REVIEW 3 major objections 6 minor 14 references
No finite NPA level is exact near the critical doubly-tilted CHSH line
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 →
T0 review · deepseek-v4-flash
2026-08-02 03:43 UTC pith:UVALQKQT
load-bearing objection Real supercritical exactness and an exact overshoot coefficient, but the phase-transition headline leans on a companion proof not included. the 3 major comments →
A phase transition in the exactness of the NPA hierarchy at the critical doubly-tilted CHSH functional
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
Core claim
For the symmetric doubly-tilted CHSH functional, the paper proves the quantum value has the expansion c_Q(s) = 4 − s + s^3/6 − s^4/36 + O(s^5), a cubic departure from the local bound, while every finite NPA level k has c_k(s) = 4 − s + a_k s^2 + O(s^3), a quadratic overshoot. Since the cubic term is eventually smaller than the quadratic one, any level with a_k > 0 is not exact for all sufficiently small s. The paper proves that the supercritical quadrant α,β ≥ 1 is exact at all levels via three explicit rational certificates realizing an affine identity, and it proves the exact value a_{1+AB} = 3/64, along with certified rational lower bounds for a_2, a_3, and a_4. Assuming the companion pap
What carries the argument
The central objects are the NPA moment matrix Γ(y) and three explicit rational dual certificates Z_0, Z_A, Z_B. Z_0 certifies the critical identity 4 − B_{1,1}·y = ⟨Z_0, Γ(y)⟩, while Z_A and Z_B certify 1 − ⟨A_0⟩ and 1 − ⟨B_0⟩; nonnegative combinations give exactness for all α,β ≥ 1. On the subcritical side, the machinery is the quadratic-versus-cubic gap: the quantum curve is cubically flat at the deterministic vertex, while each relaxation retains positive curvature 2a_k, so the convex NPA value rounds the cusp. A Puiseux expansion at the critical corner shows the limiting loss is a rescaling of the Motzkin polynomial, locating the failure in the restricted-certificate (bounded word length
Load-bearing premise
The phase-transition claim for all levels rests on the companion paper's proof that the overshoot coefficient a_k is positive for every finite level k; the present manuscript only certifies the first four levels.
What would settle it
Compute a_k for any finite NPA level k to high precision and find a_k ≤ 0, or exhibit a level-k certificate that is exact for a sequence of tilts approaching the critical line; either would disprove the phase transition.
If this is right
- In the supercritical quadrant α,β ≥ 1, the NPA hierarchy is exact at every level for the doubly-tilted CHSH functional, with explicit rational certificates.
- Near the critical line on the subcritical side, no single finite NPA level is exact (given the companion proof that a_k > 0 for all k); the required level diverges as s → 0+.
- The almost-quantum level overshoots with exact curvature 3/64, the first exact NPA overshoot coefficient computed in this scenario.
- The quantum value's cubic law c_Q = 4 − s + s^3/6 − s^4/36 + O(s^5) is derived from a corrected polynomial system and matches published numerics.
- The finite-level failure is a restricted-certificate phenomenon: the limiting obstruction is the Motzkin polynomial, and the commutative (abelianized) reduction stays bounded.
Where Pith is reading between the lines
- The phase-transition picture suggests that any Bell functional with a cubic (or higher) contact at a self-tested optimum will evade finite-level NPA exactness, while a nondegenerate quadratic contact may be the precise dividing line for finite convergence.
- The exact a_{1+AB} = 3/64 and the certified rational ladder for low levels provide benchmark instances on which to test proposed strengthenings of the NPA hierarchy or alternative certificate classes.
- If the companion proof's constants could be made effective, one would obtain quantitative lower bounds on the required NPA level as a function of s, which would be directly testable numerically.
- The Motzkin-polynomial scaling limit indicates that the obstruction is genuinely two-variable and noncommutative; a dual construction of positive pseudo-moment functionals without flat extensions might yield a self-contained proof of the phase transition.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. This paper analyzes the NPA hierarchy for the doubly-tilted CHSH functional B_αβ = CHSH + α⟨A0⟩ + β⟨B0⟩. On the symmetric slice s = 2−α−β it derives the cubic expansion of the quantum value, c_Q(s) = 4−s+s^3/6−s^4/36+O(s^5), introduces quadratic overshoot coefficients a_k for each NPA level via c_k(s)=4−s+a_k s^2+O(s^3), computes exactly a_{1+AB}=3/64, and certifies in exact arithmetic that levels 1+AB, 2, 3 and 4 are not exact at selected small tilts. It also proves, by explicit rational PSD certificates, that the hierarchy is exact at every level in the quadrant α,β≥1. The paper then claims a phase transition: exact everywhere in the supercritical quadrant, and exact at no finite level near the critical line in the subcritical region. The subcritical half is identified with Conjecture 2 (a_k>0 for all k), whose proof is delegated to a companion paper [14]. The Motzkin scaling limit and a contact-order dichotomy are offered as geometric explanations.
Significance. If the full claim were established, this would be a valuable, rare example of an exactly solved NPA phase transition, with explicit rational certificates and machine-checkable exact arithmetic. The supercritical theorem is self-contained and is a genuine contribution, as is the exact value a_{1+AB}=3/64 and the four certified low-level non-exactness instances. The Motzkin scaling limit and the contact-order discussion are insightful and suggestive. However, the central subcritical phase-transition claim is not proved in this manuscript; it rests entirely on the companion paper [14]. The abstract also overstates the logical relation between divergence of the required NPA level and positivity of (a_k). The paper is honest about its division of labor, but as a standalone submission its main advertised result is conditional.
major comments (3)
- [Abstract; Section 3, Remark 3] The central claim — that no finite NPA level is exact near the critical line in the subcritical region — is not established in this manuscript. The four exact certificates (levels 1+AB at s=1/10, 2 at 1/5, 3 at 1/20, 4 at 1/40) are isolated instances and do not imply the uniform statement. The uniform statement is exactly Conjecture 2 (a_k>0 for all k), and Remark 3 delegates its proof to the companion paper [14], including the 'no-repair arc-feasibility lemma' that converts pinned tangent-program objectives into bounds on a_k. The abstract nevertheless states the phase transition unconditionally. Since [14] is not part of the submitted manuscript, a referee cannot verify the load-bearing subcritical half. Either the proof must be included (or summarized with full certificates), or the title/abstract must present the subcritical phase transition as conditional on the companion paper.
- [Abstract; Remark 1] The claimed equivalence between divergence of the required exact NPA level D(s) and positivity of the sequence (a_k) is only one directional. The paper itself notes that a_k>0 implies non-exactness near 0. The converse does not follow: if a_k=0 but the next term in c_k(s)−(4−s) is positive (or the cubic coefficient exceeds that of c_Q), level k can still fail to be exact for small s, giving D(s)→∞. Thus 'equivalent' in the abstract is logically false. Please either prove the converse under an additional assumption or replace 'equivalent' by 'sufficient'.
- [Section 3, a_{1+AB} upper bound; Remark 4] The exact value a_{1+AB}=3/64 is a key advertised result, but the proof in the text is a sketch. The lower bound is described via rational points and Bonnans–Shapiro second-order conditions; the upper bound is said to be verified symbolically over Q(√3) on the 'entire solution set' of a linear system, with no explicit dual arc or matrices displayed. The ancillary scripts may establish the claim, but a journal reader should be able to check the proof from the manuscript. Please either include the explicit dual matrices and the linear system, or state clearly that this part is verified by the accompanying exact-arithmetic code.
minor comments (6)
- [Section 3, Remark 3] The text first says 'Conjecture 2 is now a theorem' and then continues to call it 'Conjecture 2' throughout. Please rename it consistently or explain why the conjectural framing is retained after the companion proof.
- [Section 4, Theorem 6] The statement 'for every NPA level k≥1+AB' should specify the indexing convention. In standard NPA notation, 1+AB is a specific level, but the paper uses k both as a level label and in the expansion c_k. Please clarify.
- [Section 3] The numerical values a_2≈0.0281 and a_3≈0.0096 are later said to be 'certified to 480 digits' for a_2. Please ensure the notation distinguishes the certified a_k from the tangent-program lower bounds v_2,v_3,v_4, and specify what exactly is certified to 480 digits.
- [Section 7, Proposition 11] The proposition states that the NPA hierarchy is 'exact at a finite level' if the Boundary Hessian Condition holds, but gives no bound or construction of the level. This is presumably a known theorem (Nie/Marshall), but the noncommutative-to-commutative reduction via Jordan's lemma should be made explicit, including the trigonometric-polynomial substitution, so that the reader can see exactly which finite level is guaranteed.
- [Remark 13] The claimed Puiseux expansion for the almost-quantum level is marked 'machine-verified, ancillary'. Please include the expansion itself in the text or give the ancillary filename, so that the claim is reproducible without downloading code.
- [References] Reference [14] is an unpublished companion paper. If it is under review elsewhere, please add a footnote stating its availability and status, since the central claim depends on it.
Circularity Check
No circular reduction; subcritical half is delegated to companion [14] which is described as exact-verified, but self-reliance warrants attention.
full rationale
The paper's main circularity-relevant feature is the split of the phase-transition claim: the supercritical quadrant exactness (Theorem 6) is proved entirely within this manuscript via explicit rational PSD certificates Z0, ZA, ZB and the affine identity (2+alpha+beta)-B_alpha,beta·y = <Z(alpha,beta), Gamma(y)>. The subcritical uniform claim ('exact at no level near the critical line') is not proved here; it is stated as Conjecture 2 and deferred to companion paper [14] via Remark 3: 'in the companion paper [14] we prove that for every finite level k there is an ε_k>0 with c_k(s)>c_Q(s) on (0,ε_k]'. This is a load-bearing self-citation, since [14] shares the author. However, it does not constitute a circular reduction by construction: no equation in this paper is fitted to the target result, no prediction is equivalent to an input by definition, and the paper's own exact low-level certificates (levels 1+AB, 2, 3, 4 at s = 1/10, 1/5, 1/20, 1/40) are independent computational evidence. Moreover, the manuscript describes [14] as verified 'in exact arithmetic' and 're-verified end to end against an independently written implementation', which under the present rules counts as independent support rather than circularity. The abstract's phrase 'equivalent to positivity of the single sequence (a_k)' is only partially supported in this paper (Remark 1 gives a_k>0 ⇒ non-exactness near 0, but not the converse without further argument), but this is an overstatement, not a circular derivation. Given the explicit conjectural framing retained in the narrative and the division of labor with an exact-verified companion, the appropriate finding is low: no significant circularity, with a score of 2 reflecting the residual self-citation weight in the central subcritical claim.
Axiom & Free-Parameter Ledger
free parameters (2)
- Pinned rational base point (β, δ2, δ3, ...) = (1/2, 1/4, 1/2, ...) =
1/2, 1/4, 1/2, ...
- Blend weights ε_p =
10^-54 at k=2; 10^-7 at k=3,4
axioms (6)
- standard math Standard NPA moment-matrix hierarchy: level-k value is a semidefinite program with moment matrix Γ_k(y) ⪰ 0 and y_1 = 1; sound and complete for the levels used.
- domain assumption Jordan's lemma reduction to qubit measurements for binary-input binary-output Bell functionals; the doubly-tilted CHSH quantum maximum is attained and self-tested.
- domain assumption The corrected sextic polynomial system of Gigena et al. (with the three documented errata) is the correct description of c_Q(α, β), with the largest real root equal to the quantum value.
- standard math Sturm's theorem for exact rational root-freeness checks and exact rational arithmetic for PSD matrix verification.
- standard math Nie's finite-convergence theorem and Marshall's boundary Hessian condition as stated in Section 7.
- ad hoc to paper The companion paper [14] proves Conjecture 2 (a_k > 0 for all k) and the no-repair arc-feasibility lemma.
read the original abstract
Gigena et al. [npj Quantum Inf. 11, 82 (2025)] proved the exact quantum maximum of the doubly-tilted CHSH functional $B_{\alpha\beta}=\alpha\langle A_0\rangle+\beta\langle B_0\rangle+\mathrm{CHSH}$ and observed that the NPA level required to reach it grows without evident bound toward the critical line $\alpha+\beta=2$. We quantify the mechanism on the symmetric slice $s=2-\alpha-\beta$: (i) the quantum value leaves the local bound cubically, $c_Q=4-s+s^3/6-s^4/36+O(s^5)$; (ii) each NPA level overshoots quadratically, $c_k(s)=4-s+a_k s^2+O(s^3)$, with the almost-quantum coefficient computed exactly, $a_{1+AB}=3/64$; (iii) the divergence of the required exact level is equivalent to positivity of the single sequence $(a_k)$ - proven for every $k$ in the companion paper. We prove the supercritical side completely: for all $\alpha,\beta\ge 1$ and every level the hierarchy is exact, via three explicit rational certificates realizing an affine identity. The hierarchy's exactness thus undergoes a phase transition at the critical line. On the subcritical side we certify the first four levels in exact arithmetic (rational pseudo-moments beating $c_Q$, confirmed by Sturm's theorem). We identify the exact mechanism: rescaled to the critical corner, the limiting obstruction is the Motzkin polynomial, the classical nonnegative-but-not-sum-of-squares form, so the finite-level failure sits in the restricted-certificate regime. The phase boundary has a precise geometric reading via Nie's finite-convergence theorem and Marshall's boundary Hessian condition: a self-tested optimum is finitely NPA-certifiable whenever its boundary Hessian is nondegenerate (contact order two), which holds for the single tilt and fails exactly at the doubly-tilted cubic touch. Three verified errata in the published polynomial system of Gigena et al. are documented.
Reference graph
Works this paper leans on
-
[1]
N. Gigena, E. Panwar, G. Scala, M. Ara´ ujo, M. Farkas, A. Chaturvedi, npj Quantum Inf. 11, 82 (2025), arXiv:2405.08743
Pith/arXiv arXiv 2025
-
[2]
Navascu´ es, S
M. Navascu´ es, S. Pironio, A. Ac ´ ın, New J. Phys.10, 073013 (2008)
2008
-
[3]
P.,A certified separation of the almost-quantum set from the quantum set in the minimal Bell scenario, companion note (2026)
A. P.,A certified separation of the almost-quantum set from the quantum set in the minimal Bell scenario, companion note (2026)
2026
- [4]
-
[5]
I. Klep, J. Levenson, S. McCullough,Fej´ er–Riesz factorization for positive noncommutative trigonometric polynomials, arXiv:2511.09267 (2025)
arXiv 2025
-
[6]
M. Fanizza, L. Kroell, A. Mehta, C. Paddock, D. Rochette, W. Slofstra, Y. Zhao,The NPA hierarchy does not always attain the commuting operator value, arXiv:2510.04943 (2025)
arXiv 2025
-
[7]
M. A. Dritschel,Factoring non-negative operator valued trigonometric polynomials in two variables, Math. Ann.391, 515 (2024); arXiv:1811.06005. 10
Pith/arXiv arXiv 2024
-
[8]
Scheiderer,Non-existence of degree bounds for weighted sums of squares representations, J
C. Scheiderer,Non-existence of degree bounds for weighted sums of squares representations, J. Complexity21, 823 (2005)
2005
-
[9]
Scheiderer,Sums of squares on real algebraic surfaces, Manuscripta Math.119, 395 (2006)
C. Scheiderer,Sums of squares on real algebraic surfaces, Manuscripta Math.119, 395 (2006)
2006
-
[10]
Nie,Optimality conditions and finite convergence of Lasserre’s hierarchy, Math
J. Nie,Optimality conditions and finite convergence of Lasserre’s hierarchy, Math. Program. 146, 97 (2014); arXiv:1206.0319
Pith/arXiv arXiv 2014
-
[11]
Marshall,Representations of non-negative polynomials, degree bounds and applications to optimization, Canad
M. Marshall,Representations of non-negative polynomials, degree bounds and applications to optimization, Canad. J. Math.61, 205 (2009);Positive Polynomials and Sums of Squares, Math. Surveys Monogr.146, AMS (2008)
2009
-
[12]
C. Bamps, S. Pironio, Phys. Rev. A91, 052111 (2015); arXiv:1504.06960
Pith/arXiv arXiv 2015
-
[13]
L. F. Vargas,On the hardness of deciding the finite convergence of Lasserre hierarchies, Numer. Algebra Control Optim.16, 35 (2026); arXiv:2401.12613
Pith/arXiv arXiv 2026
-
[14]
P.,No finite level of the NPA hierarchy is exact for the doubly-tilted CHSH functional near the critical tilt, companion paper (2026)
A. P.,No finite level of the NPA hierarchy is exact for the doubly-tilted CHSH functional near the critical tilt, companion paper (2026). 11
2026
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.