Pith. sign in

REVIEW 3 major objections 4 minor 4 cited by

Well-Posedness for Dean-Kawasaki Models of Vlasov-Fokker-Planck Type

T0 review · 3 major / 4 minor · reviewed 2026-08-12 · deepseek-v4-flash

Pith's one-line read The paper proves that stochastic Vlasov-Fokker-Planck (Dean-Kawasaki-type) equations for second-order Langevin systems are well-posed exactly for n-particle empirical initial data and become insoluble for smooth initial data.

desk verdict Solid extension of the Dean-Kawasaki existence/nonexistence dichotomy to second-order and hypoelliptic systems, with the main caveat being the gradient-structure assumption on interactions. read the letter →

arxiv 2411.14334 v3 pith:C4A67II5 submitted 2024-11-21 math.AP math-phmath.MPmath.PR

classification math.APmath-phmath.MPmath.PR MSC 60H1535R6082C2235Q8360J60
keywords Dean-KawasakiequationfluctuatinghydrodynamicsVlasov-Fokker-Planckmartingalesolutionsempiricalmeasuresill-posednesshypoellipticdiffusionsGirsanovtransform
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

Fluctuating-hydrodynamics equations are stochastic PDEs intended to describe finite collections of particles, but the standard physical derivation replaces a particle-dependent noise by a statistically similar one. This paper gives a rigorous martingale derivation for second-order Langevin systems and proves that the resulting stochastic Vlasov-Fokker-Planck (Dean-Kawasaki-type) equations are exact encodings of the particle dynamics, not approximations. Its main theorem extends the known first-order dichotomy: a solution exists if and only if the noise scale $\alpha$ is a positive integer $n$ and the initial measure is the empirical measure of $n$ particles, and then the solution is the empirical measure of the particle trajectories. Smoothing the initial data destroys existence, and for general total mass $m$ existence is equivalent to $\alpha m \in \mathbb{N}$. The result covers inertial Langevin dynamics and active-matter or flocking models whose interaction has the fluctuation-dissipation gradient form $F_\mu = \sigma\sigma^T\nabla \frac{\delta G}{\delta\mu}$.

What carries the argument

The machinery is the weak martingale-solution formulation on the space $\mathcal{M}_1(\mathbb{R}^k)$ of probability measures. For $F=0$, uniqueness and atomicity follow from a Laplace duality, $\mathbb{E}[e^{-\langle\mu_t,\varphi\rangle}] = e^{-\langle\mu_0,V_t\varphi\rangle}$, where $V_t\varphi = -\alpha\ln(P_{\alpha t} e^{-\varphi/\alpha})$ solves the Hamilton-Jacobi-Bellman equation $\partial_t\psi = \alpha L\psi - \Gamma(\psi)$, with $\Gamma(f)=\frac{1}{2}|\sigma^T\nabla f|^2$ the carré du champ. The Cole-Hopf form makes the moment generating function of $\alpha\mu_t(A)$ analytic near $0$, forcing $\alpha\mu_t(A)\in\mathbb{N}$ and hence atomicity, via the exhaustion condition (L.4). The interacting case is reduced to $F=0$ by the Girsanov transform of Proposition 5.9, which is why the proof needs the fluctuation-dissipation gradient condition $F_\mu = \sigma\sigma^T\nabla\frac{\delta G}{\delta\mu}$ plus the growth bounds of Assumption 3.4. Assumptions (L.1)-(L.4) supply the Feller semigroup regularity, smoothing estimates, and stability under initial data needed to run these steps for degenerate hypoelliptic generators.

What would settle it

Take $\mu_0$ to be a smooth density with finite second moment, say a Gaussian on $\mathbb{R}^{2d}$, fix $\alpha=1$, and look for a continuous $\mathcal{M}_1$-valued process satisfying the martingale problem of Definition 3.1; the theorem predicts none exists, so any explicit or numerical construction of such a process would refute Theorem 3.5. For the interacting case, a sharper test is to take a non-gradient velocity-alignment force $F$ and check whether smooth-data solutions appear, which would indicate that the gradient fluctuation-dissipation condition is genuinely necessary.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central result is Theorem 3.5. For the class of equations $\partial_t\mu_t = \alpha L^*\mu_t + \alpha\nabla\cdot(\mu_t F_{\mu_t}) + \nabla\cdot(\sqrt{\mu_t}\,\sigma\, \dot{W}_{z,t})$ with $L = b\cdot\nabla + \frac{1}{2}\sigma\sigma^T:\nabla^2$ and initial data $\mu_0$ a probability measure with finite second moment, a martingale solution exists if and only if $\alpha = n\in\mathbb{N}$ and $\mu_0 = \frac{1}{n}\sum_{i=1}^n \delta_{z_i}$; the unique-in-law solution is $\mu_t = \frac{1}{n}\sum_{i=1}^n\delta_{z_i(t)}$, where the $z_i$ solve $dz_i = (\alpha b(z_i)+\alpha F_{\mu_t}(z_i))\,dt + \sqrt{\alpha}\,\sigma(z_i)\,dW_i$. In particular, no solution exists for smooth or otherwise non-atomic initial data, so the SPDE is an exact transcription of the particle system rather than a free-standing PDE. The same dichotomy is proved for degenerate, hypoelliptic, and non-reversible generators, and for interactions of the fluctuation-dissipation gradient form.

Load-bearing premise

The load-bearing premise is that the interaction has the fluctuation-dissipation gradient form $F_\mu = \sigma\sigma^T\nabla\frac{\delta G}{\delta\mu}$ with the growth bounds of Assumption 3.4, together with the semigroup conditions (L.1)-(L.4); if any of these fails, the paper proves neither existence nor the smooth-data nonexistence for interacting systems.

Editorial extensions

If this is right

  • For every $n$-particle second-order Langevin system covered by the assumptions, the Dean-Kawasaki-type SPDE is an exact representation: its only solution is the empirical measure of the particle trajectories, so the SPDE carries no information beyond the particle system.
  • Mollifying the initial data, no matter how slightly, makes the SPDE unsolvable; numerical methods that assume a smooth density profile therefore cannot be used without regularising the noise.
  • The dichotomy extends to hypoelliptic and non-reversible generators, covering inertial Langevin dynamics, active swimmers, and flocking models with gradient-type interactions.
  • For initial data of total mass $m$, solutions exist exactly when $\alpha m\in\mathbb{N}$ and take the form $\frac{1}{\alpha}\sum_{i=1}^{\alpha m}\delta_{z_i(\alpha t)}$.
  • The martingale derivation supplies a rigorous replacement for the noise-substitution heuristic used to write down equations of fluctuating hydrodynamics.

Reading between the lines

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

  • The paper's gradient condition is sufficient; whether it is necessary is open. If a non-gradient velocity-alignment force admitted smooth-data solutions, the dichotomy would fail outside the fluctuation-dissipation class, so testing such models is a natural next step.
  • Because the empirical measure path determines the underlying particle trajectories up to relabelling, exactness suggests that inferring the interaction $F$ from fluctuating-hydrodynamics data is equivalent to inferring it from trajectories, which may simplify identifiability questions for active-matter models.
  • The same square-root multiplicative noise with fluctuation-dissipation balance appears in other conservative SPDEs; by analogy one would expect atomic-only well-posedness there as well, although the paper does not address those equations.
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

3 major / 4 minor

Summary. The paper studies the Cauchy problem for the stochastic Vlasov-Fokker-Planck / Dean-Kawasaki type SPDE ∂tµ = αL*µ + α∇·(µF_µ) + ∇·(√µ σ ˙W) over probability measures on R^k, with L a (possibly degenerate) diffusion generator and F an interaction satisfying a fluctuation-dissipation gradient condition. The main result (Theorem 3.5) asserts a dichotomy: a martingale solution exists if and only if α is a positive integer n and the initial datum is an empirical measure 1/n∑δ_{z^i}; the solution is then the empirical measure of the corresponding n-particle SDE. Consequently smooth initial data admit no solution. The argument uses a Laplace duality for the interaction-free case to derive uniqueness and rigidity of solutions, and a Girsanov reduction to transfer the result to the interacting case. Examples cover inertial Langevin dynamics, active particle models, and flocking/alignment models.

Significance. If correct, the theorem is a substantial extension of the known ill-posedness/triviality result for the Dean-Kawasaki equation [32,31] to second-order, hypoelliptic, and non-reversible settings. It gives a mathematically precise interpretation of fluctuating hydrodynamics as an exact finite-particle representation, and it provides a general martingale framework (Assumption 3.2) that is verified on several physically relevant examples. The paper is honest about the restrictive gradient condition (F.1) on interactions, and the nonexistence theorem for smooth data in the interacting case is conditional on it. The proof strategy—Laplace duality, moment-generating-function rigidity, and Girsanov reduction—is well suited to the problem and, once the factor errors noted below are fixed, is convincing.

major comments (3)
  1. [Section 5.1, Proposition 5.2] The time-dependent martingale used in the duality proof is displayed with generator L rather than αL: M_t(ψ_·) := ⟨µ_t,ψ_t⟩−⟨µ_0,ψ_0⟩−∫⟨µ_s, Lψ_s+∂_sψ_s⟩ ds. Since (14) has generator αL*, this process is generally not a martingale for solutions of (14). The subsequent Itô computation correctly uses αL V_{t−s}φ, so the fixed formula should read αLψ_s; please correct the display and check the intermediate steps.
  2. [Section 5.2, Proposition 5.6] Proposition 5.6 states the Itô formula with drift term ⟨µ_s, αL δE/δµ + F_µs·∇δE/δµ⟩, omitting the factor α in front of F. Definition 3.1 has αLφ+αF_µs·∇φ, so the displayed formula fails already for E(µ)=⟨µ,φ⟩. The proof sketch for cylinder functions produces the αF term, so the statement has a typographical omission, but as written the formula is wrong. Since the Girsanov density in Proposition 5.9 is built from this Itô formula, the factor must be corrected for the interacting reduction to be valid.
  3. [Section 5.2, Proposition 5.9] The statement defines Q by the stochastic exponential E(M^G), while the proof uses E(M^{αG}). These are not interchangeable: the cross-variation of M^G with M(φ) is ⟨µ,∇φ·F⟩, so E(M^G) changes the drift by only F·∇φ, not αF·∇φ; the correct density for removing the interaction term α∇·(µF) is E(M^{αG}). The proof's use of M^{αG} is right, but the proposition statement and the subsequent proof of Theorem 3.5 for F≠0 (which writes E(M^G(µ_·))) need to be made consistent.
minor comments (4)
  1. [Lemma 5.5] The statement writes |H(µ,z)| instead of |F(µ,z)|, and the Fatou display in the proof should take liminf over N (with the stopping time kept) rather than over R. The argument is otherwise clear.
  2. [Section 4.3] In the verification of condition (F.1) for the interacting Langevin model, the displayed expression for δG/δµ has the second term ∫ v′·f(x,x′)dµ(x′,v′); the correct functional derivative contains ∫ v′·f(x′,x)dµ(x′,v′). This does not affect the computed ∇_v derivative, but the formula should be corrected.
  3. [Theorem 3.5] The phrase 'if and only if α=n for some n∈N' should be read together with the requirement that µ0 be an n-point empirical measure of the stated form; as written, the sentence could be misread as asserting existence for every µ0 when α is an integer. Please rephrase for clarity.
  4. [Section 4.1] In the verification of (L.4), the Girsanov display should be an inequality (probability ≥ E^0[M_t 1_{...}]) rather than an equality; the intended estimate is clear but the formula as printed is not quite accurate.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; the nonexistence and rigidity directions are proved via Laplace duality and Girsanov reduction, independent of the explicit atomic construction.

full rationale

The paper has two distinct directions. The existence direction for atomic initial data is an explicit verification: the empirical measure of the particle system is inserted into the martingale formulation of Definition 3.1, and Section 5.1 (Proposition 5.1) shows directly via Itô calculus that it is a solution, with the quadratic variation matching by construction. This is transparent: Section 2 states that the noise term in the SPDE was chosen to induce the quadratic variation (5), and the rescaling to α = n is explicitly made to normalize the noise prefactor. That is an explicit construction rather than a hidden circularity. The substantive direction, nonexistence for smooth or non-atomic initial data, is proved independently: Proposition 5.2 establishes a Laplace duality, Proposition 5.4 uses the moment-generating-function expansion to force α μ_t(A) ∈ N for all measurable A, and hence atomic initial data; this argument does not assume the conclusion. The interacting case is reduced to the interaction-free case by a genuine Girsanov transform in Proposition 5.9, under the openly stated fluctuation-dissipation gradient assumption (F.1). Citations to the overlapping-author papers [31,32] supply a technical lemma on moment generating functions, approximation results, and the overall proof strategy, but the central rigidity proof is carried out in this paper, so those citations are not load-bearing. The stated limitations, such as the gradient-form condition (F.1) and the technical Assumption 3.2, are scope restrictions rather than circular inputs.

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

The central theorem is conditional on Assumption 3.2 for the generator L and, in the interacting case, on Assumption 3.4 for the interaction F. These are stated domain assumptions, not fitted parameters. The paper also borrows two technical theorems from the prior literature, one on approximation of measure functionals and one on semigroup regularity. No new physical entities are introduced.

assumptions (4)
  • domain assumption Assumption 3.2 (L.1)-(L.4): L admits a unique L-diffusion, a dense domain D closed under composition with C∞ functions vanishing at 0, gradient bounds for σᵀ∇P_tφ, and an exhaustion with P_t1_{A_n} ≤ c_n < 1.
    Postulated for the generator L; used in Proposition 5.2 for Laplace duality, in Proposition 5.4 for rigidity, and verified for the Langevin and active-matter examples in Sections 4.1 and 4.2.
  • domain assumption Assumption 3.4 (L.5, F.1, F.2): σ is bounded, b has linear growth, F has the gradient form Fµ = σσᵀ∇δG/δµ, and σᵀ∇δG/δµ is bounded with growth bounds on higher derivatives.
    The gradient structure is the mechanism for the Girsanov reduction in Proposition 5.9; without it the interacting-case dichotomy is not proved for non-gradient interactions.
  • standard math Approximation theorem from Konarovskyi-Lehmann-von Renesse [31, Theorem 5]: C²_b(M) functionals can be approximated by cylindrical polynomial functionals on measures with uniform derivative bounds.
    Imported in Proposition 5.6 to pass from cylindrical Ito formulas to general functionals on measure space; not reproved in this paper.
  • standard math Semigroup regularity results of Da Prato-Rockner [13, Propositions 2.5, 3.3, 3.6] used to verify Assumption 3.2 for the Langevin example.
    Imported to establish P_t D ⊂ D and the commutation LP_t = P_t L in Section 4.1.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Well-Posedness for Dean-Kawasaki Models of Vlasov-Fokker-Planck Type." pith.science (2026). https://pith.science/paper/C4A67II5

@misc{pith2026241114334,
  author       = {Pith},
  title        = {Pith review of: Well-Posedness for Dean-Kawasaki Models of Vlasov-Fokker-Planck Type},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/C4A67II5}},
  note         = {Machine review of arXiv:2411.14334}
}
read the original abstract

We consider systems of interacting particles which are described by a second order Langevin equation. The class of equations considered includes the situation where the particle evolution is governed by Hamiltonian dynamics with additional damping and noise satisfying a fluctuation-dissipation relation. Also covered are systems of two equations describing an evolution of interacting agents, as arising in several descriptions of active matter, including models for flocking and swarming. We first show that such particle systems can be represented exactly by so-called equations of fluctuating hydrodynamics, which in this case are stochastic versions of a Vlasov-Fokker-Planck type equation. While the derivation given here is simple, it is a blueprint for the rigorous derivation of equations of fluctuating hydrodynamics. We then show a dichotomy previously known for purely diffusive (first order) systems carries over to the second order setting considered here: Solutions exist for suitable atomic initial data, in which case the solution is, properly scaled, the empirical density describing the particle system. For smooth initial data, however, we prove that no solution exists.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 4 Pith papers

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

  1. Ill-posedness of the pure-noise Dean-Kawasaki equation

    math.PR 2025-01 conditional novelty 7.0 of 10

    The pure-noise Dean-Kawasaki equation with any bounded drift has no measure-valued martingale solutions.

  2. Massive Particle Systems, Wasserstein Brownian Motions, and the Dean-Kawasaki Equation

    math.PR 2024-11 conditional novelty 7.0 of 10

    Free massive particle systems, singular-drift Dean-Kawasaki equations, Wasserstein diffusions, and metric-measure Brownian motions are identified as a single process for any ultracontractive reversible diffusion.

  3. Cluster formation for weakly interacting kinetic Langevin dynamics

    math.NA 2025-10 conditional novelty 6.0 of 10

    For weakly interacting underdamped Langevin particles with short-range attraction, cluster onset occurs above a friction-independent critical inverse temperature and the onset time scales as (1/(2ψmax)) log N.

  4. The Dean-Kawasaki equation and stochastic density functional theory

    cond-mat.soft 2024-11 accept novelty 2.0 of 10

    A review of the Dean-Kawasaki equation and stochastic density functional theory, covering derivation, mathematical issues, extensions, solution methods, and applications.

Reference graph

Works this paper leans on

45 extracted references · 42 canonical work pages · cited by 4 Pith papers

  1. [1]

    von Renesse

    Sebastian Andres and Max-K. von Renesse. Particle approximation of the Wasser- stein diffusion. J. Funct. Anal. 258.11 (2010), 3879–3905

  2. [2]

    Approxi- mation of non-linear SPDEs with additive noise via weighted interacting particles systems: the stochastic McKean-Vlasov equation

    Letizia Angeli, Dan Crisan, Martin Kolodziejczyk, and Michela Ottobre. Approxi- mation of non-linear SPDEs with additive noise via weighted interacting particles systems: the stochastic McKean-Vlasov equation. 2024. 20

  3. [3]

    Self- organization of active particles by quorum sensing rules

    Tobias B¨ auerle, Andreas Fischer, Thomas Speck, and Clemens Bechinger. Self- organization of active particles by quorum sensing rules. Nat. Commun. 9.1 (2018), 3232

  4. [4]

    Bressloff

    Paul C. Bressloff. A generalized Dean-Kawasaki equation for an interacting Brow- nian gas in a partially absorbing medium. Proc. A. 480.2296 (2024), Paper No. 20230915, 23

  5. [5]

    M. E. Cates, D. Marenduzzo, I. Pagonabarraga, and J. Tailleur. Arrested phase separation in reproducing bacteria creates a generic route to pattern formation. PNAS 107.26 (2010), 11715–11720

  6. [6]

    Cates and Julien Tailleur

    Michael E. Cates and Julien Tailleur. Motility-Induced Phase Separation. Annu. Rev. Condens. Matter Phys. 6.1 (2015), 219–244

  7. [7]

    Multilevel Monte Carlo methods for the Dean-Kawasaki equation from fluctuating hydrodynamics.SIAM J

    Federico Cornalba and Julian Fischer. Multilevel Monte Carlo methods for the Dean-Kawasaki equation from fluctuating hydrodynamics.SIAM J. Numer. Anal. 63.1 (2025), 262–287

  8. [8]

    The Dean-Kawasaki equation and the structure of density fluctuations in systems of diffusing particles

    Federico Cornalba and Julian Fischer. The Dean-Kawasaki equation and the structure of density fluctuations in systems of diffusing particles. Arch. Ration. Mech. Anal. 247.5 (2023), Paper No. 76, 59

Show all 45 references
  1. [9]

    Den- sity fluctuations in weakly interacting particle systems via the Dean-Kawasaki equation

    Federico Cornalba, Julian Fischer, Jonas Ingmanns, and Claudia Raithel. Den- sity fluctuations in weakly interacting particle systems via the Dean-Kawasaki equation. arXiv:2303.00429. 2023

  2. [10]

    A regularised Dean- Kawasaki model: derivation and analysis.SIAM J

    Federico Cornalba, Tony Shardlow, and Johannes Zimmer. A regularised Dean- Kawasaki model: derivation and analysis.SIAM J. Math. Anal. 51.2 (2019), 1137– 1187

  3. [11]

    From weakly inter- acting particles to a regularised Dean-Kawasaki model

    Federico Cornalba, Tony Shardlow, and Johannes Zimmer. From weakly inter- acting particles to a regularised Dean-Kawasaki model. Nonlinearity 33.2 (2020), 864–891

  4. [12]

    Well-posedness for a regularised inertial Dean-Kawasaki model for slender particles in several space dimensions

    Federico Cornalba, Tony Shardlow, and Johannes Zimmer. Well-posedness for a regularised inertial Dean-Kawasaki model for slender particles in several space dimensions. J. Differential Equations 284 (2021), 253–283

  5. [13]

    Alexandru Myller

    Giuseppe Da Prato and Michael R¨ ockner. Cores for generators of some Markov semigroups. In: “Alexandru Myller” Mathematical Seminar . Vol. 1329. AIP Conf. Proc. Amer. Inst. Phys., Melville, NY, 2011, pp. 87–97

  6. [14]

    David S. Dean. Langevin equation for the density of a system of interacting Langevin processes. J. Phys. A 29.24 (1996), L613

  7. [15]

    Massive Particle Systems, Wasserstein Brownian Motions, and the Dean-Kawasaki Equation

    Lorenzo Dello Schiavo. Massive Particle Systems, Wasserstein Brownian Motions, and the Dean-Kawasaki Equation. arXiv 2411.14936. 2024

  8. [16]

    The Dirichlet-Ferguson diffusion on the space of prob- ability measures over a closed Riemannian manifold

    Lorenzo Dello Schiavo. The Dirichlet-Ferguson diffusion on the space of prob- ability measures over a closed Riemannian manifold. Ann. Probab. 50.2 (2022), 591–648

  9. [17]

    Weak error analysis for a nonlinear SPDE approximation of the Dean-Kawasaki equation

    Ana Djurdjevac, Helena Kremp, and Nicolas Perkowski. Weak error analysis for a nonlinear SPDE approximation of the Dean-Kawasaki equation. Stoch. Partial Differ. Equ. Anal. Comput. 12.4 (2024), 2330–2355. 21

  10. [18]

    Dynamic density functional theory with hydrodynamic interactions and fluctuations

    Aleksandar Donev and Eric Vanden-Eijnden. Dynamic density functional theory with hydrodynamic interactions and fluctuations. J. Chem. Phys. 140.23 (2014), 234115

  11. [19]

    Computing diffusivities from particle models out of equilibrium

    Peter Embacher, Nicolas Dirr, Johannes Zimmer, and Celia Reina. Computing diffusivities from particle models out of equilibrium. Proc. A. 474.2212 (2018), 20170694, 21

  12. [20]

    Ethier and Thomas G

    Stewart N. Ethier and Thomas G. Kurtz. Markov processes. Wiley Series in Prob- ability and Mathematical Statistics: Probability and Mathematical Statistics. Characterization and convergence. John Wiley & Sons, Inc., New York, 1986, x+534

  13. [21]

    Well-posedness of the Dean-Kawasaki and the nonlinear Dawson-Watanabe equation with correlated noise

    Benjamin Fehrman and Benjamin Gess. Well-posedness of the Dean-Kawasaki and the nonlinear Dawson-Watanabe equation with correlated noise. Arch. Ra- tion. Mech. Anal. 248.2 (2024), Paper No. 20, 60

  14. [22]

    Jack, and Michael E

    ´Etienne Fodor, Robert L. Jack, and Michael E. Cates. Irreversibility and Biased Ensembles in Active Matter: Insights from Stochastic Thermodynamics. Annu. Rev. Condens. Matter Phys. 13.Volume 13, 2022 (2022), 215–238

  15. [23]

    Gvalani, and Vitalii Konarovskyi

    Benjamin Gess, Rishabh S. Gvalani, and Vitalii Konarovskyi. Conservative SPDEs as fluctuating mean field limits of stochastic gradient descent. Probab. Theory Re- lated Fields (2025), 1–69

  16. [24]

    Thin-film flow influenced by thermal noise

    G¨ unther Gr¨ un, Klaus Mecke, and Markus Rauscher. Thin-film flow influenced by thermal noise. J. Stat. Phys. 122.6 (2006), 1261–1291

  17. [25]

    From interacting agents to density-based modeling with stochastic PDEs

    Luzie Helfmann et al. From interacting agents to density-based modeling with stochastic PDEs. Commun. Appl. Math. Comput. Sci. 16.1 (2021), 1–32

  18. [26]

    Statistical-Physics-Informed Neural Networks (Stat-PINNs): A Machine Learning Strategy for Coarse-graining Dissipative Dynamics.J

    Shenglin Huang et al. Statistical-Physics-Informed Neural Networks (Stat-PINNs): A Machine Learning Strategy for Coarse-graining Dissipative Dynamics.J. Mech. Phys. Solids 194 (2025), 105908

  19. [27]

    The Dean-Kawasaki equation and stochastic density functional the- ory

    Pierre Illien. The Dean-Kawasaki equation and stochastic density functional the- ory. Reports on Progress in Physics (2025)

  20. [28]

    Jack and Johannes Zimmer

    Robert L. Jack and Johannes Zimmer. Geometrical interpretation of fluctuating hydrodynamics in diffusive systems. J. Phys. A 47.48 (2014), 485001, 17

  21. [29]

    New Method in Non-Equilibrium Statistical Mechanics of Co- operative Systems

    Kyozi Kawasaki. New Method in Non-Equilibrium Statistical Mechanics of Co- operative Systems. In: Synergetics: Cooperative Phenomena in Multi-Component Systems. Ed. by H. Haken. Wiesbaden: Vieweg+Teubner Verlag, 1973, pp. 35–44

  22. [30]

    Scaling limits of interacting particle sys- tems

    Claude Kipnis and Claudio Landim. Scaling limits of interacting particle sys- tems. Vol. 320. Grundlehren der Mathematischen Wissenschaften [Fundamental Principles of Mathematical Sciences]. Berlin: Springer-Verlag, 1999, xvi+442

  23. [31]

    On Dean-Kawasaki dynamics with smooth drift potential

    Vitalii Konarovskyi, Tobias Lehmann, and Max von Renesse. On Dean-Kawasaki dynamics with smooth drift potential. J. Stat. Phys. 178.3 (2020), 666–681

  24. [32]

    von Renesse

    Vitalii Konarovskyi, Tobias Lehmann, and Max-K. von Renesse. Dean-Kawasaki dynamics: ill-posedness vs. triviality. Electron. Commun. Probab. 24 (2019), Pa- per No. 8, 9. 22

  25. [33]

    Dean–Kawasaki equation with initial con- dition in the space of positive distributions

    Vitalii Konarovskyi and Fenna M¨ uller. Dean–Kawasaki equation with initial con- dition in the space of positive distributions. J. Evol. Equ. 24.4 (2024), Paper No. 92

  26. [34]

    Harnessing fluctuations to discover dissipative evolution equa- tions

    Xiaoguai Li et al. Harnessing fluctuations to discover dissipative evolution equa- tions. J. Mech. Phys. Solids 131 (2019), 240–251

  27. [35]

    James F. Lutsko. A dynamical theory of nucleation for colloids and macro- molecules. J. Chem. Phys. 136 (2012), 034509

  28. [36]

    Large deviations principles for stochastic scalar conservation laws

    Mauro Mariani. Large deviations principles for stochastic scalar conservation laws. Probab. Theory Related Fields 147.3-4 (2010), 607–648

  29. [37]

    An Additive-Noise Approximation to Keller- Segel-Dean-Kawasaki Dynamics: Small-Noise Results

    Adrian Martini and Avi Mayorcas. An Additive-Noise Approximation to Keller- Segel-Dean-Kawasaki Dynamics: Small-Noise Results. arXiv 2410.17022. 2024

  30. [38]

    Beyond equilibrium thermodynamics

    Hans Christian ¨Ottinger. Beyond equilibrium thermodynamics . Wiley Online Li- brary, 2005

  31. [39]

    Peruani, A

    F. Peruani, A. Deutsch, and M. B¨ ar. A mean-field theory for self-propelled parti- cles interacting by velocity alignment mechanisms.Eur. Phys. J. Spec. Top. 157.1 (2008), 111–122

  32. [40]

    von Renesse and Karl-Theodor Sturm

    Max-K. von Renesse and Karl-Theodor Sturm. Entropic measure and Wasserstein diffusion. Ann. Probab. 37.3 (2009), 1114–1191

  33. [41]

    G. M. Rotskoff and E. Vanden-Eijnden. Trainability and accuracy of artificial neural networks: an interacting particle system approach. Comm. Pure Appl. Math. 75.9 (2022), 1889–1935

  34. [42]

    Stroock and S

    Daniel W. Stroock and S. R. Srinivasa Varadhan. Multidimensional diffusion processes. Classics in Mathematics. Reprint of the 1997 edition. Springer-Verlag, Berlin, 2006, xii+338

  35. [43]

    Classical dynamical density functional theory: from fundamentals to applications

    Michael te Vrugt, Hartmut L¨ owen, and Raphael Wittkowski. Classical dynamical density functional theory: from fundamentals to applications. Adv. Phys. 69.2 (2020), 121–247

  36. [44]

    Approximating Particle-Based Clustering Dynamics by Stochastic PDEs

    Nathalie Wehlitz et al. Approximating Particle-Based Clustering Dynamics by Stochastic PDEs. SIAM Journal on Applied Dynamical Systems 24.2 (2025), 1231–1250

  37. [45]

    Large and moderate deviations and exponential convergence for stochastic damping Hamiltonian systems

    Liming Wu. Large and moderate deviations and exponential convergence for stochastic damping Hamiltonian systems. Stochastic Process. Appl. 91.2 (2001), 205–238. 23

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.