Pith. sign in

REVIEW 3 major objections 6 minor 1 cited by

Diffusion cascade in a model of interacting random walkers

T0 review · 3 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Density waves in a nonlinear diffusive gas relax on a wavelength-independent clock at late times.

desk verdict The paper's surprise—a Q-independent stretched-exponential decay rate—is plausible but not yet proven; the data quality and analysis fall short of the claim. read the letter →

arxiv 2412.05222 v3 pith:B7X736IW submitted 2024-12-06 cond-mat.stat-mech

Abhishek Raj , Paolo Glorioso , Sarang Gopalakrishnan , Vadim Oganesyan This is my paper · ORCID
classification cond-mat.stat-mech
keywords diffusioncascadestretchedexponentialnonlinearhydrodynamicslong-timetailsfacilitatedlatticegasdensitywaverelaxationkineticallyconstrainedmodel
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper tests a theoretical prediction that finite-wavelength density perturbations in a nonlinear diffusive system relax at late times as exp(-$\sqrt$(D $Q^{2}$ t)) through a cascade of excitations to ever longer wavelengths. Using a one-dimensional hardcore lattice gas with tunable hydrodynamic nonlinearity, the authors find that relaxation is indeed subexponential at late times, but the rate is empirically independent of wavelength: log G(Q,t) ~ -C $\sqrt$(t) with C ≈ 0.4, not scaling as $\sqrt$(D $Q^{2}$). If correct, the cascade mechanism operates but the asymptotic decay is set by a lattice-scale length rather than by the perturbation's wavelength. The result matters because it identifies the high-order diagrammatic treatment of the cascade theory as the likely place where the prediction goes wrong.

What carries the argument

The central object is the nonlinear diffusion equation with a density-dependent diffusion constant, D(ρ) = D0(1+αρ), realized by a one-dimensional hardcore lattice gas updated in random trimers with tunable mobilities. The cascade mechanism is the repeated decay of a density wave at Q into pairs at Q/2, which relax more slowly, so the slowest relaxation is carried by a cascade to progressively smaller wavevectors. The paper's main quantitative tool is the numerical extraction of the density autocorrelation G(Q,t) in large lattices (up to $10^{6}$ samples), supported by exact noise-averaged dynamics in small rings and a field-theoretic path-integral calculation of the two-wave correlator at first order in the nonlinearity. The empirical form log G(Q,t) ~ -C $\sqrt$(t) with Q-independent C ≈ 0.4 is the load-bearing identity: it implies a finite length scale that replaces $Q^{2}$ in the relaxation rate.

What would settle it

Measure log G(Q,t) for several wavevectors at times well beyond the current simulation window (or push the crossover later by weakening the nonlinearity) and check whether the late-time slope of log G versus $\sqrt$(t) is still zero over Q; a slope that scales as $\sqrt$(D $Q^{2}$) would refute the paper's central claim.

Watch

Extended reading notes

Core claim

The paper's central claim is that in a facilitated lattice gas with strong hydrodynamic nonlinearities, the late-time relaxation of a weak density modulation at wavevector Q follows f(Q) exp(-C $\sqrt$(t)) with a scale C that is essentially independent of Q, contradicting the diffusion-cascade prediction exp(-$\sqrt$(D $Q^{2}$ t)). The crossover from ordinary diffusion (log G ~ -D $Q^{2}$ t) to this stretched-exponential regime is controlled by the diffusive timescale t ~ 1/$Q^{2}$, so the Q-dependence hides in the prefactor f(Q) rather than in the exponent. The paper also shows that the underlying cascade does occur: second-harmonic generation matches perturbative theory at short times, and a three-point correlation function shows coherent generation of wave pairs. The discrepancy with theory is therefore attributed to the treatment of the high-order, nonperturbative diagrams that dominate the extremely late-time asymptotics.

Load-bearing premise

The claim that the late-time decay rate does not depend on the wavelength is inferred from slopes fitted in a finite simulation window, with no error bars and residual curvature, so the result assumes those slopes are the true late-time limit.

Editorial extensions

If this is right

  • At sufficiently late times, every finite-wavevector density perturbation in a strongly nonlinear one-dimensional diffusive system relaxes at the same rate, so the wavelength controls the crossover time but not the asymptotic decay constant.
  • The crossover time from diffusive to cascade-dominated relaxation scales as 1/Q^2, meaning short-wavelength modes enter the universal tail much earlier than long-wavelength ones.
  • The cascade 'shower' mechanism is verified at the perturbative level through second-harmonic generation and coherent two-wave correlations, so the breakdown of the prediction is confined to the nonperturbative, high-order diagram sector.
  • In finite systems, the low-lying relaxation rates are not simply n times the fundamental mode rate; residual wave interactions produce corrections that vanish as 1/L^2, a signature that can be tested in exact small-system simulations.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A testable extension would be to measure the same autocorrelation in a model where the nonlinear coupling can be tuned independently of the bare diffusion constant, to see whether the Q-independent scale C shifts with the lattice-scale physics as the paper's dimensional argument suggests.
  • The paper's finding implies that other one-dimensional stochastic lattice gases with density-dependent mobility should show the same universal stretched-exponential tail with a model-dependent C; this could be checked with existing code by scanning the update rule.
  • If C is indeed set by the lattice scale, then the same relaxation form should appear in the response of conserved higher-charge densities, not just the density, which would provide a sharp consistency test.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 6 minor

Summary. The paper studies relaxation of finite-wavevector density perturbations in a one-dimensional facilitated classical lattice gas with trimer updates. It tests Delacretaz's prediction that nonlinear diffusion produces a stretched-exponential tail log G ~ -sqrt(D Q^2 t). Using stochastic simulations with L=99999 and exact small-system evolution, the authors find stretched-exponential decay but report that the extracted scale C in log G ~ f(Q) - C sqrt(t) is approximately independent of Q, contradicting the predicted Q dependence. They also characterize second-harmonic generation, mesoscopic relaxation spectra, and coherent two-wave correlators via the Martin-Siggia-Rose formalism.

Significance. If the Q-independent C is correct, the result is significant: it would mean a finite-Q density perturbation in a diffusive system relaxes with a wavevector-independent stretched-exponential rate, implying an emergent length scale beyond the diffusive fixed point and indicating a failure of the high-order diagrammatic argument in Ref. [1]. The paper is careful in several respects: the model is explicitly defined, D0 and D1 are calibrated from independent measurements (Fig. 2), second-harmonic generation is validated at short times, and small systems are treated by exact noise-averaged evolution. However, the central quantitative claim currently rests on slope extraction without error bars or convergence checks, so its significance cannot be fully assessed until the asymptotics are established with controlled statistics.

major comments (3)
  1. [III A, Fig. 4] The claim that C is independent of Q is not supported by the reported analysis: the slopes are obtained from data binned on the sqrt(t) axis without quoted uncertainties, the caption does not state t_max, and Fig. 4(c) itself shows weak residual Q-dependent curvature at the largest times. Please report bootstrap or jackknife confidence intervals for C(Q), show that the extracted slopes are stable as t_max is increased and as the binning scheme is varied, and verify that the smallest Q value is followed well beyond the diffusive-to-stretched crossover Q^2 t ~ 1. Without such a convergence check, C approximately 0.4 could be an effective finite-window slope rather than the true asymptote.
  2. [III A] The functional form G(Q,t) = f(Q) exp(-C sqrt(t)) is not tested as a full model: f(Q) is allowed to be rapidly varying and is not fitted simultaneously with C, so the apparent momentum-independence of C could be an artifact of underfitting a crossover with a Q-dependent prefactor. Please perform joint fits of f(Q) and C over a stated time window, report parameter correlations and goodness of fit for each Q, and repeat the procedure for the parameter sets in Appendix C to demonstrate that the conclusion is robust.
  3. [III A, III C] The dimensional argument that C is set by lattice-scale physics is not backed by a quantitative comparison: the manuscript does not report D0, D1, or the lattice scale for the main x=0, y=1 run, nor does it quote C for the Appendix C models beyond saying they are consistent. A table of C(Q) with uncertainties together with D0 and D1 for each parameter set would make the Q-independence claim falsifiable and would clarify whether C is indeed nonperturbative in D1.
minor comments (6)
  1. [II B, Eq. (6)] The notation D(t) is defined as a Q-derivative limit, but the relation <r^2> = 2D(t)t is only valid for a constant diffusion coefficient; please clarify whether D(t) is an integrated or instantaneous quantity.
  2. [Fig. 4 caption] Please state t_max and specify which panels are plotted versus sqrt(t) and which versus sqrt(Q^2 t), and include the values of Q used in the figure.
  3. [Fig. 9 caption] The caption contains a typo: 'A glimps' should be 'A glimpse'.
  4. [Acknowledgments and Appendix C] There are minor typographical errors: 'delacr` etaz' in the acknowledgments should be 'Delacretaz', and 'Floquet patter' in Appendix C should be 'Floquet pattern'.
  5. [II D] The sentence 'the signal decays slower than predicted this short time analysis' should read 'the signal decays slower than predicted by this short-time analysis'.
  6. [III C, Fig. 8] Fig. 8 is described as a qualitative comparison; please state explicitly which discrepancies are expected from the first-order truncation, since the plots are normalized to their maxima and amplitude information is therefore not displayed.

Circularity Check

0 steps flagged · score 2.0 of 10

No load-bearing circularity: central claim is a direct numerical observation; the only self-reference is a non-essential companion-paper citation and some calibrated analytic comparisons.

full rationale

The paper's central claim—that late-time relaxation follows f(Q) exp(-sqrt(C t)) with approximately Q-independent C—is an empirical fit to direct stochastic and exact simulations, not a consequence of the theory being tested (Ref. [1]). It is therefore not circular by construction. The analytic MSR correlators in Sec. III C use D0, D1, and noise couplings extracted from the same model (Fig. 2), but the plotted comparisons are normalized by their maximum value, so they test the qualitative shape in (q, t) rather than forcing the amplitude; this is calibrated consistency checking, not an input-output equivalence. Similarly, the second-harmonic generation check in Sec. II D is a short-time cross-check of the nonlinear hydrodynamic coupling, not a renamed fit that produces the main conclusion. The only self-citation is to the companion paper [4] for model-specific facts (e.g., stuck configurations); this is not load-bearing for the stretched-exponential claim. Concerns about finite-time slope convergence, binning, and residual curvature in Fig. 4 are correctness risks, not circularity, because they do not amount to the derivation reducing to its inputs. Thus no specific circular step can be quoted and exhibited; the appropriate score is low and reflects only the minor, non-load-bearing self-citation.

Assumptions & free parameters 2 free parameters · 6 assumptions · 0 invented entities

The numerical central claim itself is not fitted: C is an extracted exponent, not an input. Fitted inputs D0 and D1 are needed only for analytic side computations. The main unfunded assumptions are hydrodynamic reduction, linear response, and the asymptotic-fit procedure.

free parameters (2)
  • D0 = D0 = 1 - rho0 for (x,y)=(0,1), from Fig. 2
    Base diffusion constant extracted from density autocorrelation and current ratio; enters hydrodynamics, second harmonic, and MSR equations.
  • D1 = D1 = -1 for (x,y)=(0,1), from Fig. 2
    Nonlinear coupling D1 = alpha D0 extracted from the density dependence of D; used as the interaction strength in Eq. 7 and the MSR action.
assumptions (6)
  • domain assumption Hydrodynamic description: the lattice gas is captured by Fick's law with Gaussian white noise and a density-dependent diffusion constant (Eqs. 5 and 11).
    Used throughout to define D, D1, and the MSR action; ignores higher-order gradient terms and possible lattice-scale memory.
  • domain assumption Weak-quench linear response: amplitude A <= 0.2 makes the response equal to the equilibrium correlation function (Eq. 2).
    The authors assume and spot-check linearity; if nonlinear response contaminates the data, the extracted decay of G(Q,t) would not be an equilibrium two-time correlator.
  • domain assumption Thermodynamic-limit representativeness: L=99999 random-update runs and exact L=12,18 Floquet runs capture the infinite-system asymptotics.
    Finite-size effects are argued to be 1/L^2 using only two small sizes; stuck configurations in the x=0 model are asserted negligible at large L.
  • ad hoc to paper Asymptotic extraction: binning log G on the sqrt(t) axis and reading late-time slopes gives the true t to infinity exponent.
    The central Q-independent C rests on this fitting procedure, which has no error bars and residual curvature in Fig. 4(c).
  • domain assumption MSR perturbative expansion is valid to first order in D1 and sigma1, with sigma1 = D1 chi at half-filling.
    Used to derive Eqs. 14-15; higher-order diagrams are neglected and the comparison with simulations is qualitative.
  • domain assumption The trimer model with x=0,y=1 is a representative realization of nonlinear diffusion and is not dominated by model-specific jamming.
    Companion paper [4] identifies extra jamming features; appendix x=0.2,y=0.8 checks are used to claim the Q-independent behavior is generic.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Diffusion cascade in a model of interacting random walkers." pith.science (2026). https://pith.science/paper/B7X736IW

@misc{pith2026241205222,
  author       = {Pith},
  title        = {Pith review of: Diffusion cascade in a model of interacting random walkers},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/B7X736IW}},
  note         = {Machine review of arXiv:2412.05222}
}
read the original abstract

We consider the relaxation of finite-wavevector density waves in a facilitated classical lattice gas. Linear hydrodynamics predicts that such perturbations should relax exponentially, but nonlinear effects were predicted to cause subexponential relaxation via nonperturbative long-time tails. We present a detailed numerical study of this effect. While our results clearly indicate the importance of nonlinear effects, we find that the wavevector-dependence of the late-time relaxation is clearly inconsistent with theoretical predictions. We discuss manifestations of hydrodynamic nonlinearities in mesoscopic samples and at short times.

Figures

Figures reproduced from arXiv: 2412.05222 by the authors.

Figure 1
Figure 1. FIG. 1. The three triplets represented in red, green and blue [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2. Diffusion constant for symmetric(gold) and asym [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. FIG. 3. 2 [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figures from the paper (9 more)
Figure 4
Figure 4. Figure 4: FIG. 4. (panel a) Cross-over from diffusive relaxation [PITH_FULL_IMAGE:figures/full_fig_p004_4.png]
Figure 5
Figure 5. Figure 5: FIG. 5. Relaxation in small ( [PITH_FULL_IMAGE:figures/full_fig_p005_5.png]
Figure 6
Figure 6. Figure 6: FIG. 6. Fourier cascade in real time: initial single wave ex [PITH_FULL_IMAGE:figures/full_fig_p005_6.png]
Figure 7
Figure 7. Figure 7: FIG. 7. Time- and average denity- dependence of the diffusion [PITH_FULL_IMAGE:figures/full_fig_p006_7.png]
Figure 8
Figure 8. Figure 8: FIG. 8. Heatmap of the cascade [PITH_FULL_IMAGE:figures/full_fig_p007_8.png]
Figure 9
Figure 9. Figure 9: FIG. 9. A glimps of nongaussian stats for Q= [PITH_FULL_IMAGE:figures/full_fig_p007_9.png]
Figure 10
Figure 10. Figure 10: FIG. 10. Cascade plots for update rules ( [PITH_FULL_IMAGE:figures/full_fig_p011_10.png]
Figure 11
Figure 11. Figure 11: FIG. 11. Late-time dynamics in large lattices to compare with Fig. 4. Upper panel shows results for x=0.2, y=0.8 with random [PITH_FULL_IMAGE:figures/full_fig_p011_11.png]
Figure 12
Figure 12. Figure 12: FIG. 12. D vs [PITH_FULL_IMAGE:figures/full_fig_p012_12.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. A kinetically constrained model exhibiting non-linear diffusion and jamming

    cond-mat.stat-mech 2024-12 conditional novelty 7.0 of 10

    A kinetically constrained model on a triangular ladder is shown to have mean-field diffusion coefficient D = 3(1 - ρ), with a jamming transition at density ρ = 2/3.

Reference graph

Works this paper leans on

16 extracted references · 14 canonical work pages · cited by 1 Pith paper

  1. [1]

    by observing exponential decay of density autocor- relation at momentum Q with rate Γ Q = DQ2t (as implied by combining Eqs.4, 5 to eliminate cur- rent)

    work page

  2. [2]

    3) The configurations 000 and 111 remain unchanged

    similarly, 101 has probability x/2 of transitioning into either 011 or 110, and probability 1 − x of staying 101, and reverse transitions have the same probability. 3) The configurations 000 and 111 remain unchanged. Writing the Markov matrix for a trimer in the basis of configura- tions {cj} = {(000), (001), (010),(011), (100) . . .(111)} ↓ {m} = {0, 1, ...

    work page

  3. [3]

    conven- tional

    by observing both density and current at momen- tum Q and computing the appropriate ratio as im- plied by Eq. 5 (this assumes the noise vanishes on average). Note, these implicitly require taking thermodynamic L → ∞limit first. As we explore in Sec. III B and a companion paper[4] the diffusion constant may be ex- tracted from asymptotic late time decay al...

    work page

  4. [4]

    Onset and decay of wave coherences in the diffusion cascade FIG. 6. Fourier cascade in real time: initial single wave ex- pectation value at Q seeds a coherent wave pair with the same Fourier momentum. Density dependent diffusion constant is natural[7]. With an eye towards experimental detection of the dif- fusion cascade in such systems we now explore co...

    work page

  5. [5]

    Mukerjee, V

    S. Mukerjee, V. Oganesyan, and D. Huse, Statistical the- ory of transport by strongly interacting lattice fermions, Phys. Rev. B 73, 035113 (2006)

    2006

  6. [6]

    diffusion cas- cade

    Non-gaussian statistics Another observable and potentially interesting sig- nature of the diffusive cascade is its (transient) non- gaussianity. Inspired by basic observables used in fluid turbulence we examined variance, skewness and kurtosis of the initially excited wave. Local equlibrium at t = 0 and t → ∞is characterized by strictly Gaussian (normal) ...

    work page

  7. [7]

    We perform this this path integral in the fourier space to first order in the interaction

    There are two distinct types of diagrams contributing to the first order in perturbation. We perform this this path integral in the fourier space to first order in the interaction. For the quench setup we have ⟨ρq(t)ρQ−q(t)⟩µ = Tr [Poρq(t)ρQ−q(t)] (B10) Po = Y x eµ(x)ρ(x,0) 1 + eµ(x) (B11) = Y x 1 2 1 − µ(x) 2 δ(ρ(x, 0)) + 1 + µ(x) 2 δ(ρ(x, 0) − 1) (B12) ...

    work page

  8. [8]

    L. V. Delacretaz, Heavy operators and hydrodynamic tails, SciPost Phys. 9, 034 (2020)

    work page 2020

Show all 16 references
  1. [9]

    Collier, A

    S. Collier, A. Maloney, H. Maxfield, and I. Tsiares, Uni- versal dynamics of heavy operators in cft2, Journal of High Energy Physics 2020, 1 (2020)

    2020

  2. [10]

    Karlsson, A

    R. Karlsson, A. Parnachev, and P. Tadi´ c, Thermalization in large-n cfts, Journal of High Energy Physics 2021, 1 (2021)

    2021

  3. [11]

    A. Raj, V. Oganesyan, and A. Scardicchio, A kinetically constrained model exhibiting non-linear diffusion and jam- ming, Journal of Statistical Mechanics: Theory and Ex- periment 2025, 073208 (2025)

    2025

  4. [13]

    However, if the diffusion constant is random in space, even the finite- Q response is dominated by a power-law long- time tail [8, 9]

  5. [14]

    Upadhyay, D

    P. Upadhyay, D. G. Su´ arez-Forero, T.-S. Huang, M. J. Mehrabad, B. Gao, S. Sarkar, D. Session, K. Watanabe, T. Taniguchi, Y. Zhou, et al. , Giant enhancement of ex- citon diffusion near an electronic mott insulator, arXiv preprint arXiv:2409.18357 (2024)

    2024 arXiv

  6. [15]

    and also extended to higher orders if necessary. Cru- cially, these multi-point correlators owe their existence to perfect phase coherence among generated waves, thus any of the individual wave observables with k ̸= Q is zero due to random phases, which are nevertheless ex- 6 ...

  7. [16]

    Ernst, J

    M. Ernst, J. Machta, J. Dorfman, and H. Van Beijeren, Long time tails in stationary random media. i. theory, Journal of statistical physics 34, 477 (1984)

    1984

  8. [17]

    P. J. H. Denteneer and M. H. Ernst, Diffusion in systems with static disorder, Phys. Rev. B 29, 1755 (1984). 11 FIG. 10. Cascade plots for update rules ( x, y) = (0.2, 0.8). Floquet updates FIG. 11. Late-time dynamics in large lattices to compare with Fig. 4. Upper panel shows...

    1984

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.