Pith. sign in

REVIEW 3 major objections 5 minor 29 references

Stochastic oscillators out of equilibrium: scaling limits and correlation estimates

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 For a stochastic harmonic oscillator chain, this paper proves two autonomous heat equations when the Hamiltonian part decays faster than 1/N, a coupled system at the critical rate, and a martingale characterization of volume fluctuations.

arxiv 2505.10256 v1 pith:OVIEIPQ4 submitted 2025-05-15 math.PR

Patrícia Gonçalves , Kohei Hayashi , João Pedro Mangi This is my paper · ORCID
classification math.PR MSC 60H1560J2782C2260K50
keywords Bernardin–Stoltzmodelharmonicoscillatorshydrodynamiclimitnon-equilibriumfluctuationsOrnstein–UhlenbeckprocessEdwards–Wilkinsonuniversalitytwo-pointcorrelationdecayfourth-momentestimates
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

The paper studies the Bernardin–Stoltz model, a harmonic chain of oscillators with random nearest-neighbor exchanges, under diffusive scaling and with Hamiltonian strength $\alpha_N=\alpha N^{-\kappa}$. It proves that when $\kappa>1$ the two conserved macroscopic fields—volume and energy—each converge to the solution of a plain heat equation, $\partial_t v=\Delta v$ and $\partial_t e=\Delta e$, so the conservation laws decouple. At the critical value $\kappa=1$ the limiting system is coupled: $\partial_t v=\Delta v+2\alpha\nabla v$ and $\partial_t e=\Delta e+2\alpha\nabla v^2$, with volume convergence in probability and energy convergence in expectation, plus an $O(\log N/N)$ bound for smooth profiles. It also characterizes non-equilibrium volume fluctuations for general initial states satisfying a short-range correlation condition: limit points have the form $V_t(f)=V_0(T_t f)+M_t(f)$, where $M_t$ is a mean-zero martingale whose mean quadratic variation is $\int_0^t\int 2\chi(s,u)(\nabla f)^2\,du\,ds$, with $\chi=e-v^2$. A sympathetic reader would care because this gives a rigorous, parameter-dependent derivation of macroscopic heat and drift equations from a microscopic unbounded-state-space dynamics, together with the correlation estimates needed to control fluctuations out of equilibrium.

What carries the argument

The argument is carried by the two-point volume correlation function $\varphi^N_t(x,y)=E[\eta_t(x)\eta_t(y)]-v^N_t(x)v^N_t(y)$, defined off the diagonal of $\mathbb{T}_N^2$. Its evolution equation is governed by a two-dimensional random walk that is reflected on the diagonal lines $y=x\pm1$; via Duhamel's principle the correlation is written as an expectation of that walk, and the key estimate bounds the walk's local time at the diagonal by $O(1/N)$, yielding $\sup_t\|\varphi^N_t\|_{\ell^\infty}\le C/N$ for $\kappa\ge1$. A second load-bearing object is the discrete $H^{-1}$ norm $\|\eta^2\|_{-1,N}$, whose generator computation produces the uniform fourth-moment bound $\int_0^T N^{-1}\sum_x E[\eta_s(x)^4]\,ds\le C(1+\alpha_N N)$. Together these estimates replace the one- and two-block replacement steps that are unavailable for unbounded real-valued variables, and they power both the hydrodynamic-limit proofs and the fluctuation field.

What would settle it

Start the $\kappa=1$ dynamics from a smooth profile with nonzero $\nabla v_0$ and measure the energy profile: the coupled equation predicts a visible $2\alpha\nabla v^2$ source term, so its absence would disprove Theorem 2.8. Independently, begin from a product Gaussian with $O(1)$ long-range covariance: Theorem 2.14 predicts $\sup_{0\le t\le T}\|\varphi^N_t\|_{\ell^\infty}\le C/N$, so a numerical violation of that decay would falsify the correlation estimate and the tightness arguments built on it.

Watch

Extended reading notes

Core claim

The central discovery is a decoupling–coupling transition controlled by $\kappa$, the exponent in $\alpha_N=\alpha N^{-\kappa}$. For $\kappa>1$, Theorem 2.5 shows the volume and energy empirical measures converge in probability to $v$ and $e$ solving the two autonomous heat equations. For $\kappa=1$, Theorem 2.7 shows the volume field converges in probability to $\partial_t v=\Delta v+2\alpha\nabla v$, while Theorem 2.8 shows the energy profile converges at rate $\sup_t\max_x |e^N_t(x)-e(t,x/N)|\le C\log N/N$ to $\partial_t e=\Delta e+2\alpha\nabla v^2$. For fluctuations, Theorem 2.11 proves tightness of the volume fluctuation field and identifies every limit point as $V_t(f)=V_0(T_t f)+M_t(f)$ with a mean-zero martingale $M_t$ of mean quadratic variation $\int_0^t\int 2\chi(s,u)(\nabla f)^2\,du\,ds$; from invariant Gaussian product measures the limit is the Ornstein–Uhlenbeck equation $dV_t=\Delta V_t\,dt+\sqrt{2/\beta}\,\nabla\dot W_t$.

Load-bearing premise

The proof needs the initial two-point volume correlations to decay as $C/N$, together with an initial $H^{-1}$ bound on the squared configuration; an initial state with order-one long-range correlations is not covered, and the correlation estimate and fluctuation characterization would likely break down.

Editorial extensions

If this is right

  • For $\kappa>1$, volume and energy evolve macroscopically as two independent heat equations; no coupling term survives in the limit.
  • At $\kappa=1$, energy transport is slaved to volume: a nonzero volume gradient produces a source term $2\alpha\nabla v^2$ in the energy equation.
  • Out-of-equilibrium volume fluctuations are universal in form for the allowed initial measures: every limit point is a heat-evolved initial field plus a mean-zero martingale, with the noise intensity fixed by the local static compressibility $\chi=e-v^2$.
  • Two-point volume correlations decay as $C/N$ uniformly up to a fixed time for $\kappa\ge1$, so second-order structure is tractable even though the state space is unbounded.
  • For smooth data at $\kappa=1$, the discrete energy profile is within $O(\log N/N)$ of the solution of the coupled PDE uniformly in space and time.

Reading between the lines

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

  • If the correlation decay extends to four-point functions, the volume fluctuation limit should be fully characterized rather than merely tight, giving the noise term in the volume equation; the paper explicitly leaves this as an open problem.
  • The same machinery suggests that for $\kappa<1$, where the proof's absorption step fails, the volume–energy coupling may produce superdiffusive or anomalous transport; that is an implicit prediction, not a theorem of this paper.
  • One could test the role of the initial-state assumption by running the dynamics from a correlated Gaussian initial measure with $O(1)$ long-range covariances: the predicted $C/N$ decay of $\varphi^N_t$ should fail, separating the contribution of the initial condition from the dynamics.
  • The $O(\log N/N)$ energy-profile rate is likely not optimal; a sharper estimate would require better short-time control of the diagonal local time and could be checked numerically against the actual sup-norm discrepancy.
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 studies the Bernardin-Stoltz model of a harmonic chain perturbed by an exchange noise, with the Hamiltonian strength scaled by α_N = α N^{-κ} and in the diffusive time scale N^2. The main results are: (i) for κ>1, a hydrodynamic limit in probability for both volume and energy empirical measures, converging to two autonomous heat equations (Theorem 2.5); (ii) for κ=1, a hydrodynamic limit in probability for the volume only, with a drift term (Theorem 2.7), and a quantitative L∞-rate estimate for the energy profile under stronger smooth/product initial data (Theorem 2.8); (iii) a non-equilibrium fluctuation result for the volume field, tightness and partial characterization of the limit martingale (Theorem 2.11). The proofs are based on entropy-method tightness, Dynkin martingales, a uniform fourth-moment bound (Theorem 2.13), and a decay estimate for the two-point volume correlation (Theorem 2.14), the latter being the main technical ingredient.

Significance. If the stated results are correct, the paper would extend rigorous hydrodynamic limit and non-equilibrium fluctuation results for the Bernardin-Stoltz model in a regime where the two conserved quantities do not evolve autonomously, and it would provide quantitative energy-profile estimates. The proof strategy is clearly organized and uses standard tools (entropy method, Duhamel representations, random-walk estimates), with explicitly stated assumptions and no fitted parameters. The paper also honestly identifies several open problems, such as the lack of a full characterization of the limit martingale and of energy correlation decay. These strengths make the work potentially useful to the stochastic-particle-systems community. However, two load-bearing technical issues in the correlation estimate and in the fluctuation theorem need to be resolved before the central claims can be accepted.

major comments (3)
  1. [Section 6, Lemma 6.2, Eq. (6.4)] The estimate (6.4) cannot hold as stated for κ>1. For a non-constant smooth initial profile v0, the term -(∇_N v^N_0(x))^2 in the definition of g^N_0(x), Eq. (6.1), is generically of order 1, while α_N^2 N^2 = α^2 N^{2-2κ} tends to 0. Even if the α_N N^2 h^N_0 term is small, this leaves the left-hand side of (6.4) of order 1 and the right-hand side tending to 0, a contradiction. The proof of Lemma 6.2 does not repair this: from the bound on h^N in the proof one obtains ‖g^N‖ bounded by C + Cα_N N + Cα_N^2 N^{8/3}‖φ^N‖, not by (6.4). Since (6.4) is the input to the bootstrap (6.5) proving Theorem 2.14, the proof of the key correlation decay estimate is invalid as written. A corrected bound of the form ‖g^N‖ ≤ C(1 + N^{2/3}‖φ^N‖) might still imply Theorem 2.14, so the result may be salvageable, but the present argument does not establish it.
  2. [Section 4.4, proof of Theorem 2.11(b), Eq. (2.16)] The proof of item (b) explicitly invokes (2.14), i.e. the quantitative energy-profile estimate of Theorem 2.8, to replace e^N_s by e(s,·) in the quadratic variation. However Theorem 2.8 requires v0,e0 ∈ C∞_b, a product initial measure of the form (2.9), and initial data (2.13); none of these assumptions appear in Theorem 2.11, whose stated hypotheses are only Assumption 2.1, bounded measurable profiles, and convergence of the initial fluctuation field. Moreover, Theorem 2.7 explicitly leaves the energy hydrodynamic limit open for κ=1 in the general setting. Consequently, the convergence needed for (2.16) is not proved under the hypotheses of Theorem 2.11. The statement must either restrict Theorem 2.11 to the smooth product setting of Theorem 2.8 or supply a new energy-profile estimate valid under Assumption 2.1.
  3. [Section 6, proof of Theorem 2.14] Related to the previous comment, the bootstrap in (6.5) also relies on the specific form of (6.4). If (6.4) is replaced by the correct order-1 bound for ‖g^N‖, the factor (1/N)‖g^N‖ would give a contribution of order 1/N, which is still compatible with the desired conclusion; but the manuscript does not provide that argument. As written, the proof of Theorem 2.14 is incomplete, and Theorem 2.14 is used in the proofs of Theorems 2.5, 2.7, 2.8, and 2.11.
minor comments (5)
  1. [Theorem 2.5] There is a typo: 'weal solution' should read 'weak solution'.
  2. [Corollary 3.6] The proof refers to 'Theorem 3.5' when the intended statement is Lemma 3.5.
  3. [Abstract and Introduction] The phrase 'general initial measures' overstates the hypotheses: Assumption 2.1 requires the initial two-point volume correlation to be O(1/N) pointwise and imposes uniform bounds on discrete gradients of the initial profiles; Example 2.2 exhibits only a subclass of such measures. The wording should be adjusted to match the stated assumptions.
  4. [Section 6] The text refers to 'Theorem 6.1' and 'Theorem 6.2' where the objects are labelled as Lemma 6.1 and Lemma 6.2; this should be corrected for consistency.
  5. [Lemma 3.11 proof] In the Aldous-criterion estimate, the displayed probability incorrectly mixes the volume martingale and the energy martingale: it should read |M^{e,N}_{τ+θ}(f)-M^{e,N}_τ(f)|, not |M^{v,N}_{τ+θ}(f)-M^{e,N}_τ(f)|.

Circularity Check

0 steps flagged · score 1.0 of 10

No circularity: the macroscopic equations are derived from the generator, and the only flagged issue is a hypotheses gap in the proof of Theorem 2.11(b), not a circular reduction.

full rationale

The derivation chain is not circular. The hydrodynamic equations are obtained by applying the generator to the empirical measures through Dynkin martingales in Section 3, and the limiting equations (2.12) are not used as inputs: the discrete profile equations (3.1) are derived directly from the generator, and the correlation evolution (A.2) is computed explicitly in Appendix A. The key estimates, namely the fourth-moment bound (Theorem 2.13) and the two-point correlation decay (Theorem 2.14), are proved in Sections 5 and 6 rather than assumed. Assumption 2.1 imposes initial-data bounds only and does not contain the target limits; Example 2.2 shows that nontrivial initial measures satisfy those bounds. Self-citations such as [13] appear in the introduction as background and are not load-bearing, and the uniqueness of the weak solution is justified by standard arguments. The skeptic's concern is a genuine proof gap: the proof of Theorem 2.11(b) in Section 4.4 invokes the quantitative energy estimate (2.14), which is Theorem 2.8 and is stated under stronger hypotheses (smooth profiles and product initial measures of the form (2.9)) than those assumed in Theorem 2.11. This is a missing-support or completeness issue, not a circular reduction, because the estimate is independently proved and not equivalent to the statement of Theorem 2.11. Therefore the circularity burden remains low.

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

No new entities and no fitted parameters. The model parameters α and κ and the initial profiles v_0 and e_0 are inputs from the problem statement rather than quantities fitted to close the derivation. Proof-internal optimization choices (e.g., t_0 = N^{-4/3} in Lemma 6.2) do not enter the final estimates and are not counted. The two key estimates (fourth-moment bound and correlation decay) are proved inside the paper.

assumptions (4)
  • domain assumption Initial measures satisfy Assumption 2.1, including the C/N decay of two-point volume correlations and a bounded H^{-1} norm of the squared initial configuration.
    Needed for Theorems 2.5, 2.7, 2.8, and 2.11; Example 2.2 shows Gaussian product measures with smooth profiles satisfy it. Section 2.2.1.
  • domain assumption Diffusive scaling N^2 and weak asymmetry α_N = α N^{-κ} with κ ≥ 1, on the one-dimensional torus T_N.
    Defines the regime: the Hamiltonian part is subcritical (κ>1) or critical (κ=1) relative to the exchange noise. Section 2.1.1.
  • standard math Standard probabilistic machinery: Dynkin's martingale formula, Aldous's tightness criterion, Duhamel's principle, and heat kernel estimates for continuous-time random walks.
    Used throughout Sections 3-6 and Appendix B without proof.
  • domain assumption Harmonic potential V(η)=η^2/2 and exchange noise with uniform rates.
    Restricts the BS model to the purely harmonic, spatially homogeneous exchange-noise case; the paper does not treat anharmonic potentials or boundary-driven versions.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Stochastic oscillators out of equilibrium: scaling limits and correlation estimates." pith.science (2026). https://pith.science/paper/OVIEIPQ4

@misc{pith2026250510256,
  author       = {Pith},
  title        = {Pith review of: Stochastic oscillators out of equilibrium: scaling limits and correlation estimates},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/OVIEIPQ4}},
  note         = {Machine review of arXiv:2505.10256}
}
read the original abstract

We consider a purely harmonic chain of oscillators which is perturbed by a stochastic noise. Under this perturbation, the system exhibits two conserved quantities: the volume and the energy. At the level of the hydrodynamic limit, under diffusive scaling, we show that depending on the strength of the Hamiltonian dynamics, energy and volume evolve according to either a system of autonomous heat equations or a non-linear system of coupled parabolic equations. Moreover, for general initial measures, under diffusive scaling, we can characterize the non-equilibrium volume fluctuations. The proofs are based on precise bounds on the two-point volume correlation function and a uniform fourth-moment estimate.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

29 extracted references · 22 canonical work pages

  1. [1]

    Ahmed, C

    R. Ahmed, C. Bernardin, P. Gon¸ calves, and M. Simon. A microscopic derivation of coupled SPDE’s with a KPZ flavor.Annales de l’Institut Henri Poincar´ e, Probabilit´ es et Statistiques, 58(2):890–915, 2022

    2022

  2. [2]

    Bernardin

    C. Bernardin. Hydrodynamics for a system of harmonic oscillators perturbed by a conservative noise.Sto- chastic processes and their applications, 117(4):487–513, 2007

    work page 2007

  3. [3]

    Bernardin and P

    C. Bernardin and P. Gon¸ calves. Anomalous fluctuations for a perturbed Hamiltonian system with exponential interactions.Communications in Mathematical Physics, 325(1):291–332, 2014

    work page 2014

  4. [4]

    Bernardin, P

    C. Bernardin, P. Gon¸ calves, and M. Jara. 3/4-fractional superdiffusion in a system of harmonic oscillators perturbed by a conservative noise.Archive for rational mechanics and analysis, 220(2):505–542, 2016

    work page 2016

  5. [5]

    Bernardin, P

    C. Bernardin, P. Gon¸ calves, and M. Jara. Weakly harmonic oscillators perturbed by a conservative noise. The Annals of Applied Probability, 28(3):1315–1355, 2018

    work page 2018

  6. [6]

    Bernardin, P

    C. Bernardin, P. Gon¸ calves, M. Jara, and M. Simon. Interpolation process between standard diffusion and fractional diffusion.Annales de l’Institut Henri Poincar´ e, Probabilit´ es et Statistiques, 54(3):1731–1757, 2018

    2018

  7. [7]

    Bernardin, P

    C. Bernardin, P. Gon¸ calves, M. Jara, and M. Simon. Nonlinear perturbation of a noisy Hamiltonian lattice field model: universality persistence.Communications in Mathematical Physics, 361(2):605–659, 2018

    work page 2018

  8. [8]

    Bernardin and G

    C. Bernardin and G. Stoltz. Anomalous diffusion for a class of systems with two conserved quantities.Non- linearity, 25(4):1099, 2012

    2012

Show all 29 references
  1. [9]

    Bertini, D

    L. Bertini, D. Gabrielli, and C. Landim. Concurrent Donsker–Varadhan and hydrodynamical large deviations. The Annals of Probability, 51(4):1298–1341, 2023

    2023

  2. [10]

    Franceschini, P

    C. Franceschini, P. Gon¸ calves, M. Jara, and B. Salvador. Non-equilibrium fluctuations for SEP(α) with open boundary.Stochastic Processes and their Applications, 178, 104463, 2024

    2024

  3. [11]

    Franco, P

    T. Franco, P. Gon¸ calves, and A. Neumann. Non-equilibrium and stationary fluctuations of a slowed boundary symmetric exclusion.Stochastic Processes and their Applications, 129(4):1413–1442, 2019

    2019

  4. [12]

    Gess and V

    B. Gess and V. Konarovskyi. A quantitative central limit theorem for the simple symmetric exclusion process. arXiv preprint arXiv:2408.01238, 2024

    2024 arXiv

  5. [13]

    Gon¸ calves and K

    P. Gon¸ calves and K. Hayashi. Derivation of anomalous behavior from interacting oscillators in the high- temperature regime.Communications in Mathematical Physics, 403(3):1193–1243, 2023

    2023

  6. [14]

    Gon¸ calves and M

    P. Gon¸ calves and M. Jara. Nonlinear fluctuations of weakly asymmetric interacting particle systems.Archive for Rational Mechanics and Analysis, 212(2):597–644, 2014

    2014

  7. [15]

    Gon¸ calves, M

    P. Gon¸ calves, M. Jara, R. Marinho, and O. Menezes. CLT for NESS of a reaction-diffusion model.Probability Theory and Related Fields, 190(1):337–377, 2024

    2024

  8. [16]

    Gon¸ calves, M

    P. Gon¸ calves, M. Jara, O. Menezes, and A. Neumann. Non-equilibrium and stationary fluctuations for the SSEP with slow boundary.Stochastic Processes and their Applications, 130(7):4326–4357, 2020

    2020

  9. [17]

    Gon¸ calves, C

    P. Gon¸ calves, C. Landim, and A. Milan´ es. Nonequilibrium fluctuations of one-dimensional boundary driven weakly asymmetric exclusion processes.The Annals of Applied Provaility, 27(1):140–177, 2017

    2017

  10. [18]

    Gon¸ calves, N

    P. Gon¸ calves, N. Perkowski, and M. Simon. Derivation of the stochastic Burgers equation with Dirichlet boundary conditions from the WASEP.Annales Henri Lebesgue, 3:87–167, 2020

    2020

  11. [19]

    M. Z. Guo, G. C. Papanicolaou, and S. S. Varadhan. Nonlinear diffusion limit for a system with nearest neighbor interactions.Communications in Mathematical Physics, 118(1):31–59, 1988

    1988

  12. [20]

    M. Jara, T. Komorowski, and S. Olla. Superdiffusion of energy in a chain of harmonic oscillators with noise. Communications in Mathematical Physics, 339(2):407–453, 2015

    2015

  13. [21]

    Kipnis and C

    C. Kipnis and C. Landim.Scaling limits of interacting particle systems, volume 320. Springer Science & Business Media, 1999

    1999

  14. [22]

    Kipnis and S

    C. Kipnis and S. S. Varadhan. Central limit theorem for additive functionals of reversible Markov processes and applications to simple exclusions.Communications in Mathematical Physics, 104(1):1–19, 1986

    1986

  15. [23]

    Komorowski and S

    T. Komorowski and S. Olla. Kinetic limit for a chain of harmonic oscillators with a point langevin thermostat. Journal of Functional Analysis, 279(12):108764, 2020

    2020

  16. [24]

    Komorowski and S

    T. Komorowski and S. Olla. Thermal boundaries in kinetic and hydrodynamic limits. InRecent Advances in Kinetic Equations and Applications, pages 253–288. Springer, 2021

    2021

  17. [25]

    Komorowski, S

    T. Komorowski, S. Olla, and L. Ryzhik. Fractional diffusion limit for a kinetic equation with an interface. The Annals of Probabability, 48(5):2290–2322, 2020

    2020

  18. [26]

    Komorowski, S

    T. Komorowski, S. Olla, and M. Simon. Macroscopic evolution of mechanical and thermal energy in a harmonic chain with random flip of velocities.Kinetic and Related Models, 11(3), 2018

    2018

  19. [27]

    Komorowski, S

    T. Komorowski, S. Olla, and M. Simon. Hydrodynamic limit for a chain with thermal and mechanical boundary forces.Electron. J. Probab., 26:1–49, 2021

    2021

  20. [28]

    Landim, A

    C. Landim, A. Milan´ es, and S. Olla. Stationary and nonequilibrium fluctuations in boundary driven exclusion processes.Markov Processes and Related Fields, 14(2):165–184, 2008

    2008

  21. [29]

    Suzuki and K

    Y. Suzuki and K. Uchiyama. Hydrodynamic limit for a spin system on a multidimensional lattice.Probability theory and related fields, 95:47–74, 1993. STOCHASTIC OSCILLATORS OUT OF EQUILIBRIUM 39 Center for Mathematical Analysis, Geometry and Dynamical Systems, Instituto Superio...

    1993

Pith tools

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