pith. sign in

arxiv: 2505.02636 · v3 · submitted 2025-05-05 · 🧮 math.OC · math.ST· stat.TH

Phase retrieval and matrix sensing via benign and overparametrized nonconvex optimization

Andrew D. McRae This is my paper

Pith reviewed 2026-05-22 16:05 UTC · model grok-4.3

classification 🧮 math.OC math.STstat.TH
keywords phase retrievalmatrix sensingnonconvex optimizationBurer-Monteiro factorizationoverparametrizationsecond-order critical pointssemidefinite programmingdual certificates
0
0 comments X

The pith

Mild overparametrization in a quartic nonconvex formulation recovers the ground truth with optimal sample complexity for phase retrieval and low-rank matrix sensing.

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

The paper analyzes the landscape of a quartic Burer-Monteiro factored least-squares objective for phase retrieval and semidefinite matrix sensing. It introduces a framework that exploits the underlying semidefinite structure to show when second-order critical points recover the unknown matrix. Mild overparametrization, increasing the factorization rank by at most a logarithmic factor in the dimension, ensures these critical points are benign. This yields the optimal statistical sample complexity and error rates previously obtained only via convex semidefinite programming, for sub-Gaussian phase retrieval and rank-1 Gaussian matrix sensing. The analysis draws on convex dual certificates and recovers prior non-overparametrized guarantees as special cases.

Core claim

Overparametrizing the rank in the quartic Burer-Monteiro formulation by a factor at most logarithmic in the dimension ensures that second-order critical points approximately recover the ground truth matrix, achieving optimal statistical sample complexity and error for phase retrieval with sub-Gaussian measurements and for semidefinite matrix sensing with rank-1 Gaussian measurements.

What carries the argument

A new analysis framework that exploits the semidefinite problem structure to characterize the properties of second-order critical points in the mildly overparametrized quartic Burer-Monteiro formulation, combined with convex dual certificates.

Load-bearing premise

The framework correctly identifies when second-order critical points in the overparametrized quartic formulation recover the ground truth.

What would settle it

An explicit counterexample with sub-Gaussian measurements where a second-order critical point of the log-overparametrized quartic problem fails to recover the ground truth matrix up to the claimed error.

Figures

Figures reproduced from arXiv: 2505.02636 by Andrew D. McRae.

Figure 1
Figure 1. Figure 1: Error histograms of 100 random trials with [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
read the original abstract

We study a nonconvex optimization algorithmic approach to phase retrieval and the more general problem of semidefinite low-rank matrix sensing. Specifically, we analyze the nonconvex landscape of a quartic Burer-Monteiro factored least-squares optimization problem. We develop a new analysis framework, taking advantage of the semidefinite problem structure, to understand the properties of second-order critical points -- specifically, whether they (approximately) recover the ground truth matrix. We show that it can be helpful to (mildly) overparametrize the problem, that is, to optimize over matrices of higher rank than the ground truth. We then apply this framework to several well-studied problem instances: in addition to recovering existing state-of-the-art phase retrieval landscape guarantees (without overparametrization), we show that overparametrizing by a factor at most logarithmic in the dimension allows recovery with optimal statistical sample complexity and error for the problems of (1) phase retrieval with sub-Gaussian measurements and (2) more general semidefinite matrix sensing with rank-1 Gaussian measurements. Previously, such statistical results had been shown only for estimators based on semidefinite programming. More generally, our analysis is partially based on the powerful method of convex dual certificates, suggesting that it could be applied to a much wider class of problems.

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 manuscript analyzes the nonconvex landscape of a quartic Burer-Monteiro factored least-squares formulation for phase retrieval and semidefinite low-rank matrix sensing. It introduces a new analysis framework that exploits semidefinite problem structure and convex dual certificates to characterize second-order critical points (SOCPs). The central results are that (i) existing landscape guarantees for phase retrieval are recovered without overparametrization and (ii) overparametrizing the factorization rank by a factor at most logarithmic in the dimension yields recovery of the ground truth (up to optimal statistical error) with information-theoretically optimal sample complexity for sub-Gaussian phase retrieval and rank-1 Gaussian semidefinite matrix sensing.

Significance. If the claims hold, the work is significant because it supplies the first landscape analysis showing that mildly overparametrized nonconvex methods achieve the same optimal statistical rates previously known only for SDP relaxations. The explicit use of dual certificates to certify SOCPs is a strength that may extend to other problems; the logarithmic overparametrization factor is parameter-free and therefore particularly attractive. No machine-checked proofs or reproducible code are mentioned.

major comments (2)
  1. [Abstract, paragraph on new analysis framework] Abstract, paragraph on new analysis framework: the claim that the semidefinite-structure-aware framework certifies that all SOCPs approximately recover the ground truth once the factorization rank exceeds the true rank by a logarithmic factor requires explicit verification that the dual-certificate construction absorbs the quartic Hessian structure and sub-Gaussian concentration tails without inflating the sample complexity beyond the information-theoretic optimum.
  2. [Main theorem for phase retrieval] Main theorem for phase retrieval (presumably Theorem 1 or 2): the stated logarithmic overparametrization must be shown to be sufficient for the residual terms arising from the quartic objective to remain controlled; if the certificate strictness requires a poly-log or dimension-dependent factor, the optimal-complexity claim would not hold.
minor comments (2)
  1. [Introduction] Notation for the overparametrization factor r = k + log(d) should be introduced with a precise definition of k (true rank) and d (ambient dimension) at first use.
  2. [Abstract] The abstract states that the framework 'could be applied to a much wider class of problems'; a concrete example or remark on the required structural assumptions would improve clarity.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their thoughtful review and positive evaluation of our work's significance. We address each major comment below and plan to incorporate clarifications in the revised manuscript.

read point-by-point responses

  1. Referee: [Abstract, paragraph on new analysis framework] Abstract, paragraph on new analysis framework: the claim that the semidefinite-structure-aware framework certifies that all SOCPs approximately recover the ground truth once the factorization rank exceeds the true rank by a logarithmic factor requires explicit verification that the dual-certificate construction absorbs the quartic Hessian structure and sub-Gaussian concentration tails without inflating the sample complexity beyond the information-theoretic optimum.

    Authors: We appreciate this observation. The explicit verification is provided in the proof of Theorem 1 (for phase retrieval) and the corresponding theorem for matrix sensing. Specifically, in the dual certificate construction (see Section 4), we bound the quartic Hessian contributions using the mild logarithmic overparametrization, which ensures that the residual terms are dominated by the linear terms without requiring additional sample factors. The sub-Gaussian tails are handled via standard concentration inequalities that match the information-theoretic bounds. To make this more prominent, we will revise the abstract to include a brief reference to this control. revision: yes

  2. Referee: [Main theorem for phase retrieval] Main theorem for phase retrieval (presumably Theorem 1 or 2): the stated logarithmic overparametrization must be shown to be sufficient for the residual terms arising from the quartic objective to remain controlled; if the certificate strictness requires a poly-log or dimension-dependent factor, the optimal-complexity claim would not hold.

    Authors: We agree that this control is crucial. In the analysis of the second-order critical points, the logarithmic factor in the overparametrization is precisely chosen to absorb the quartic residuals. Our calculations show that no additional poly-log factors are needed beyond the logarithmic overparametrization, preserving the optimal sample complexity. We will add a dedicated remark or lemma highlighting this control in the revised version to address any potential ambiguity. revision: yes

Circularity Check

0 steps flagged

No significant circularity; derivation relies on external convex dual certificates and semidefinite structure

full rationale

The paper introduces a new analysis framework that leverages the semidefinite problem structure and the method of convex dual certificates to characterize second-order critical points of the quartic Burer-Monteiro objective. This framework is applied to recover existing landscape guarantees for phase retrieval without overparametrization and to establish new results for mildly overparametrized cases with sub-Gaussian and rank-1 Gaussian measurements. No equations or steps reduce the claimed recovery guarantees or optimal sample complexity to a fitted quantity defined from the same data, nor do they depend on self-citations that are load-bearing or uniqueness theorems imported from the authors' prior work. The central claims remain self-contained against external benchmarks from convex optimization.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

Abstract-only view limits visibility; the central claims rest on standard sub-Gaussian and Gaussian measurement assumptions plus the semidefinite structure of the underlying problem, with no free parameters or invented entities explicitly introduced.

axioms (2)
  • domain assumption Measurements are sub-Gaussian (phase retrieval) or rank-1 Gaussian (matrix sensing).
    Invoked to obtain optimal statistical rates; stated in the abstract when describing the problem instances.
  • domain assumption The quartic Burer-Monteiro objective admits a semidefinite lift whose dual certificates control second-order critical points.
    Core of the new analysis framework; referenced in the abstract paragraph on the analysis method.

pith-pipeline@v0.9.0 · 5765 in / 1522 out tokens · 42440 ms · 2026-05-22T16:05:07.735395+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

64 extracted references · 64 canonical work pages

  1. [1]

    Array imaging using intensity-only measurements,

    A. Chai, M. Moscoso, and G. Papanicolaou, “Array imaging using intensity-only measurements,” Inverse Probl., vol. 27, no. 015005, 2011

    work page 2011

  2. [2]

    PhaseLift: Exact and stable signal recovery from magnitude measurements via convex programming,

    E. J. Cand` es, T. Strohmer, and V. Voroninski, “PhaseLift: Exact and stable signal recovery from magnitude measurements via convex programming,”Commun. Pure Appl. Math., vol. 66, no. 8, pp. 1241–1274, 2013

    work page 2013

  3. [3]

    Concise complexity analyses for trust region meth- ods,

    F. E. Curtis, Z. Lubberts, and D. P. Robinson, “Concise complexity analyses for trust region meth- ods,”Optim. Lett., vol. 12, pp. 1713–1724, 2018

    work page 2018

  4. [4]

    First-order methods almost always avoid strict saddle points,

    J. D. Lee, I. Panageas, G. Piliouras, M. Simchowitz, M. I. Jordan, and B. Recht, “First-order methods almost always avoid strict saddle points,”Math. Program., vol. 176, pp. 311–337, 2019

    work page 2019

  5. [5]

    Phase retrieval with transverse translation diversity: a non- linear optimization approach,

    M. Guizar-Sicairos and J. R. Fienup, “Phase retrieval with transverse translation diversity: a non- linear optimization approach,”Opt. Express, vol. 16, no. 10, 2008

    work page 2008

  6. [6]

    Phase retrieval via Wirtinger flow: Theory and algo- rithms,

    E. J. Cand` es, X. Li, and M. Soltanolkotabi, “Phase retrieval via Wirtinger flow: Theory and algo- rithms,”IEEE Trans. Inf. Theory, vol. 61, no. 4, pp. 1985–2007, 2015

    work page 1985

  7. [7]

    ROP: Matrix recovery via rank-one projections,

    T. T. Cai and A. Zhang, “ROP: Matrix recovery via rank-one projections,”Ann. Stat., vol. 43, no. 1, pp. 102–138, 2015

    work page 2015

  8. [8]

    A geometric analysis of phase retrieval,

    J. Sun, Q. Qu, and J. Wright, “A geometric analysis of phase retrieval,”Found. Comput. Math., vol. 18, no. 5, pp. 1131–1198, 2018

    work page 2018

  9. [9]

    Nearly optimal bounds for the global geometric landscape of phase retrieval,

    J.-F. Cai, M. Huang, D. Li, and Y. Wang, “Nearly optimal bounds for the global geometric landscape of phase retrieval,”Inverse Probl., vol. 39, no. 7, 2023

    work page 2023

  10. [10]

    Solving quadratic equations via PhaseLift when there are about as many equations as unknowns,

    E. J. Cand` es and X. Li, “Solving quadratic equations via PhaseLift when there are about as many equations as unknowns,”Found. Comput. Math., vol. 14, no. 5, pp. 1017–1026, 2013. 27

    work page 2013

  11. [11]

    Phase retrieval: From computational imaging to machine learning: A tutorial,

    J. Dong, L. Valzania, A. Maillard, T.-a. Pham, S. Gigan, and M. Unser, “Phase retrieval: From computational imaging to machine learning: A tutorial,”IEEE Signal Process. Mag., vol. 40, no. 1, pp. 45–57, 2023

    work page 2023

  12. [12]

    The numerics of phase retrieval,

    A. Fannjiang and T. Strohmer, “The numerics of phase retrieval,”Acta Numer., vol. 29, pp. 125–228, 2020

    work page 2020

  13. [13]

    Complex dynamics in simple neural networks: Understanding gradient flow in phase retrieval,

    S. Sarao Mannelli, G. Biroli, C. Cammarota, F. Krzakala, P. Urbani, and L. Zdeborov´ a, “Complex dynamics in simple neural networks: Understanding gradient flow in phase retrieval,” inProc. Conf. Neural Inf. Process. Syst. (NeurIPS), vol. 33, Virtual conference, Dec. 2020, pp. 3265–3274

    work page 2020

  14. [14]

    Local and global linear convergence of general low-rank matrix recovery problems,

    Y. Bi, H. Zhang, and J. Lavaei, “Local and global linear convergence of general low-rank matrix recovery problems,” inProc. AAAI Conf. Artif. Intell. (AAAI), vol. 36, no. 9, Virtual conference, Feb. 2022, pp. 10 129–10 137

    work page 2022

  15. [15]

    Geometric analysis of noisy low-rank matrix recovery in the exact parametrized and the overparametrized regimes,

    Z. Ma, Y. Bi, J. Lavaei, and S. Sojoudi, “Geometric analysis of noisy low-rank matrix recovery in the exact parametrized and the overparametrized regimes,”INFORMS J. Opt., vol. 5, no. 4, pp. 356–375, 2023

    work page 2023

  16. [16]

    Improved global guarantees for the nonconvex Burer-Monteiro factorization via rank overparameterization,

    R. Y. Zhang, “Improved global guarantees for the nonconvex Burer-Monteiro factorization via rank overparameterization,”Math. Program., vol. 213, pp. 1009–1038, 2025

    work page 2025

  17. [17]

    Sharp global guarantees for nonconvex low-rank recovery in the noisy overparameterized regime,

    ——, “Sharp global guarantees for nonconvex low-rank recovery in the noisy overparameterized regime,”SIAM J. Optim., vol. 35, no. 3, pp. 2128–2154, 2025

    work page 2025

  18. [18]

    Manopt, a Matlab toolbox for optimization on manifolds,

    N. Boumal, B. Mishra, P.-A. Absil, and R. Sepulchre, “Manopt, a Matlab toolbox for optimization on manifolds,”J. Mach. Learn. Res., vol. 15, no. 42, pp. 1455–1459, 2014

    work page 2014

  19. [19]

    Sensor network localization has a benign landscape after low-dimensional relaxation,

    C. Criscitiello, A. D. McRae, Q. Rebjock, and N. Boumal, “Sensor network localization has a benign landscape after low-dimensional relaxation,” 2025

    work page 2025

  20. [20]

    Complex phase retrieval from subgaussian measurements,

    F. Krahmer and D. St¨ oger, “Complex phase retrieval from subgaussian measurements,”J. Fourier Anal. Appl., vol. 26, no. 89, 2020

    work page 2020

  21. [21]

    Vershynin,High-Dimensional Probability

    R. Vershynin,High-Dimensional Probability. Cambridge University Press, 2018

    work page 2018

  22. [22]

    Phase retrieval without small-ball probability assumptions,

    F. Krahmer and Y.-K. Liu, “Phase retrieval without small-ball probability assumptions,”IEEE Trans. Inf. Theory, vol. 64, no. 1, pp. 485–500, 2018

    work page 2018

  23. [23]

    Stable optimizationless recovery from phaseless linear measurements,

    L. Demanet and P. Hand, “Stable optimizationless recovery from phaseless linear measurements,” J. Fourier Anal. Appl., vol. 20, pp. 199–221, 2014

    work page 2014

  24. [24]

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

    E. J. Cand` es and Y. Plan, “Tight oracle inequalities for low-rank matrix recovery from a minimal number of noisy random measurements,”IEEE Trans. Inf. Theory, vol. 57, no. 4, pp. 2342–2359, 2011

    work page 2011

  25. [25]

    Estimation of (near) low-rank matrices with noise and high- dimensional scaling,

    S. Negahban and M. J. Wainwright, “Estimation of (near) low-rank matrices with noise and high- dimensional scaling,”Ann. Stat., vol. 39, no. 2, pp. 1069–1097, 2011

    work page 2011

  26. [26]

    Estimation of high-dimensional low-rank matrices,

    A. Rohde and A. B. Tsybakov, “Estimation of high-dimensional low-rank matrices,”Ann. Stat., vol. 39, no. 2, pp. 887–930, 2011

    work page 2011

  27. [27]

    Nonconvex matrix factorization from rank-one measurements,

    Y. Li, C. Ma, Y. Chen, and Y. Chi, “Nonconvex matrix factorization from rank-one measurements,” IEEE Trans. Inf. Theory, vol. 67, no. 3, pp. 1928–1950, 2021

    work page 1928

  28. [28]

    Solving systems of random quadratic equations via truncated amplitude flow,

    G. Wang, G. B. Giannakis, and Y. C. Eldar, “Solving systems of random quadratic equations via truncated amplitude flow,”IEEE Trans. Inf. Theory, vol. 64, no. 2, pp. 773–794, 2018

    work page 2018

  29. [29]

    Robust phase retrieval by alternating minimization,

    S. Kim and K. Lee, “Robust phase retrieval by alternating minimization,”IEEE Trans. Signal Process., vol. 73, pp. 40–54, 2025

    work page 2025

  30. [30]

    Preconditioned gradient descent for overparameterized nonconvex Burer–Monteiro factorization with global optimality certification,

    G. Zhang, S. Fattahi, and R. Y. Zhang, “Preconditioned gradient descent for overparameterized nonconvex Burer–Monteiro factorization with global optimality certification,”J. Mach. Learn. Res., vol. 24, no. 163, pp. 1–55, 2023. 28

    work page 2023

  31. [31]

    Phase retrieval from coded diffraction patterns,

    E. J. Cand` es, X. Li, and M. Soltanolkotabi, “Phase retrieval from coded diffraction patterns,”Appl. Comput. Harmon. Anal., vol. 39, no. 2, pp. 277–299, 2015

    work page 2015

  32. [32]

    Improved recovery guarantees for phase retrieval from coded diffraction patterns,

    D. Gross, F. Krahmer, and R. Kueng, “Improved recovery guarantees for phase retrieval from coded diffraction patterns,”Appl. Comput. Harmon. Anal., vol. 42, no. 1, pp. 37–64, 2017

    work page 2017

  33. [33]

    Sampling complexity on phase retrieval from masked Fourier measurements via Wirtinger flow,

    H. Li, S. Li, and Y. Xia, “Sampling complexity on phase retrieval from masked Fourier measurements via Wirtinger flow,”Inverse Probl., vol. 38, no. 10, 2022

    work page 2022

  34. [34]

    Truncated amplitude flow with coded diffraction patterns,

    H. Li and J. Li, “Truncated amplitude flow with coded diffraction patterns,”Inverse Probl., vol. 41, no. 1, 2025

    work page 2025

  35. [35]

    Structured random model for fast and robust phase retrieval,

    Z. Hu, J. Tachella, M. Unser, and J. Dong, “Structured random model for fast and robust phase retrieval,” inProc. IEEE Int. Conf. on Acoustics, Speech, and Signal Processing (ICASSP), Hyder- abad, India, Apr. 2025

    work page 2025

  36. [36]

    Phase retrieval with application to optical imaging: A contemporary overview,

    Y. Shechtman, Y. C. Eldar, O. Cohen, H. N. Chapman, J. Miao, and M. Segev, “Phase retrieval with application to optical imaging: A contemporary overview,”IEEE Signal Process. Mag., vol. 32, no. 3, pp. 87–109, 2015

    work page 2015

  37. [37]

    Noisy phase retrieval from subgaussian measurements,

    H. Peng, D. Han, L. Li, and M. Huang, “Noisy phase retrieval from subgaussian measurements,” 2024

    work page 2024

  38. [38]

    Phase retrieval for sub-Gaussian measurements,

    B. Gao, H. Liu, and Y. Wang, “Phase retrieval for sub-Gaussian measurements,”Appl. Comput. Harmon. Anal., vol. 53, pp. 95–115, 2021

    work page 2021

  39. [39]

    Nonconvex optimization meets low-rank matrix factorization: An overview,

    Y. Chi, Y. M. Lu, and Y. Chen, “Nonconvex optimization meets low-rank matrix factorization: An overview,”IEEE Trans. Signal Process., vol. 67, no. 20, pp. 5239–5269, 2019

    work page 2019

  40. [40]

    No spurious local minima in nonconvex low rank problems: A unified geometric analysis,

    R. Ge, C. Jin, and Y. Zheng, “No spurious local minima in nonconvex low rank problems: A unified geometric analysis,” inProc. Int. Conf. Mach. Learn. (ICML), Sydney, Australia, Aug. 2017, pp. 1233–1242

    work page 2017

  41. [41]

    The local landscape of phase retrieval under limited samples,

    K. Liu, Z. Wang, and L. Wu, “The local landscape of phase retrieval under limited samples,”IEEE Trans. Inf. Theory, vol. 70, no. 12, pp. 9012–9035, 2024

    work page 2024

  42. [42]

    The nonsmooth landscape of phase retrieval,

    D. Davis, D. Drusvyatskiy, and C. Paquette, “The nonsmooth landscape of phase retrieval,”IMA J. Numer. Anal., vol. 40, no. 4, pp. 2652–2695, 2020

    work page 2020

  43. [43]

    Toward the optimal construction of a loss function without spurious local minima for solving quadratic equations,

    Z. Li, J.-F. Cai, and K. Wei, “Toward the optimal construction of a loss function without spurious local minima for solving quadratic equations,”IEEE Trans. Inf. Theory, vol. 66, no. 5, pp. 3242– 3260, 2020

    work page 2020

  44. [44]

    Solving phase retrieval with random initial guess is nearly as good as by spectral initialization,

    J.-F. Cai, M. Huang, D. Li, and Y. Wang, “Solving phase retrieval with random initial guess is nearly as good as by spectral initialization,”Appl. Comput. Harmon. Anal., vol. 58, pp. 60–84, 2022

    work page 2022

  45. [45]

    The global landscape of phase retrieval I: Perturbed amplitude models,

    ——, “The global landscape of phase retrieval I: Perturbed amplitude models,”Ann. Appl. Math., vol. 37, no. 4, pp. 437–512, 2021

    work page 2021

  46. [46]

    The global landscape of phase retrieval II: Quotient intensity models,

    ——, “The global landscape of phase retrieval II: Quotient intensity models,”Ann. Appl. Math., vol. 38, no. 1, pp. 62–114, 2022

    work page 2022

  47. [47]

    Phase retrieval: Stability and recovery guarantees,

    Y. C. Eldar and S. Mendelson, “Phase retrieval: Stability and recovery guarantees,”Appl. Comput. Harmon. Anal., vol. 36, pp. 473–494, 2014

    work page 2014

  48. [48]

    Generalized phase retrieval: Measurement number, matrix recovery and beyond,

    Y. Wang and Z. Xu, “Generalized phase retrieval: Measurement number, matrix recovery and beyond,”Appl. Comput. Harmon. Anal., vol. 47, no. 2, pp. 423–446, 2019

    work page 2019

  49. [49]

    Kaczmarz method for solving quadratic equations,

    Y. Chi and Y. M. Lu, “Kaczmarz method for solving quadratic equations,”IEEE Signal Processing Letters, vol. 23, no. 9, pp. 1183–1187, 2016

    work page 2016

  50. [50]

    Phase retrieval via randomized Kaczmarz: theoretical guarantees,

    Y. S. Tan and R. Vershynin, “Phase retrieval via randomized Kaczmarz: theoretical guarantees,” Inform. Inference., vol. 8, pp. 97–123, 2019. 29

    work page 2019

  51. [51]

    Exact and stable covariance estimation from quadratic sampling via convex programming,

    Y. Chen, Y. Chi, and A. J. Goldsmith, “Exact and stable covariance estimation from quadratic sampling via convex programming,”IEEE Trans. Inf. Theory, vol. 61, no. 7, pp. 4034–4059, 2015

    work page 2015

  52. [52]

    Low rank matrix recovery from rank one measurements,

    R. Kueng, H. Rauhut, and U. Terstiege, “Low rank matrix recovery from rank one measurements,” Appl. Comput. Harmon. Anal., vol. 42, no. 1, pp. 88–116, 2017

    work page 2017

  53. [53]

    Lipschitz analysis of generalized phase retrievable matrix frames,

    R. Balan and C. B. Dock, “Lipschitz analysis of generalized phase retrievable matrix frames,”SIAM J. Matrix Anal. Appl., vol. 43, no. 3, pp. 1518–1571, 2022

    work page 2022

  54. [54]

    Deterministic guarantees for Burer-Monteiro factor- izations of smooth semidefinite programs,

    N. Boumal, V. Voroninski, and A. S. Bandeira, “Deterministic guarantees for Burer-Monteiro factor- izations of smooth semidefinite programs,”Commun. Pure Appl. Math., vol. 73, no. 3, pp. 581–608, 2019

    work page 2019

  55. [55]

    The Burer-Monteiro SDP method can fail even above the Barvinok-Pataki bound,

    L. O’Carroll, V. Srinivas, and A. Vijayaraghavan, “The Burer-Monteiro SDP method can fail even above the Barvinok-Pataki bound,” inProc. Conf. Neural Inf. Process. Syst. (NeurIPS), New Or- leans, Louisiana, Dec. 2022, pp. 31 254–31 264

    work page 2022

  56. [56]

    Expander graphs are globally synchronizing,

    P. Abdalla, A. S. Bandeira, M. Kassabov, V. Souza, S. H. Strogatz, and A. Townsend, “Expander graphs are globally synchronizing,”Adv. Math., vol. 488, no. 110773, 2026

    work page 2026

  57. [57]

    Local geometry determines global landscape in low-rank factorization for synchronization,

    S. Ling, “Local geometry determines global landscape in low-rank factorization for synchronization,” Found. Comput. Math., 2025

    work page 2025

  58. [58]

    Benign landscape for Burer-Monteiro factorizations of MaxCut-type semidefinite programs,

    F. Rakoto Endor and I. Waldspurger, “Benign landscape for Burer-Monteiro factorizations of MaxCut-type semidefinite programs,” 2024

    work page 2024

  59. [59]

    Benign landscapes for synchronization on spheres via normalized Laplacian matri- ces,

    A. D. McRae, “Benign landscapes for synchronization on spheres via normalized Laplacian matri- ces,” 2025

    work page 2025

  60. [60]

    Optimal convex lifted sparse phase retrieval and PCA with an atomic matrix norm regularizer,

    A. D. McRae, J. Romberg, and M. A. Davenport, “Optimal convex lifted sparse phase retrieval and PCA with an atomic matrix norm regularizer,”IEEE Trans. Inf. Theory, vol. 69, no. 3, pp. 1866–1882, 2023

    work page 2023

  61. [61]

    M. J. Wainwright,High-Dimensional Statistics: A Non-Asymptotic Viewpoint. Cambridge Uni- versity Press, 2019

    work page 2019

  62. [62]

    Boucheron, G

    S. Boucheron, G. Lugosi, and P. Massart,Concentration Inequalities. Oxford University Press, 2013

    work page 2013

  63. [63]

    Convex recovery of a structured signal from independent random linear measure- ments,

    J. A. Tropp, “Convex recovery of a structured signal from independent random linear measure- ments,” inSampling Theory, a Renaissance, G. E. Pfander, Ed. Springer, 2015, pp. 67–101

    work page 2015

  64. [64]

    Koltchinskii,Oracle Inequalities in Empirical Risk Minimization and Sparse Recovery Problems

    V. Koltchinskii,Oracle Inequalities in Empirical Risk Minimization and Sparse Recovery Problems. Springer, 2011. 30

    work page 2011