Pith. sign in

REVIEW 4 minor 44 references

The minimum number of intensity measurements for phase retrieval in C4 is exactly eleven, and pure-state tomography in C4 needs exactly four orthonormal bases.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review · deepseek-v4-flash

2026-08-01 02:40 UTC pith:34GFKUQU

load-bearing objection A novel and sound topological proof that N_min(C^4)=11; the only blemish is a two-line fix in Corollary 1.4.

arxiv 2607.27719 v1 pith:34GFKUQU submitted 2026-07-30 quant-ph cs.ITmath.DGmath.IT

Phase Retrieval in mathbb C⁴ Requires Exactly Eleven Measurements

Meng Huang This is my paper
classification quant-ph cs.ITmath.DGmath.IT MSC 42C1557R20
keywords phase retrievalC4minimal measurementscharacteristic classesPontryagin classesnormal bundlepure state tomographyorthonormal bases
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

The paper settles the long-open question of the minimal number of intensity measurements needed to recover a vector in C4 up to global phase. It proves that no ten measurement vectors can have the phase retrieval property, so ten measurements are impossible. Together with a known explicit eleven-vector construction, this fixes the exact minimum at eleven. The same result implies that three orthonormal bases cannot distinguish all pure quantum states in C4, so exactly four bases are necessary and sufficient for pure-state tomography. The proof is topological: a hypothetical ten-vector frame would create an embedding of CP3 into R9 whose normal bundle obeys incompatible constraints on its Pontryagin class.

Core claim

The central claim is that no family of ten vectors in C4 possesses the phase retrieval property. Assuming such a frame, normalization produces a Parseval frame whose projective intensity map smoothly embeds CP3 into R9. Each nonzero measurement vector has a zero locus isomorphic to CP2; the rank-one structure of that intensity coordinate supplies a nowhere-zero section of the normal bundle over this hyperplane, forcing a splitting of the rank-three normal bundle into a trivial line bundle and an oriented rank-two bundle. The first Pontryagin class of the restricted normal bundle is fixed by the stable tangent-normal identity to be -4h̄², while the splitting forces it to be k²h̄² for some int

What carries the argument

The carrying mechanism is the normal-bundle splitting induced by a rank-one intensity coordinate. For an embedding G: CP3→R9 coming from a ten-vector Parseval frame, the zero locus Hj={bⱼ*x=0} is a CP2 inside CP3, and the gradient of the corresponding intensity coordinate gives a global nowhere-zero section of the normal bundle over Hj. This splits the restricted rank-three normal bundle as R⊕η, where η is an oriented real rank-two bundle over CP2. The contradiction comes from comparing two evaluations of the first Pontryagin class: the stable tangent-normal relation gives p1(ν|Hj)=-4h̄², whereas for oriented rank-two bundles p1(η)=e(η)²=k²h̄², forcing k²=-4.

Load-bearing premise

The load-bearing premise is the standard characteristic-class identity that for an oriented real rank-two bundle over CP2 the first Pontryagin class equals the square of the Euler class; if an orientation or torsion subtlety made that identity fail here, the contradiction k²=-4 would not follow.

What would settle it

Produce an explicit set of ten vectors in C4 whose intensity map |aⱼ*x|² is injective on C4 modulo global phase; equivalently, exhibit a ten-vector Parseval frame whose measurement map has no nonzero Hermitian rank-≤2 matrix in its kernel. Either construction would immediately disprove the theorem.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • The phase retrieval problem in C4 is fully resolved: the exact minimum is 11 measurements, not 10.
  • Any rank-one POVM on C4 needs exactly 11 outcomes to be informationally complete for pure states.
  • Three orthonormal bases cannot distinguish all pure states in C4; four bases are necessary and sufficient.
  • The generic lower bound of 10 for C4 is not achievable, so the known 11-vector frame is genuinely optimal.
  • Since 11=4M-5 with M=4, the result confirms that the conjectured generic bound 4M-4 is not the true minimum in this exceptional dimension.

Where Pith is reading between the lines

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

  • The same normal-bundle strategy may yield stronger lower bounds in other even dimensions M≥6, where the exact minimum is still open, if analogous zero-loci splittings can be built.
  • The obstruction separates rank-one from general measurements: ten-outcome non-rank-one POVMs on C4 were already known to distinguish pure states, so the paper shows the ten-vector failure is specifically a rank-one phenomenon.
  • A testable extension is to ask whether every embedding CP^{M-1}→R^{4M-5} forces a similar Pontryagin-class contradiction; if it does not, the C4 result may be a low-dimensional special case rather than a general pattern.
  • The claim is directly falsifiable by construction: finding any ten-vector frame in C4 with an injective intensity map would overturn the theorem.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

0 major / 4 minor

Summary. The paper settles the long-standing question of the minimal number of intensity measurements for phase retrieval in C^4. Section 3 first reduces any hypothetical 10-vector phase-retrieval frame to a Parseval frame (Lemma 3.1) and uses the standard rank-≤2 kernel criterion (Lemma 3.2). For a Parseval frame, the projective intensity map defines a smooth map F: CP^3 → Y, and after centering, a map G: CP^3 → E_0 ≅ R^9; Proposition 3.4 proves G is a smooth embedding. For each nonzero b_j, the hyperplane H_j = {b_j^* x = 0} ≅ CP^2 is shown to have the constant vector γ_j = e_j − (1/10)1 as a nowhere-zero normal section (Proposition 3.5), so ν|_{H_j} ≅ R ⊕ η_j. Characteristic-class arguments give p_1(TCP^3) = 4h^2, hence p_1(ν) = −4h^2 and p_1(ν|_{H_j}) = −4\bar h^2; the splitting forces p_1(η_j) = e(η_j)^2 = k^2\bar h^2. The resulting equation k^2 = −4 is impossible, proving Theorem 1.2. Corollary 1.3 then gives N_min(C^4) = 11 via Vinzant's construction, and Corollary 1.4 shows exactly four orthonormal bases are needed.

Significance. This is a strong and definitive result. It resolves the 10-versus-11 gap in phase retrieval and, as a byproduct, the number of orthonormal bases needed for pure-state tomography in C^4. The proof is internally consistent and self-contained; the lower-bound argument is parameter-free and does not rely on numerical searches. The topological method—using a normal-bundle section over a projective hyperplane and a Pontryagin-class contradiction—is novel and likely to be influential for higher-dimensional cases. The external inputs are clearly delineated: Vinzant's 11-vector frame and the four-basis sufficiency of [13] serve only for upper bounds. The critical identity p_1(η) = e(η)^2 for oriented rank-2 bundles is correctly cited to Milnor–Stasheff and is applied in a torsion-free integral cohomology group; the stress-test concern about orientation or torsion does not materialize.

minor comments (4)
  1. [Corollary 1.4] The proof omits one explicit step: after the norm equality is obtained from the first basis, the missing fourth-basis intensities for ℓ = 2, 3 are determined by |⟨u_{ℓ,4}, x⟩|^2 = ∥x∥^2 − Σ_{j=1}^3 |⟨u_{ℓ,j}, x⟩|^2, and similarly for y. Since equality holds on the ten listed vectors and the norms agree, equality also holds on the two omitted basis vectors; this justifies the assertion that the normalized states agree on all three orthonormal bases. Please add this sentence.
  2. [Proposition 3.5] The notation (b_j^*x)^* in the displayed derivative is equal to x^*b_j; the vanishing is clearer if written as 2Re(x^*b_j b_j^*z). Also, 'because b_j acts as a nonzero complex linear functional' should read 'because b_j^* acts as a nonzero complex linear functional.'
  3. [Lemma 3.2] In the necessity direction, the statement that a nonzero rank-≤2, trace-zero Hermitian matrix has exactly two nonzero eigenvalues could use a one-sentence justification that rank-one matrices with trace zero are zero.
  4. [Abstract/Title] Minor typographical issues: 'Phase Retrieval inC4' lacks a space, and 'resolving the problem left in [13]' should read 'resolving the problem left open in [13]'.

Circularity Check

0 steps flagged

No significant circularity: the lower-bound proof is self-contained and external results are used only as upper bounds or standard background.

full rationale

The derivation chain of Theorem 1.2 is self-contained. Lemma 3.1 normalizes an arbitrary phase-retrieval frame to a Parseval frame via the frame operator, an invertible transformation that preserves injectivity modulo global phase; this is a reduction, not a fitted input. Lemma 3.2 (rank-kernel criterion) is restated from Bandeira–Cahill–Mixon–Nelson and proved in the paper, and it is an independent algebraic fact. Proposition 3.4 constructs the embedding CP^3 -> R^9 from injectivity and the immersion argument; the only input is the assumed phase-retrieval property. Proposition 3.5 uses the zero locus of a rank-one coordinate to produce a nowhere-zero normal section; the calculation that gamma_j is orthogonal to dG[T CP^3] follows from the Parseval identity and is not assumed. The topological contradiction equates p1(nu|Hj) = -4 hbar^2 (computed from Euler sequence and stable normal relation) with k^2 hbar^2 from the splitting plus the standard Milnor–Stasheff identity p1(eta)=e(eta)^2; the impossibility k^2=-4 is arithmetic. Vinzant's 11-vector frame is used only for the upper bound, and the four-bases sufficiency is cited from Carmeli et al. as an external construction; neither is used to force the lower bound. Corollary 1.4's three-bases-to-ten-vectors argument is a valid reduction: the missing fourth intensities are determined by normalization and the included intensities, so no hidden input is imported. The self-citations in the bibliography concern algorithmic convergence and are contextual, not load-bearing. No fitted parameter is renamed as a prediction, and no uniqueness theorem by the author is invoked. Therefore no circularity is present.

Axiom & Free-Parameter Ledger

0 free parameters · 6 axioms · 0 invented entities

The lower-bound proof introduces no free parameters and no new entities. It relies on standard characteristic-class identities and on cited external constructions for the matching upper bounds.

axioms (6)
  • standard math Whitney sum formula for Pontryagin classes is valid over Z when the relevant cohomology is torsion-free.
    Used in Section 3.3 to obtain p1(R^9) = p1(TCP^3) + p1(ν), hence p1(ν) = −4h^2.
  • standard math For an oriented real rank-2 vector bundle η, p1(η) = e(η)^2 in integral cohomology.
    Load-bearing identity converting the splitting ν|Hj ≅ R ⊕ η into k^2 hbar^2; cited to Milnor-Stasheff Ch. 15.
  • standard math The Euler sequence for CP^3: 0 → O → O(1)^{⊕4} → T^{1,0}CP^3 → 0, giving c(TCP^3) = (1+h)^4.
    Used to compute p1(TCP^3) = 4h^2 in Section 3.3.
  • standard math A smooth injective immersion from a compact manifold into a Euclidean space is a smooth embedding.
    Used in Proposition 3.4 to obtain the embedding and a well-defined normal bundle.
  • domain assumption Vinzant's 11-vector frame in C^4 has the phase retrieval property.
    External upper-bound result from [39] used in Corollary 1.3; not used in the lower-bound proof.
  • domain assumption Four orthonormal bases suffice to distinguish all pure states in any finite dimension.
    External sufficiency result from [13] used in Corollary 1.4.

pith-pipeline@v1.3.0-daily-deepseek · 12841 in / 24440 out tokens · 237416 ms · 2026-08-01T02:40:56.669642+00:00 · methodology

0 comments
read the original abstract

Determining the minimal number of intensity measurements required for phase retrieval in $\mathbb{C}^4$ has been a long-standing open problem. Prior to this work, the best-known results implied that this minimum was either $10$ or $11$. In this paper, we leverage characteristic classes and cohomology groups from differential topology to prove that no family of $10$ vectors in $\mathbb{C}^4$ possesses the phase retrieval property. Combining our lower bound with Vinzant's explicit eleven-vector construction establishes that the exact minimum is $11$. Our result yields a significant consequence for pure state quantum tomography, namely, a rank-one POVM on $\mathbb{C}^4$ requires exactly $11$ elements to be informationally complete for pure states. This further implies that three orthonormal bases are insufficient to uniquely distinguish all pure states in $\mathbb{C}^4$. Because four orthonormal bases are already known to be sufficient, we conclude that exactly four bases are required, thereby completely resolving the problem left in [C. Carmeli, T. Heinosaari, J. Schultz, A. Toigo, Eur. Phys. J. D].

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

44 extracted references · 1 linked inside Pith

  1. [1]

    Balan, P

    R. Balan, P. G. Casazza, D. Edidin. On signal reconstruction without phase.Appl. Comput. Harmon. Anal., 2006, 20(3): 345–356

    2006

  2. [2]

    A. S. Bandeira, J. Cahill, D. G. Mixon, A. A. Nelson. Saving phase: Injectivity and stability for phase retrieval.Appl. Comput. Harmon. Anal., 2014, 37(1): 106–125

    2014

  3. [3]

    B. G. Bodmann, N. Hammen. Stable phase retrieval with low-redundancy frames.Adv. Comput. Math., 2015, 41(2): 317–331

    2015

  4. [4]

    B. G. Bodmann, N. Hammen. Algorithms and error bounds for noisy phase retrieval with low-redundancy frames.Appl. Comput. Harmon. Anal., 2017, 43(3): 482–503

    2017

  5. [5]

    Botelho-Andrade, P

    S. Botelho-Andrade, P. G. Casazza, H. Van Nguyen, J. C. Tremain. Phase retrieval versus phaseless reconstruction.J. Math. Anal. Appl., 2016, 436(1): 131–137

    2016

  6. [6]

    T. Cai, A. Zhang. ROP: Matrix recovery via rank-one projections.Ann. Stat., 2015, 43(1): 102–138

    2015

  7. [7]

    T. T. Cai, X. Li, Z. Ma. Optimal rates of convergence for noisy sparse phase retrieval via thresholded Wirtinger flow.Ann. Stat., 2016, 44(5): 2221–2251

    2016

  8. [8]

    J. F. Cai, M. Huang, D. Li, Y. Wang. Solving phase retrieval with random initial guess is nearly as good as by spectral initialization.Appl. Comput. Harmon. Anal., 2022, 58: 60–84. 14

    2022

  9. [9]

    E. J. Candès, X. Li. Solving quadratic equations via PhaseLift when there are about as many equations as unknowns.Found. Comput. Math., 2014, 14(5): 1017–1026

    2014

  10. [10]

    E. J. Candès, X. Li, M. Soltanolkotabi. Phase retrieval via Wirtinger flow: Theory and algorithms.IEEE Trans. Inf. Theory, 2015, 61(4): 1985–2007

    2015

  11. [11]

    E. J. Candès, T. Strohmer, V. Voroninski. Phaselift: Exact and stable signal recovery from magnitude measurements via convex programming.Commun. Pure Appl. Math., 2013, 66(8): 1241–1274

    2013

  12. [12]

    Carmeli, T

    C. Carmeli, T. Heinosaari, M. Kech, J. Schultz, A. Toigo. Stable pure state quantum tomography from five orthonormal basesEurophysics Letters, 2016, 115(3): 30001

    2016

  13. [13]

    Carmeli, T

    C. Carmeli, T. Heinosaari, J. Schultz, A. Toigo. How many orthonormal bases are needed to distinguish all pure quantum states?Eur. Phys. J. D, 2015, 69(7): 1–11

    2015

  14. [14]

    A. Chai, M. Moscoso, G. Papanicolaou. Array imaging using intensity-only measurements. Inverse Probl., 2011, 27(1): 015005

    2011

  15. [15]

    Y. Chen, Y. Chi, J. Fan, C. Ma. Gradient descent with random initialization: Fast global convergence for nonconvex phase retrieval.Math. Program., 2019, 176: 5–37

    2019

  16. [16]

    Conca, D

    A. Conca, D. Edidin, M. Hering, C. Vinzant. An algebraic characterization of injectivity in phase retrieval.Appl. Comput. Harmon. Anal., 2015, 38(2): 346–356

    2015

  17. [17]

    J. C. Dainty, J. R. Fienup. Phase retrieval and image reconstruction for astronomy.Image Recovery: Theory and Appl., 1987, 231: 275

    1987

  18. [18]

    J. C. Duchi, F. Ruan. Solving (most) of a set of quadratic equalities: Composite optimization for robust phase retrieval.Inf. Inference, 2019, 8(3): 471–529

    2019

  19. [19]

    Finkelstein

    J. Finkelstein. Pure-state informationally complete and ‘really’ complete measurements.Phys. Rev. A, 2004, 70(5): 052107

    2004

  20. [20]

    S. T. Flammia, A. Silberfarb, C. M. Caves. Minimal informationally complete measurements for pure states.Found. Phys., 2005, 35(12): 1985–2006

    2005

  21. [21]

    Griffiths, J

    P. Griffiths, J. Harris. Principles of Algebraic Geometry.John Wiley and Sons, 2014

    2014

  22. [22]

    Goyeneche, G

    D. Goyeneche, G. Cañas, S. Etcheverry, E. S. Gómez, G. B. Xavier, G. Lima, A. Delgado. Five measurement bases determine pure quantum states on any dimension.Phys. Rev. Lett., 2015, 115(9): 090401

    2015

  23. [23]

    Grohs, S

    P. Grohs, S. Koppensteiner, M. Rathmair. Phase Retrieval: Uniqueness and Stability.SIAM Rev., 2020, 62(2): 301–350

    2020

  24. [24]

    R. W. Harrison. Phase problem in crystallography.J. Opt. Soc. Am. A, 1993, 10(5): 1046–1055

    1993

  25. [25]

    Heinosaari, L

    T. Heinosaari, L. Mazzarella, M. M. Wolf. Quantum tomography under prior information. Commun. Math. Phys., 2013, 318(2): 355–374

    2013

  26. [26]

    M. Huang. Near-Quadratic Convergence of the Gauss–Newton Method for Complex Phase Retrieval.IEEE Trans. Inf. Theory, 2026, 72(1): 222–245. 15

    2026

  27. [27]

    Huang, Y

    M. Huang, Y. Wang. Linear convergence of randomized Kaczmarz method for solving complex- valued phaseless equations.SIAM J. Imaging Sci., 2022, 15(2): 989–1016

    2022

  28. [28]

    J. M. Lee. Manifolds and Differential Geometry.Grad. Stud. Math., 2009, 107: American Mathematical Society

    2009

  29. [29]

    J. M. Lee. Introduction to Smooth Manifolds (2nd ed.).Graduate Texts in Mathematics, Springer, New York, 2013, 218

    2013

  30. [30]

    Z. Li. On injectivity of phase retrieval.arXiv preprint, 2026, arXiv:2606.17922

    Pith/arXiv arXiv 2026

  31. [31]

    J. Miao, T. Ishikawa, Q. Shen, T. Earnest. Extending X-ray crystallography to allow the imaging of noncrystalline materials, cells, and single protein complexes.Annu. Rev. Phys. Chem., 2008, 59: 387–410

    2008

  32. [32]

    R. J. Milgram. Immersing projective spaces.Ann. Math., 1967, 85(3): 473–482

    1967

  33. [33]

    R. P. Millane. Phase retrieval in crystallography and optics.J. Opt. Soc. Am. A, 1990, 7(3): 394–411

    1990

  34. [34]

    J. W. Milnor, J. D. Stasheff. Characteristic Classes.Ann. Math. Stud., 1974, 76: Princeton University Press

    1974

  35. [35]

    Shechtman, Y

    Y. Shechtman, Y. C. Eldar, O. Cohen, H. N. Chapman, J. Miao, M. Segev. Phase retrieval with application to optical imaging: a contemporary overview.IEEE Signal Process. Mag., 2015, 32(3): 87–109

    2015

  36. [36]

    L. L. Sun, S. Yu, Z.-B. Chen. Minimal determination of a pure qutrit state and four-measurement protocol for pure qudit state.J. Phys. A Math. Theor., 2020, 53, 075305

    2020

  37. [37]

    J. Sun, Q. Qu, J. Wright. A geometric analysis of phase retrieval.Found. Comput. Math., 2018, 18(5): 1131–1198

    2018

  38. [38]

    Y. S. Tan, R. Vershynin. Phase retrieval via randomized kaczmarz: theoretical guarantees.Inf. Inference, 2019, 8(1): 97–123

    2019

  39. [39]

    C. Vinzant. A small frame and a certificate of its injectivity.Proc. SampTA, 2015, 197–200

    2015

  40. [40]

    Waldspurger, A

    I. Waldspurger, A. d’Aspremont, S. Mallat. Phase recovery, maxcut and complex semidefinite programming.Math. Program., 2015, 149(1-2): 47–81

    2015

  41. [41]

    Waldspurger

    I. Waldspurger. Phase retrieval with random gaussian sensing vectors by alternating projections. IEEE Trans. Inf. Theory, 2018, 64(5): 3301–3312

    2018

  42. [42]

    A. Walther. The question of phase retrieval in optics.J. Mod. Opt., 1963, 10(1): 41–49

    1963

  43. [43]

    Y. Wang, Y. Shang. Pure state ‘really’ informationally complete with rank-1 POVM.Quantum Inf. Process., 2018, 17(3): 1–11

    2018

  44. [44]

    Y. Wang, Z. Xu. Generalized phase retrieval: measurement number, matrix recovery and beyond.Appl. Comput. Harmon. Anal., 2019, 47(2): 423–446. 16

    2019