{"id":"64d212a6-88de-4f87-8830-65bb6334c347","arxiv_id":"2507.07600","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"The paper introduces a first-order nontrivial-resonance correction to the Rayleigh-Jeans distribution and reports strong-nonlinearity agreement with simulations on three model systems.","lead":"This paper derives a beyond-random-phase-approximation correction to the equilibrium mode distribution for strongly nonlinear systems, and validates it numerically on nonlinear Schrodinger, MMT, and FPUT-beta models. The corrected formula matches simulations at nonlinear strengths where the standard Rayleigh-Jeans and trivial-resonance predictions clearly fail.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"First-order expansion of the generating functional is not controlled at b≳200; Eq.15 resums the O(βλ) vertex into a frequency shift without a small-parameter estimate, so the claimed strong-nonlinearity validity needs a second-order check.","rationale":"The central claim is that Eq.15 gives the equilibrium occupation for strongly nonlinear systems. For that claim to follow from the derivation, the generating-functional expansion in Eqs.12–13 must be controlled at the parameters used in the numerical checks. The reader identified this as the weakest assumption, and my reading reinforces it: the paper never states the expansion parameter, never estimates the omitted O((βλ)²) terms, and the n0→n substitution in Eq.14 is a resummation that goes beyond first order. At b=200 with L=256 and T=0.1, βλ≈3.9, so linearizing exp(−βλV) is not obviously valid; the true small parameter must include occupation factors and resonance-phase-space suppression, but the text provides no bound. The claimed b>200 accuracy is therefore an empirical finding, not a consequence of the derivation. This is not grounds for rejection: the numerics are consistent and cover three model families, and a hidden small parameter could justify the expansion. A second-order computation would settle the issue. I also note that the paper repeatedly routes derivations to the Supplemental Material, including Eqs.12–13, the n0→n error claim in Eq.14, the average-frequency identity, and the full FPUT-β case; the missing SM is a completeness problem that reinforces the conditional status. I agree with the reader that CONDITIONAL is the right verdict, with the condition being an explicit small-parameter estimate or second-order check.","tokens_in":7580,"tokens_out":16797,"duration_ms":170994,"concrete_test":"Compute the second-order (two-vertex) correction to the four-point correlation in Eq.8 for the discrete NSE at L=256, T=0.1, μ=−0.2, b=200, using the same Gaussian contraction rules and the self-consistent tildeω_k. If this O((βλ)²) contribution is not small, say below 10% of the O(βλ) term of Eq.13 for the k-modes shown in Fig.1, then Eq.15 is not justified by the stated first-order derivation; if it is small, the truncation is controlled and the concern is resolved.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Eqs.12–13 truncate the generating functional at first order in βλ, with λ=b/2L. For the NSE checks the text claims validity up to b>200; with L=256 and T=0.1, βλ≈3.9, so βλ is not small. The actual expansion parameter must be βλ multiplied by resonance-weighted occupation factors, but no such combination is defined or bounded. Equation 14 then replaces n0 by n, saying this introduces only a higher-order error; that is valid only if the correction 8βλ²Σn0³ is small relative to tildeω_k−μ, and no estimate is supplied. Moreover, replacing n0 by n converts a first-order vertex correction into a fully resummed denominator, so Eq.15 is not strictly first-order in λ. The omitted second-order diagrams carry an extra βλ and an internal convolution sum; for low-k modes in Fig.1 the additional factor can be O(1), making the truncation unjustified from the text alone. Thus the stated validity at b≳200 is supported by the numerics, not by the derivation. The FPUT-β derivation is also deferred entirely to the absent Supplemental Material, so the analogous truncation cannot be audited.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a framework for computing equilibrium mode occupation distributions in strongly nonlinear systems, going beyond the random phase approximation (RPA). For the discrete nonlinear Schrödinger equation (NSE), the authors use the generalized equipartition theorem and a generating functional to derive a correction term ω*_k that appears in the denominator: n_k = k_B T / (tilde_ω_k − μ − ω*_k). The correction is attributed to nontrivial four-wave resonances and is analytically computed from the first-order expansion in the nonlinear coupling λ. Analogous formulas are presented for the Majda–McLaughlin–Tabak (MMT) model and the FPUT-β model, and numerical simulations are shown to agree with the theoretical predictions in strongly nonlinear regimes, including a frequency-splitting phenomenon in MMT34.","tokens_in":7825,"tokens_out":9207,"duration_ms":95590,"significance":"If the result holds, it would extend equilibrium distribution theory beyond the near-integrable and weak-nonlinearity limits where RPA-based approaches operate. The paper gives explicit, analytically computable formulas and demonstrates impressive agreement with simulations across three distinct models, including the prediction of asymmetric spectra and frequency splitting. The use of the generalized equipartition theorem as a starting point is elegant and potentially widely applicable. However, the derivation's perturbative truncation is not controlled, and the supporting derivations for two of the three models are deferred to an absent Supplemental Material, so the breadth of the claim currently rests on the numerical evidence plus an un-audited calculation.","major_comments":[{"comment":"Equations (12)–(15): The generating functional is expanded to first order in βλ, but the numerical tests reach b = 200, where βλ ≈ 3.9 for L = 256 and T = 0.1; thus βλ is not small. No dimensionless small parameter is identified, and the neglected O((βλ)^2) terms are not estimated. The replacement of n0 by n in Eq. (14) is asserted to introduce 'only a higher-order error' without a bound. Consequently, the claimed validity up to b ≳ 200 is not justified by the derivation itself; it is an empirical observation. Please provide a second-order calculation or a quantitative estimate of the neglected terms, and specify the actual expansion parameter (e.g., βλ multiplied by resonance-weighted occupation factors).","section":"RJ distribution corrected by nontrivial resonances"},{"comment":"The derivations of Eqs. (19), (20), and the FPUT-β expression are entirely deferred to the Supplemental Material, which is not included in the manuscript. Since the paper's claim of generality across distinct nonlinear systems depends on these results, the referee cannot audit the truncation, the coefficients, or the treatment of different resonance types. The Supplemental Material must be provided for review; without it, these sections are unverifiable.","section":"MMT system / FPUT-β system"},{"comment":"The transition from the first-order moment correction (Eq. (13)) to the modified equipartition relation (Eq. (14)) is not shown. In particular, how the coefficient 8 arises and why the sum over n0_1 n0_2 n0_3 appears linearly in the denominator are not derived. Please include the intermediate algebra in the main text or in the (provided) Supplemental Material.","section":"Eq. (14)"}],"minor_comments":[{"comment":"The notation 'F123' in Eq. (20) is confusing; since the right-hand side involves a sum over modes 1 and 2 with δ12_k, the quantity should be written as F_k or defined clearly with indices.","section":"Eq. (20)"},{"comment":"The caption says 'the theoretical prediction from Eq.19 (or Fig.20)'; this should be 'Eq. 20' rather than 'Fig. 20'.","section":"Fig. 3 caption"},{"comment":"The statement that under RPA the averaged contribution of residual nonlinear interactions to the frequency shift vanishes is not proven or referenced in the text; please add a brief derivation or a citation to justify this claim.","section":"Introduction"},{"comment":"The symbol b is used both for the nonlinearity parameter in the NSE Hamiltonian (b/2 |ψ|^4) and for the quartic coupling in the MMT34 model; this overloads notation and may confuse readers comparing formulas.","section":"Notation across models"},{"comment":"Figure 1 shows only b = 30 and b = 200; the claim that Eq. (15) remains accurate 'up to b > 200' would be more convincing with additional intermediate values (e.g., b = 50, 100, 150).","section":"Fig. 1"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is timely and the numerical results are striking, but the central derivation has a load-bearing gap: the first-order expansion in βλ is not controlled at the claimed strong-nonlinearity parameters, and the Supplemental Material containing key derivations is not provided. I recommend major revision with a clear request for a second-order check or an error estimate, and for the missing Supplemental Material to be supplied. The paper fits the journal's scope and the approach is original enough to merit further consideration, but the validity claim 'up to b > 200' currently rests entirely on simulation agreement."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here is my take. The genuinely new thing is the claim that the random phase approximation omits a specific class of nontrivial four-wave resonances, and that these produce a computable correction to the equilibrium occupation: n_k = k_B T / (tilde_omega_k - mu - omega*_k), with omega*_k given by Eq. 15 and analogous formulas for MMT and FPUT. As far as I know, the cited wave-turbulence and self-consistent-phonon literature stops at trivial resonances, so this is a real step. The method is also refreshingly direct: use exact equipartition, compute the fourth-order moment from a generating functional to first order in the nonlinearity, substitute back. The numerical agreement across three very different models—NSE, MMT4/MMT34, FPUT-beta—is remarkably good, up to b~200 for NSE and with a convincing qualitative feature, the frequency splitting in MMT34. Credit where due: that is a strong empirical case.\n\nThe soft spots are real and concentrated in the derivation. The expansion is truncated at first order in beta*lambda, but at the claimed validity boundary, b=200, L=256, T=0.1, beta*lambda is about 3.9, not small. The actual small parameter must be beta*lambda weighted by occupation factors and resonance sums, but the paper neither defines nor bounds it. The replacement of n0 by n in Eq. 14 is said to introduce only a higher-order error, but no estimate is given, and since the correction is then absorbed into the denominator, Eq. 15 is effectively a resummation of the first-order vertex. That can be legitimate—mean-field resummations work that way—but the authors need to show the second-order terms are subleading, or at least test the truncation numerically. As it stands, the claimed validity range rests on the numerics, not on the derivation.\n\nTwo smaller things. The FPUT-beta derivation and the detailed steps leading to Eqs. 12–14 are in the Supplemental Material, which is absent from the arXiv posting; that makes the paper hard to audit. And the theoretical curves depend on the total particle number N through the trivial-resonance shift; the text doesn't say whether N is taken from the simulations, computed self-consistently, or adjusted. Some error bars or ensemble statistics on the numerical spectra would also help.\n\nBottom line: the central physical idea is sound and the numerics are unusually persuasive. What's missing is a controlled truncation estimate plus the supplement. I would send this to a serious referee rather than desk-reject, and I would cite the nontrivial-resonance formula in my own work, with a caveat about the missing error analysis. Bring it to reading group.","headline":"A real candidate for beyond-RPA equilibrium distributions, with surprisingly good numerics, but the perturbation control and missing supplement need work before I'd trust the claimed range.","tokens_in":8361,"tokens_out":4215,"would_cite":true,"duration_ms":48744,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["05.20.-y","05.45.-a"],"model":"deepseek-v4-flash","headline":"A single correction term from nontrivial four-wave resonances fixes the equilibrium mode occupation formula for strongly nonlinear systems.","keywords":["equilibrium distribution","strong nonlinearity","random phase approximation","nontrivial resonances","generalized equipartition theorem","nonlinear Schrödinger equation","MMT model","FPUT-beta model"],"falsifier":"Directly compute the fourth-order moment $\\langle a_1 a_2 a_3^* a_4^*\\rangle$ in equilibrium NSE simulations at $b=200$ and compare it with the first-order expression (13); a deviation of more than a few percent would indicate that the truncation is not controlling the error even where Eq. (15) fits the distribution. Alternatively, measure the product $n_k(\\tilde{\\omega}_k - \\mu - \\omega^*_k)$ across many modes: if it depends noticeably on $k$, the reported closed form is not the whole story.","tokens_in":7343,"feed_emoji":"⚛️","tokens_out":6176,"duration_ms":56284,"temperature":0.7,"pith_summary":"The paper claims that the equilibrium occupation of a mode in a strongly nonlinear system is not the Rayleigh–Jeans form with a shifted frequency, but a rational expression with an extra correction $\\omega^*_k$ coming from nontrivial four-wave resonances. For the discrete nonlinear Schrödinger equation this takes the explicit form $n_k = k_B T / ( \\tilde{\\omega}_k - 8 \\beta \\lambda^2 \\sum^\\ast_{123} n^0_1 n^0_2 n^0_3 \\delta^{12}_{3k} - \\mu )$, which the paper finds accurate in numerical simulations up to $b \\gtrsim 200$. The same framework yields corrected distributions for the Majda–McLaughlin–Tabak models and the FPUT-$\\beta$ chain, and it predicts average mode frequencies in agreement with simulation. A sympathetic reader would care because this extends equilibrium statistical mechanics beyond the weakly nonlinear regimes where the random phase approximation and Wick's theorem apply.","feed_headline":"Equilibrium distribution formula now extends into strong nonlinearity","feed_subtitle":"The first-order four-wave correction keeps mode occupations accurate up to b = 200 in NSE, MMT, and FPUT-beta models.","key_machinery":"The machinery is a generating functional $Z(J)$ built by coupling the Hamiltonian to two real external sources per mode, then expanding $e^{-\\beta \\lambda H_{\\rm int}}$ to first order in $\\lambda$. The first-order term gives the fourth-order moment $\\langle a_1 a_2 a_3^* a_4^* \\rangle_1 \\simeq -4\\beta\\lambda\\,\\delta^{12}_{34} \\prod_{j=1}^4 [\\beta(\\tilde{\\omega}_j-\\mu)]^{-1}$, which is then inserted into the generalized equipartition identity. This converts the otherwise vanishing nontrivial-resonance contribution into a computable occupation-number correction. The method defines the correction as an averaged distortion of the spectral structure, not a uniform frequency shift.","core_discovery":"The central discovery is that nontrivial four-wave resonances contribute a nonzero, analytically computable term $\\omega^*_k$ to the denominator of the equilibrium distribution, so that $n_k = k_B T / (\\tilde{\\omega}_k - \\mu - \\omega^*_k)$. For the discrete NSE, $\\omega^*_k = 8\\beta\\lambda^2 \\sum^\\ast_{123} n^0_1 n^0_2 n^0_3 \\delta^{12}_{3k}$, where the star excludes trivial resonances. The paper demonstrates numerically that this formula fits the exact equilibrium occupation over a wide range of nonlinear strengths, whereas the trivial-resonance renormalization fails when $b$ becomes large. In the MMT34 model the correction even reproduces a three-peaked frequency spectrum, indicating that the correction is a genuine spectral-structure effect rather than a uniform frequency shift.","pith_inferences":["Editorial inference: the same first-order generating-functional closure could be applied to other quartic Hamiltonians with different dispersion relations, predicting where non-Lorentzian multi-peaked spectra should appear.","Editorial inference: the absence of a small parameter suggests there may be a resummation of the expansion—for instance, a self-consistent replacement of $n^0$ with $n$ in the numerator—that would extend validity beyond the tested range, a testable extension.","Editorial inference: the deviation from Wick's theorem implied by Eq. (13) can be measured in the same simulations by computing the excess kurtosis of the mode amplitudes; observing that it tracks $\\omega^*_k$ would corroborate the colored-noise interpretation."],"forward_implications":["If correct, the formula gives an explicit closed-form equilibrium distribution for discrete NSE, MMT4, MMT34, and FPUT-$\\beta$ in strongly nonlinear regimes.","Because the correction is computed from the trivial-resonance-renormalized occupations $n^0_k$, the method turns the strong-nonlinearity problem into a self-consistent algebraic calculation.","The average mode frequency $\\bar{\\omega}_k$ follows from the same formula via $\\bar{\\omega}_k = k_B T / n_k + \\mu$, giving a prediction with no free parameters.","For the MMT34 model the observed three-peaked spectrum indicates that effective mode counting changes in strongly nonlinear systems; the corrected distribution nevertheless keeps ensemble averages valid."],"supporting_citations":[{"why":"supplies the renormalized-resonance-quartet baseline (trivial-resonance frequency shift) that the new nontrivial-resonance correction is compared against.","marker":"[24]"},{"why":"is the wave-turbulence textbook whose random-phase/renormalization treatment gives Eq. (2), the starting point.","marker":"[21]"},{"why":"defines the Majda–McLaughlin–Tabak model used for the MMT4 and MMT34 tests.","marker":"[28]"},{"why":"defines the FPUT-$\\beta$ chain used as the real-valued test system.","marker":"[29]"},{"why":"represents the variational/self-consistent phonon approaches that the paper contrasts with its own distribution formula.","marker":"[13]"},{"why":"is an example of the phonon Green's function method whose perturbative corrections are limited to weak nonlinearity.","marker":"[5]"}],"fun_headline_variants":["Four-wave correction extends equilibrium to strong nonlinearity","New analytic term fixes strong-nonlinearity equilibrium","Beyond RPA: accurate equilibrium for strong nonlinearity","Strong-nonlinear equilibrium from four-wave resonance term","Nontrivial resonances correct equilibrium in strong regime"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The derivation truncates the expansion of the Boltzmann weight at first order in $\\lambda$ and replaces $n^0_k$ by $n_k$ in Eq. (14) with no controlled error estimate, so the accuracy at strong coupling (where $\\beta\\lambda$ is of order one) rests on an unquantified approximation.","fun_headline_variants_meta":{"raw":{"variants":["Four-wave correction extends equilibrium to strong nonlinearity","New analytic term fixes strong-nonlinearity equilibrium","Beyond RPA: accurate equilibrium for strong nonlinearity","Strong-nonlinear equilibrium from four-wave resonance term","Nontrivial resonances correct equilibrium in strong regime"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000228,"raw_usage":{"total_tokens":1420,"prompt_tokens":838,"completion_tokens":582,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":454,"completion_tokens_details":{"reasoning_tokens":506}},"tokens_in":454,"tokens_out":582,"duration_ms":6831,"temperature":1.0,"reasoning_tokens":506,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T18:38:56.048054+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Directly compute the fourth-order moment $\\langle a_1 a_2 a_3^* a_4^*\\rangle$ in equilibrium NSE simulations at $b=200$ and compare it with the first-order expression (13); a deviation of more than a few percent would indicate that the truncation is not controlling the error even where Eq. (15) fits the distribution. Alternatively, measure the product $n_k(\\tilde{\\omega}_k - \\mu - \\omega^*_k)$ across many modes: if it depends noticeably on $k$, the reported closed form is not the whole story.","supporting_citations":[{"cited_title":"Renor- malized resonance quartets in dispersive wave turbulence","cited_arxiv_id":null,"evidence_quote":"supplies the renormalized-resonance-quartet baseline (trivial-resonance frequency shift) that the new nontrivial-resonance correction is compared against."},{"cited_title":"Wave turbulence , volume 825","cited_arxiv_id":null,"evidence_quote":"is the wave-turbulence textbook whose random-phase/renormalization treatment gives Eq. (2), the starting point."},{"cited_title":"A one-dimensional model for dispersive wave tur- bulence","cited_arxiv_id":null,"evidence_quote":"defines the Majda–McLaughlin–Tabak model used for the MMT4 and MMT34 tests."},{"cited_title":"Studies of the nonlinear problems","cited_arxiv_id":null,"evidence_quote":"defines the FPUT-$\\beta$ chain used as the real-valued test system."},{"cited_title":"Renormalized phonons in nonlinear lat- tices: A variational approach","cited_arxiv_id":null,"evidence_quote":"represents the variational/self-consistent phonon approaches that the paper contrasts with its own distribution formula."},{"cited_title":"Self-consistent phonon formulation of anharmonic lattice dynamics","cited_arxiv_id":null,"evidence_quote":"is an example of the phonon Green's function method whose perturbative corrections are limited to weak nonlinearity."}],"review_version":1}