Pith. sign in

REVIEW 1 major objections 5 minor 1 cited by

Noise sensitivity for stochastic heat and Schr\"odinger equation

T0 review · 1 major / 5 minor · reviewed 2026-08-08 · deepseek-v4-flash

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

Pith's one-line read For the stochastic heat and Schrödinger equations, the Wiener chaos spectrum of the solution is asymptotically Gaussian, placing the onset of noise chaos at perturbation scale $1/t$.

desk verdict A correct and useful short paper: the chaos-order CLT for two linear SPDEs, with one asserted approximation that needs a referee's attention. read the letter →

arxiv 2502.04587 v1 pith:A5IRKUCH submitted 2025-02-07 math.PR

Yu Gu , Tomasz Komorowski This is my paper · ORCID
classification math.PR MSC 60H1560F0560H0735R60
keywords chaosnoisesensitivityFourierspectrumstochasticheatequationSchrödingerWienercentrallimittheorem
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 shows that for two canonical equations with multiplicative Gaussian noise—the Itô–Schrödinger equation with smooth spatial covariance and the one-dimensional stochastic heat equation driven by spacetime white noise—the distribution of the Wiener chaos order contributing to the solution's second moment becomes Gaussian after centering and rescaling by $\sqrt{t}$. Concretely, the Fourier spectrum $N_t$ satisfies $(N_t - \mu t)/(\sigma \sqrt{t}) \Rightarrow N(0,1)$, with $\mu = \sigma^2 = R(0)$ in the Schrödinger case and $\mu = \beta^4/2$, $\sigma^2 = \beta^4$ in the heat case with inverse temperature $\beta$. Because $N_t$ behaves like a Poisson random variable of order $t$, the correlation between the solution and the solution driven by a perturbed noise of strength $s$ decays like $e^{-sN_t}$, so perturbations far larger than $1/t$ destroy correlation while those far smaller leave it intact. The paper's contribution is to make the onset of chaos quantitative in two models where every coefficient of the Wiener chaos expansion is explicit.

What carries the argument

The central object is the Fourier spectrum $N_t$, the probability distribution on non-negative integers that assigns to each $n$ the fraction of the solution's second moment carried by the $n$-th Wiener chaos. For the Schrödinger case the load-bearing computation is the exact coefficient $c^2_{t,n} = (t^n/n!) e^{-R(0)t} \int R(y)^n \Phi_0(y)\,dy$, which makes the characteristic function of $N_t$ a ratio of integrals that can be evaluated by Laplace asymptotics. For the stochastic heat equation the load-bearing mechanism is the identity $P_s Z_\beta(t,0) = Z_{\beta e^{-s}}(t,0)$ for the Ornstein–Uhlenbeck semigroup—the semigroup that damps the $n$-th chaos by $e^{-ns}$—which converts the generating function $E e^{-sN_t}$ into a ratio of explicit second moments. The normalization comes from the exact formula $E[Z_\beta(t,0)^2] = 2e^{\beta^4 t/4} \int_{-\infty}^{\beta^2 \sqrt{t/2}} (1/\sqrt{2\pi}) e^{-y^2/2}\,dy$.

What would settle it

Approximate spacetime white noise by smooth mollified noise and compute the ratio $E[Z_\beta(t,0) P_s Z_\beta(t,0)]/E[Z_\beta(t,0)^2]$ as the mollification width goes to zero; if the limit is not $f(\beta e^{-s/2},t)/f(\beta,t)$, where $f(\beta,t)=E[Z_\beta(t,0)^2]$, the load-bearing identity fails and the central limit theorem for $N_t$ does not follow.

Watch

Extended reading notes

Core claim

The paper establishes a central limit theorem for the Fourier spectrum $N_t$, defined by $P(N_t=n)=c^2_{t,n}/E X_t^2$, where $c^2_{t,n}$ is the contribution of the $n$-th Wiener chaos to the second moment of the solution. For $X_t = \hat{\phi}(t,0)$ of the Itô–Schrödinger equation, the exact formula $c^2_{t,n} = (t^n/n!) e^{-R(0)t} \int R(y)^n \Phi_0(y)\,dy$ turns the characteristic function of $N_t$ into a ratio of exponential integrals, and a Taylor expansion around $y=0$ yields $(N_t - R(0)t)/\sqrt{R(0)t} \Rightarrow N(0,1)$. For $X_t = Z(t,0)$ of the stochastic heat equation, the paper proves the Ornstein–Uhlenbeck semigroup identity $P_s Z_\beta(t,0) = Z_{\beta e^{-s}}(t,0)$, giving the exact Laplace transform $E e^{-sN_t} = E Z_{\beta e^{-s/2}}(t,0)^2 / E Z_\beta(t,0)^2$; analytic continuation of this formula gives $(N_t - \beta^4 t/2)/(\beta^2 \sqrt{t}) \Rightarrow N(0,1)$. In both cases $N_t$ is asymptotically equivalent to a Poisson variable with intensity of order $t$, so the solution's second moment is dominated by Wiener chaos of order $t$.

Load-bearing premise

The stochastic heat equation proof assumes, with the limiting argument left out, that applying the noise-smoothing semigroup to the solution with inverse temperature $\beta$ gives exactly the solution with inverse temperature $\beta e^{-s}$ even for spacetime white noise.

Editorial extensions

If this is right

  • If the noise is perturbed with strength $s\sim t^{-\alpha}$, the correlation $\mathrm{Cor}[X_t(0),X_t(s)]$ tends to $1$ for $\alpha>1$ and to $0$ for $\alpha<1$; the onset of chaos is exactly at $s\sim t^{-1}$.
  • The dominant Wiener chaos order grows linearly in $t$, so the second moment of the solution is concentrated on chaos of order comparable to $t$, not on low-order chaos.
  • Since $N_t$ becomes a Poisson variable of intensity $R(0)t$ or $\beta^4 t/2$, the Gaussian limit is the usual Poisson-to-normal transition for a large intensity parameter.
  • In the Schrödinger picture, $N_t$ can be read as the number of scatterings of an underlying compound Poisson process, tying the chaos spectrum to the kinetic equation for the second moment.

Reading between the lines

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

  • The identity $P_s Z_\beta = Z_{\beta e^{-s}}$ is likely to hold for other Gaussian noises and initial data, in which case the same Poisson-type CLT should hold in any strong-disorder regime with exponentially growing second moment; the paper conjectures similar behavior but does not prove it.
  • The geometric-Brownian analogy suggests viewing $N_t$ as a count of branching or collision events in a hidden particle picture, a viewpoint that could link chaos spectra of SPDEs with overlap statistics in last-passage percolation.
  • One could test the paper's closing conjecture that the free energy $\log Z(t,0)$ has Fourier spectrum of order $t^{1/3}$ by numerically estimating its low-order chaos coefficients, providing evidence for or against a $t^{1/3}$ chaos time scale.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

1 major / 5 minor

Summary. The paper studies the Fourier spectrum N_t of the spatial Fourier mode at the origin of two linear SPDEs with multiplicative Gaussian noise: the stochastic Schrödinger equation in Stratonovich form and the 1+1-dimensional stochastic heat equation. The Fourier spectrum is defined as the probability distribution on Wiener chaos orders determined by the normalized second-moment contributions. Theorem 1.1 claims a central limit theorem: (N_t - μt)/(σ√t) converges in law to a standard normal, with explicit constants in the proofs (μ=σ²=R(0) for the Schrödinger equation; μ=β⁴/2, σ²=β⁴ for the stochastic heat equation). The proof for the Schrödinger case uses exact chaos coefficients and Laplace's method after a change of variables, while the heat-equation case combines the Chen–Dalang second-moment formula with an Ornstein–Uhlenbeck semigroup identity. The paper derives as a consequence that the onset of chaos occurs at perturbation strength s∼1/t.

Significance. The result gives a sharp, quantitative description of noise-induced chaos for two canonical SPDEs: the second moment of the solution is asymptotically dominated by Wiener chaos of order proportional to t, and the Fourier spectrum is approximately Poisson with mean of order t. The proofs are explicit and use no fitted parameters: the Schrödinger computation is a direct saddle-point calculation from exact chaos coefficients, and the heat-equation part reduces to known exact second-moment asymptotics. The derived 1/t decorrelation threshold is a clear, falsifiable prediction. The main shortcoming is a missing justification for the white-noise identity (3.3), but the statement is true and a short chaos-expansion proof would close the gap; after that revision the paper would be a clean and self-contained note.

major comments (1)
  1. [Section 3, Eq. (3.3)] The identity P_s Z_β(t,0)=Z_{βe^{-s}}(t,0) for spacetime white noise is asserted after a smooth-noise Feynman–Kac verification, with the sentence 'An approximation leads to the same conclusion' and no limiting argument. This identity is the only bridge between the Ornstein–Uhlenbeck semigroup and the explicit second-moment formula f(β,t), so the exact Laplace transform (3.5) and the SHE half of Theorem 1.1 rest on it. The omitted limit is nontrivial because for white noise the Feynman–Kac expression involves the divergent constant R(0)t and one must track how the Wick renormalization interacts with the Gaussian average over the independent noise copy. The claim is nevertheless true: for the mild solution, the n-th Wiener chaos of Z_β(t,0) is β^n times a β-independent chaos coefficient, and the OU semigroup P_s multiplies that chaos by e^{-ns}. I recommend that the authors add this direct chaos-expansion verification or a complete approximation argument before acceptance.
minor comments (5)
  1. [Theorem 1.1] The theorem states only that 'there exist σ, μ>0', but the proofs determine them explicitly (μ=σ²=R(0) for the Schrödinger case; μ=β⁴/2, σ²=β⁴ for the SHE case). Stating these values in the theorem would make the result more precise and would make the correlation corollary in §1.3 immediate.
  2. [Section 2, near Eq. (2.5)] The sentence 'which we assume is strictly negative definite' introduces a hypothesis on ∇²R(0) that is not listed in Theorem 1.1 or in the case-1 assumptions. Under the stated Schwartz-class and positive-definiteness assumptions, this strictness is in fact automatic for a nondegenerate covariance because the continuous nonnegative spectral density cannot be supported on a hyperplane; however, the paper should either justify this or move the condition into the theorem statement to avoid the appearance of an unstated assumption.
  3. [Section 3, Eq. (3.2)] The phrase 'where F(⋅) is arbitrary square integrable function' should read 'an arbitrary square-integrable function'.
  4. [Section 2, characteristic-function display] The display 'E exp{iθ Nt−R(0)t√t }' contains a parenthesis typo and should be written as 'E exp{iθ (N_t−R(0)t)/√t}'.
  5. [Section 3, proof of (3.3) for smooth noise] The interchange of the Brownian expectation and the Gaussian expectation over the independent copy of the noise is not explicitly justified; a brief appeal to stochastic Fubini would make the smooth-noise argument more transparent.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity; the derivation is self-contained apart from a cited Fourier-domain equation and one unproved approximation, neither of which reduces the result to its inputs.

full rationale

The paper obtains the Fourier spectrum N_t explicitly from Wiener chaos expansions. For the Schrödinger case, equation (2.4) gives exact chaos coefficients, from which the characteristic function and CLT are derived directly. The self-citation [5] is used only to write the Fourier-domain form of the equation; it does not contain the theorem and is not fitted. For the stochastic heat equation, the second moment formula is quoted from [2] (Chen and Dalang, independent external work), and the key identity (3.3) is verified for smooth noise; the white-noise extension is asserted by 'An approximation leads to the same conclusion,' which is a missing justification but not a circular reduction: the identity does not assume the target CLT. The proof of (3.1) uses the standard OU semigroup action on Wiener chaos. No parameter is fitted to the target quantity, no prediction is defined in terms of the result, and no uniqueness claim is imported. The remaining concern is rigor, not circularity.

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

No fitted parameters: the centering mu and variance sigma^2 are explicit functions of the input data, R(0) for Schrodinger and beta for SHE, not chosen to make the theorem work. The proof imports three external results: the Fourier-domain form of the Schrodinger equation from [5], the exact second moment of the SHE from [2], and the Ornstein-Uhlenbeck semigroup calculus from [6]. No new entities are introduced.

assumptions (4)
  • domain assumption The solutions phi(t,0) and Z(t,0) lie in L^2 and admit the Wiener chaos expansion with the second-moment split (1.2).
    The definition of N_t in (1.3) requires this; for the Schrodinger equation the paper relies on [5, Section 5.1] for the transformed equation and on [6] for chaos expansion background.
  • domain assumption The Hessian of R at 0 is strictly negative definite, so the Gaussian integral in the Laplace-method evaluation converges.
    Invoked in Section 2's Taylor expansion to evaluate the denominator limit; not stated in Theorem 1.1. It is automatic for non-degenerate Schwartz covariance functions but remains an unstated premise.
  • standard math Chen and Dalang's exact second-moment formula for the SHE: E[Z_beta(t,0)^2] = 2 exp(beta^4 t/4) Phi(beta^2 sqrt(t)/2), Corollary 2.5 of [2].
    Quoted in Section 3 and used to write (3.5); the entire SHE half of the theorem depends on this exact formula.
  • standard math The Ornstein-Uhlenbeck semigroup action P_s X = sum_n e^{-ns} I_n, and the Feynman-Kac representation of the SHE, including its extension from smooth noise to white noise.
    Used to prove (3.3) and (3.4); the smooth-to-white-noise passage is only sketched in Section 3.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Noise sensitivity for stochastic heat and Schr\"odinger equation." pith.science (2026). https://pith.science/paper/A5IRKUCH

@misc{pith2026250204587,
  author       = {Pith},
  title        = {Pith review of: Noise sensitivity for stochastic heat and Schr\"odinger equation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/A5IRKUCH}},
  note         = {Machine review of arXiv:2502.04587}
}
abstract

In this note, we consider the stochastic heat and Schr\"odinger equation, and show that, at time $t$, the onset of the chaos occurs on the scale of $1/t$, and the Fourier spectrum of the solution is asymptotically Gaussian after centering and rescaling.

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. Enhanced noise sensitivity, 2D directed polymers and Stochastic Heat Flow

    math.PR 2025-07 conditional novelty 7.0 of 10

    A general, rate-optimal BKS noise-sensitivity criterion is proven, and it yields the independence of the critical 2D Stochastic Heat Flow from the disorder white noise.

Reference graph

Works this paper leans on

6 extracted references · 6 canonical work pages · cited by 1 Pith paper

  1. [5]

    Gaussian fluctuations from random Schr¨ odinger equation, Communications in Partial Differential Equations 46.2 (2021): 201-232

    Yu Gu and Tomasz Komorowski. Gaussian fluctuations from random Schr¨ odinger equation, Communications in Partial Differential Equations 46.2 (2021): 201-232. 3

    work page 2021

  2. [1]

    Superconcentration and related topics, Vol

    Sourav Chatterjee. Superconcentration and related topics, Vol. 15. Cham: Springer, 2014. 3

    work page 2014

  3. [2]

    Le Chen and Robert C. Dalang. Moments and growth indices for the nonlinear stochastic heat equation with rough initial conditions , Annals of Probability (2015): 3006-3051. 6

    work page 2015

  4. [3]

    Stability and chaos in dynamical last passage percolation, Communications of the American Mathematical Society 4.09 (2024): 387-479

    Shirshendu Ganguly and Alan Hammond. Stability and chaos in dynamical last passage percolation, Communications of the American Mathematical Society 4.09 (2024): 387-479. 3

    work page 2024

  5. [4]

    Christophe Garban and Jeffrey E. Steif. Noise sensitivity of Boolean functions and percolation , Vol. 5. Cambridge University Press, 2014. 1

    work page 2014

  6. [6]

    Malliavin calculus and related topics , (2006)

    David Nualart. Malliavin calculus and related topics , (2006). 1, 6 (Yu Gu) Department of Mathematics, University of Maryland, College Park, MD 20742, USA. Email address : yugull05@gmail.com (Tomasz Komorowski) Institute of Mathematics, Polish Academy of Sciences, ul. ´Sniadeckich 8, 00-636 W arsaw, Poland. Email address : tkomorowski@impan.pl

    work page 2006

Pith tools

Reviewed August 8, 2026 · model on record in the stance chip above.