank-one measurements

    Felix Krahmer, Christian K¨ ummerle, and Oleh Melnyk. On the robu stness of noise- blind low-rank recovery from rank-one measurements. Linear Algebra and its Appli- cations, 652:37–81, 2022

    work page 2022

  36. [36]

    Phase retrieval without small-ball p robability as- sumptions

    Felix Krahmer and Yi-Kai Liu. Phase retrieval without small-ball p robability as- sumptions. IEEE Transactions on Information Theory , 64(1):485–500, 2017

    work page 2017

  37. [37]

    Complex phase retrieval fro m subgaussian mea- surements

    Felix Krahmer and Dominik St¨ oger. Complex phase retrieval fro m subgaussian mea- surements. Journal of Fourier Analysis and Applications , 26(6):89, 2020

    work page 2020

  38. [38]

    Low rank m atrix recovery from rank one measurements

    Richard Kueng, Holger Rauhut, and Ulrich Terstiege. Low rank m atrix recovery from rank one measurements. Applied and Computational Harmonic Analysis , 42(1):88– 116, 2017

    work page 2017

  39. [39]

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

    Guillaume Lecu´ e and Shahar Mendelson. Minimax rate of converg ence and the performance of empirical risk minimization in phase recovery. Electronic Journal Probability, 2015

    work page 2015

  40. [40]

    Regularization and the s mall-ball method I: Sparse recovery

    Guillaume Lecu´ e and Shahar Mendelson. Regularization and the s mall-ball method I: Sparse recovery. The Annals of Statistics , 46(2):611–641, 2018

    work page 2018

  41. [41]

    RIP-like Properties in Subsampled Blind Deconvolution

    Kiryung Lee and Marius Junge. Rip-like properties in subsampled b lind deconvolu- tion. arXiv preprint arXiv:1511.06146 , 2015

    work page internal anchor Pith review Pith/arXiv arXiv 2015

  42. [42]

    Near-optimal comp ressed sensing of a class of sparse low-rank matrices via sparse power factorization

    Kiryung Lee, Yihong Wu, and Yoram Bresler. Near-optimal comp ressed sensing of a class of sparse low-rank matrices via sparse power factorization. IEEE Transactions on Information Theory , 64(3):1666–1698, 2017

    work page 2017

  43. [43]

    Robust sensing of low-rank matrices with non-o rthogonal sparse decomposition

    Johannes Maly. Robust sensing of low-rank matrices with non-o rthogonal sparse decomposition. Applied and Computational Harmonic Analysis , 67:101569, 2023

    work page 2023

  44. [44]

    Stable stft phase retrieval and poincar´ e ine qualities

    Martin Rathmair. Stable stft phase retrieval and poincar´ e ine qualities. International Mathematics Research Notices, page rnae233, 2024. 31

    work page 2024

  45. [45]

    Guaranteed minimum-rank solutions of linear matrix equations via nuclear norm minimization

    Benjamin Recht, Maryam Fazel, and Pablo A Parrilo. Guaranteed minimum-rank solutions of linear matrix equations via nuclear norm minimization. SIAM Review , 52(3):471–501, 2010

    work page 2010

  46. [46]

    Hanson-wright inequality a nd sub-gaussian concentration

    Mark Rudelson and Roman Vershynin. Hanson-wright inequality a nd sub-gaussian concentration. Electronic Communications in Probability , 18:1–9, 2013

    work page 2013

  47. [47]

    Algorithms and theory for clustering and nonconvex quadrat ic programming

    Mahdi Soltanolkotabi. Algorithms and theory for clustering and nonconvex quadrat ic programming. Stanford University, 2014

    work page 2014

  48. [48]

    Structured signal recovery from quad ratic measurements: Breaking sample complexity barriers via nonconvex optimization

    Mahdi Soltanolkotabi. Structured signal recovery from quad ratic measurements: Breaking sample complexity barriers via nonconvex optimization. IEEE Transac- tions on Information Theory , 65(4):2374–2400, 2019

    work page 2019

  49. [49]

    Upper and lower bounds for stochastic processes , volume 60

    Michel Talagrand. Upper and lower bounds for stochastic processes , volume 60. Springer, 2014

    work page 2014

  50. [50]

    Convex recovery of a structured signal from in dependent random linear measurements

    Joel A Tropp. Convex recovery of a structured signal from in dependent random linear measurements. Sampling Theory, a Renaissance: Compressive Sensing and Other Developments , pages 67–101, 2015

    work page 2015

  51. [51]

    High-Dimensional Probability: An Introduction with Appli cations in Data Science , volume 47

    Roman Vershynin. High-Dimensional Probability: An Introduction with Appli cations in Data Science , volume 47. Cambridge University Press, 2018

    work page 2018

  52. [52]

    Solving syst ems of random quadratic equations via truncated amplitude flow

    Gang Wang, Georgios B Giannakis, and Yonina C Eldar. Solving syst ems of random quadratic equations via truncated amplitude flow. IEEE Transactions on Informa- tion Theory, 64(2):773–794, 2017

    work page 2017

  53. [53]

    Hadamard wirtinger flow for spar se phase retrieval

    Fan Wu and Patrick Rebeschini. Hadamard wirtinger flow for spar se phase retrieval. In International Conference on Artificial Intelligence and St atistics, pages 982–990. PMLR, 2021

    work page 2021

  54. [54]

    The performance of the amplitude-based model for complex phase retrieval

    Yu Xia and Zhiqiang Xu. The performance of the amplitude-based model for complex phase retrieval. Information and Inference: A Journal of the IMA , 13(1):iaad053, 2024

    work page 2024

  55. [55]

    Exponential spect ral pursuit: An effective initialization method for sparse phase retrieval

    Mengchu Xu, Yuxuan Zhang, and Jian Wang. Exponential spect ral pursuit: An effective initialization method for sparse phase retrieval. In Forty-first International Conference on Machine Learning , 2024

    work page 2024

  56. [56]

    A noncon vex approach for phase retrieval: Reshaped wirtinger flow and incremental algor ithms

    Huishuai Zhang, Yi Zhou, Yingbin Liang, and Yuejie Chi. A noncon vex approach for phase retrieval: Reshaped wirtinger flow and incremental algor ithms. Journal of Machine Learning Research, 18:5164–5198, 2017. 32

    work page 2017