{"id":"b96e5621-67c8-4d40-9de9-a1ed8d45cf3e","arxiv_id":"2502.01622","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Perturbations around nonthermal fixed points obey a quasinormal-mode eigenvalue problem, and for a Fokker-Planck collision kernel the spectrum is a tower of purely imaginary frequencies.","lead":"The paper proposes that the approach to a nonthermal fixed point, a transient self-similar state in far-from-equilibrium systems, is governed by quasinormal modes analogous to black hole ringdown. For a Fokker-Planck kinetic theory, it computes these modes and finds a tower of decaying power-law perturbations.","discovery_kind":"first_principles","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Claimed damped tower may be an artifact of the UV cutoff and δϵ=0 condition; full-grid spectrum differs and no time-domain check is provided.","rationale":"The paper's framework is elegant and the two analytically derived modes (Ω=−7i from t_* shifts and Ω=0 from t_ref shifts) are reproduced numerically, which is genuine internal support. However, those modes are guaranteed by the scaling structure and do not validate the three non-analytic frequencies that carry the paper's main quantitative claim. The load-bearing step is the extraction of the spectrum from a finite matrix after imposing δϵ=0 and a UV cutoff. The manuscript itself shows that the infinite-domain calculation gives a qualitatively different spectrum (an integer imaginary tower, Appendix 4), and the justification for discarding those modes is the unboundedness of δf_Ω/f_s—the same unboundedness that motivates the cutoff. Since the eigenvalue problem with a hard cutoff is not a restriction of the original operator, there is no guarantee that the surviving eigenvalues are physical; only a boundary-condition-independent or time-domain verification would settle this. This is exactly the reader's weakest assumption, and it remains unresolved. The verdict stays CONDITIONAL: the framework and analytic modes are solid, but the numerical tower should not be taken as a definitive prediction until the regularization independence is demonstrated.","tokens_in":16228,"tokens_out":8515,"duration_ms":78932,"concrete_test":"Run a direct time-domain simulation of the linearized equation (10) on a large but finite grid with the same UV cutoff, starting from a compactly supported perturbation with δϵ=0; at each fixed scaled momentum p̄, fit log|δf(t,p̄)| versus log B(t) at late times and compare the slope to Re(iΩ) for the claimed modes, in particular −4.29i. As a complementary spectral check, repeat the eigenvalue calculation at a fixed cutoff but replace the δϵ=0 row with a zero-flux boundary condition at p_UV; if the low-lying eigenvalues move by O(1), the spectrum is not regularization-independent.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central numerical claim—the damped tower Ω≈−4.29i, −6.49i, −7i, −8.59i for the energy-conserving FP kernel—rests on a finite-grid eigenvalue calculation with an artificial UV cutoff and a δϵ=0 boundary condition. The paper's own Appendix 4 (Fig. 7) shows that on the full compact grid (no UV cutoff) the spectrum contains an additional set of modes at every negative integer imaginary multiple, and these are discarded because δf_Ω/f_s diverges as p→∞. Once a finite UV cutoff is introduced, the generalized eigenvalue problem is no longer a restriction of the original infinite-domain operator: the boundary condition at the cutoff selects a discrete subset of modes. The paper asserts robustness to the cutoff value and gridpoints, but it does not quantify the cutoff dependence, does not test other admissible boundary conditions at the cutoff (for example zero flux at p_UV or matching to the known Gaussian tail of f_s), and does not provide a time-domain simulation showing that these modes govern the relaxation of generic δϵ=0 initial data. If the surviving modes are an artifact of the regularization, the central claim that a physical decaying tower governs the approach to the NTFP is unsupported.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper proposes that the approach to a nonthermal fixed point (NTFP) in kinetic theory can be described by a quasinormal-mode expansion. Writing f(t,p)=A(t)[fs(B(t)p)+δf(t,p̄)], the authors derive the linear eigenvalue problem in Eqs. (5) and (10) for perturbations of the form δf=B(t)^{iΩ}δf_Ω(p̄). They identify two universal modes: a shift in t* with Ω=i/β and a zero mode from rescaling t_ref, and they show how energy conservation constrains the allowed modes. For the overoccupied Fokker-Planck kernel with energy cascade (σ=4, β=−1/7), they compute the scaling function fs by a pseudospectral method, find a Gaussian UV tail, and solve the eigenvalue problem numerically on grids with a UV cut-off, reporting a tower of purely imaginary frequencies Ω≈−4.29i, −6.49i, −7i, −8.59i. They verify the analytic Ω=−7i mode and compare the zero-mode eigenfunction with numerics, and they include appendices on a particle-number-conserving cascade toy model.","tokens_in":16400,"tokens_out":11312,"duration_ms":100845,"significance":"If the central numerical claim is correct, the paper introduces a genuinely new diagnostic for the attractive nature of NTFPs and connects far-from-equilibrium kinetic theory with the well-developed quasinormal-mode toolkit of black-hole perturbation theory. The analytic derivation of the t* mode and the zero mode is clean and largely independent of the collision kernel, and the verification of the Ω=−7i mode is a strong internal consistency check. The self-consistent determination of the scaling function with a Gaussian tail corroborates earlier ab initio results [40], and the inclusion of a Mathematica notebook supporting the analytic solution in Appendix 2 is a reproducibility plus. The main reservation is that the reported damped tower—the quantitative centerpiece—rests on a finite cut-off and a boundary-condition selection whose validity is not yet demonstrated; if that selection is not robust, the physical claim of a power-law tower governing the approach to the NTFP remains unproven.","major_comments":[{"comment":"The central numerical claim—the damped tower Ω≈−4.29i, −6.49i, −7i, −8.59i—rests on the finite-grid generalized eigenvalue problem with a UV cut-off and the condition δϵ=0. Appendix 4 (Fig. 7) shows that the full compact-grid spectrum contains an additional set of modes at every negative integer imaginary multiple, and these are discarded because δf_Ω/f_s diverges for large p. A finite cut-off changes the operator domain, so the surviving modes are selected by the boundary condition at p_UV rather than by the infinite-domain physics. The paper asserts robustness to the cut-off and grid size but does not quantify the convergence of Ω with p_UV, does not test alternative boundary conditions at p_UV (zero flux, or matching to the Gaussian tail of f_s), and does not explain the full-grid modes. If the surviving tower is an artifact of the regularization, the assertion that a power-law tower governs the approach to the NTFP is unsupported. I request convergence data over a sequence of cut-offs and a time-domain evolution of δϵ=0 perturbations whose extracted decay exponents can be compared with the reported Ω values.","section":"Far from equilibrium QNMs spectrum; Appendix 4"},{"comment":"The procedure of eliminating all modes with Im Ω>0 by imposing δϵ=0 is not validated. Energy conservation only forces δϵ(t) to be constant, i.e., either Ω=0 or δϵ_Ω=0 (Eq. (14)); the unstable modes seen without the condition must therefore be shown to have δϵ_Ω≠0 and to be numerical artifacts rather than genuine dynamics. Because the attractor property of the NTFP is exactly the absence of physical growing modes, this selection is load-bearing and needs a direct consistency check, for example by reporting δϵ_Ω for the discarded eigenvectors and by demonstrating their disappearance under a controlled regularization.","section":"QNMs and conservation laws; spectrum section"},{"comment":"The input scaling function is itself selected by imposing Ĩa[gs]=Ĩb[gs] on a compact grid, which forces a Gaussian UV tail, while the presence of an IR cut-off is shown in Fig. 2 to produce a 1/p̄⁴ tail with divergent energy density. The paper motivates a UV cut-off from this discussion, but the QNM eigenfunctions have relative growth in the UV even for the Gaussian-tail scaling function, so the spectrum may depend on the tail selection and on the cut-off beyond what is asserted. A concrete test would be to repeat the eigenvalue calculation on the finite interval with the power-law-tail scaling function and a physical boundary condition, and to show either that the reported Ω values are unchanged or to quantify the change.","section":"Accurate determination of the scaling function; Fig. 2"}],"minor_comments":[{"comment":"\"Focker-Planck\" should be \"Fokker-Planck\"; similar typos include \"descretize\" in the Setup section, \"Gayuss-Lobatto\" in Appendix 3, and \"caclulated\" in the conservation-laws section.","section":"Abstract and Eq. (16)"},{"comment":"Reference [38] is incomplete (no publisher or year), and the arXiv field in reference [18] appears malformed; please correct the bibliographic entries.","section":"References"},{"comment":"The statement that the boundary condition \"disposes of the predicted zero-mode Ω=0\" and the subsequent calculation comparing the Ω=0 eigenfunction are confusing: if the mode is removed from the spectrum, its status as a physical perturbation should be clarified explicitly.","section":"Zero-mode discussion"},{"comment":"The caption says \"the y axis multiplied by p̄^{-1}/√(D2 Ĩa[gs])\", which reads like a division rather than a multiplication; please clarify how the plotted quantity relates to p̄ fs.","section":"Fig. 1 caption"}],"recommendation":"major_revision","confidential_remarks":"The analytic part of the paper is solid and the presentation is generally clear, but the numerical spectrum—the quantitative centerpiece—needs to be hardened. The authors' own admission in Appendix 4 that the full-grid spectral structure is not understood is more than a cosmetic caveat; it directly concerns whether the cut-off-selected tower is physical. I would recommend requiring the convergence and time-domain checks outlined in the major comments before publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The two things you should know: this paper puts a genuinely new tool on the table—quasinormal modes for the approach to nonthermal fixed points—and the core derivation is clean. But the concrete spectrum for the Fokker–Planck kernel is conditional on a boundary-condition choice that the paper itself only partially justifies. I read the stress-test note before the manuscript, and it overstates the risk: the authors don't hide the full-grid spectrum; they show it in Fig. 7 and explain why they discard those modes. The worry about no time-domain check is fair, though.\n\nWhat's actually new: the B(t)^{iΩ} ansatz and the eigenvalue problem (Eq. 10) give a spectral language for relaxation to NTFPs, parallel to black hole QNMs. Two universal modes, Ω=i/β from a shift in t* and Ω=0 from a shift in tref, are derived by pure symmetry and are hard to argue with. The conservation-law argument that δϵ=0 removes growing modes is physically motivated. A byproduct—the numerical determination of the scaling function with a Gaussian UV tail matching [40]—is a nice result in its own right.\n\nWhere it gets soft: the FP spectrum (Ω ≈ −4.29i, −6.49i, −7i, −8.59i) comes from a finite grid with a UV cutoff and δϵ=0 imposed. The authors show robustness to the cutoff value, but they don't quantify how quickly the spectrum converges, don't test other admissible boundary conditions at the cutoff (zero flux, matching the Gaussian tail), and don't present a time-domain simulation showing these modes actually govern relaxation of generic δϵ=0 initial data. The appendix's full-grid spectrum has extra modes at every negative integer imaginary multiple, which are dropped because δf_Ω/f_s diverges in the UV. The paper's defense—overoccupation only holds between cutoffs, and the Gaussian tail makes the UV boundary contribution negligible—is reasonable but not airtight. If that reasoning is wrong, the tower is an artifact of the regularization and the attractor claim loses its support. So the right verdict is conditional, not accept.\n\nWho it's for: anyone working on kinetic theory, prescaling, or nonthermal fixed points; also people who enjoy seeing black hole technology exported. I'd cite it for the framework, and I think it deserves a serious referee. The referee should push for a time-domain check and a more careful treatment of the UV boundary condition.","headline":"New, clean framework for quasinormal modes of nonthermal fixed points, but the concrete FP spectrum rests on a regularization choice the authors only partially justify; worth a serious referee.","tokens_in":17008,"tokens_out":2549,"would_cite":true,"duration_ms":23750,"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":"Perturbations of a nonthermal fixed point decay through a tower of power-law quasinormal modes.","keywords":["nonthermal fixed points","quasinormal modes","Fokker-Planck kinetic theory","self-similar scaling","Boltzmann equation","power-law relaxation","thermalization","prescaling"],"falsifier":"Solve the linearized eigenvalue problem (10) on successively finer grids with larger ultraviolet cutoffs but without imposing $\\delta\\epsilon=0$; if any mode with $\\mathrm{Im}\\,\\Omega>0$ converges with a cutoff-independent eigenfunction carrying nonzero energy density, the attractor picture is wrong. Equivalently, an ab initio kinetic-theory simulation initialized close to the Fokker-Planck scaling solution should show late-time deviations in the static frame decaying as a sum of the predicted power laws, with the least-damped mode giving a decay $(t-t_*)^{4.29\\beta}\\approx (t-t_*)^{-0.61}$ for $\\beta=-1/7$; a different late-time behavior would falsify the tower.","tokens_in":15952,"feed_emoji":"⚛️","tokens_out":15391,"duration_ms":113476,"temperature":0.7,"pith_summary":"The paper proposes that the approach to a nonthermal fixed point—a transient attractor in which the distribution function factorizes as $f(t,p)=A(t)f_s(B(t)p)$—is organized by a discrete spectrum of quasinormal modes, complex eigenfrequencies whose imaginary parts set decay rates, in the same spirit as black-hole ringdown. Its central claim is that a small perturbation of the self-similar scaling solution evolves as $\\delta f(t,\\bar p)=B(t)^{i\\Omega}\\delta f_\\Omega(\\bar p)$, with eigenfrequencies $\\Omega$ fixed by a linearized eigenvalue problem around the scaling function. For the Fokker-Planck collision kernel with energy conservation the spectrum is a tower of purely imaginary frequencies, so each mode contributes a decaying power law in time; the least-damped mode sets the late-time approach. If correct, this gives a coupling-independent description of how overoccupied gluon plasmas and cold-atom gases lock onto self-similar cascades, and it explains the observed attractiveness of nonthermal fixed points. A byproduct is a precise determination of the scaling function, whose ultraviolet tail is Gaussian.","feed_headline":"Power-law tower governs approach to nonthermal fixed points","feed_subtitle":"Kinetic-theory calculation predicts the quasinormal-mode spectrum behind universal self-similar cascades.","key_machinery":"The key object is the quasinormal-mode ansatz $\\delta f(t,\\bar p)=B(t)^{i\\Omega}\\delta f_\\Omega(\\bar p)$, inserted into the Boltzmann equation linearized around the nonthermal fixed point. This converts a time-dependent approach to self-similarity into a time-independent generalized eigenvalue problem, Eq. (10), whose eigenvalues $\\Omega$ encode power-law decay when $\\mathrm{Im}\\,\\Omega<0$ and logarithmic oscillations when $\\mathrm{Re}\\,\\Omega\\neq0$. The spectrum is obtained by Chebyshev pseudospectral discretization on a compact momentum grid, with the scaling function $f_s$ determined self-consistently from the Fokker-Planck kernel; conservation of energy is enforced through the $\\delta\\epsilon=0$ boundary condition, and an ultraviolet cutoff is included because perturbation eigenfunctions grow relative to $f_s$ at large momentum.","core_discovery":"Equations (5) and (10) are the paper's main results. In the static frame of the scaling solution, defined by $\\bar p=B(t)p$ and $A(t)^{-1}f(t,p)\\approx f_s(\\bar p)+\\delta f(t,\\bar p)$, any perturbation obeys $\\delta f(t,\\bar p)=B(t)^{i\\Omega}\\delta f_\\Omega(\\bar p)$, where $\\Omega$ is an eigenvalue of the linearized Boltzmann operator. For the isotropic, homogeneous Fokker-Planck kernel with energy conservation ($\\sigma=4$, $\\beta=-1/7$), the computed spectrum consists of purely imaginary frequencies, approximately $\\Omega\\approx -4.29i$, $-6.49i$, $-7i$, $-8.59i$, with the $-7i$ mode following analytically from a shift of the time offset $t_*$. Because $\\mathrm{Im}\\,\\Omega<0$, every mode decays as a power law in $t-t_*$, producing a tower of progressively more decaying contributions. The authors impose the boundary condition that perturbations carry no energy density, $\\delta\\epsilon=0$, which removes modes with positive imaginary part and makes the fixed point an attractor; they also identify a zero mode coming from a shift of the reference time $t_{\\rm ref}$, the analogue of a variation of a thermodynamic variable in equilibrium.","pith_inferences":["The coupling independence suggests the same power-law tower should appear in any homogeneous isotropic system governed by Fokker-Planck-type small-angle scattering, which makes the spectrum a target for cold-atom experiments that track deviations from scaling in the static frame.","The treatment of growing modes as numerical artifacts is a choice; if grid refinement ever produced a convergent eigenmode with $\\mathrm{Im}\\,\\Omega>0$ and nonzero energy density, the nonthermal fixed point would be a saddle rather than an attractor, and only the transient wedge before the instability would be described by the tower.","The analogy with black-hole quasinormal modes invites transferring pseudospectrum and mode-sum techniques to this setting, which could quantify how robust the computed spectrum is to changing cutoffs, adding nonlinear corrections, or including the time dependence of the Coulomb logarithm."],"forward_implications":["At late times the approach to the nonthermal fixed point is a sum over power laws, with the least-damped mode dominating; the self-consistency condition $\\mathrm{Re}(i\\Omega)\\beta<0$ is satisfied by the computed spectrum.","The eigenfrequencies are independent of the coupling constant at leading order, so the relaxation rates are fixed by the scaling exponents rather than by the interaction strength.","The two universal modes predicted from the symmetries of the scaling ansatz — $\\Omega=i/\\beta$ for a shift of $t_*$ and $\\Omega=0$ for a shift of $t_{\\rm ref}$ — appear in the numerical spectrum, with the zero mode removed by the $\\delta\\epsilon=0$ condition.","The scaling function for the Fokker-Planck kernel has a Gaussian ultraviolet tail $\\propto \\bar p\\,e^{-\\bar p^2/2}$, corroborated by ab initio kinetic-theory simulations, and is essentially insensitive to an ultraviolet cutoff once that tail is resolved.","If any mode had $\\mathrm{Re}\\,\\Omega\\neq0$, it would induce oscillations in $\\log(t-t_*)$ in deviations from scaling; the energy-conserving Fokker-Planck case has none, while the toy particle-number-conserving cascade in the appendices does."],"supporting_citations":[{"why":"Supplies the scaling ansatz $A(t)=B(t)^\\sigma$, the exponent relation $B(t)=((t-t_*)/t_{\\rm ref})^\\beta$, and the prescaling framework that defines the nonthermal fixed point.","marker":"[28]"},{"why":"Establishes that homogeneous functional collision kernels produce exact self-similar nonthermal fixed points, which is the premise for the QNM construction.","marker":"[34]"},{"why":"Provides the near-equilibrium kinetic-theory treatment of linearized perturbations with complex eigenfrequencies and the sign convention used here.","marker":"[35]"},{"why":"Supplies the Chebyshev-Gauss-Lobatto pseudospectral method used to discretize the scaling-function equation and the eigenvalue problem.","marker":"[38]"},{"why":"Demonstrates the pseudospectral approach for quasinormal modes of holographic black holes, the numerical template adapted here.","marker":"[39]"},{"why":"Gives an ab initio kinetic-theory scaling function with $1/\\bar p$ infrared behavior and a Gaussian ultraviolet tail, used to corroborate the present scaling function.","marker":"[40]"},{"why":"Defines the QCD effective kinetic theory that the Fokker-Planck kernel approximates, fixing the physical setting for the energy cascade.","marker":"[41]"},{"why":"Earlier study of stable and unstable perturbations near nonthermal fixed points with 2PI techniques, the stability context the paper connects to in the outlook.","marker":"[44]"}],"fun_headline_variants":["Quasinormal modes of nonthermal fixed points","Power-law tower for off-equilibrium fixed points","Kinetic theory predicts fixed-point quasinormal decay","Nonthermal fixed points: quasinormal tower emerges"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing assumption is that the modes with positive imaginary part found numerically are artifacts that can be discarded by imposing $\\delta\\epsilon=0$ together with an ultraviolet cutoff; if those modes are physical, the nonthermal fixed point is not an attractor and the tower of decaying power laws does not govern the late-time approach.","fun_headline_variants_meta":{"raw":{"variants":["Quasinormal modes of nonthermal fixed points","Power-law tower for off-equilibrium fixed points","Kinetic theory predicts fixed-point quasinormal decay","Nonthermal fixed points: quasinormal tower emerges"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000148,"raw_usage":{"total_tokens":1193,"prompt_tokens":956,"completion_tokens":237,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":572,"completion_tokens_details":{"reasoning_tokens":176}},"tokens_in":572,"tokens_out":237,"duration_ms":3277,"temperature":1.0,"reasoning_tokens":176,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-09T14:47:42.043926+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Solve the linearized eigenvalue problem (10) on successively finer grids with larger ultraviolet cutoffs but without imposing $\\delta\\epsilon=0$; if any mode with $\\mathrm{Im}\\,\\Omega>0$ converges with a cutoff-independent eigenfunction carrying nonzero energy density, the attractor picture is wrong. Equivalently, an ab initio kinetic-theory simulation initialized close to the Fokker-Planck scaling solution should show late-time deviations in the static frame decaying as a sum of the predicted power laws, with the least-damped mode giving a decay $(t-t_*)^{4.29\\beta}\\approx (t-t_*)^{-0.61}$ for $\\beta=-1/7$; a different late-time behavior would falsify the tower.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the Chebyshev-Gauss-Lobatto pseudospectral method used to discretize the scaling-function equation and the eigenvalue problem."},{"cited_title":"Linearized nonequilibrium dynamics in nonconformal plasma","cited_arxiv_id":"1503.07149","evidence_quote":"Demonstrates the pseudospectral approach for quasinormal modes of holographic black holes, the numerical template adapted here."},{"cited_title":"Stable and unstable perturbations in universal scaling phenomena far from equilibrium","cited_arxiv_id":"2209.14883","evidence_quote":"Earlier study of stable and unstable perturbations near nonthermal fixed points with 2PI techniques, the stability context the paper connects to in the outlook."}],"review_version":1}