Pith. sign in

REVIEW 3 major objections 4 minor 50 references

Quantum Chaos, Thermalization, and Non-locality

T0 review · 3 major / 4 minor · reviewed 2026-08-15 · deepseek-v4-flash

Pith's one-line read The paper argues that a non-local Hamiltonian with an integrable spectrum can drive a subsystem into the Haar-typical state, thermalizing without quantum chaos.

desk verdict A clean unitary-dressing construction and a real numerical drift toward Haar values, but the thermalization claim outruns the evidence: three scalar measures at N=10 cannot distinguish Haar-typical behavior from a dressed GGE. read the letter →

arxiv 2508.19556 v1 pith:PPSP6QTR submitted 2025-08-27 hep-th cond-mat.stat-mechnlin.CDquant-ph

classification hep-thcond-mat.stat-mechnlin.CDquant-ph MSC 81P4081Q5081-08
keywords quantumthermalizationtypicalstateentanglemententropymutualinformationlogarithmicnegativitynon-localHamiltonianchaosintegrablesystem
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

This paper tries to establish that quantum thermalization—a subsystem losing memory of the initial state and looking random—can be produced by a Hamiltonian whose energy spectrum is integrable, provided the Hamiltonian is sufficiently non-local. The authors construct $H_{\mathrm{NL}}(\tau)=u^{-1}(\tau)Hu(\tau)$ from an integrable transverse Ising chain and show numerically that for $\tau=10$ the time-averaged entanglement entropy and mutual information move toward their Haar-random values while logarithmic negativity is strongly suppressed. If correct, this decouples thermalization from quantum chaos: level statistics are not the decisive ingredient. The reason to care is that scrambling and thermalization are usually tied to chaotic spectra, and this construction exhibits both in a manifestly integrable spectrum.

What carries the argument

The central object is the similarity-transformed Hamiltonian $H_{\mathrm{NL}}(\tau)=u^{-1}(\tau)Hu(\tau)$ with $u(\tau)=e^{-i\tau H_1}$. Because similarity transformations preserve the characteristic polynomial, $H_{\mathrm{NL}}$ inherits $H$'s eigenvalues and hence its integrable spectral statistics, while the eigenstates are rotated to $u^{-1}|E_a\rangle$. The Baker–Campbell–Hausdorff expansion $H_{\mathrm{NL}}=H+(i\tau)[H_1,H]+\frac{(i\tau)^2}{2!}[H_1,[H_1,H]]+\cdots$ shows that increasing $\tau$ generates commutators of growing range, producing non-local couplings; equivalently, local operators are delocalized as $O\mapsto uOu^{-1}$. This mechanism separates global spectral properties from local entanglement dynamics: the spectrum says 'integrable' while the eigenstates can still behave thermally.

What would settle it

Compute the long-time reduced density matrix $\rho_A$ under $H_{\mathrm{NL}}(\tau=10)$ and compare it with the Haar-typical reduced state, e.g., via fidelity or R\'enyi-2 entropy; if $\rho_A$ is far from the typical state while the entanglement entropy, mutual information, and logarithmic negativity match the Haar values, the thermalization claim collapses. A complementary check is to test whether the approach to Page values improves with system size ($N=8,10,12$) while level statistics remain Poisson.

Watch

Extended reading notes

Core claim

The central claim is that non-locality alone can thermalize. For the integrable pair $H$ and $H_1$ (both transverse Ising with transverse fields in $x$ and $y$), the non-local Hamiltonian $H_{\mathrm{NL}}(\tau)=e^{i\tau H_1}He^{-i\tau H_1}$ has exactly the same eigenvalues as $H$, so all spectral diagnostics—level spacing, spectral form factor, resolvent—remain those of an integrable system. Yet its time evolution sends a product state into a regime where, at $\tau=10$, the half-chain entanglement entropy averages to about $3.636$ (Haar random value $4.279$), mutual information averages to about $0.2857$ (Haar value $0.1585$), and logarithmic negativity drops to about $0.0527$, well below the chaotic Ising comparison. The authors read this as the subsystem relaxing to the typical state while quantum correlation between distant intervals is destroyed—quantum thermalization in the absence of quantum chaoticity.

Load-bearing premise

The argument rests on equating thermalization with matching the Haar-random values of three scalar entanglement measures; if a non-thermal state such as a generalized Gibbs ensemble reproduced those same numbers, the conclusion that thermalization occurs would not follow.

Editorial extensions

If this is right

  • If the claim holds, the standard association between quantum chaos and thermalization is broken: integrable spectra can thermalize subsystems when the Hamiltonian is non-local.
  • The construction supplies a tunable dial, $\tau$: at small $\tau$ the dynamics resemble the integrable chain, and as $\tau$ grows the entanglement approaches Haar values while fluctuations shrink, so $\tau$ acts as a non-locality-strength control.
  • The same dressing can be applied to other integrable Hamiltonians, potentially turning any integrable model into a thermalizing one without changing its spectrum.
  • Because $e^{-iH_{\mathrm{NL}}t}=e^{iH_1\tau}e^{-iHt}e^{-iH_1\tau}$, the non-local evolution can be implemented by three sequential quenches, making the effect testable in analog quantum simulators.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The reported match is approximate at $N=10$ ($3.636$ vs $4.279$ for entropy), so the paper's own data suggest convergence is asymptotic; a stronger test would be to verify that the gap closes with increasing $N$, which could also reveal whether the mechanism is true thermalization or finite-size proximity.
  • A sharper diagnostic would compare the full reduced density matrix, not just three scalar measures; if it converges to the Haar-typical state, the mechanism is genuine eigenstate thermalization for twisted eigenstates, otherwise it is a weaker form of scrambling.
  • The 'non-locality is the key property' reading predicts a family of models ordered by interaction range: longer-range or all-to-all interacting models should thermalize faster than nearest-neighbor models even at fixed spectral statistics, which is a testable extension of the paper's logic.
  • The symmetry remark—$H_{\mathrm{NL}}$ can have charges that $H$ does not—suggests a natural follow-up: systems where the dressed Hamiltonian has extra conserved quantities might thermalize only partially, delimiting when non-locality can and cannot produce typical states.
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.

Referee Report

3 major / 4 minor

Summary. The paper constructs a family of non-local Hamiltonians H_NL(τ)=u^{-1}(τ)H u(τ) by a unitary dressing of the integrable transverse-field Ising chain, and compares the entanglement dynamics under H_NL with the dynamics of a chaotic Ising reference model. The central claim is that for sufficiently strong non-locality (τ=10) the system evolves to the Haar-random 'typical' state, as indicated by time-averaged entanglement entropy and mutual information approaching the Page and Sen values, while logarithmic negativity is suppressed; the authors conclude that thermalization does not require quantum chaoticity.

Significance. If established, the central claim would be a compact and explicit counterexample to the common association between quantum chaos and thermalization, with a simple experimental realization. The exact identity e^{-itH_NL}=u^{-1}e^{-itH}u is elegant, and the Haar-random comparison values are computed independently from Page and Sen formulas with no fitted parameters. However, the current evidence consists of three scalar entanglement measures at modest system sizes, the numerical agreement is incomplete, and the construction inherits the conserved charges of the integrable H, so the late-time state is not automatically Haar-typical. The idea is worth pursuing, but additional diagnostics are needed before the claim can be endorsed.

major comments (3)
  1. [Section III, Figs. 1 and 2] The numerical support for the statement that the entanglement entropy 'almost coincides with the Haar random value' is quantitatively weak. At N=10 the time-averaged entropy is 3.636 versus the Page value 4.279, a deficit of about 15%, and the mutual information is 0.2857 versus the Haar value 0.1585, which is about 80% higher. A relative deviation of this size in a central observable does not establish that the reduced state is typical; the text should report these deviations explicitly and justify why they are consistent with a finite-size approach to Haar values, or temper the wording accordingly.
  2. [Section III, typical-state definition and Section II, Eq. (1)] The paper equates thermalization with the matching of three scalar entanglement measures to their Haar-random averages. Because H_NL(τ)=u^{-1}(τ) H u(τ) has the same spectrum as the integrable H and inherits N dressed conserved charges u^{-1}Q_i u, the long-time averaged state is the diagonal ensemble in these eigenstates, generically a generalized Gibbs ensemble (GGE) rather than a microcanonical or Haar-typical state. A GGE with non-local charges can produce near-Page subsystem entanglement while differing from Haar-typical states in other probes. The authors should test the full reduced density matrix, for example by computing the trace distance or fidelity to the Haar-average reduced state, the second Rényi entropy, or expectation values of local operators, and compare against the GGE prediction. Without such a test, the central claim is underdetermined.
  3. [Section IV, experimental feasibility and Section II, non-locality] The exact identity e^{-itH_NL}=u^{-1}e^{-itH}u is used for the experimental realization, but its implications for the interpretation are not discussed. Since the dynamics is unitarily equivalent to integrable dynamics in a rotated basis, the manuscript should clarify what notion of 'information scrambling' is being claimed. It would be helpful to state what this identity does and does not imply for operator spreading or out-of-time-order correlators, so that the phrase 'evolve the system into the typical state' is not read as implying genuine chaotic scrambling.
minor comments (4)
  1. [Appendix A, Fig. 6 caption] The caption of Fig. 6 refers to 'The entanglement entropy for a random pure state' while the plot shows mutual information; this should be corrected.
  2. [Appendix A, Figs. 5 and 7] In the τ-dependence plots, the curves are not clearly identified; please specify whether each curve corresponds to H_NL or to the chaotic Ising model, and add a legend.
  3. [Section III, logarithmic negativity] The Haar-random value for logarithmic negativity is not reported in the main text or figures, although the appendix says the results are compared with random pure states; please state the value and the partition used.
  4. [Section II, non-locality] The text allows for complex τ and non-Hermitian H_NL, but all numerical work uses real τ and a unitary u(τ); the restriction to the Hermitian case should be stated explicitly before the numerics.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the Haar-random comparisons are independent benchmarks, and the only self-citation is not load-bearing.

full rationale

The paper's derivation chain is self-contained at the level of its numerical claims. The non-local Hamiltonian is explicitly constructed as H_NL(τ)=u^{-1}(τ) H u(τ) with u(τ)=e^{-iτ H_1}, and the time evolution is computed directly as e^{-iH_NL t}|ψ⟩; no parameter is fitted to the target observables (entanglement entropy, mutual information, logarithmic negativity). The Haar-random reference values are obtained from the independent Page and Sen formulas, not from the same numerics, so the comparison is an external benchmark rather than a fitted input. The identification of 'typical state' with matching Haar values of selected entanglement measures is a stated interpretive standard, not a circular derivation: the paper reports approximate agreement at τ=10 for entropy and mutual information and interprets it as thermalization, which is an underdetermination/correctness concern (three scalar measures do not fix the reduced density matrix), not a reduction of the conclusion to its inputs. The only self-citation, Ref. [40] by author Yoshii, appears in a passing remark about known thermalizing models and is not load-bearing for the central claim. Therefore no circular step is exhibited.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The central claim rests on a small set of modeling choices: two integrable transverse Ising Hamiltonians, a large rotation strength, and a Haar-random benchmark for typicality. No novel entities are introduced, and no parameters are fitted to the reported data.

free parameters (3)
  • tau (non-locality strength) = 10 (main text); scanned up to 100 in Appendix A
    Chosen by hand. The central demonstration that subsystems approach Haar-typical values is reported for tau=10; the Appendix shows tau-dependence for N=7.
  • hx = hy transverse fields = hx = hy = 1
    Chosen so that both H and H1 are integrable transverse Ising chains. The result is shown at this special point, not across the integrable parameter manifold.
  • h~x, h~z for the chaotic Ising reference = 1.05, -0.5
    Standard chaotic Ising parameters used as a reference; they do not enter the central non-local Hamiltonian but affect the baseline comparison.
assumptions (4)
  • domain assumption H and H1 are integrable transverse Ising Hamiltonians.
    Section III defines H and H1 and cites Pfeuty [38]; integrability of H_NL is inherited from the similarity transformation.
  • domain assumption Spectral statistics determine quantum chaos: Poisson level spacing for integrable spectra, Wigner-Dyson for chaotic.
    Introduction and Section III use this identification to argue that H_NL is not chaotic because its spectrum equals that of integrable H (Eq. 2).
  • domain assumption Haar-random entanglement values define the typical state.
    Section III defines the typical state as coincidence of entanglement measures with Haar random averages (citing Page [31] and Sen [33]); this is the operational criterion for thermalization.
  • domain assumption Time averaging over [t0,T] with t0=10 and T=1000 captures the late-time behavior.
    Appendix B removes the initial transient and uses discrete time averaging; no extrapolation to infinite time or larger N is provided.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Quantum Chaos, Thermalization, and Non-locality." pith.science (2026). https://pith.science/paper/PPSP6QTR

@misc{pith2026250819556,
  author       = {Pith},
  title        = {Pith review of: Quantum Chaos, Thermalization, and Non-locality},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PPSP6QTR}},
  note         = {Machine review of arXiv:2508.19556}
}
read the original abstract

In this paper, we numerically investigate whether quantum thermalization occurs during the time evolution induced by a non-local Hamiltonian whose spectra exhibit integrability. This non-local and integrable Hamiltonian is constructed by combining two types of integrable Hamiltonians. From the time dependence of entanglement entropy and mutual information, we find that non-locality can evolve the system into the typical state. On the other hand, the time dependence of logarithmic negativity shows that the non-locality can destroy the quantum correlation. These findings suggest that the quantum thermalization induced by the non-local Hamiltonian does not require the quantum chaoticity of the system.

Figures

Figures reproduced from arXiv: 2508.19556 by the authors.

Figure 2
Figure 2. FIG. 2: Time dependence of the mutual information. The [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. FIG. 3: The logarithmic negativities as functions of the tim [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figure 4
Figure 4. FIG. 4: Time-averaged entanglement entropy as a function of [PITH_FULL_IMAGE:figures/full_fig_p006_4.png] view at source ↗
Figures from the paper (3 more)
Figure 5
Figure 5. Figure 5: FIG. 5: Long-time average value of the entanglement entropy [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: FIG. 6: Long-time average value of the mutual information [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]
Figure 8
Figure 8. Figure 8: FIG. 8: Time-averaged logarithmic negativity as a function [PITH_FULL_IMAGE:figures/full_fig_p007_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

50 extracted references · 27 canonical work pages

  1. [40]

    Pfeuty, Ann

    P. Pfeuty, Ann. Phys. 57, 79 (1970)

  2. [1]

    Sustainability with Knotted Chiral Meta Matter (SKCM 2)

    05 and −0. 5, respectively. The system considered in this paper is in the time-evolved pure state, |Ψ(t,τ )⟩ =e− itHNL(τ ) |Ψ ⟩ (7) where we assume that all spins are up in the initial nor- malized state, |Ψ ⟩. The parameter, t, is the time associ- ated with HNL(τ). For later convenience, we labels subsystems in two ways. In the case of the bipartite part...

  3. [2]

    Xu and B

    S. Xu and B. Swingle, Phys. Rev. X 9, 031048 (2019) , arXiv:1805.05376 [cond-mat.str-el]

  4. [3]

    R. J. Lewis-Swan, A. Safavi-Naini, A. M. Kauf- man, and A. M. Rey, Nat. Rev. Phys. 1, 627 (2019) , arXiv:1908.11747 [quant-ph]

  5. [4]

    is that u(τ), that induces the non- locality of HNL(τ), delocalizes the local operators, Oi. III. SETUP AND OUR FINDINGS Here, we will present the setup considered in this paper and our findings. The non-local Hamiltonian considered in this paper is constructed of two kinds of transverse Ising Hamiltonians with the open boundary condition, H =Hint + N∑ i=1...

  6. [5]

    M. V. Berry and M. Tabor, Proceedings of the Royal So- ciety of London. A. Mathematical and Physical Sciences 356, 375 (1977)

  7. [6]

    Bohigas, M

    O. Bohigas, M. J. Giannoni, and C. Schmit, Phys. Rev. Lett. 52, 1 (1984)

  8. [7]

    Bhattacharya, S

    T. Bhattacharya, S. Habib, and K. Jacobs, Phys. Rev. Lett. 85, 4852 (2000)

Show all 50 references
  1. [8]

    Maldacena, S

    J. Maldacena, S. H. Shenker, and D. Stanford, J. High Energy Phys. 8, 106 (2016)

  2. [9]

    D. A. Roberts and D. Stanford, Phys. Rev. Lett. 115, 131603 (2015)

  3. [10]

    Popescu, A

    S. Popescu, A. J. Short, and A. Winter, Nat. Phys. 2, 754 (2006)

  4. [11]

    Goldstein, J

    S. Goldstein, J. L. Lebowitz, R. Tumulka, and N. Zangh ` ı, Phys. Rev. Lett. 96, 050403 (2006)

  5. [12]

    Goldstein, D

    S. Goldstein, D. A. Huse, J. L. Lebowitz, and R. Tu- mulka, Annalen der Physik 529, 1600301 (2017)

  6. [13]

    Goldstein, D

    S. Goldstein, D. A. Huse, J. L. Lebowitz, and R. Tu- mulka, Phys. Rev. Lett. 115, 100402 (2015)

  7. [14]

    Goldstein, J

    S. Goldstein, J. L. Lebowitz, C. Mastrodonato, R. Tu- mulka, and N. Zanghi, Phys. Rev. E 81, 011109 (2010)

  8. [15]

    Tasaki, J

    H. Tasaki, J. Stat. Phys. 163, 937 (2016)

  9. [16]

    Sugita, arXiv e-prints , cond-mat/0602625 (2006)

    A. Sugita, arXiv e-prints , cond-mat/0602625 (2006)

  10. [17]

    M. V. Berry, J. Phys. A: Math. Gen. 10, 2083 (1977)

  11. [18]

    Feingold and A

    M. Feingold and A. Peres, Phys. Rev. A 34, 591 (1986)

  12. [19]

    D’Alessio, Y

    L. D’Alessio, Y. Kafri, A. Polkovnikov, and M. Rigol, Advances in Physics 65, 239 (2016)

  13. [20]

    Mondaini and M

    R. Mondaini and M. Rigol, Phys. Rev. E 96, 012157 (2017) . 6

  14. [21]

    Berry, Physica Scripta 40, 335 (1989)

    M. Berry, Physica Scripta 40, 335 (1989)

  15. [22]

    M¨ uller, S

    S. M¨ uller, S. Heusler, P. Braun, F. Haake, and A. Alt- land, Phys. Rev. Lett. 93, 014103 (2004)

  16. [23]

    M¨ uller, S

    S. M¨ uller, S. Heusler, P. Braun, F. Haake, and A. Alt- land, Phys. Rev. E 72, 046207 (2005)

  17. [24]

    P. Kos, M. Ljubotina, and T. Prosen, Phys. Rev. X 8, 021062 (2018)

  18. [25]

    Bertini, P

    B. Bertini, P. Kos, and T. Prosen, Phys. Rev. Lett. 121, 264101 (2018)

  19. [26]

    Braun, D

    P. Braun, D. Waltner, M. Akila, B. Gutkin, and T. Guhr, Phys. Rev. E 101, 052201 (2020)

  20. [27]

    Nahum, J

    A. Nahum, J. Ruhman, S. Vijay, and J. Haah, Phys. Rev. X 7, 031016 (2017)

  21. [28]

    C. W. von Keyserlingk, T. Rakovszky, F. Pollmann, and S. L. Sondhi, Phys. Rev. X 8, 021013 (2018)

  22. [29]

    A. Chan, A. De Luca, and J. T. Chalker, Phys. Rev. Lett. 121, 060601 (2018)

  23. [30]

    A. J. Friedman, A. Chan, A. De Luca, and J. T. Chalker, Phys. Rev. Lett. 123, 210603 (2019)

  24. [31]

    Bertini, P

    B. Bertini, P. Kos, and T. Prosen, Phys. Rev. X 9, 021033 (2019)

  25. [32]

    A. Chan, A. De Luca, and J. T. Chalker, Phys. Rev. X 8, 041019 (2018)

  26. [33]

    D. N. Page, Phys. Rev. Lett. 71, 1291 (1993)

  27. [34]

    Srednicki, Phys

    M. Srednicki, Phys. Rev. E 50, 888 (1994)

  28. [35]

    Sen, Phys

    S. Sen, Phys. Rev. Lett. 77, 1 (1996)

  29. [36]

    Srednicki, J

    M. Srednicki, J. Phys. A: Math. Gen. 32, 1163 (1999)

  30. [37]

    M. C. Ba˜ nuls, J. I. Cirac, and M. B. Hastings, Phys. Rev. Lett. 106, 050405 (2011)

  31. [38]

    J. S. Cotler, G. Gur-Ari, M. Hanada, J. Polchinski, P. Saad, S. H. Shenker, D. Stanford, A. Streicher, and M. Tezuka, J. High Energy Phys. 5, 118 (2017)

  32. [39]

    Gharibyan, M

    H. Gharibyan, M. Hanada, S. H. Shenker, and M. Tezuka, J. High Energy Phys. 2018, 124 (2018)

  33. [41]

    Iyoda and T

    E. Iyoda and T. Sagawa, Phys. Rev. A 97, 042330 (2018)

  34. [42]

    Yoshii, S

    R. Yoshii, S. Yamashika, and S. Tsuchiya, J. Phys. Soc. Jap. 91, 054601 (2022)

  35. [43]

    Nakata and M

    Y. Nakata and M. Tezuka, Phys. Rev. Res. 6, L022021 (2024)

  36. [44]

    Groisman, S

    B. Groisman, S. Popescu, and A. Winter, Phys. Rev. A 72, 032317 (2005)

  37. [45]

    Vidal and R

    G. Vidal and R. F. Werner, Phys. Rev. A 65, 032314 (2002)

  38. [46]

    M. B. Plenio, Phys. Rev. Lett. 95, 090503 (2005)

  39. [47]

    Peres, Phys

    A. Peres, Phys. Rev. Lett. 77, 1413 (1996)

  40. [48]

    Horodecki, P

    M. Horodecki, P. Horodecki, and R. Horodecki, Phys. Lett. A 223, 1 (1996)

  41. [49]

    G¨ arttner, J

    M. G¨ arttner, J. G. Bohnet, A. Safavi-Naini, M. L. Wall, J. J. Bollinger, and A. M. Rey, Nat. Phys. 13, 781 (2017)

  42. [50]

    Haar random

    J.-H. Wang, T.-Q. Cai, X.-Y. Han, Y.-W. Ma, Z.-L. Wang, Z.-H. Bao, Y. Li, H.-Y. Wang, H.-Y. Zhang, L.-Y. Sun, Y.-K. Wu, Y.-P. Song, and L.-M. Duan, Phys. Rev. Res. 4, 043141 (2022) . Appendix A: System size dependence In this appendix, we present the analysis of entangle- ment...

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.