{"id":"f220677a-d57d-4885-9ae0-b86114bdfafc","arxiv_id":"2608.07242","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Spatiotemporal coherence measurements in continuous 2D polariton condensates show stretched-exponential decays with KPZ exponents on the excitonic side and power-law decays with the characteristic EW exponent ratio on the photonic side.","lead":"Experiments on 2D polariton condensates find coherence decays that switch between two universal scaling laws as the detuning and pump power are tuned. The results point to a crossover between Edwards-Wilkinson and Kardar-Parisi-Zhang universality in a continuous driven-dissipative system.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The KPZ/EW identification rests on an untested, partly manual analysis pipeline (hand-picked windows, free-threshold dispersive-branch filtering, tuned κ/C_0/y_0); no null test shows it would reject non-universal decays.","rationale":"Reading in good faith: the paper measures g(1) via Michelson interferometry, identifies stretched-exponential decays with β≈0.24 and χ≈0.39 in the excitonic-detuning regime and algebraic decays with a_s/a_t≈2 in the photonic-detuning regime, and shows collapses onto the KPZ and EW universal scaling functions. Independent support is real: raw and filtered data are both shown, collapses are repeated at several powers and at an independent detuning (δ=−6.8 meV), and the numerical simulations in Fig. S15 verify, in silico, the phase-dominance assumption behind Eq. 14. The strongest_claim, a crossover between two universality classes, therefore rests on two conditions: (i) the measured decays equal exp(−C_θθ/2) for the condensate phase in the fitted windows, and (ii) the extraction is robust to analysis choices. The reader identified (i); I judge it partially secured by the numerics and by the fact that at p=1.21 the raw data already show the KPZ window. Less secure is (ii): the pipeline contains hand-picked windows, a nonlinear FFT filter with free thresholds, and tuned constants κ, C_0, y_0, and the manuscript nowhere demonstrates that a non-universal decay injected into this pipeline would be rejected. The paper's own SI reports a systematic bias in β at δ=−6.8 meV that exceeds pure fit statistics, confirming that analysis choices matter. This is load-bearing because a sufficiently flexible pipeline could manufacture apparent KPZ/EW signatures from generic decays, which would falsify both legs of the claim; conversely, a passing blind-injection test would substantially strengthen it. For these reasons the concern is genuine, but the existing multi-dataset evidence and the in-silico checks keep the appropriate verdict at CONDITIONAL; the condition is that the pipeline survives the proposed null test and sensitivity sweep.","tokens_in":41522,"tokens_out":14879,"duration_ms":149112,"concrete_test":"Blind-injection control through the real pipeline: synthesize g(1) maps with experiment-realistic noise, short-time transient, and dispersive branches, but with non-universal decay laws—(i) pure exponential, (ii) power law with a_s/a_t=1, (iii) stretched exponential with β=0.5—and run them through the exact analysis (FFT filter, κ normalization, window selection, AIC classification, universal-function collapse), blind to the injected law. If any null model is classified as KPZ/EW with exponents near β≈0.24/χ≈0.39 or a_s/a_t≈2, the identification is underdetermined. Then sweep the real-data analysis by varying window boundaries and the filter thresholds of Fig. S8c by ±20%; if β, χ, a_s, a_t shift by more than the quoted ±1σ fit errors, the error bars understate the dominant systematic.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim requires that the decays fitted in the chosen windows are the condensate's universal phase correlations, and that the pipeline used to reveal them does not itself imprint the reported scalings. The weakest link is the latter. Three steps are manual or tuned: (1) the grey-shaded KPZ/EW windows in Figs. 2-3 are hand-picked, and the temporal KPZ window spans only ~2-35 ps (about 1.2 decades); (2) the dispersive-branch filter (SI S3.5) has free thresholds (green/red lines in Fig. S8c), and the main text states it can 'reveal otherwise masked KPZ stretched exponential decays' and 'shift the oblique asymptote and align it with the KPZ universal function'; (3) the KPZ collapse uses a renormalization κ (SI S3.6 warns an incorrect κ 'can lead to misleading conclusions') plus tuned scale factors C_0 and y_0 (Fig. 2 caption). The reader's concern about g(1)≈exp(−C_θθ/2) (Eq. 14) is partially mitigated by the numerics of Fig. S15; what is never tested is whether the pipeline, applied to genuinely non-universal inputs (exponentials, power laws with a_s/a_t≠2, stretched exponentials with wrong β), would still yield apparent KPZ/EW collapses and exponents. The known systematic bias at δ=−6.8 meV (Fig. S13: fitted β larger than 0.24, attributed to weak oscillations) shows fit-only error bars understate systematics. Without a null test and a window/filter sensitivity sweep, the extracted β≈0.24, χ≈0.39 and a_s/a_t≈2 may reflect the flexibility of the selection procedure rather than universality.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports interferometric measurements of the first-order coherence g(1)(Δr,Δt) of continuous two-dimensional polariton condensates at two cavity-exciton detunings and several pump powers. For a more excitonic detuning (ℏδ=−5.2 meV and −6.8 meV), the authors observe stretched-exponential decays with fitted exponents β≈0.24 and χ≈0.39, which they attribute to the 2D KPZ universality class, and they show collapses of the spatiotemporal data onto the KPZ universal scaling function. For a more photonic detuning (ℏδ=−12.6 meV), they observe algebraic decays with spatial and temporal exponents satisfying a_s/a_t≈2, which they attribute to the EW universality class. A statistical phase diagram as a function of detuning and pump power is constructed by comparing power-law and stretched-exponential fits. Numerical simulations of the generalized Gross-Pitaevskii equation with realistic parameters reproduce the two regimes and show that vortex-antivortex pairs dominate, supporting the interpretation that the observed phases are the nonequilibrium analogue of the ordered BKT phase.","tokens_in":41916,"tokens_out":4212,"duration_ms":46528,"significance":"If the central claim is correct, this is a substantial experimental advance: it would establish continuous two-dimensional driven-dissipative condensates as a platform for observing 2D KPZ universality and a crossover to EW scaling, going beyond prior 1D demonstrations and addressing the debate about vortex-induced destruction of KPZ order. The paper has genuine strengths: multiple independent datasets at several detunings and powers, raw and filtered collapses, comparison with independent theoretical exponents and scaling functions, a parameter-free linearized model for the threshold determination, and numerical simulations with realistic parameters that explicitly check the phase-only approximation and the pairing of vortices. The main weakness is that the identification of the universal regimes depends at several steps on manually chosen analysis choices: the grey-shaded scaling windows, the free-threshold dispersive-branch filter, the renormalization constant κ, and the scale factors C_0 and y_0. These choices are not accompanied by null tests or a systematic sensitivity analysis, so the claimed exponents may partly reflect the flexibility of the analysis pipeline.","major_comments":[{"comment":"The dispersive-branch filtering procedure is load-bearing for the KPZ claim, especially at lower powers, but it is not blinded or shown to be harmless. The main text states that filtering can 'reveal otherwise masked KPZ stretched exponential decays' and can 'shift the oblique asymptote and align it with the KPZ universal function'; the SI procedure uses red and green threshold lines chosen by inspection of the Fourier phase maps (Fig. S8c). A null test is needed: the same filtering and window-selection pipeline should be applied to synthetic g(1)(Δr,Δt) maps with known non-universal decays (e.g., simple exponentials, power laws with a_s/a_t≠2, stretched exponentials with β≠0.24) and the authors should show that the pipeline does not produce apparent KPZ exponents and collapses. Without such a test, the filtered spatial scaling in Fig. 2d-e and the asymptotic alignment in Fig. 2f cannot be fully distinguished from artifacts of the filter.","section":"Main text Fig. 2d-f; SI S3.5"},{"comment":"The KPZ and EW scaling windows are hand-picked grey-shaded regions, and the temporal KPZ window spans only about 2–35 ps, roughly 1.2 decades. The reported fit uncertainties on β_exp and χ_exp in the insets of Fig. 2 are only statistical fit errors and do not include uncertainty from window choice. This matters because the SI itself shows a systematic effect: at ℏδ=−6.8 meV (Fig. S13), the fitted β is larger than 0.24 and the deviation is attributed to weak oscillations, demonstrating that fit-only error bars understate the systematics. The authors should state the objective criteria for window selection and provide a sensitivity sweep of the fitted exponents as the window bounds are varied by plausible amounts.","section":"Fig. 2a-b and Fig. 3a-b; SI S3.8"},{"comment":"The collapse onto the KPZ universal scaling function uses the renormalization constant κ, which is obtained by extrapolating stretched-exponential fits to Δt=0, and the scale factors C_0 and y_0, which are adjusted to the experimental data. The SI correctly warns that an incorrect κ 'can lead to misleading conclusions', but the present test has at least two free parameters whose values are not independently measured. The authors should quantify the effective number of independent data points entering each collapse and demonstrate that alternative scaling functions or wrong exponents fail the same collapse test. As it stands, the agreement in Figs. 2c and 2f, while visually good, has an unknown false-positive rate.","section":"SI S3.6; Fig. 2 caption and Fig. S13 caption"},{"comment":"The extraction of universal exponents from g(1) relies on the approximation g(1)≈g_n^(1)⟨e^{iΔθ}⟩≈exp(−C_θθ/2), i.e., that density fluctuations and uncondensed-polariton emission are negligible in the selected windows. This is validated only numerically (Fig. S15 and related discussion), not by a direct experimental decomposition of the measured coherence into density and phase contributions. Since the filtering step removes the dispersive branches associated with uncondensed polaritons, the residual filtered signal is assumed to be the condensate phase contribution. This is a reasonable assumption, but it would be strengthened by analyzing the sensitivity of the extracted exponents to the filter strength and by showing that the conclusions are stable when the filtering is varied within a physically motivated range.","section":"SI Eq. (14); SI S5.2"}],"minor_comments":[{"comment":"The line for the KPZ dataset lists ℏδ=−15.2 meV, whereas the main text and the rest of the SI use ℏδ=−5.2 meV for the KPZ regime; this appears to be a typographical error and should be corrected.","section":"SI Table I"},{"comment":"The name Berezinskii is misspelled as 'Berezinski' in the abstract and introduction; the standard transliteration should be used consistently.","section":"Throughout"},{"comment":"The statistical comparison fixes the stretched-exponential exponent to the KPZ value β=0.24 while leaving the power-law exponent free; this asymmetry should be stated explicitly in the main text so that the phase diagram is not misinterpreted as a model selection over equally parameterized hypotheses.","section":"Fig. 4 and SI S3.7"},{"comment":"The statement that 'the KPZ regime is accessed starting from the EW regime by simultaneously increasing the interaction strength g_R via μ and the loss rate γ_2' would benefit from a small quantitative summary of how g_KPZ varies across the two parameter sets in Table II, since this is the physical mechanism behind the observed crossover.","section":"SI S5.3"}],"recommendation":"major_revision","confidential_remarks":"This is a strong and well-executed experimental paper with a conditional but defensible central claim. The main risk is that the analysis pipeline—hand-picked windows, free-threshold Fourier filtering, and tuned normalization constants—could imprint the reported KPZ/EW scalings. I would recommend requesting null tests on synthetic data and a systematic sensitivity analysis before considering acceptance. The paper is well within the scope of the journal and, if the robustness checks pass, would be an important contribution."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The thing to know about this paper is that it is the first experimental observation of 2D KPZ scaling in a continuous planar polariton condensate, and it also shows a tunable crossover to EW scaling. That is a genuinely important step: the earlier 1D result was from the same group, the 2D lattice attempt was inconclusive, and this work goes after the continuous system that had been theoretically debated. If the claim holds, it settles a real controversy.\n\nWhat it does well: the coherence measurements are extensive, with multiple powers and detunings, spatial and temporal decays, and data collapses onto the theoretical KPZ and EW scaling functions. The numerical simulations reproduce the main features and convincingly show that density fluctuations are subdominant in the selected windows, supporting the g(1)≈exp(−Cθθ/2) link. The vortex analysis is a real plus: it shows defects are short-lived, paired, and mostly outside the condensate center, which explains why they do not destroy quasi-ordering. The paper is honest about its systematic bias at one detuning and about the need to filter dispersive modes.\n\nSoft spots, in proportion: the KPZ temporal window is roughly 2–35 ps, about 1.2 decades. That is short for a universal exponent claim. The dispersive-branch filter has free thresholds, and the main text admits it can “reveal otherwise masked KPZ stretched exponential decays” and shift the oblique asymptote. The KPZ collapse uses a tuned κ plus adjusted C0 and y0. None of these individually kill the claim, but together they mean the pipeline is flexible. The strongest missing piece is a null test: apply the same filtering and window selection to synthetic non-universal decays (stretched exponentials with wrong β, power laws with a_s/a_t≠2, plain exponentials) and show you do not get apparent KPZ/EW collapses. Without that, the fitted exponents may partly reflect the selection procedure. The reported β=0.24 and χ=0.39 are reasonable, but the error bars from fits alone understate systematics, as the δ=−6.8 meV bias shows.\n\nThe central claim is plausible and well-supported, not fully established. This paper deserves a serious referee. For me, acceptance would require raw data deposition, a sensitivity sweep over windows and filter thresholds, and ideally a null test. I would bring it to a reading group because it is exactly the kind of result people will build on or push back on, and the experimental effort is impressive.\n\nRecommendation: engage with it. Send it to referees, but ask for the robustness analysis. If the authors can close the null-test gap, this becomes a landmark paper.","headline":"First real experimental claim of 2D KPZ scaling in a continuous planar polariton condensate, with an EW-KPZ crossover; the evidence is substantial but the partly manual analysis pipeline needs a null test before the identification is air-tight.","tokens_in":42500,"tokens_out":1318,"would_cite":true,"duration_ms":16923,"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":"Two-dimensional polariton condensates can be tuned between KPZ and Edwards–Wilkinson universal coherence decays.","keywords":["polariton condensates","Kardar–Parisi–Zhang universality","Edwards–Wilkinson universality","driven-dissipative systems","first-order coherence","two-dimensional quantum fluids","universal scaling","vortex pairing"],"falsifier":"Measure the phase correlation function C_θθ(Δr,Δt) directly in the same sample—for instance by off-axis digital holography that retrieves the optical phase—and check whether it obeys $Δt^{{2β}}$ or logarithmic growth with the fitted exponents; if the directly measured C_θθ does not show the same scaling, the coherence exponents are not evidence for KPZ or EW universality.","tokens_in":41332,"feed_emoji":"⚛️","tokens_out":7276,"duration_ms":68251,"temperature":0.7,"pith_summary":"Two-dimensional polariton condensates are driven-dissipative fluids of light whose phase fluctuations should fall into nonequilibrium universality classes, not the equilibrium Berezinskii–Kosterlitz–Thouless class. This paper reports that by tuning the cavity-exciton detuning of a planar semiconductor microcavity, the measured first-order coherence of the condensate crosses over between two scaling laws: stretched-exponential decays with exponents β≈0.24 and χ≈0.39, attributed to the 2D Kardar–Parisi–Zhang (KPZ) class, and algebraic decays whose spatial-to-temporal exponent ratio is about 2, attributed to the Edwards–Wilkinson (EW) class. The data collapse onto the universal KPZ and EW scaling functions after filtering out dispersive modes and properly normalizing the coherence. If correct, this establishes polariton microcavities as a controllable platform for exploring 2D KPZ physics and shows that the universal behavior of open condensates is richer than the equilibrium BKT paradigm.","feed_headline":"Polariton condensates flip between two universal scaling laws","feed_subtitle":"By tuning cavity-exciton detuning, coherence decays switch from KPZ stretched exponentials to EW power laws.","key_machinery":"The load-bearing object is the mapping of the condensate phase dynamics onto a two-dimensional Kardar–Parisi–Zhang equation, ∂tθ = ν∇²θ + (λ/2)(∇θ)² + √D η, where ν, λ, D are effective parameters built from the microscopic gain, loss, mass, and interaction coefficients. The dimensionless effective nonlinearity g_KPZ = λ²D/ν³ decides the class: weak nonlinearity gives the Edwards–Wilkinson diffusive regime with logarithmic phase correlations (algebraic coherence decays), and strong nonlinearity gives the KPZ super-diffusive regime with power-law phase correlations (stretched-exponential coherence decays). The link to the experiment is the identity g(1)(Δr,Δt) ∝ exp(−C_θθ(Δr,Δt)/2), relating the measured first-order coherence to the two-point phase correlation C_θθ; the paper verifies numerically that density correlations are negligible in the scaling windows. Interferometric measurement of g(1) followed by dispersion-branch filtering and the κ renormalization converts the measured fringe visibility into a test of the universal scaling functions.","core_discovery":"The paper's central claim is that a continuous two-dimensional polariton condensate in a planar microcavity can realize both the KPZ and EW universality classes as its microscopic parameters are varied, and that the crossover between them appears in the spatio-temporal decay of the first-order coherence. In the more excitonic regime (detuning ℏδ=−5.2 meV) the coherence decays as stretched exponentials g(1)∝exp(−A_t $Δt^{{2β}}$) in time and g(1)∝exp(−A_r $Δr^{{2χ}}$) in space, with fitted exponents agreeing with the numerical KPZ values β≈0.24 and χ≈0.39; the full space-time data collapse onto the 2D KPZ scaling function. In the more photonic regime (ℏδ=−12.6 meV) the same measurements show power-law decays, with spatial exponent a_s>0.25 and ratio a_s/a_t≈2, the signature of the EW diffusive phase. A statistical comparison of stretched-exponential versus power-law fits across detuning and pump power yields a phase diagram showing the EW-to-KPZ crossover, and numerical simulations with experimentally calibrated parameters reproduce both regimes and show that vortex-antivortex pairs remain bound, so the quasi-ordered KPZ/EW phases are the nonequilibrium analog of the ordered BKT phase.","pith_inferences":["Beyond the paper, the same coherence-measurement protocol could be applied to other dissipative condensates of light—photon condensates, plasmonic lattices, or vertical-cavity lasers—to test whether the same detuning-controlled crossover appears there.","The central relation g(1) ≈ exp(−C_θθ/2) is validated here mainly by numerics; a future experiment that directly images the phase field (for example by off-axis holography) could independently extract C_θθ and check the exponent attribution without relying on window selection.","If the finite-size explanation for the EW regime is right, systematically varying the pump spot diameter should move the onset of the Schawlow–Townes exponential tail and change the extent of the EW window; this is a quantitative prediction the paper does not make explicitly.","The vortex statistics suggest that increasing the noise or the pump spot beyond the studied range should eventually favor free vortices and a transition to the spiral-vortex or disordered phase that the paper lists as future work."],"forward_implications":["Tuning a single accessible parameter—the cavity-exciton detuning—moves a driven-dissipative condensate across a universality-class crossover, giving experimental control over KPZ versus EW behavior.","The measured collapse onto the KPZ universal scaling function in a continuous 2D system resolves the earlier debate about whether vortices necessarily destroy KPZ order in planar polariton condensates.","The phase diagram in detuning and pump power provides a map for choosing parameters that exhibit KPZ or EW scaling, which can guide future experiments.","Because the EW regime is only transient for a finite system (in infinite size, 2D KPZ is the stable fixed point), the observed EW window is a finite-size effect that should shrink when the pump spot is enlarged.","Vortex-antivortex pairing in numerics implies that topological defects do not spoil the universal scaling in the measured window, so the KPZ/EW phases are the nonequilibrium counterpart of the ordered BKT phase."],"supporting_citations":[{"why":"Supplies the theoretical phase diagram and the prediction that 2D polariton condensates can show EW algebraic decays with as/at=2 and KPZ stretched exponentials.","marker":"[15]"},{"why":"Predicted algebraic coherence decays with the characteristic spatial-to-temporal exponent ratio as/at=2 in nonequilibrium condensates.","marker":"[16]"},{"why":"Defines the Kardar–Parisi–Zhang equation whose 2D exponents and scaling function are the target of the coherence analysis.","marker":"[22]"},{"why":"Establishes the phase-to-KPZ mapping for polariton condensates and the interferometric methodology, including the κ normalization, used here in 2D.","marker":"[29]"},{"why":"Underlines the importance of the renormalization factor κ in data collapses onto the KPZ scaling function, motivating the paper's normalization protocol.","marker":"[31]"},{"why":"Provides the numerical 2D KPZ exponents β≈0.24 and χ≈0.39 against which the measured stretched-exponential exponents are compared.","marker":"[46]"},{"why":"Provides the nonperturbative renormalization-group calculation of the stationary 2D KPZ universal scaling function used for the data collapse.","marker":"[47]"},{"why":"Supplies the Edwards–Wilkinson scaling function in 2D used to collapse the power-law coherence data.","marker":"[48]"}],"fun_headline_variants":["Polariton condensates cross KPZ–EW scaling","Tunable 2D polaritons switch between KPZ and EW laws","Polariton condensates reveal KPZ–EW crossover","Detuning flips polariton coherence from KPZ to EW","2D polariton condensates hit KPZ and EW scaling"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The classification stands or falls on whether, after removing the dispersive branches and excluding short-time and short-distance transients, the measured g(1) decay is governed by condensate phase fluctuations through g(1)∝exp(−C_θθ/2), with density fluctuations and uncondensed-polariton emission negligible in the selected windows.","fun_headline_variants_meta":{"raw":{"variants":["Polariton condensates cross KPZ–EW scaling","Tunable 2D polaritons switch between KPZ and EW laws","Polariton condensates reveal KPZ–EW crossover","Detuning flips polariton coherence from KPZ to EW","2D polariton condensates hit KPZ and EW scaling"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000536,"raw_usage":{"total_tokens":2622,"prompt_tokens":1042,"completion_tokens":1580,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":658,"completion_tokens_details":{"reasoning_tokens":1492}},"tokens_in":658,"tokens_out":1580,"duration_ms":10580,"temperature":1.0,"reasoning_tokens":1492,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T11:58:25.950385+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the phase correlation function C_θθ(Δr,Δt) directly in the same sample—for instance by off-axis digital holography that retrieves the optical phase—and check whether it obeys $Δt^{{2β}}$ or logarithmic growth with the fitted exponents; if the directly measured C_θθ does not show the same scaling, the coherence exponents are not evidence for KPZ or EW universality.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the theoretical phase diagram and the prediction that 2D polariton condensates can show EW algebraic decays with as/at=2 and KPZ stretched exponentials."},{"cited_title":"(24) of Sec","cited_arxiv_id":null,"evidence_quote":"Predicted algebraic coherence decays with the characteristic spatial-to-temporal exponent ratio as/at=2 in nonequilibrium condensates."},{"cited_title":"The result is shown in Fig","cited_arxiv_id":null,"evidence_quote":"Defines the Kardar–Parisi–Zhang equation whose 2D exponents and scaling function are the target of the coherence analysis."},{"cited_title":"We plot the spatial (temporal) experimental decay in logarithmic scales as a function of the expected KPZ scaling ∆r 2χ (∆t2β) (see Fig","cited_arxiv_id":null,"evidence_quote":"Establishes the phase-to-KPZ mapping for polariton condensates and the interferometric methodology, including the κ normalization, used here in 2D."},{"cited_title":"We first demonstrate additional data collapses onto the universal KPZ scaling function for different pump powers, using the dataset presented in Fig","cited_arxiv_id":null,"evidence_quote":"Underlines the importance of the renormalization factor κ in data collapses onto the KPZ scaling function, motivating the paper's normalization protocol."},{"cited_title":"The probability distribution of vortex (respectively antivortex) lifetimes is displayed in Fig","cited_arxiv_id":null,"evidence_quote":"Provides the numerical 2D KPZ exponents β≈0.24 and χ≈0.39 against which the measured stretched-exponential exponents are compared."},{"cited_title":"Results in the KPZ regime are presented in Figs","cited_arxiv_id":null,"evidence_quote":"Provides the nonperturbative renormalization-group calculation of the stationary 2D KPZ universal scaling function used for the data collapse."}],"review_version":1}