Pith. sign in

REVIEW 3 major objections 5 minor 61 references

Operator Spreading, Duality, and the Noisy Long-Range FKPP Equation

T0 review · 3 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read This paper establishes that discrete operator-growth population dynamics and the continuum noisy long-range FKPP equation describe the same butterfly light cone in power-law quantum systems, by proving an exact duality through a…

desk verdict The exact soft-constraint duality is real and the numerics are serious, but the hard-soft universality is numerically supported only, and the Levy-index inconsistency in the Conclusion needs fixing before the unification claim is taken as established. read the letter →

arxiv 2505.06353 v1 pith:ZOH3TWUD submitted 2025-05-09 cond-mat.stat-mech cond-mat.dis-nn

Tianci Zhou , Éric Brunet , Xiaolin Qi This is my paper · ORCID
classification cond-mat.stat-mechcond-mat.dis-nn
keywords operatorspreadingbutterflylightconelong-rangeFKPPequationpopulationdynamicsdualitycoalescencequantumchaospower-lawinteractions
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 tries to close the gap between two standard ways of describing operator spreading in chaotic quantum systems: a discrete population process that counts how many nontrivial operators live on each site, and a continuum stochastic reaction-diffusion equation of FKPP type. The authors construct an exact duality between a population process and the stochastic FKPP equation by replacing the hard cap on local population with a soft constraint, namely pairwise coalescence that keeps the equilibrium population at $N$. Under that soft model the duality is a mathematical identity, a time-independent combined correlator that forces the butterfly light cones of the two descriptions to coincide. The hard-cap model is not exactly dualizable, so the final step is numerical: in one dimension the soft and hard models show matching light-cone scalings for $\alpha>0.5$. A sympathetic reader would care because if the chain holds, the discrete growth picture and the noisy PDE are equivalent descriptions of the same physics, and one can choose whichever is easier to compute.

What carries the argument

The load-bearing object is the duality identity between the soft-constraint population model and the stochastic long-range FKPP equation. On one side is an integer height $h(x,t)$ with birth-at-distance rate $G(\|y-x\|)h(x)$ and on-site coalescence rate $\tfrac{\beta}{2}h(h-1)$; on the other side is a continuum field $f(x,t)\in[0,1]$ satisfying $\partial_t f = -(1-f)\tilde D(-\Delta)^{\alpha-1/2}f + \tilde\lambda f(1-f) + \sqrt{\tilde\beta f(1-f)}\,\eta$. The identity $\langle \prod_x (1-f(x,\tau))^{h(x,t-\tau)}\rangle = \text{const}$ transfers the light cone of one process to the other. It is the exact, parameter-free link between the discrete and continuum perspectives; the hard constraint has no such dual, so the soft model is the hinge of the whole argument.

What would settle it

Simulate hard- and soft-constraint models in one dimension at several $\alpha$ values in $(0.5,1.5)$ with the reported no-finite-size algorithm, run until the iterative integral equation plateaus, and compare the inferred exponents $d\log\ell/d\log t$; any difference outside numerical error would falsify the claimed equivalence.

Watch

Extended reading notes

Core claim

The central claim is that the butterfly light cone of operator growth in a system with power-law interactions is the same object whether computed from a discrete birth-at-distance population process or from a stochastic long-range FKPP equation. The bridge is an intermediate population model in which the hard local maximum $N$ is replaced by a coalescence process $A+A\to A$ with equilibrium population $N$. For this soft model the paper proves a duality: a combined correlator $\langle \prod_x (1-f(x,\tau))^{h(x,t-\tau)}\rangle$ is independent of $\tau$, which equates the event that spreading has reached distance $L$ with the event that the dual field $f$ has traveled distance $L$. The paper then shows numerically, with a no-finite-size-effect algorithm that can track up to $10^9$ sites, that the soft and hard constraint models have overlapping light-cone curves for $d=1$ and $\alpha>0.5$, including the critical crossover near $\alpha=1$, and concludes that the two perspectives share identical butterfly light-cone scaling.

Load-bearing premise

The chain of equivalence assumes that replacing the hard local population cap by a soft coalescence constraint with the same equilibrium population leaves the butterfly light-cone scaling unchanged; the paper proves the soft model is dual to FKPP but cannot construct a dual for the hard model.

Editorial extensions

If this is right

  • The butterfly light-cone scalings of discrete population dynamics and the noisy long-range FKPP equation are identical for one-dimensional power-law interactions with $\alpha>0.5$.
  • The dual FKPP equation has a Levy index $\mu=2\alpha-2$, rather than $2\alpha-1$ as obtained by cutoff mean-field approximations, showing that noise and discreteness alter the effective transport.
  • The iterative integral equation $\ell(t)^{2\alpha}\sim \int_0^t \ell(\tau)\ell(t-\tau)\,d\tau$ organizes the stretched-exponential, power-law, and linear regimes, including the critical point $\alpha=1.5$ and the crossover near $\alpha=1$.
  • The higher-order analytic scaling near $\alpha=1$ can serve as a benchmark for times where asymptotic scaling is not numerically or experimentally reachable.
  • The sparse Gillespie algorithm with no finite-size effect can probe light cones over $10^9$ sites and supplies a practical tool for both constraint models.

Reading between the lines

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

  • Beyond the paper: the same duality should extend to higher dimensions and to the dissipative, non-unitary settings the paper mentions as future directions.
  • Beyond the paper: the duality gives a microscopic reading of the FKPP noise term as coalescence noise, so finite-$N$ corrections to the light cone could be tested in power-law quantum simulators and should scale as $1/\sqrt{N}$.
  • Beyond the paper: the fact that soft and hard constraint light cones overlap even at early times hints at a stronger coupling between the two processes, possibly a domination or coupling that would convert the numerical agreement into a proof.
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 / 5 minor

Summary. The paper proposes to unify two complementary descriptions of operator spreading in chaotic long-range interacting systems: the discrete stochastic population-dynamics picture with a hard local height cap, and the continuum noisy long-range FKPP equation. The authors introduce an intermediate soft-constraint model in which the hard cap is replaced by a coalescence process that fixes the equilibrium population. They prove, in Appendix D, an exact duality between this soft-constraint model and a coupled system of diffusions that, in the continuum limit, becomes the stochastic long-range FKPP equation. The light-cone equivalence follows from the duality. Since the hard-constraint model is the one directly tied to the operator-growth picture, the authors then compare hard- and soft-constraint models numerically in one dimension, finding overlapping butterfly light-cone curves for the tested parameters, and they derive a third-order perturbative scaling around α=1. The central claim is that, for α>0.5 in 1D, the discrete operator-growth picture and the noisy long-range FKPP description share identical butterfly light-cone scaling.

Significance. The exact duality in Appendix D is a genuine contribution: it is mathematically clean, holds for arbitrary population size N, and gives a direct bridge between a discrete Markov process and a stochastic PDE with multiplicative noise. The numerical algorithm is also a strength: it simulates up to 10^9 sites with event selection in constant time and no finite-size space cutoff, which is necessary for the stretched-exponential regime near α=1. The third-order perturbative solution of the iterative light-cone equation around α=1 is a useful benchmark for a regime that is extremely hard to access asymptotically. If the soft-hard equivalence were proven or convincingly established over the full claimed range, the unification would be an important conceptual step for the operator-spreading literature. However, the paper currently overreaches in its headline range α>0.5 and contains an internal inconsistency in the Lévy index, so the significance is conditional on those points being fixed.

major comments (3)
  1. [Conclusion, Eq. (4), and Appendix D.5.b] There is an internal inconsistency in the Lévy index of the claimed FKPP equation. Equation (4) states μ=2α−1, and Appendix D.5.b derives the same index by writing ∂t f = −(1−f) D (−Δ)^{α−1/2} f + λ f(1−f) + sqrt(β f(1−f)) η. The Conclusion, however, states that the resulting long-range FKPP equation has Lévy index μ=2α−2 and contrasts this with Ref. [34], which supposedly gives μ=2α−1. The last sentence would mean the paper's own derivation gives the same index as the work it claims to contradict. This must be resolved before the FKPP equation is unambiguously identified, because the title and abstract present that equation as one of the two unified perspectives.
  2. [Introduction and Appendix D.4.f; Figs. 4 and 5] The load-bearing link in the unification is the equivalence between the hard-constraint model (the operator-growth picture) and the soft-constraint coalescence model. The authors state in the Introduction that they cannot implement the hard constraint exactly in the dual model, and Appendix D.4.f explicitly concludes that no dual process exists for arrival-dependent hard-constraint births. The soft-hard equivalence is therefore a numerical conjecture, not a theorem. The numerical evidence covers α=0.85–1.05 and α=1.3–2.0, while the Conclusion announces the result for all α>0.5 in 1D. The interval 0.5<α<0.85, where the predicted stretched-exponential exponent diverges most quickly, is not tested. Moreover, Fig. 5 shows that other front statistics (Np, Ns, M) do not exactly overlap between the two models, so the ℓ(t) overlap in Fig. 4 is the only direct scaling-level evidence. I recommend either restricting the claim to the tested α range or adding simulations and, if possible, an analytic argument for the missing interval.
  3. [Fig. 4(b), Eq. (7), and Fig. 4(c)] The validation of the α=1 prediction is partly circular. The constant c1 in Eq. (7) is computed from K, and K is read from the plateau in Fig. 4(c), which is obtained from the same simulation data used to evaluate the prediction in Fig. 4(b); in addition, the constant C is adjusted by a vertical shift. The authors disclose this, and the test is still meaningful for the functional form, but the text's phrase 'no fitting of unknown constants' (near Fig. 4) is too strong. The agreement should be described as a consistency check of the predicted scaling form with a constant determined from the same data, not as an independent parameter-free confirmation.
minor comments (5)
  1. [Appendix E.1] The word 'varitable' appears in the sentence introducing the change of variable τ=t/2+ut; it should be 'variable'.
  2. [Fig. 4 caption and text near Eq. (7)] The caption says the dashed line is Eq. (7) with an adjusted value for C, while the text says the only fitting is a global vertical shift; please reconcile these descriptions so the reader knows exactly which constants are free.
  3. [Appendix A and Fig. 5] The observation that Np, Ns, and M do not exactly overlap between soft and hard constraints is important and should be moved or at least highlighted in the main text, because it delimits the claimed equivalence to ℓ(t) scaling rather than to full statistical equivalence of the two stochastic processes.
  4. [Appendix D.5.b] The passage from the lattice SDE (D38)–(D39) to the continuum fractional Laplacian equation (D40) is heuristic; a sentence clarifying the sense in which this limit is taken (e.g., Fourier/normalization of the kernel near the origin) would help the reader assess the status of the resulting FKPP equation.
  5. [Fig. 2] The scaling functions in Fig. 2 for the hard constraint model are stated without derivation; since the paper relies on them for the phase boundaries, a brief derivation or a precise citation to the exact results of Refs. [8,31,33] for each regime would improve self-containedness.

Circularity Check

2 steps flagged · score 3.0 of 10

Duality and hard/soft comparison are non-circular; only subleading 'prediction' curves reuse fitted constants (K from Fig. 4(c), C hand-tuned).

  1. fitted input called prediction [Numerical results, Fig. 4(b)-(c), page 4]
    "In Fig. 4(b), we observe an agreement with the numerical data at late times when corrections up to O(ln ln t) are included. When plotting the prediction, we used the value of K obtained from the plateau in Fig. 4(c), so that there was no fitting of unknown constants in Fig. 4(b) except for a global vertical shift in logarithmic scale."

    The plotted 'prediction' Eq. 7 contains c1 = 1/2 + (1/2) ln ln 2 + ln(K sqrt(pi)), while K is read from the plateau of the iterative integral Eq. 6 evaluated on the same l(t) data (Fig. 4(c), where K is reported as roughly 24.2 at alpha = 1). The curve compared with simulation is therefore constrained by a constant measured from that same simulation, so the match is a self-consistency check on the subleading constant rather than an independent prediction of the light cone. The leading stretched-exponential shape is still fixed analytically, so the circularity is partial and confined to prefactor-level agreement.

  2. fitted input called prediction [Appendix E.4, Numerical checks, page 21]
    "To make these figures, we needed values of c and of C for each of the ten data-points (five values of delta in two models). The values of c were obtained as in Fig. 4(c). The values of C were actually hand-tuned so that the curves in Fig. 6(b) fall together between the black lines corresponding to two-order and three-order corrections."

    The comparison in Fig. 6 between measured phi = log2 l(t) and the analytic scaling (E34)/(E37) uses per-model unknown constants C that are hand-tuned to make the curves fall in the expected window. The plotted 'prediction' therefore has its vertical offset adjusted to the very data it is compared with; only the shape of the 1/z correction terms remains non-fitted. This is an admitted fitting of constants, and it affects subleading corrections rather than the leading light-cone exponents.

full rationale

The central derivation is not circular. The exact duality between the soft-constraint coalescence process and the stochastic long-range FKPP equation is established by a self-contained moment-duality argument in Appendix D, following the earlier independent Doering-Mueller-Smereka construction [30], and the hard-constraint scaling functions are taken from Hallatschek-Fisher [31] and Chatterjee-Dey [33], not from the present authors. The one self-citation to prior work by an author [8] is backed by those independent derivations, so no load-bearing self-citation chain is present. The key claim that soft and hard constraint models share identical light cone scaling rests on direct Gillespie simulations of both models (Fig. 4), a numerical comparison that does not presuppose the conclusion. The circularity that does exist is confined to the analytic 'prediction' plots: the constant K entering Eq. 7 is read from the plateau of Eq. 6 computed on the same l(t) data (Fig. 4(c)), and the constant C in Fig. 6(b) is explicitly hand-tuned, so those curve comparisons are partially self-consistent fits rather than fully independent predictions. These fitted constants affect subleading prefactors, not the leading stretched-exponential or power-law exponents, so the main unification claim retains independent content. Separately, but not as circularity, the Conclusion's Levy index mu = 2 alpha - 2 contradicts Eq. (4) and Appendix D.5.b (mu = 2 alpha - 1), and the claimed alpha > 0.5 range is numerically probed only for alpha >= 0.85; both are correctness or completeness issues, not definitional circularity.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The central claim rests on the domain assumption that operator spreading maps to the hard-constraint population model, and on the iterative scaling equation from prior work. The new duality requires only standard stochastic calculus and the heuristic continuum limit. The free parameters K and C appear in the higher-order scaling predictions and are partly fitted to the simulation data.

free parameters (3)
  • K = ~24.2 at alpha=1
    Constant in iterative scaling equation Eq. 6, measured from the numerical plateau in Fig. 4(c) and used to evaluate the higher-order prediction Eq. 7. Extracted from the same simulation data rather than derived.
  • C = hand-tuned
    Integration constant in the perturbative expansion near alpha=1 (Eq. 7, Eq. E34). In Appendix E.4 the authors state 'The values of C were actually hand-tuned so that the curves in Fig. 6(b) fall together.'
  • beta (coalescence rate) = beta=1.44649*lambda for N=2
    Chosen so the soft-constraint equilibrium height equals the hard-constraint cap N (Appendix C). This is a model calibration to make the comparison fair, not fitted to light cone scaling.
assumptions (4)
  • domain assumption The hard-constraint birth-at-distance process with rate G(||x-y||)h(x)(1-h(y)/N) is an exact mapping of operator spreading in chaotic Brownian circuits.
    Used at the start of the paper to justify the population dynamics model; inherited from Refs [8,13,15-19].
  • domain assumption The iterative scaling equation Eq. 5 (from Hallatschek-Fisher [31]) describes light cone growth via seeding and merging of clusters.
    Used to derive the leading order scalings and as the starting point for the perturbative expansion (Eq. 6, Eq. E1).
  • domain assumption The continuum limit of the lattice dual process is a fractional Laplacian PDE with multiplicative white noise (Eq. D40/D41).
    The replacement of a lattice sum by (-Delta)^(alpha-1/2) is stated as 'A continuous version ... would look like' and is not rigorously justified.
  • standard math The scaling ansatz phi(z,delta)=delta^{-2} f(delta z/2, delta) with f of order one (Eq. E5).
    Assumed to perform the small-delta expansion; proven consistent within the expansion by balancing terms.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Operator Spreading, Duality, and the Noisy Long-Range FKPP Equation." pith.science (2026). https://pith.science/paper/ZOH3TWUD

@misc{pith2026250506353,
  author       = {Pith},
  title        = {Pith review of: Operator Spreading, Duality, and the Noisy Long-Range FKPP Equation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZOH3TWUD}},
  note         = {Machine review of arXiv:2505.06353}
}
read the original abstract

Operator spreading provides a new characterization of quantum chaos beyond the semi-classical limit. There are two complementary views of how the characteristic size of an operator, also known as the butterfly light cone, grows under chaotic quantum time evolution: A discrete stochastic population dynamics or a stochastic reaction-diffusion equation in the continuum. When the interaction decays as a power function of distance, the discrete population dynamics model features superlinear butterfly light cones with stretched exponential or power-law scaling. Its continuum counterpart, a noisy long-range Fisher-Kolmogorov-Petrovsky-Piscunov (FKPP) equation, remains less understood. We use a mathematical duality to demonstrate their equivalence through an intermediate model, which replaces the hard local population limit by an equilibrium population. Through an algorithm with no finite size effect, we demonstrate numerically remarkable agreements in their light cone scalings.

Figures

Figures reproduced from arXiv: 2505.06353 by the authors.

Figure 1
Figure 1. Two perspectives of understanding operator spread [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Scalings functions of the light cones for the hard [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Butterfly light cones. (a) A greymap of a [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Numerical results of ℓ(t) for (a) α on the two sides of 1.5 (b) α close to 1 [standard log scale for the vertical axis and | ln(t)| 2 scale for the horizontal axis], and (c) consistent the check of the iterative equation for α close to 1. The dashed lines in (a) repres…
Figure 5
Figure 5. Figure 5: Consistency check of the iterative integral equation against (a) [PITH_FULL_IMAGE:figures/full_fig_p005_5.png]
Figure 6
Figure 6. Figure 6: Simulations of the soft constraint and hard constraint models for [PITH_FULL_IMAGE:figures/full_fig_p021_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

61 extracted references · 45 canonical work pages

  1. [34]

    A. De, U. Borla, X. Cao, and S. Gazit, Physical Review B 110, 155135 (2024), arXiv:2401.06215 [cond-mat]

    arXiv 2024

  2. [1]

    Np(t) =⟨P∞ x=−∞h(x,t )⟩: the total population

    work page

  3. [2]

    The superlinear light cone, especially the stretched exponential type, poses a significant challenge for numerical verification

    light cones as α increases. The superlinear light cone, especially the stretched exponential type, poses a significant challenge for numerical verification. clustersx t h(x, t) (a) x ⋆𝓁(t)t Bf (τ ) Bb(t −τ )(b) Figure 3. Butterfly light cones. (a) A greymap of a h(x,t ) instance for the coalescence model withα = 1 and equilibrium height≈ 2 for x∈ [−105, 1...

    work page

  4. [3]

    Numerical results

    We will present the higher order scalings in numerical results, but defer its details to SM[32]. Numerical results . To bridge the missing link in our roadmap (Fig. 1), we numerically compare the light cone scalings of two models in d = 1 dimension: the hard constraint model with a maximal height N = 2, and the soft constraint model where the equilibrium ...

    work page

  5. [4]

    4(b), we observe an agreement with the numerical data at late times when corrections up to O(ln lnt) are included

    In Fig. 4(b), we observe an agreement with the numerical data at late times when corrections up to O(ln lnt) are included. When plotting the prediction, we used the value ofK obtained from the plateau in Fig. 4(c), so that there was no fitting of un- known constants in Fig. 4(b) except for a global vertical shift in logarithmic scale. We further perform a...

    work page

  6. [5]

    Ns(t) =⟨P∞ x=−∞θ(h(x,t ))⟩: the total number of occupied sites, where θ equals 1 for positiveh and 0 otherwise

    work page

  7. [6]

    one of the sites with height equal to h gives birth

    M(t) = D max n |x| h(x,t )> 0 oE typical : the distance to the furthest occupied site. 101 102 t 0.04 0.06 0.08 0.10 0.850 0.900 0.950 1.0001.050 (a) ∫ t 0 Np(τ)Np(t − τ)dτ /Np(t)2α soft constraint, Np(t) hard constraint, Np(t) 101 102 t 0.04 0.06 0.08 0.10 0.850 0.9000.9501.000 1.050 (b) ∫ t 0 Ns(τ)Ns(t − τ)dτ /Ns(t)2α soft constraint, Ns(t) hard constra...

    work page arXiv

  8. [7]

    To do this, compute the total rate, equal to R = X h h inc rate[h]×λ + h dec rate[h]×β (B1) and increase the time by a random exponential number of average 1 /R

    Update the time of the system after the current step. To do this, compute the total rate, equal to R = X h h inc rate[h]×λ + h dec rate[h]×β (B1) and increase the time by a random exponential number of average 1 /R

    work page

Show all 61 references
  1. [8]

    In this step, we do not yet select on which site the event starts

    Randomly select an event with probability proportional to the rates a) We randomly select whether we do a birth or a coalescence, and the height h of the site where the event starts, according to the total rates for that height from h inc rate and h dec rate. In this step, we ...

  2. [9]

    We must update the height status in all the recording structures a) We update the height in h current

    We know which site to increase or decrease; in the example, we increase the site at position 5. We must update the height status in all the recording structures a) We update the height in h current. 2 0 3 1 3 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2· · · 1 1 · · ·h current h in...

  3. [10]

    At any time t, the state of the system is characterized by the heights (or numbers of particles) h1(t) and h2(t) on each of the two sites

    Master equation for a particle system We consider a particle system living on two sites, labeled 1 and 2. At any time t, the state of the system is characterized by the heights (or numbers of particles) h1(t) and h2(t) on each of the two sites. To shorten notations, we write p...

  4. [11]

    Coupled diffusive processes We now consider coupled diffusive processes Z1(t) and Z2(t) (one process for each of the two lattice sites) which are real numbers evolving according to stochastic differential equations of the form dZi =µi(Z1,Z 2) dt +σi(Z1,Z 2) dBi. (D5) In this e...

  5. [12]

    The processes are independent

    Duality We now consider, on the same two-site lattice, a particle system hi(t) described by the matrix MP and coupled diffusive processes Zi(t) with moments described by the matrix MQ. The processes are independent. Let t >0 be some arbitrary positive time, and introduce, for ...

  6. [13]

    birth at distance

    Simple examples a. Birth on site Consider a particle system on a single site, where the number h of particles on the site increases by one with rate γh. (This is a Galton-Watson process in continuous time.) The non-zero elements of the matrix MP are MP (n + 1,n ) =γn, M P (n,n...

  7. [14]

    Adding correlations to the Brownian processes do not help us, and we are unable to find a dual process

  8. [15]

    hard constraint

    Height Evolution of OTOC a. On the lattice We now consider a d-dimensional lattice indexed by vector x, and write h(x,t ) for the height at position x and time t. The dual process will be made of one diffusion Z(x,t ) per lattice site x. In the original “hard constraint” model...

  9. [16]

    6 in dimension d = 1, which is assumed to be valid for large times

    Main equation We start from Eq. 6 in dimension d = 1, which is assumed to be valid for large times. We write 2 α = 2 +δ and, for later convenience, we introduce c byK = 2−c/√π. Thus, our starting point is ℓ(t)2+δ = 2−c √π Z t 0 ℓ(τ)ℓ(t−τ) dτ. (E1) The goal is to study the beha...

  10. [17]

    equal up to δ

    The scaling limit We now assume that δ is small but non-zero, and that, simultaneously, z = log 2t is large. As shown in [31], the correct scaling is φ(z,δ ) = 1 δ2f(y,δ ) with y = δz 2 (E5) where the function f(y,δ ) remains of order 1 when δ goes to zero while y remains fixe...

  11. [18]

    Let Q(y,δ ) = Z y 0 dy′ log2 ey′ − 1 δ

    There remains a term with a complicated primitive. Let Q(y,δ ) = Z y 0 dy′ log2 ey′ − 1 δ . (E25) Second term in Eq. E23: We have, from Eq. E19, f′′′ 0 (y) =−2e−y. Hence Z dyeyf′′′ 0 (y) =−2y. (E26) Third term in Eq. E23: One has, f′ 0(y) = 2(1−e−y) and so 1 f′ 0(y) = 1 2 ey e...

  12. [19]

    E34 by taking the limit δ→ 0 after replacing y by zδ/2

    The critical case The critical case α = 1, or δ = 0 can be recovered from Eq. E34 by taking the limit δ→ 0 after replacing y by zδ/2. One obtains φ(z, 0) = z2 4 + h 1 4 + 1 4 ln 2 − 1 4 log2 z 2− 1 2(c + 1) i z + h C + 1 16 2 ln 2− 1 + 2c + log2 z 2 2 − 3 8 ln 2 i +··· = z2 4 ...

  13. [20]

    E34 and Eq

    Numerical checks We compare the results Eq. E34 and Eq. E37 to numerical simulation of the model withδ close to zero, in Fig. 6(a). Because of the quick spreading of the system, we are only able to simulate the system up to t around 100 (depending on the value of δ), which cor...

  14. [21]

    A. I. Larkin and Y. N. Ovchinnikov, Soviet Journal of Experimental and Theoretical Physics 28, 1200 (1969)

    1969

  15. [22]

    S. H. Shenker and D. Stanford, J. High Energ. Phys. 2014, 67 (2014)

    2014

  16. [23]

    D. A. Roberts and D. Stanford, Physical Review Letters 115, 131603 (2015)

    2015

  17. [24]

    Kitaev, (2015), talks at KITP, April 7, 2015 and May 27, 2015

    A. Kitaev, (2015), talks at KITP, April 7, 2015 and May 27, 2015

    2015

  18. [25]

    Maldacena, S

    J. Maldacena, S. H. Shenker, and D. Stanford, J. High Energ. Phys. 2016, 106 (2016)

    2016

  19. [26]

    Nahum, S

    A. Nahum, S. Vijay, and J. Haah, Phys. Rev. X 8, 021014 (2018)

    2018

  20. [27]

    C. W. von Keyserlingk, T. Rakovszky, F. Pollmann, and S. L. Sondhi, Phys. Rev. X 8, 021013 (2018)

    2018

  21. [28]

    T. Zhou, S. Xu, X. Chen, A. Guo, and B. Swingle, Phys. Rev. Lett. 124, 180601 (2020)

    2020

  22. [29]

    Xu and B

    S. Xu and B. Swingle, Phys. Rev. X 9 (2019), 10.1103/physrevx.9.031048, arXiv: 1805.05376

    2019 arXiv

  23. [30]

    Chen and T

    X. Chen and T. Zhou, Phys. Rev. B 100, 064305 (2019). 22

    2019

  24. [31]

    Zhou and X

    T. Zhou and X. Chen, Phys. Rev. E 99, 052212 (2019)

    2019

  25. [32]

    D. A. Roberts, D. Stanford, and A. Streicher, J. High Energ. Phys. 2018, 122 (2018)

    2018

  26. [33]

    Qi and A

    X.-L. Qi and A. Streicher, J. High Energ. Phys. 2019 (2019), 10.1007/jhep08(2019)012, arXiv:1810.11958 [cond-mat, physics:hep-th, physics:quant-ph]

    2019 arXiv

  27. [35]

    Lashkari, D

    N. Lashkari, D. Stanford, M. Hastings, T. Osborne, and P. Hayden, J. High Energ. Phys. 2013 (2013), 10.1007/jhep04(2013)022

    2013 doi

  28. [36]

    Agarwal and S

    L. Agarwal and S. Xu, arXiv:2108.05810 [cond-mat, physics:hep-th, physics:quant-ph] (2021), arXiv:2108.05810 [cond-mat, physics:hep-th, physics:quant-ph]

    2021 arXiv

  29. [37]

    Erd˝ os and D

    L. Erd˝ os and D. Schr¨ oder, Mathematical Physics, Analysis and Geometry17, 441 (2014), arXiv:1407.1552

    2014 arXiv

  30. [38]

    Lucas, arXiv:1903.01468 [cond-mat, physics:hep-th, physics:quant-ph] (2019), arXiv:1903.01468 [cond-mat, physics:hep- th, physics:quant-ph]

    A. Lucas, arXiv:1903.01468 [cond-mat, physics:hep-th, physics:quant-ph] (2019), arXiv:1903.01468 [cond-mat, physics:hep- th, physics:quant-ph]

    2019 arXiv

  31. [39]

    Entanglement dynamics from universal low-lying modes,

    S. Vardhan and S. Moudgalya, “Entanglement dynamics from universal low-lying modes,” (2024), arXiv:2407.16763 [cond- mat, physics:hep-th, physics:quant-ph]

    2024

  32. [40]

    I. L. Aleiner, L. Faoro, and L. B. Ioffe, Ann. Phys-new. York. 375, 378 (2016), arXiv:1609.01251 [cond-mat, physics:hep-th, physics:quant-ph]

    2016 arXiv

  33. [41]

    Aron, ´E

    C. Aron, ´E. Brunet, and A. Mitra, Physical Review B 108, L241106 (2023), arXiv:2305.04958 [cond-mat, physics:quant- ph]

    2023 arXiv

  34. [42]

    C. Aron, E. Brunet, and A. Mitra, SciPost Physics 15, 042 (2023), arXiv:2212.13265 [cond-mat, physics:quant-ph]

    2023 arXiv

  35. [43]

    Kolmogorov, I

    A. Kolmogorov, I. Petrovsky, and N. Piskunov, Bulletin of Moscow State University Series A: Mathematics and Mechanics , 1 (1937)

    1937

  36. [44]

    R. A. Fisher, Ann. Eugenic. 7, 355 (1937), eprint: https://onlinelibrary.wiley.com/doi/pdf/10.1111/j.1469- 1809.1937.tb02153.x

    1937

  37. [45]

    Ablowitz and A

    M. Ablowitz and A. Zeppetella, B. Math. Biol. 41, 835 (1979)

    1979

  38. [46]

    Xu and B

    S. Xu and B. Swingle, Nat. Phys. 16, arXiv:1802.00801 (2019), arXiv:1802.00801 [quant-ph]

    2019 arXiv

  39. [47]

    Bentsen, S

    G. Bentsen, S. Sahu, and B. Swingle, arXiv:2104.07688 [cond-mat, physics:hep-th, physics:quant-ph] (2021), arXiv:2104.07688 [cond-mat, physics:hep-th, physics:quant-ph]

    2021 arXiv

  40. [48]

    D. A. Roberts and B. Swingle, Phys. Rev. Lett. 117, 091602 (2016)

    2016

  41. [49]

    C. R. Doering, C. Mueller, and P. Smereka, Physica A: Statistical Mechanics and its Applications Stochastic Systems: From Randomness to Complexity, 325, 243 (2003)

    2003

  42. [50]

    Hallatschek and D

    O. Hallatschek and D. S. Fisher, Proc. Natl. Acad. Sci. U.S.A. 111, E4911 (2014)

    2014

  43. [51]

    See Supplemental Material for an overview of the algorithm, estimation of β given equilibrium height, derivation of the duality relation and the scaling function close to α = 1

  44. [52]

    Chatterjee and P

    S. Chatterjee and P. S. Dey, Commun. Pur. Appl. Math. 69, 203 (2015)

    2015

  45. [53]

    T. Zhou, A. Y. Guo, S. Xu, X. Chen, and B. Swingle, Physical Review B 107, 014201 (2023), arXiv:2208.01649 [cond-mat, physics:physics, physics:quant-ph]

    2023 arXiv

  46. [54]

    G. A. ´Alvarez and D. Suter, Phys. Rev. A 84, 012320 (2011)

    2011

  47. [55]

    X. Mi, P. Roushan, C. Quintana, S. Mandr` a, J. Marshall, C. Neill, F. Arute, K. Arya, J. Atalaya, R. Babbush, J. C. Bardin, R. Barends, J. Basso, A. Bengtsson, S. Boixo, A. Bourassa, M. Broughton, B. B. Buckley, D. A. Buell, B. Burkett, N. Bushnell, Z. Chen, B. Chiaro, R. Col...

    2021

  48. [56]

    Spectral gaps of local quantum channels in the weak-dissipation limit,

    J. A. Jacoby, D. A. Huse, and S. Gopalakrishnan, “Spectral gaps of local quantum channels in the weak-dissipation limit,” (2024), arXiv:2409.17238 [cond-mat, physics:quant-ph]

    2024 arXiv

  49. [57]

    Absorbing state transitions with long-range annihila- tion,

    N. O’Dea, S. Bhattacharjee, S. Gopalakrishnan, and V. Khemani, “Absorbing state transitions with long-range annihila- tion,” (2024)

    2024

  50. [58]

    D. T. Gillespie, Journal of Computational Physics 22, 403 (1976)

    1976

  51. [59]

    D. T. Gillespie, The Journal of Physical Chemistry 81, 2340 (1977)

    1977

  52. [60]

    Y. Cao, D. T. Gillespie, and L. R. Petzold, The Journal of Chemical Physics 124, 044109 (2006)

    2006

  53. [61]

    D. T. Gillespie, The Journal of Chemical Physics 115, 1716 (2001)

    2001

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.