Nonlocal nonstabilizerness for slightly entangled quantum many-body states

Jian Cui; Lei-Yi-Nan Liu

REVIEW 3 major objections 5 minor 1 cited by

Nonlocal nonstabilizerness for slightly entangled quantum many-body states

T0 review · 3 major / 5 minor · reviewed 2026-07-14 · grok-4.5

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

Pith's one-line read Nonlocal magic equals the stabilizer Rényi entropy of a sorted Schmidt reference state, turning hard local-unitary minimization into a direct spectral formula.

desk verdict Useful spectral shortcut for nonlocal SRE of weakly entangled states; the equality is still a well-supported conjecture, not a theorem. read the letter →

arxiv 2607.10714 v2 pith:WE2BKW7U submitted 2026-07-12 quant-ph

Lei-Yi-Nan Liu , Jian Cui This is my paper · ORCID

classification quant-ph

keywords nonlocalnonstabilizernessstabilizerRényientropySchmidtspectrummagicresourcequantummany-bodystatesentanglementPXPmodelcriticalspinchains

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper claims that the irreducible magic resource locked inside bipartite entanglement can be read straight off the sorted Schmidt spectrum. Instead of minimizing the stabilizer Rényi entropy over all local unitaries—an exponentially hard, nonconvex search—one builds a canonical reference state that simply assigns the ordered Schmidt coefficients to computational-basis product states. The authors conjecture that the nonlocal SRE of any pure bipartite state equals the ordinary SRE of this reference state. They prove that the reference state is a stationary point of the local-unitary landscape and show numerical agreement for rank-4 spectra and small many-body states. The resulting formula is practical precisely when entanglement is modest: once the entanglement spectrum is known (from free-fermion methods, DMRG, or tensor networks), nonlocal magic becomes a controlled spectral calculation, including a rigorous truncation bound. Applied to Haar states, critical Ising and XXZ chains, and scarred PXP dynamics, the quantity exposes nonstabilizer structure that ordinary entanglement entropy misses—logarithmic critical scaling whose coefficient varies inside a fixed central-charge phase, and a clear separation between entanglement growth and irreducible-magic growth.

What carries the argument

The Schmidt reference state: the pure state whose Schmidt coefficients are the ordered entanglement spectrum of the original state, assigned to computational-basis product vectors. Its SRE supplies both an upper bound and, by conjecture, the exact nonlocal SRE (Eqs. 6 and 10).

What would settle it

Find a bipartite pure state whose manifold-optimized nonlocal SRE is strictly smaller than the SRE of its descending-order Schmidt reference state, within numerical precision, for any Rényi index α ≥ 2.

Watch

Extended reading notes

Core claim

The nonlocal stabilizer Rényi entropy of a bipartite pure state is conjectured to equal the SRE of the descending-order Schmidt reference state that encodes the same spectrum in the computational basis; thus nonlocal nonstabilizerness is completely determined by a direct spectral expression that bypasses local-unitary optimization.

Load-bearing premise

That the proven stationary point of the local-unitary landscape is the global minimum rather than a saddle or local minimum; the global claim remains a conjecture supported by numerics.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Desk Editor's Note

read the letter

The real advance here is practical: once you have the sorted Schmidt spectrum, they give a direct formula for nonlocal SRE (the descending-order reference-state SRE) and show how to evaluate it for the states we already compute with DMRG/VUMPS or free-fermion methods. Huang et al. already proved spectral dependence and introduced the same canonical encoding; this paper’s new claim is that the descending reference saturates the local-unitary minimum.

What they do well is solid. Appendix A cleanly proves first-order stationarity under any local Hermitian generator. Appendix B gives a controlled truncation error in discarded weight η. Stiefel-manifold optimization matches the reference SRE to numerical precision for the full rank-4 family and several α, and they extract a simple closed relation to low-order Rényi entropies for χ=4. The FNL ≥ NL inequality is proven without the conjecture and is useful. The many-body applications (Ising log scaling, XXZ coefficient varying at fixed c=1, PXP scar vs empty separation of entanglement growth from nonlocal magic) are the payoff and look carefully done.

The soft spot is exactly the one the stress-test flags, and it is real but not fatal: stationarity plus “not a maximizer” plus low-rank numerics do not prove global minimality for generic χ. Large-system plots therefore inherit the conjecture. That is stated as a conjecture in the paper; they do not hide it. Code/data are not public, which is a minor annoyance for a computational claim.

This is for people who already compute entanglement spectra and want a nonlocal-magic diagnostic without nonconvex optimization. It deserves a serious referee. I would engage with it and cite the framework when I need the quantity; I would flag the conjecture when I use the large-system numbers.

Referee Report

3 major / 5 minor

Summary. The manuscript introduces a Schmidt-reference-state construction for bipartite nonlocal stabilizer Rényi entropy (SRE): sorted Schmidt coefficients are assigned to a canonical computational-basis state, and the authors conjecture that the nonlocal SRE equals the SRE of this reference state (Eqs. 5–6, 10). They prove that any such reference-form state is a stationary point of the SRE under arbitrary local unitaries (Appendix A, Eq. 8), bound the truncation error of the spectrum by the discarded weight η (Appendix B), and prove that fermionic nonlocal magic (FNL) upper-bounds the reference-state SRE without using the conjecture (Appendix D). Numerical Riemannian optimization on the complex Stiefel manifold agrees with the reference SRE for rank-4 spectra across several Rényi indices (Sec. III, Fig. 1). Under the conjecture they evaluate nonlocal SRE for Haar-random states, critical Ising and XXZ chains, and PXP quench dynamics, arguing that it probes entanglement-spectrum structures invisible to ordinary entanglement measures and to FNL.

Significance. If the conjecture holds, the paper supplies a practical spectral formula for nonlocal nonstabilizerness that is especially useful for weakly entangled many-body states (MPS/VUMPS, free-fermion spectra). The stationarity proof, truncation bound, and FNL ≥ NL inequality are solid contributions independent of global minimality. The applications are physically interesting: an O(1) Haar nonlocal SRE, logarithmic critical scaling with a coefficient that varies across the XXZ Luttinger-liquid phase at fixed c=1, and a clear dynamical separation between entanglement growth and nonlocal-magic growth in scarred PXP dynamics. These results would open a concrete route from entanglement spectra to irreducible magic resources in and out of equilibrium. The work is transparent that the global claim is a conjecture and backs it with first-order optimality plus rank-4 optimization; that honesty is a strength, but the large-system conclusions still rest on the unproven step.

major comments (3)

Sec. II, Eq. (6) and Appendix A: The central identification M_NL,α = M_α(|ψ̃⟩) is a conjecture. Appendix A only proves that every reference-form state (not necessarily sorted) is a critical point: ∂_t F_α|t=0 = 0 under every local Hermitian generator. Compactness guarantees extrema exist but does not identify the global minimizer. The claim that descending order yields the global minimum among stationary points is observational (citing Ref. [33]), not proven. For a load-bearing claim used throughout Sec. IV, the manuscript needs either (i) substantially stronger numerical evidence that local Stiefel optimization never finds a lower value for χ>4 (e.g., random rank-8/16 spectra with multi-start Riemannian optimization and reported success rates), or (ii) a clear, repeated demarcation in Sec. IV that all large-system results are for the reference-state upper bound M_NL,ref, with the equali
Sec. III and Fig. 1: Numerical support for the conjecture is essentially confined to Schmidt rank χ=4 (one-parameter family and the (r,θ,φ) scan) plus unspecified “representative many-body states” of small size. For χ>4 the Stiefel manifold is high-dimensional and non-convex; local optimizers can miss lower critical points. The paper should report explicit optimization-vs-reference comparisons for at least a few higher-rank spectra (e.g., χ=8) and for the small Ising/XXZ ground states where manifold optimization is still feasible, with quantitative residuals. Without that, the leap from rank-4 agreement to thermodynamic-limit applications is under-supported.
Sec. IV B–C, Eqs. (23), (30), (32): Logarithmic scaling of nonlocal SRE (β_NL_ξ, β_NL_ℓ) is fitted over limited ranges of ξ and ℓ, and the authors themselves note that eventual saturation at large ξ cannot be excluded. The claim that β_NL varies across the XXZ critical phase at fixed c=1 is the most distinctive many-body result; it should be accompanied by a controlled check that the variation survives changes in bond dimension, truncation threshold η, and fitting window, and by an explicit statement that the coefficient is for M_NL,ref under the conjecture. Presenting β_NL as a robust interaction-dependent diagnostic without those controls overstates the evidence.

minor comments (5)

Typos and spelling: “Schimidt” (Sec. II), “interralation” (Introduction), “tenser-network” (Introduction), “OPTIMIZA TION” / “SYTEMS” (section titles), “subsustem” (Fig. 5 caption). Clean these systematically.
Eq. (10) and the padding rule for non-power-of-2 χ should be stated once with an explicit algorithm (sort, pad zeros to 2^⌈log2 χ⌉, evaluate the XOR sum); the present wording is easy to mis-implement.
Fig. 2 is hard to read (overlapping LaTeX labels). A cleaner parameterization plot or a table of maximizers/minimizers would help.
The low-rank formula Eq. (17) is useful for experiment; state clearly that it is exact only for χ≤4 and that the PXP orange curves are therefore early-time approximations, not full nonlocal SRE.
Data availability: “available upon reasonable request” is weak for a methods paper whose main deliverable is a spectral formula. Depositing the rank-4 optimization scripts and the VUMPS/DMRG spectra used in Figs. 4–6 would strengthen reproducibility.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: nonlocal SRE equals reference-state SRE is an openly stated conjecture, not a derivation forced by definition or self-citation.

full rationale

The paper's load-bearing claim (Eq. 6) is explicitly labeled a conjecture: the reference-state SRE is only an upper bound by construction (Eq. 5, via Huang et al. and local-unitary invariance of the Schmidt spectrum), and saturation is not derived from the definition of nonlocal SRE. Appendix A proves only first-order stationarity under local Hermitian generators for any reference-form state (not necessarily sorted); the authors do not claim this implies global minimality, and they support the conjecture with independent Riemannian Stiefel-manifold optimization that does not presuppose the reference state (Sec. III, Fig. 1). The FNL ≥ NL inequality (Appendix D) is proven without the conjecture. Large-system applications (Haar, Ising/XXZ, PXP) rest on the conjecture, which is a correctness/scope risk, not circularity: nothing reduces by construction to a fitted input or to a self-citation uniqueness theorem. Eq. (10) is properly attributed to prior work in a different context. No self-definitional loop, no fitted-parameter-as-prediction, and no load-bearing self-citation chain by the present authors.

Assumptions & free parameters 2 free parameters · 4 assumptions · 1 invented entities

The load-bearing novelty is a single unproven equality (the conjecture). Everything else is either standard resource-theory definitions, a first-order calculus proof, or numerical evidence. Free parameters appear only in application-level fits (β coefficients, truncation cutoffs) and do not underwrite the central claim. No new physical entity is postulated; the reference state is a canonical encoding already introduced by Huang et al.

free parameters (2)

β_NL_ξ / β_NL_ℓ logarithmic scaling coefficients = ≈0.1077 (Ising ξ), ≈0.1037–0.163 (ℓ fits)
Fitted from VUMPS or subsystem-size data for Ising and XXZ; used to claim interaction-dependent scaling inside the c=1 phase, but not required for the main conjecture.
Schmidt truncation threshold / discarded weight η = 10^{-15} (typical)
Practical cutoff (e.g., retain λ_i > 10^{-15}) used for large-ℓ NL evaluation; error is bounded but the concrete threshold is a numerical choice.

assumptions (4)

domain assumption Nonlocal α-SRE is defined by minimization of SRE over local unitaries U_A ⊗ U_B (Eq. 3).
Standard definition from the nonlocal-magic literature (Qian-Wang, Cao et al.); taken as given.
domain assumption Bipartite nonlocal magic depends only on the nonzero Schmidt spectrum (Huang et al.).
Cited as proven; used to reduce the problem to the reference state.
ad hoc to paper The descending-order Schmidt reference state realizes the global minimum of SRE under local unitaries (the central conjecture, Eq. 6).
Only first-order stationarity is proven; global minimality is conjectured and checked numerically for small ranks.
domain assumption SRE with α≥2 is a magic monotone; focus on α=2 is legitimate.
Standard in the stabilizer-Rényi literature (Leone et al., Haug-Piroli).

invented entities (1)

Schmidt reference state |ψ̃⟩ (descending computational-basis encoding of the ordered Schmidt spectrum)
purpose: Canonical state whose SRE is conjectured to equal nonlocal SRE, enabling direct spectral evaluation.
Coincides with the canonical encoding of Huang et al.; the paper’s contribution is the saturation claim, not the object itself.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Nonlocal nonstabilizerness for slightly entangled quantum many-body states." pith.science (2026). https://pith.science/paper/WE2BKW7U

@misc{pith2026260710714,
  author       = {Pith},
  title        = {Pith review of: Nonlocal nonstabilizerness for slightly entangled quantum many-body states},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/WE2BKW7U}},
  note         = {Machine review of arXiv:2607.10714}
}

read the original abstract

Nonlocal nonstabilizerness quantifies the irreducible magic resource encoded in bipartite entanglement, but its evaluation is generally hindered by a highly nonconvex optimization over local unitary transformations. Here we propose a Schmidt-reference-state framework that replaces this optimization by a direct construction from the sorted Schmidt spectrum. We conjecture that the nonlocal stabilizer R\'enyi entropy (SRE) is given by the SRE of the corresponding reference state, and support this conjecture through analytical and numerical evidences. Our framework makes nonlocal nonstabilizerness efficiently accessible for weakly entangled many-body states whenever the entanglement spectrum is available. Applying it to Haar-random states, critical spin chains, and PXP dynamics, we show that nonlocal SRE captures nonstabilizer structures in the entanglement spectrum that are invisible to conventional entanglement measures. Our results establish entanglement spectra as a powerful window into irreducible nonstabilizer correlations, opening a broadly applicable route to studying nonlocal magic resources of quantum many-body systems in and out of equilibrium.

Figures

Figures reproduced from arXiv: 2607.10714 by the authors.

**Figure 3.** FIG. 3. Average nonlocal SRE of Haar-random states as a [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗

**Figure 4.** FIG. 4. Nonlocal SRE in the Ising model. For finite systems, we obtain the entanglement spectrum using DMRG, while [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗

Figures from the paper (4 more)

**Figure 5.** Figure 5: FIG. 5. FNL and NL as the function of subsustem size [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]

**Figure 6.** Figure 6: FIG. 6. Scaling coefficient of the nonlocal SRE in the critical [PITH_FULL_IMAGE:figures/full_fig_p009_6.png]

**Figure 7.** Figure 7: FIG. 7. Entanglement and nonlocal magic dynamics in [PITH_FULL_IMAGE:figures/full_fig_p010_7.png]

**Figure 8.** Figure 8: FIG. 8. Long time evolution for PXP model with different [PITH_FULL_IMAGE:figures/full_fig_p011_8.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

Schmidt-Gauge Non-Local Magic: Representation, Optimality, and Mathematical Properties
quant-ph 2026-08 conditional novelty 6.0 of 10

For every 1xN bipartition, the Schmidt-gauge non-local magic equals the true local-unitary minimum, and the Walsh-Hadamard representation yields bounds, selection rules, and an inverse-participation-ratio interpretation.

Pith tools

Reviewed July 14, 2026 · model on record in the stance chip above.