{"id":"a6bc69ab-070a-48eb-8896-e20fbabc9a09","arxiv_id":"1908.02596","paper_version":2,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"For quenches in the fully connected transverse-field Ising model, the variance of the order parameter shows four distinct regimes, including a new periodically enhanced squeezing effect, and the bipartite entanglement Hamiltonian is a time-dependent harmonic oscillator to leading order in 1/N.","lead":"This paper analyzes what happens after a sudden change of magnetic field in a fully connected quantum Ising model, deriving exact leading-order formulas for how quantum fluctuations and entanglement grow. The main result is that the entanglement structure is captured by a single time-dependent harmonic oscillator, and the quality of simple mean-field descriptions fails on timescales that shrink as the system size grows.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The load-bearing assumption is persistence of the large-deviation Gaussian form; the paper's t* estimate is heuristic, and the harmonic-oscillator entanglement Hamiltonian inherits this assumption.","rationale":"The reader's weakest assumption is correctly identified: Gaussianity/localization is the load-bearing premise. The paper explicitly acknowledges the breakdown at the Ehrenfest time and provides numerical confirmation of the variance and entropy dynamics for early times, so this is a limitation rather than an internal inconsistency. My stress-test sharpens the concern: the t* estimate is heuristic, and the entanglement Hamiltonian claim requires the Gaussian ansatz to hold not just for second moments but for the full reduced density matrix. A concrete check of the fourth cumulants would settle whether the Gaussian truncation is self-consistently valid up to t*. Since the paper already delimits the validity window and the ED comparisons are consistent with it, I do not believe the verdict should change; the check is a worthwhile verification of the paper's central conditional claim.","tokens_in":40171,"tokens_out":18669,"duration_ms":210930,"concrete_test":"In regime IV (e.g., Γ_i=0.4, Γ_f=0.8), evolve the rate-function hierarchy of Eq. (31) to order n=3,4 alongside Eq. (7), and compute the next-order correction to var(n+) in Eq. (32b) for N=10^4. Then extract the fourth cumulant of the subsystem magnetization a+ from exact diagonalization in the Dicke subspace at times t ~ 0.5 sqrt(N), sqrt(N), and 2 sqrt(N). If the normalized fourth cumulant (κ4/var^2) is not small at t ≲ sqrt(N), the Gaussian reduced density matrix underlying Eq. (19) and the resulting Rényi entropies are not justified within the claimed validity window.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claims (four variance regimes, t* scaling, harmonic-oscillator entanglement Hamiltonian) all rest on the statement that the time-evolved Dicke-subspace wave function remains, to leading order in 1/N, a Gaussian wave packet with covariance evolving under the linearized flow (Eq. (8); equivalent to Eq. (7) by Appendix C). The variance dynamics alone only needs the second-order saddle point; but the entanglement claims are stronger: Eq. (19) treats the reduced density matrix as exactly Gaussian, so the 'complete' entanglement Hamiltonian is a harmonic oscillator only if the full rate function f_AB is quadratic to the required accuracy. The paper's own Sec. 4.5 estimates the breakdown time t* using a heuristic anharmonicity scale chi ~ sqrt(∂V/∂^3V), not from a controlled bound on the neglected f_3, f_4 terms in Eq. (31). If the fourth-order cumulants of n+ (or of the subsystem variable a+) become comparable to the square of the variance before t*~sqrt N in regime IV, the symplectic eigenvalue lambda in Eq. (24) and hence all Rényi entropies acquire O(1) corrections, and the harmonic-oscillator entanglement Hamiltonian is not valid up to the claimed time.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper analyzes quantum quenches in the fully connected transverse-field Ising model (Lipkin-Meshkov-Glick) using a rate-function (WKB) expansion in the inverse system size. The authors derive a leading-order ODE for the curvature f2 of the rate function (Eq. (7)), prove its equivalence to the nearby-orbit covariance evolution (Appendix C), and use it to classify the time dependence of the order-parameter variance into four regimes: exponential, periodic, quadratic, and periodically enhanced squeezing. They further estimate the breakdown time of the mean-field approximation (t* ~ sqrt(N) for polynomial spreading, t* ~ log N near hyperbolic points) and derive the bipartite entanglement Hamiltonian, which in the semiclassical limit is a time-dependent harmonic oscillator whose symplectic eigenvalue determines all Rényi entropies. The analytical results are compared with exact diagonalization for systems up to N=10^4 spins.","tokens_in":40363,"tokens_out":20560,"duration_ms":233083,"significance":"If correct, the paper provides a parameter-free leading-order analytical description of a benchmark mean-field model, with no fitted parameters: initial conditions are fixed by the WKB ground state, the variance follows from Eq. (7), and the entanglement spectrum follows from the Gaussian covariance. The equivalence proof in Appendix C and the replica calculation in Appendix G are careful and internally consistent, and the numerical checks for the variance and entropy support the central claims. The identification of the fourth regime (periodically enhanced squeezing) and the connection between spin squeezing and the entire entanglement spectrum are valuable contributions. The main limitation, as the paper itself states in Sec. 4.5, is that the Gaussian (nearby-orbit) approximation is valid only up to the Ehrenfest time, so the harmonic-oscillator entanglement Hamiltonian is a leading-order statement in 1/N; the stress-test concern about uncontrolled f3/f4 corrections is real but is explicitly acknowledged by the authors and does not undermine the central derivation.","major_comments":[],"minor_comments":[{"comment":"The paper should state explicitly that the harmonic-oscillator form (19) is a leading-order-in-1/N statement valid for t much less than t*, and that corrections from f3 and f4 (Eq. (32)) to the symplectic eigenvalue (24) are not computed; the t* estimate is heuristic and threshold-dependent (Fig. 5).","section":"Sec. 4.5 and Sec. 5.2"},{"comment":"The assertion that P(t)^T v traverses all directions in each period is used to explain the variance minima, but no proof is given; please add a short argument or qualify the statement as an observation supported by the numerical examples.","section":"Sec. 4.4 and Appendix D"},{"comment":"The notation N0 omega is unclear; it should be n omega with n in N0, or N0 should be defined explicitly.","section":"Eq. (20)"},{"comment":"The caption says 'rho_A(t1) and rho_A(t1)' and should read 'rho_A(t1) and rho_A(t2)'.","section":"Fig. 10 caption"},{"comment":"The citation gives a placeholder title 'E. Heller, title'; a complete bibliographic entry is needed.","section":"Reference [54]"},{"comment":"There are several typos, e.g. 'fascilitated' in Sec. 5.4; the manuscript would benefit from a final proofread.","section":"Throughout"}],"recommendation":"minor_revision","confidential_remarks":"The paper is well within the scope of cond-mat.stat-mech and is a solid contribution. I do not think the heuristic nature of the t* estimate or the absence of a rigorous error bound on the Gaussian approximation require a major revision; the authors are candid about the leading-order nature of the claims. The main things I would ask in revision are a clarifying sentence about the status of corrections to the entanglement Hamiltonian and a proof or qualification for the direction-coverage claim in Sec. 4.4."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this paper deserves a serious referee, and the reader's accept recommendation is about right. The genuinely new content is the four-regime classification of the magnetization variance after a quench (especially periodically enhanced squeezing with its 1/t^2 envelope) and the construction of the entanglement Hamiltonian as a time-dependent harmonic oscillator whose spectrum determines all Rényi entropies. The rate-function ODE for f2, the equivalence with nearby-orbit covariance in Appendix C, and the Gaussian reduced-density-matrix calculation in Appendix G are all internally consistent, and the exact-diagonalization checks in the Dicke subspace up to N=10^4 give real support. The paper is also honest about provenance: it credits Sciolla-Biroli for the rate-function machinery and Lerose-Pappalardi for the entropy results, so the novelty claim is appropriately scoped.\n\nWhere are the soft spots? The main one is the breakdown-time estimate in Sec. 4.5. The scale chi ~ sqrt(dV/d^3V) is heuristic, borrowed from the Moyal-bracket argument, and the paper does not provide a controlled bound on the neglected f3, f4 terms in Eq. (31). The stress-test concern that fourth-order cumulants could become comparable to the variance before t*~sqrt(N) in regime IV is a legitimate logical gap, and the harmonic-oscillator entanglement Hamiltonian does inherit the Gaussianity assumption. However, the authors explicitly flag this limitation, and their numerical t* data (Fig. 5) plus the ED agreement in Figs. 4, 7, and 9 give no indication that the stress-test scenario actually occurs in the parameter range studied. It remains a plausible but unproven possibility, not a demonstrated flaw. A careful referee should ask for a sharper statement about the neglected-cumulant terms in regime IV, but this should not block the paper.\n\nThe citation pattern looks fair, and there is no parameter fitting disguised as prediction: the initial curvature f2(0) comes from the WKB ground-state condition, and the dynamics follows from the Hamiltonian.\n\nWho should read this? People working on mean-field quench dynamics, spin squeezing and entanglement in fully connected spin models, and semiclassical methods for collective spins. I would bring it to a reading group, and I would cite it in my own work on semiclassical entanglement growth. Send it to peer review with a request to tighten the breakdown estimate rather than to redo the core analysis.","headline":"Solid semiclassical quench analysis for the LMG model: the variance classification and harmonic-oscillator entanglement Hamiltonian are genuine advances, and the main caveat is a heuristic rather than controlled breakdown-time estimate.","tokens_in":40912,"tokens_out":2256,"would_cite":true,"duration_ms":27663,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"After a quantum quench in the fully connected transverse-field Ising model, the leading 1/N corrections put the magnetization variance into four distinct regimes, set mean-field validity at times ~sqrt(N), and reduce the bipartite…","keywords":["transverse-field Ising model","fully connected lattice","quantum quench","mean-field dynamics","rate function expansion","spin squeezing","entanglement Hamiltonian","Rényi entanglement entropy"],"falsifier":"Use numerically exact time evolution for $N=10^3$ and $N=10^4$ on quenches in each of the four regimes and locate the first time $t_*$ at which $N\\,\\mathrm{var}(n_+)$ deviates from the semiclassical result by a fixed threshold; the claim is falsified if $t_*$ does not scale as $\\sqrt{N}$ in regimes III and IV and as $\\log N$ in regime I. A second check is the entanglement spectrum of the reduced density matrix: it must be equidistant with spacing $\\omega=2\\,\\mathrm{arctanh}[1/(2\\lambda)]$ as long as the semiclassical approximation holds.","tokens_in":39934,"feed_emoji":"🧲","tokens_out":14199,"duration_ms":139360,"temperature":0.7,"pith_summary":"This paper tries to establish that out-of-equilibrium mean-field dynamics in the fully connected transverse-field Ising model can be computed beyond the mean, and that the corrections carry the physics. It shows that the leading $1/N$ correction to the order-parameter variance obeys a single ordinary differential equation and organizes into four regimes — exponential, periodic, quadratic, and periodically enhanced squeezing — each invisible in the mean magnetization. It further claims that the entanglement Hamiltonian of any bipartition is, to leading order, a time-dependent harmonic oscillator whose frequency determines the full entanglement spectrum and all Rényi entropies. These results matter because they give a controlled, analytically solvable benchmark for when mean-field approximations fail out of equilibrium and how spin squeezing quantifies entanglement.","feed_headline":"Four variance regimes rule quench dynamics in a mean-field magnet","feed_subtitle":"Variance, not the mean, exposes the regimes; mean-field validity ends at sqrt(N) or log N times.","key_machinery":"The engine is the inverse-system-size expansion of the rate function $f(n_+)$ in a large-deviation ansatz $\\psi(n_+)\\asymp e^{-Nf(n_+)}$ (Eq. (3)). Its time-dependent curvature $f_2(t)$ obeys a closed Riccati-type ordinary differential equation (Eq. (7)) equivalent to evolving a Gaussian covariance under the linearized classical flow, $C(t)=S(t)C(0)S(t)^T$ (Eq. (8)). For periodic orbits, Floquet theory splits $S(t)=P(t)M(t)$ into a periodic part and a shear matrix with shear factor $\\alpha=\\|H'\\|^2 T'(E_0)$; this shear is the mechanism behind periodically enhanced squeezing and spreading. On the entanglement side, the same covariance, reduced to a subsystem, has a Gaussian Wigner function, and the star-exponential of a quadratic form converts this into a quadratic entanglement Hamiltonian whose spectrum is fixed by the symplectic eigenvalue $\\lambda=\\sqrt{\\det\\Sigma_A}$ through the Williamson decomposition.","core_discovery":"The paper claims that for a sudden quench $\\Gamma_i \\to \\Gamma_f$ in the fully connected transverse-field Ising model, the leading $1/N$ corrections to mean-field dynamics are captured by a quadratic rate-function curvature $f_2$ obeying $i df_2/dt = -(1, if_2) H''(1, if_2)$ (Eq. (7)), equivalently by a Gaussian covariance $C(t)=S(t)C(0)S(t)^T$ (Eq. (8)) transported along the classical orbit. From this single object the magnetization variance falls into four qualitatively different regimes — exponential growth (I), periodic oscillations (II), quadratic growth (III), and periodically enhanced squeezing (IV) — that cannot be distinguished by the mean magnetization alone. The same Gaussian covariance, after tracing out a subsystem, yields a reduced density matrix whose entanglement Hamiltonian is exactly quadratic, i.e. a time-dependent harmonic oscillator whose frequency $\\omega(t)$ is fixed by the symplectic eigenvalue $\\lambda = \\sqrt{\\det \\Sigma_A}$ of the covariance matrix; all Rényi entropies follow from $\\lambda$, and $\\lambda$ is itself a function of the spin-squeezing parameter $\\xi_S$. Because the wave packet spreads, the semiclassical description ceases to be valid at an Ehrenfest time $t_* \\sim \\sqrt{N}$ in regimes III and IV and $t_* \\sim \\log N$ in regime I, with $t_* \\sim N$ near a stable fixed point.","pith_inferences":["Going beyond the paper, the Ehrenfest-time scalings imply that effective single-particle or semiclassical descriptions of fully connected quenches should be treated as reliable only up to $O(\\sqrt{N})$ ($O(\\log N)$ near hyperbolic points), even when the mean magnetization looks stable; finite-size dephasing is not a small correction but sets the practical validity window.","A cold-atom dimer experiment that can prepare the effective Hamiltonian and read out $N\\,\\mathrm{var}(n_+)$ could test the sharpest prediction here: in regime IV the squeezing minima should decay like $1/t^2$ before the Ehrenfest time, a signature that does not depend on model details.","The same single-scalar description opens the door to designing explicit LOCC conversion protocols for collective-spin states, since the target majorization ordering is controlled entirely by $\\lambda(t)$."],"forward_implications":["In regimes III and IV the mean-field approximation is trustworthy only up to $t_*\\sim\\sqrt{N}$, and in regime I only up to $t_*\\sim\\log N$, so large finite fully connected spin systems should show measurable departures from mean-field predictions on these timescales.","Because the variance distinguishes four regions of the dynamical phase diagram that the mean order parameter alone cannot, measuring $\\mathrm{var}(n_+)$ (or a squeezing parameter built from it) is a sharper probe of dynamical phase transitions than the magnetization.","Entanglement growth after a quench is linear when the variance grows exponentially (regime I), logarithmic when the variance grows quadratically (regimes III and IV), and bounded when the variance oscillates (regime II); all cases follow from one formula for the symplectic eigenvalue.","Since the entanglement Hamiltonian is a harmonic oscillator, all Rényi entropies are known analytically from the single number $\\lambda$, and $\\lambda$ is itself determined by the spin-squeezing parameter $\\xi_S$.","The periodically enhanced squeezing mechanism requires only a periodic reference orbit with nearby orbits of different periods, so the same four-regime variance phenomenology should appear in other fully connected mean-field models such as the Bose-Hubbard dimer."],"supporting_citations":[{"why":"Supplies the rate-function expansion, the classical equations for the mean, and the dynamical phase diagram that this paper extends by adding the variance.","marker":"[39]"},{"why":"Provides the nearby-orbit approximation that produces the Gaussian covariance evolution $C(t)=S(t)C(0)S(t)^T$ used for the variance and entanglement.","marker":"[53, 54]"},{"why":"Establishes the spin-squeezing–entanglement link and the standard quantum limit, which the paper makes quantitative through the spin-squeezing parameter.","marker":"[45]"},{"why":"Gives the star-exponential formula for a quadratic form, used to turn the Gaussian reduced Wigner function into the quadratic (harmonic-oscillator) entanglement Hamiltonian.","marker":"[85]"},{"why":"Provides the Williamson symplectic decomposition and the uncertainty bound needed to obtain the equidistant entanglement spectrum from the covariance matrix.","marker":"[86]"}],"fun_headline_variants":["Variance reveals four quench regimes in mean-field Ising model","Quench dynamics split into four variance regimes, mean-field fails at sqrt(N)","Four variance regimes in quenched Ising magnet, mean-field breaks down","Order parameter variance exposes four quench regimes; entanglement is harmonic","Variance, not mean, defines four phases in collective Ising quench"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The predictions rest on the assumption that the time-evolved state remains, to leading order in $1/N$, a Gaussian wave packet whose phase-space width stays of order $1/\\sqrt{N}$, so that the flow around the classical orbit can be linearized and anharmonic terms neglected for the times considered.","fun_headline_variants_meta":{"raw":{"variants":["Variance reveals four quench regimes in mean-field Ising model","Quench dynamics split into four variance regimes, mean-field fails at sqrt(N)","Four variance regimes in quenched Ising magnet, mean-field breaks down","Order parameter variance exposes four quench regimes; entanglement is harmonic","Variance, not mean, defines four phases in collective Ising quench"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001068,"raw_usage":{"total_tokens":4488,"prompt_tokens":973,"completion_tokens":3515,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":589,"completion_tokens_details":{"reasoning_tokens":3430}},"tokens_in":589,"tokens_out":3515,"duration_ms":22810,"temperature":1.0,"reasoning_tokens":3430,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:40:40.766007+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Use numerically exact time evolution for $N=10^3$ and $N=10^4$ on quenches in each of the four regimes and locate the first time $t_*$ at which $N\\,\\mathrm{var}(n_+)$ deviates from the semiclassical result by a fixed threshold; the claim is falsified if $t_*$ does not scale as $\\sqrt{N}$ in regimes III and IV and as $\\log N$ in regime I. A second check is the entanglement spectrum of the reduced density matrix: it must be equidistant with spacing $\\omega=2\\,\\mathrm{arctanh}[1/(2\\lambda)]$ as long as the semiclassical approximation holds.","supporting_citations":[{"cited_title":"Sciolla and G","cited_arxiv_id":null,"evidence_quote":"Supplies the rate-function expansion, the classical equations for the mean, and the dynamical phase diagram that this paper extends by adding the variance."}],"review_version":1}