pith. sign in

arxiv: 2604.11800 · v1 · submitted 2026-04-13 · ❄️ cond-mat.stat-mech

Diffusing diffusivity model with dichotomous noise

Dongho Lee , Jae-Hyung Jeon , Pascal Viot , Gleb Oshanin This is my paper

Pith reviewed 2026-05-10 15:53 UTC · model grok-4.3

classification ❄️ cond-mat.stat-mech
keywords diffusing diffusivitydichotomous noiseLangevin dynamicsstochastic diffusivityprobability density functionshort-time asymptoticsself-averagingOrnstein-Uhlenbeck process
0
0 comments X

The pith

Modeling diffusivity as an Ornstein-Uhlenbeck process driven by symmetric dichotomous noise yields exact short-time displacement PDFs with logarithmic origin divergence and power-law modulated Gaussian tails.

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

The authors introduce a minimal model in which the particle's diffusivity fluctuates as an Ornstein-Uhlenbeck process forced by symmetric dichotomous noise, keeping the diffusivity bounded between two values. They obtain closed-form expressions for the short-time probability density of the particle's displacement and show that it diverges logarithmically at zero displacement. The far tails of this distribution decay as a Gaussian multiplied by a power law whose exponent depends on the noise switching rate, unlike the pure exponential tails found when the noise is Gaussian white. At long times the particle position follows ordinary Gaussian diffusion, and the time-averaged diffusivity is self-averaging with finite variance. This setup supplies an analytically solvable case of stochastic transport in media whose fluctuations are bounded rather than unbounded.

Core claim

We derive analytical expressions for the short-time probability density function (PDF) of the particle displacement and analyse its asymptotic behaviour. While the PDF retains the characteristic logarithmic divergence at the origin, its tails differ from the Gaussian white-noise case: exponential tails are replaced by Gaussian ones modulated by a power-law with a switching-rate-dependent exponent. At long times, the dynamics converges to ordinary Gaussian diffusion. We determine the variance and covariance of the time-averaged stochastic diffusivity and show that it is self-averaging.

What carries the argument

The Ornstein-Uhlenbeck process for diffusivity driven by symmetric dichotomous noise, which confines diffusivity to a finite interval and enables analytical tractability.

If this is right

  • Exact closed-form short-time PDFs are available for this bounded-diffusivity class of models.
  • The tail exponent varies explicitly with the dichotomous switching rate, providing a direct experimental signature.
  • Long-time dynamics recovers standard Gaussian diffusion with self-averaging statistics.
  • The time-averaged diffusivity possesses finite variance and covariance.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The model offers a baseline for comparing against unbounded diffusivity fluctuations in heterogeneous media.
  • Single-particle tracking experiments with abrupt environmental switches could detect the predicted power-law prefactors in displacement tails.
  • Replacing the symmetric dichotomous driver with asymmetric or multi-state noise might generate further analytically solvable diffusion cases.

Load-bearing premise

The diffusivity is modeled as an Ornstein-Uhlenbeck process driven by symmetric dichotomous noise that confines it to a finite interval.

What would settle it

Measuring short-time particle displacement distributions in a system whose diffusivity switches between two bounded values and finding pure exponential tails rather than Gaussian tails multiplied by a power law would contradict the model's analytical predictions.

Figures

Figures reproduced from arXiv: 2604.11800 by Dongho Lee, Gleb Oshanin, Jae-Hyung Jeon, Pascal Viot.

Figure 1
Figure 1. Figure 1: Stationary probability density function Pst(D) for τ = 1, A = 1 and three values of the switching rate λ: λ = 3 (left), λ = 1 (center), and λ = 0.1 (right). The thin magenta lines represent the analytical predictions from Eq. (17), while the blue shaded regions show histograms obtained from numerical simulations (see Appendix A for more details). numerical simulations. This Figure highlights the three dist… view at source ↗
Figure 2
Figure 2. Figure 2: Short-time evolution of P(X, t) in the DD model driven by dichotomous noise for τ = 1, A = 1 and three values of the switching rate λ: λ = 3 (left), λ = 1 (center) and λ = 0.1 (right). Thin red curves correspond to our theoretical prediction in Eq. (22). The blue histograms depict our numerical simulations results. more strongly concentrated around the origin, as intuitively expected for fluctuations of bo… view at source ↗
Figure 3
Figure 3. Figure 3: The position PDF P(X, t) in the DD model driven by dichotomous noise for τ = 1, A = 1, λ = 0.1 and three moderately short values of time t : t = 3 (left), t = 5 (center) and t = 20 (right). The thin black curves correspond to the theoretical prediction in Eq. (33), including the first two terms of the Edgeworth expansion, while the red curve represents the leading Gaussian contribution only. The histograms… view at source ↗
Figure 4
Figure 4. Figure 4: The position PDF P(X, t) in the DD model driven by dichotomous noise for τ = 1, A = 1, t = 100 and three values of the switching rate λ: λ = 3 (left), λ = 1 (center) and λ = 0.1 (right). Thin red curves correspond to our theoretical prediction in Eq. (33). The blue histograms depict our numerical simulations results. We next examine the behaviour of the position PDF at sufficiently large times, focusing on… view at source ↗
read the original abstract

We study Langevin dynamics with stochastic diffusivity arising from fluctuations of the surrounding medium. The diffusivity is modeled as Ornstein-Uhlenbeck process driven by symmetric dichotomous noise, which confines it to a finite interval. We derive analytical expressions for the short-time probability density function (PDF) of the particle displacement and analyse its asymptotic behaviour. While the PDF retains the characteristic logarithmic divergence at the origin, its tails differ from the Gaussian white-noise case: exponential tails are replaced by Gaussian ones modulated by a power-law with a switching-rate-dependent exponent. At long times, the dynamics converges to ordinary Gaussian diffusion. We determine the variance and covariance of the time-averaged stochastic diffusivity and show that it is self-averaging. The model provides a minimal analytically tractable framework for stochastic transport in environments with bounded or switching fluctuations.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

0 major / 3 minor

Summary. The manuscript studies Langevin dynamics with stochastic diffusivity modeled as an Ornstein-Uhlenbeck process driven by symmetric dichotomous noise that confines the diffusivity to a finite interval. It derives analytical expressions for the short-time PDF of particle displacement, analyzes the asymptotics (retaining logarithmic divergence at the origin but replacing exponential tails with Gaussian ones modulated by a switching-rate-dependent power law), shows convergence to ordinary Gaussian diffusion at long times, and proves self-averaging of the time-averaged diffusivity via variance and covariance calculations.

Significance. If the derivations hold, the work supplies a minimal, exactly solvable model for stochastic transport under bounded or switching fluctuations. The explicit short-time PDF asymptotics and the analytical demonstration of self-averaging constitute concrete, falsifiable advances over generic diffusing-diffusivity phenomenology.

minor comments (3)
  1. [Abstract and §3] The abstract and introduction state that analytical expressions and asymptotic analysis were derived, yet the main text should explicitly display the key steps of the coupled Fokker-Planck solution for the joint (position, diffusivity-state) process at short times, even if relegated to an appendix, to allow direct verification of the tail exponent.
  2. [Throughout] Notation for the switching rate and the diffusivity bounds should be introduced once with a single symbol table or consistent definition list; repeated re-definition of the same parameters across sections reduces readability.
  3. [§4] The long-time self-averaging proof relies on ergodicity of the two-state Markov chain; a brief remark on the mixing time relative to the observation window would clarify the regime of validity.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for the careful summary of our manuscript and the positive assessment of its significance. The recommendation for minor revision is noted. However, the report lists no specific major comments, so we have no individual points requiring point-by-point rebuttal or revision.

Circularity Check

0 steps flagged

Derivation follows directly from model equations with no circular reductions

full rationale

The paper solves the joint Markov process (particle position coupled to the two-state dichotomous driver of the OU diffusivity) via coupled Fokker-Planck equations to obtain exact short-time PDFs and their asymptotics, then invokes ergodicity of the finite-state Markov chain to establish long-time Gaussian diffusion and self-averaging of the time-averaged diffusivity. These steps are direct consequences of the stochastic process definition and do not reduce to pre-fitted parameters, self-citations, or ansatzes; the modeling choice is stated as enabling tractability rather than being justified by prior results from the same authors.

Axiom & Free-Parameter Ledger

2 free parameters · 2 axioms · 0 invented entities

The central claim rests on the domain assumption that particle motion obeys a Langevin equation with time-dependent diffusivity governed by an OU process switched by symmetric dichotomous noise. No explicit free parameters are fitted to data in the abstract; the switching rate and diffusivity bounds are model parameters. No new physical entities are postulated beyond the stochastic process definition.

free parameters (2)
  • switching rate
    Controls the frequency of dichotomous flips and appears in the power-law exponent of the PDF tails; treated as a free model parameter.
  • diffusivity bounds
    Upper and lower limits to which the OU process is confined by the dichotomous noise; chosen to keep diffusivity finite.
axioms (2)
  • domain assumption Particle obeys overdamped Langevin dynamics with stochastic diffusivity D(t)
    Standard starting point in statistical mechanics for Brownian motion in fluctuating media; invoked to set up the displacement PDF.
  • domain assumption D(t) is an Ornstein-Uhlenbeck process driven by symmetric dichotomous noise
    Defines the bounded stochastic diffusivity; enables the claimed analytical short-time solution.

pith-pipeline@v0.9.0 · 5439 in / 1682 out tokens · 35609 ms · 2026-05-10T15:53:32.368054+00:00 · methodology

Review history (2 revisions) →

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

52 extracted references · 52 canonical work pages

  1. [1]

    Wang B, Anthony S M, Bae S C and Granick S 2009PNAS10615160–15164

    work page

  2. [2]

    Jeon J H, Javanainen M, Martinez-Seara H, Metzler R and Vattulainen I 2016Phys. Rev. X6 021006

    work page

  3. [3]

    Guan J, Wang B and Granick S 2014ACS Nano83331–3336

    work page

  4. [4]

    Matse M, Chubynsky M V and Bechhoefer J 2017Phys. Rev. E96042604

    work page

  5. [5]

    Ciarlo A, Pastore R, Greco F, Sasso A and Pesce G 2023Sci Rep137408

    work page

  6. [6]

    Slezak J and Burov S 2021Sci Rep115101

    work page

  7. [7]

    Brizioli M, Sentjabrskaja T, Egelhaaf S U, Laurati M, Cerbino R and Giavazzi F 2022Front. Phys. 10

    work page

  8. [8]

    Kwon G, Sung B J and Yethiraj A 2014J. Phys. Chem. B1188128–8134

    work page

  9. [9]

    Akimoto T, Jeon J H, Metzler R, Miyaguchi T, Uneyama T and Yamamoto E 2026Rep. Prog. Phys.89014602

    work page

  10. [10]

    Chubynsky M V and Slater G W 2014Phys. Rev. Lett.113098302

    work page

  11. [11]

    Chechkin A V, Seno F, Metzler R and Sokolov I M 2017Phys. Rev. X7021002

    work page

  12. [12]

    Lanoisel´ ee Y and Grebenkov D S 2018J. Phys. A: Math. Theor.51145602

    work page

  13. [13]

    Phys.20043044

    Sposini V, Chechkin A V, Seno F, Pagnini G and Metzler R 2018New J. Phys.20043044

    work page

  14. [14]

    Lanoisel´ ee Y, Moutal N and Grebenkov D S 2018Nat Commun94398

    work page

  15. [15]

    Phys.22063056

    Sposini V, Grebenkov D S, Metzler R, Oshanin G and Seno F 2020New J. Phys.22063056

    work page

  16. [16]

    Jain R and Sebastian K L 2018Phys. Rev. E98052138

    work page

  17. [17]

    Grebenkov D S, Sposini V, Metzler R, Oshanin G and Seno F 2021J. Phys. A: Math. Theor.54 04LT01

    work page

  18. [18]

    Phys.23023014

    Grebenkov D S, Sposini V, Metzler R, Oshanin G and Seno F 2021New J. Phys.23023014

    work page

  19. [19]

    Hidalgo-Soria M, Barkai E and Burov S 2021Entropy23231

    work page

  20. [20]

    Dean D S, Majumdar S N and Sabhapandit S 2025J. Phys. A: Math. Theor.58235002

    work page

  21. [21]

    Weron A, Burnecki K, Akin E J, Sol´ e L, Balcerek M, Tamkun M M and Krapf D 2017Sci Rep7 5404

    work page

  22. [22]

    Bressloff P C 2017J. Phys. A: Math. Theor.50133001

    work page

  23. [23]

    Immunol.14

    Song T, Choi Y, Jeon J H and Cho Y K 2023Front. Immunol.14

    work page

  24. [24]

    Park S, Yang I, Lee J, Kim S, Mart´ ın-L´ opez J, Fishel R, Lee J B and Jeon J H 2025Preprint arXiv:2506.21943

    work page arXiv

  25. [25]

    Hachiya Y, Uneyama T, Kaneko T and Akimoto T 2019J. Chem. Phys.151034502

    work page

  26. [26]

    Pastore R, Coniglio A and Ciamarra M P 2015Sci Rep511770

    work page

  27. [27]

    Miyaguchi T, Akimoto T and Yamamoto E 2016Phys. Rev. E94012109

    work page

  28. [28]

    Grebenkov D S 2019J. Phys. A: Math. Theor.52174001

    work page

  29. [29]

    Horsthemke W and Lefever R 2006Noise-Induced Transitions: Theory and Applications in Physics, Chemistry, and Biology(Springer Science & Business Media)

    work page

  30. [30]

    H¨ aunggi P and Jung P 1994 Colored Noise in Dynamical SystemsAdvances in Chemical Physics (John Wiley & Sons, Ltd) pp 239–326 ISBN 978-0-470-14148-9

    work page 1994

  31. [31]

    Bena I 2006Int. J. Mod. Phys. B202825–2888

    work page

  32. [32]

    Gardiner C 2009Stochastic Methods: A Handbook for the Natural and Social Sciences(Springer Berlin Heidelberg)

    work page

  33. [33]

    Um J, Song T and Jeon J H 2019Front. Phys.7

    work page

  34. [34]

    Cox J C, Ingersoll J E and Ross S A 1985Econometrica53385–407

    work page

  35. [35]

    Stein E M and Stein J C 1991The Review of Financial Studies4727–752

    work page

  36. [36]

    Heston S L 1993Rev Financ Stud6327–343

    work page

  37. [37]

    Black F and Scholes M 1972The Journal of Finance27399–417

    work page

  38. [38]

    Black F and Scholes M 1973Journal of Political Economy81637–654

    work page

  39. [39]

    Bouchaud J P and Potters M 2003Theory of Financial Risk and Derivative Pricing: From Statistical Physics to Risk Management2nd ed (Cambridge: Cambridge University Press) Diffusing diffusivity model with dichotomous noise16

    work page

  40. [40]

    Tailleur J and Cates M E 2008Phys. Rev. Lett.100218103

    work page

  41. [41]

    Gu´ eneau M, Majumdar S N and Schehr G 2023J. Phys. A: Math. Theor.56475002

    work page

  42. [42]

    Sancho J M 1984J. Math. Phys.25354–359

    work page

  43. [43]

    Herbeau T, Pastur L, Viot P and Oshanin G 2026J. Stat. Mech.2026013205

    work page

  44. [44]

    Beck C and Cohen E G D 2003Physica A: Statistical Mechanics and its Applications322267–275

    work page

  45. [45]

    Bateman H 1955Higher Transcendental Functions

    work page

  46. [46]

    Feller W 1968An Introduction to Probability Theory and Its Applications, Volume 1(Wiley)

    work page

  47. [47]

    Kloeden P E and Platen E 1992Numerical Solution of Stochastic Differential Equations(Berlin, Heidelberg: Springer)

    work page

  48. [48]

    Higham D J 2001SIAM Rev.43525–546

    work page

  49. [49]

    Devroye L 1986Non-Uniform Random Variate Generation(New York, NY: Springer)

    work page

  50. [50]

    Brent R P 2013Algorithms for Minimization Without Derivatives(Courier Corporation)

    work page

  51. [51]

    Press W H 2007Numerical Recipes 3rd Edition: The Art of Scientific Computing(Cambridge University Press)

    work page

  52. [52]

    2 ∆t,t′ α= (1 + 2λτ) 2 −(1 + 2λτ) ∆ t,t′ α= 2 τ # , t≤t ′ , (B.16) where Var(Dt) is defined in Eq. (B.14), while ∆t,t′(α) = 1 α2

    Shapiro V E and Loginov V M 1978Physica A: Statistical Mechanics and its Applications91 563–574 Appendix A. Numerical simulations In this appendix, we describe the numerical procedures used to simulate the stochastic dynamics and to evaluate the statistical properties of the position probability density function. Numerical simulations are performed using ...

    work page