{"id":"181aebf6-7334-4378-8a17-94c53d4caf97","arxiv_id":"1908.06007","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"Direct simulations show the chaotic Kuramoto-Sivashinsky equation's steady height fluctuations obey the Tracy-Widom and Baik-Rains limit distributions, placing it in the one-dimensional Kardar-Parisi-Zhang universality class.","lead":"The one-dimensional, deterministic Kuramoto-Sivashinsky equation is shown to produce height fluctuations that follow the same universal Tracy-Widom and Baik-Rains distributions as the stochastic Kardar-Parisi-Zhang equation. The result extends a central universality law of nonequilibrium statistical mechanics to a chaotic system with no external noise.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The reported Γ values for IC1 and IC2 violate the single-amplitude requirement of Eq. (1), so the IC-dependent TW/BR matches may not be on a common KPZ scaling trajectory.","rationale":"The reader's weakest assumption was the heuristic mapping from IC1-IC6 to the KPZ subuniversality geometries. I agree that this mapping is under-justified, but the more load-bearing concern is internal: the paper's own fitted amplitude Γ differs between IC1 and IC2. In the standard KPZ limit-distribution statement, Γ is a property of the equation, not of the initial condition; a single Γ must collapse all subuniversality classes after the appropriate χβ is used. The reported values 0.358 and 0.496 imply that the underlying measured variances are nearly equal, whereas TW-GUE and TW-GOE variances differ by about 27%. This suggests that the apparent agreement with the two TW distributions is achieved by an extra IC-dependent rescaling, which is not what Eq. (1) claims. This does not by itself prove the claim false, because the discrepancy could be within statistical error, but no error bars are given and no fixed-Γ test is reported. The proposed re-analysis of the existing data would settle the issue directly. I therefore keep the reader's CONDITIONAL verdict unchanged, with this specific internal check as the primary condition.","tokens_in":11758,"tokens_out":14260,"duration_ms":162992,"concrete_test":"Take the existing DNS time series for IC1 and IC2. Estimate Γ once, e.g., from the two-point correlation function or the flat initial condition, and form χ=(h−v∞t)/(Γt)^{1/3} for both. Compute the variance ratio and the Kolmogorov-Smirnov distance of each empirical CDF to its claimed TW distribution. If both ICs cannot simultaneously pass a KS test at 95% confidence with the same Γ, then the separate Γ values reported in Supplemental §2 are masking a failure of the single-amplitude scaling in Eq. (1). This requires no new runs, only re-analysis of the saved data.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Eq. (1) states that v∞ and Γ are model-dependent constants for the equation, with the initial condition selecting only χβ. Supplemental §2 reports Γ≈0.358 for IC1 (TW-GUE) and Γ≈0.496 for IC2 (TW-GOE), a 28% difference. With a common Γ, the raw variances of (h−v∞t)/t^{1/3} for IC1 and IC2 should be in the ratio Var(GUE)/Var(GOE)≈0.813/0.638≈1.27. The reported Γ pair implies only a ≈1.03 ratio of measured variances, so the two datasets do not collapse under one amplitude. If Γ is allowed to vary per initial condition, Eq. (1) is not the KPZ universality statement and the observed PDF matches contain an extra fitted parameter. This is an internal check, independent of the heuristic IC-to-geometry mapping; it calls into question whether KS actually exhibits the same limit distributions, or merely distributions that can be rescaled to resemble them.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper claims that height fluctuations in the spatiotemporally chaotic nonequilibrium steady state (NESS) of the deterministic one-dimensional Kuramoto-Sivashinsky (KS) equation obey the same limit distributions as the one-dimensional Kardar-Parisi-Zhang (KPZ) stochastic PDE. Specifically, for six initial conditions chosen to mimic wedge, flat, stationary, and mixed geometries of the ASEP/PNG models, the rescaled fluctuations χ = (h − v∞ t)/(Γ t)^{1/3} are reported to follow TW-GUE (IC1), TW-GOE (IC2), BR F0 (IC3), and other sub-universality distributions (IC4–IC6). The evidence is drawn from extensive GPU pseudospectral DNS at L = 2^20 with t up to 2–6×10^5, and includes semilog PDF comparisons, skewness/kurtosis plateaus, Family-Vicsek scaling, and a collapse of the two-point time-dependent correlation S(k,δt) against the Prähofer-Spohn prediction.","tokens_in":11958,"tokens_out":10326,"duration_ms":104417,"significance":"If the result holds, it would extend the KPZ universality class to the NESS of a deterministic chaotic PDE, going beyond the usual stochastic settings and establishing that the entire limit-distribution structure, not just scaling exponents, is universal. The numerical effort is substantial and the S(k,δt) comparison is new. However, as written the central claim is undercut by an internal inconsistency in the extracted amplitude Γ, and by missing specification of the initial conditions and quantitative goodness-of-fit tests.","major_comments":[{"comment":"The reported values Γ≈0.358 for IC1 and Γ≈0.496 for IC2 violate the single-amplitude requirement of Eq. (1), where v∞ and Γ are model-dependent constants independent of the initial condition. Under a common Γ, the ratio of the raw variances of (h−v∞t)/t^{1/3} for IC1 and IC2 should equal Var(TW-GUE)/Var(TW-GOE) ≈ 0.813/0.638 ≈ 1.27. The quoted Γs imply a raw variance ratio of only (0.358/0.496)^{2/3} × 1.27 ≈ 1.03, meaning the GOE/GUE identification is achieved only by allowing an extra IC-dependent amplitude. This internal check directly bears on whether Eq. (1) in the KPZ sense holds. The authors must reconcile this, e.g., by showing that Γ converges to a common value at longer times or by explicitly reformulating Eq. (1) to permit IC-dependent Γ and discussing how that is compatible with the KPZ universality claim.","section":"Supplement §2 (Fig. 3) and Eq. (1)"},{"comment":"The functional forms of IC1–IC6 are never specified. The text states that these initial conditions are chosen to mimic wedge, flat, stationary, wedge-to-stationary, wedge-to-flat, and flat-to-stationary geometries, but no formulas, tables, or precise descriptions of h(x,0) are given in either the main text or the Supplemental Material. Because the assignment of each IC to a particular sub-universality distribution (TW-GUE, TW-GOE, BR F0, etc.) depends on those forms, the comparison cannot be reproduced or critically evaluated. Please provide explicit definitions for all six initial conditions.","section":"Main text, initial conditions IC1–IC6 (Fig. 1)"},{"comment":"The paper provides no quantitative goodness-of-fit test for the PDF comparisons. Given the stated sample size (≈5×10^8 data points), the reported error bars are smaller than the symbols, so visual agreement over three decades does not establish that deviations from TW/BR are statistically negligible. Furthermore, the text says the PDFs are computed when the skewness and kurtosis are 'close to their standard values,' which is a selection criterion; the authors should demonstrate that the PDF shape is stable over a range of times and report, for example, Kolmogorov-Smirnov statistics or chi-square values for each IC.","section":"Figs. 1(d,h,l) and Fig. 4; text near Eq. (4)"},{"comment":"The claims for IC4–IC6 are not supported by comparisons to the relevant predicted distributions. For IC4 the text mentions a comparison with (FGOE)^2, but no such plot is shown for IC5 and IC6, and the spatial averaging over the intervals [7L/32,9L/32] and [23L/32,25L/32] is an inadequate substitute for sampling exactly at the meeting points, as the authors admit. If the central claim includes these mixed-geometry sub-classes, the evidence for them must be strengthened or the claim narrowed accordingly.","section":"Figs. 1(p,t,x); text near 'Stricly speaking'"}],"minor_comments":[{"comment":"There are numerous typographical errors: 'Stricly speaking' appears in both the main text and the Supplemental Material; 'unversal limit distributions' appears in the Conclusion; 'couterparts' appears in the Introduction; and 'the the Kardar-Parisi-Zhang' appears in the Introduction. These should be corrected.","section":"Throughout"},{"comment":"The sentence 'the skewness and kurtosis shown in Fig. 2' should refer to Fig. 3, not Fig. 2, since Fig. 2 displays S(k,δt) while the skewness/kurtosis are plotted in Fig. 3.","section":"Conclusion (last paragraph)"},{"comment":"The free-energy function F(h) is defined with F(χ,t) without a clear specification of the cumulative distribution used; the relation to the TW large-deviation functions would be clearer if the notation were defined explicitly and a reference to the KPZ large-deviation literature were added.","section":"Eq. (10)"},{"comment":"The Family-Vicsek plots for IC4–IC6 in the Supplemental Material show regimes labeled 'slope 1/2' in addition to 'slope 1/3', but this is not explained in the text. It would be helpful to clarify whether these are expected sub-leading corrections or artefacts of the spatial averaging procedure.","section":"Supplemental Material, §3 (Fig. 4)"},{"comment":"The plots of Σ(t) versus t in the Supplemental Material do not show the saturation plateaus from which Γ is extracted. Adding the plateau lines and error estimates would make the determination of Γ more transparent and would allow readers to assess the claimed convergence.","section":"Supplemental Material, §2 (Fig. 3)"}],"recommendation":"major_revision","confidential_remarks":"The paper addresses a high-interest problem and the numerical investment is substantial, with a very large system size and long integration times. However, the IC-dependent Γ is a serious internal inconsistency that, if unresolved, reduces the claim to 'PDF shapes can be rescaled to mimic TW/BR' rather than establishing the KPZ universality class in the full Eq. (1) sense. The missing specification of the initial conditions is also a significant reproducibility concern. I recommend major revision with a request that the authors provide a detailed response to the Γ check, full initial-condition definitions, and quantitative statistical tests."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The punchline first: this paper claims that the 1D Kuramoto-Sivashinsky equation, a deterministic chaotic PDE, falls in the 1D KPZ universality class at the level of the full limit distributions—TW-GUE, TW-GOE, and Baik-Rains—selected by the initial condition. If true, that extends the KPZ universality story well beyond stochastic growth models. The evidence for the first three initial conditions is visually strong: the rescaled height PDFs sit on the analytic curves over three decades, and the skewness and kurtosis level off at the TW/BR values. That is a genuinely new numerical result, as is the comparison of the KS two-point correlation function to the Prähofer-Spohn scaling form. The simulation effort is real (L=2^20, ETDRK4 on GPUs), so the paper is not a toy calculation.\n\nThe soft spots are real too. The most concrete one is internal. The supplement reports Γ≈0.358 for IC1 and Γ≈0.496 for IC2. Equation (1) says Γ is one constant for the model, with the initial condition only selecting χ_β. The measured variances then imply Var(GUE)/Var(GOE)≈1.27, but the two Γ's imply only ≈1.03. So the data are not collapsing under a single amplitude. If Γ is allowed to depend on the initial condition, the claim in Eq. (1) is not what is being tested; you have an extra fitted parameter for each PDF match. That does not kill the paper—finite-time corrections could explain the discrepancy—but the paper neither notices nor addresses it.\n\nSecondary issues: there are no goodness-of-fit statistics or error bars on the PDF comparisons, and v∞ and Γ are fitted from the same data the collapse is tested on, so the match is not parameter-free. No code or data are released. For IC4–IC6, the paper admits the spatial averaging it needs to get statistics already makes the PDFs fail to match the target distributions; those cases are not supporting evidence. And the mapping from the six KS initial conditions to the KPZ initial-condition geometries is asserted, not derived or tested.\n\nTaken together, the central claim may well be right, but the current manuscript overstates it. The cleanest fix is to test for a single Γ across all initial conditions, and to provide the statistical tools that let a reader judge the collapses.\n\nMy recommendation: send it to peer review; a knowledgeable referee might well be able to push this into a solid result or a clear negative. I would not cite it as established—treat it as a promising numerical observation. It is a good reading-group paper, though.\n\nBest,","headline":"KS may be in the KPZ universality class, but the case is weakened by an inconsistent amplitude parameter and overfitted collapses.","tokens_in":12515,"tokens_out":4791,"would_cite":false,"duration_ms":46816,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["82C31","82C05","60B20"],"pacs":["02.30.Jr","05.10.-a","47.70.-n","68.35.Rh","74.40.Gh"],"model":"deepseek-v4-flash","headline":"Height fluctuations in a deterministic chaotic equation match KPZ's universal laws.","keywords":["Kardar-Parisi-Zhang equation","Kuramoto-Sivashinsky equation","Tracy-Widom distribution","Baik-Rains distribution","universality class","nonequilibrium steady state","pseudospectral simulation","height fluctuations"],"falsifier":"Take the same IC1-IC3 runs at a larger linear size, for example $L = 2^{21}$, extend $t_{\\max}$ beyond $4\\times 10^5$, and re-extract $v_\\infty$ and $\\Gamma$ from the data; if the rescaled height PDF moves away from the Tracy-Widom and Baik-Rains curves in the tails, the claimed matches were finite-size or fitting artifacts.","tokens_in":11462,"feed_emoji":"📈","tokens_out":12737,"duration_ms":112599,"temperature":0.7,"pith_summary":"The paper sets out to show that the one-dimensional Kuramoto-Sivashinsky equation, a deterministic partial differential equation with no noise term, reproduces the universal limit distributions of the one-dimensional Kardar-Parisi-Zhang universality class once it settles into its spatiotemporally chaotic, statistically steady state. In that state the height difference $h(x,t)-h(x,0)$ grows as $v_\\infty t + (\\Gamma t)^{1/3}\\chi_\\beta$, with $\\chi_\\beta$ distributed according to the Tracy-Widom GUE law for a wedge-like initial condition, the Tracy-Widom GOE law for a flat interface, and the Baik-Rains $F_0$ law for a stationary interface. The consequence a curious reader should care about is that the full single-point distribution, not just scaling exponents, is shared by a deterministic chaotic system and by stochastic growth models.","feed_headline":"Deterministic chaotic equation matches KPZ's exact universal laws","feed_subtitle":"In its turbulent steady state, the Kuramoto-Sivashinsky equation reproduces Tracy-Widom and Baik-Rains height statistics.","key_machinery":"The load-bearing object is the one-point height-fluctuation relation Eq. (1) of the 1D KPZ class, applied to the deterministic Kuramoto-Sivashinsky equation (2), $\\partial_t h + \\Delta h + \\Delta^2 h + \\tfrac12(\\nabla h)^2 = 0$. The argument is carried by pseudospectral direct numerical simulation with a fourth-order exponential time-differencing scheme at $L = 2^{20}$, a set of initial conditions crafted to mimic wedge, flat, stationary, and mixed interface geometries, and the rescaling $\\chi = (h(x,t)-v_\\infty t)/(\\Gamma t)^{1/3}$, with $v_\\infty$ and $\\Gamma$ extracted from the same runs. The distributions being matched are the Tracy-Widom laws from the largest-eigenvalue statistics of GOE and GUE random matrices and the Baik-Rains $F_0$ distribution.","core_discovery":"On the paper's own terms, the central discovery is that the spatiotemporally chaotic nonequilibrium steady state of the deterministic 1D KS PDE lies in the 1D KPZ universality class in the strong sense of Eq. (1): $h(x,t)-h(x,0) = v_\\infty t + (\\Gamma t)^{1/3}\\chi_\\beta + o(t^{1/3})$. By choosing six initial conditions IC1-IC6 to mimic the wedge, flat, stationary, and mixed geometries studied in the asymmetric simple exclusion process, and running pseudospectral simulations at linear size $L = 2^{20}$ to times $t_{\\max} \\geq 2 \\times 10^5$, the authors obtain rescaled height PDFs that collapse onto TW-GUE (IC1), TW-GOE (IC2), and BR $F_0$ (IC3) over more than three orders of magnitude, with skewness and kurtosis approaching the corresponding random-matrix values. They also report the compensated spectrum, interface-width scaling with $\\beta \\simeq 0.32$, and a time-dependent two-point correlation function $S(k,\\delta t)$ consistent with the analytic 1D KPZ scaling form. This is presented as the first observation of these limit distributions in a spatiotemporally chaotic steady state of a deterministic PDE.","pith_inferences":["A testable extension would be to measure the large-deviation tails of $P(\\chi)$ for the KS equation; the paper's formal free-energy argument suggests a third-order phase transition in the deterministic height field, which would connect deterministic chaos to random-matrix large-deviation theory.","If the claim holds generally, other deterministic chaotic equations whose height variable is governed by the same combination of nonlinear advection-like growth and dissipation should display the same limit distributions, making the laws a property of chaotic dynamics rather than of stochastic forcing.","For the mixed initial conditions IC4-IC6, the PDFs were assembled from spatial patches near the two meeting points; a sharper test would sample exactly at those points across many independent runs and compare with the predicted mixtures such as $(F_{\\rm GOE})^2$."],"forward_implications":["The 1D KS equation can serve as a deterministic laboratory for probing 1D KPZ universality, since the same limit distributions arise without averaging over stochastic noise realizations.","The initial-condition geometry, not the presence of noise, is what selects the Tracy-Widom GUE, GOE, or Baik-Rains law in this class.","The correspondence between the deterministic and stochastic systems extends beyond exponents to the complete scaling form of the time-dependent two-point correlation function.","The authors conjecture that the phase-chaos regime of the one-dimensional complex Ginzburg-Landau equation will show the same KPZ limit distributions."],"supporting_citations":[{"why":"It defines the stochastic Kardar-Parisi-Zhang equation whose universality class is the object of comparison.","marker":"[33]"},{"why":"It established the Tracy-Widom and Baik-Rains limit distributions for the poly-nuclear growth model, the curves the DNS is matched against.","marker":"[40]"},{"why":"It supplies the Tracy-Widom distribution functions for GUE and GOE used as target PDFs for the wedge and flat initial conditions.","marker":"[48]"},{"why":"It supplies the Baik-Rains F0 distribution used as the target PDF for the stationary initial condition.","marker":"[49]"},{"why":"It reviews the initial-condition-dependent subuniversality classes of the 1D KPZ class that motivate the choice of IC1-IC6.","marker":"[44]"},{"why":"It provides the analytic scaling form of the time-dependent two-point correlation function used for comparison with the KS data.","marker":"[50]"},{"why":"It is an earlier KS direct numerical simulation showing KPZ scaling of correlations, the result the present paper extends to full distributions.","marker":"[28]"},{"why":"It is an earlier KS simulation of long-wavelength properties that established the crossover scales the present runs must exceed.","marker":"[29]"}],"fun_headline_variants":["Chaotic PDE hits KPZ's universal limits","Deterministic chaos obeys KPZ's Tracy-Widom laws","Steady-state chaos matches KPZ random growth stats","KS turbulence reproduces KPZ's Baik-Rains and Tracy-Widom","One chaotic equation matches KPZ universal distributions"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the chosen initial conditions impose the same long-time interface geometries—wedge, flat, stationary—that select the Tracy-Widom and Baik-Rains laws in stochastic growth models; if the chaotic equation does not preserve those geometries, the observed distribution matches would be coincidental.","fun_headline_variants_meta":{"raw":{"variants":["Chaotic PDE hits KPZ's universal limits","Deterministic chaos obeys KPZ's Tracy-Widom laws","Steady-state chaos matches KPZ random growth stats","KS turbulence reproduces KPZ's Baik-Rains and Tracy-Widom","One chaotic equation matches KPZ universal distributions"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000767,"raw_usage":{"total_tokens":3415,"prompt_tokens":974,"completion_tokens":2441,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":590,"completion_tokens_details":{"reasoning_tokens":2358}},"tokens_in":590,"tokens_out":2441,"duration_ms":17323,"temperature":1.0,"reasoning_tokens":2358,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:19:52.065586+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take the same IC1-IC3 runs at a larger linear size, for example $L = 2^{21}$, extend $t_{\\max}$ beyond $4\\times 10^5$, and re-extract $v_\\infty$ and $\\Gamma$ from the data; if the rescaled height PDF moves away from the Tracy-Widom and Baik-Rains curves in the tails, the claimed matches were finite-size or fitting artifacts.","supporting_citations":[{"cited_title":"Pr\\\"ahofer \\ and\\ author H","cited_arxiv_id":null,"evidence_quote":"It established the Tracy-Widom and Baik-Rains limit distributions for the poly-nuclear growth model, the curves the DNS is matched against."},{"cited_title":"Pr \\\"a hofer \\ and\\ author H","cited_arxiv_id":null,"evidence_quote":"It provides the analytic scaling form of the time-dependent two-point correlation function used for comparison with the KS data."},{"cited_title":"Sneppen , author J","cited_arxiv_id":null,"evidence_quote":"It is an earlier KS direct numerical simulation showing KPZ scaling of correlations, the result the present paper extends to full distributions."},{"cited_title":"Hayot , author C","cited_arxiv_id":null,"evidence_quote":"It is an earlier KS simulation of long-wavelength properties that established the crossover scales the present runs must exceed."}],"review_version":1}