Pith. sign in

REVIEW 4 major objections 5 minor 83 references

Mean-Field Stochastic PDEs: Well-posedness and Quantitative Dimension-Free Propagation of Chaos

T0 review · 4 major / 5 minor · reviewed 2026-08-01 · deepseek-v4-flash

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

Pith's one-line read This paper proves quantitative, dimension-free propagation-of-chaos rates for mean-field stochastic PDEs, with the rate set by the solution space's modulus of convexity.

desk verdict Serious, original result with one load-bearing gap: the existence theorem for the interacting particle system is asserted, not proved, and it must be fixed before the PoC rates are trustworthy. read the letter →

arxiv 2607.17195 v1 pith:5FEUYZG3 submitted 2026-07-19 math.PR

Wei Hong , Wei Liu , Luhan Yang This is my paper · ORCID
classification math.PR MSC 60H1560H1060K35
keywords mean-fieldstochasticPDEpropagationofchaospseudo-monotoneoperatoruniformlyconvexBanachspacedimension-freeratesSteinvariationalgradientdescentAllen-CahnLagrangian-averagedBurgers
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 aims to show that a broad class of mean-field stochastic PDEs with super-linear, measure-dependent coefficients has well-posed strong and weak solutions, and that the associated N-particle systems converge to the mean-field limit at explicit rates that do not degrade with dimension. The central new claim is quantitative dimension-free propagation of chaos: in an α-uniformly convex Banach space, the pathwise strong error is of order N^{-1/α}, improving to near N^{-1/(α-1)} when the drift has extra dissipation. The authors identify the geometry of the solution space—specifically its modulus of convexity—as the intrinsic driver of the convergence rate, supported by an i.i.d. example in ℓ^q where the law-of-large-numbers rate exhibits the same convexity exponent. A sympathetic reader would care because these appear to be the first quantitative convergence rates for general infinite-dimensional interacting particle systems, with applications to stochastic Stein variational gradient descent, mean-field Allen-Cahn equations, and Lagrangian-averaged Burgers equations. A caveat internal to the paper: the existence and uniqueness theorem for the N-particle system is asserted with a one-sentence proof that appeals to an external product-space result and omits the verification.

What carries the argument

The central object is the measure-dependent pseudo-monotone operator A(t,u,μ): V×M→V*, a generalization of the classical pseudo-monotone operator to coefficients that depend on the law μ as well as the state u; with it, the paper obtains weak solutions via Galerkin approximation and stochastic compactness. Uniqueness and strong solvability rest on a decoupled local monotonicity condition that separates the state-derivative term from the measure-derivative term, making pathwise uniqueness feasible in the mean-field setting. For the convergence rates, the load-bearing tools are martingale-difference estimates in the dual of an α-uniformly convex Banach space and a stopping-time argument that t

What would settle it

Verify directly whether the empirical-measure-coupled coefficients of (2.9) satisfy the product-space theorem cited in Theorem 2.13, with bounds uniform in N; if any N-dependent constant appears, the rate estimates have no grounding. A second check: simulate a weakly interacting SPDE satisfying the dissipation condition—for instance, the mean-field Allen-Cahn equation on [0,1] with additive noise—for N=100, 400, 1600, and measure sup_i E sup_t ||X^{i,N}-X^i||_{L^2}^2; Theorem 2.16 predicts a near-N^{-1} decay for α=2, so a slope of -1/4 or -1/2 would contradict the claimed rate.

Watch

Extended reading notes

Core claim

The paper's central claim is that the variational theory for stochastic PDEs extends to mean-field equations whose drift and diffusion coefficients may grow faster than linearly in both state and measure, provided the coefficients satisfy a measure-dependent pseudo-monotonicity condition and, for uniqueness, a decoupled local monotonicity condition. On top of this, the paper proves quantitative dimension-free propagation of chaos: for an α-uniformly convex Banach space V, sup_i E sup_t ||X^{i,N}-X^i||_H^2 is O(N^{-1/α}) under the basic condition, and under a strengthened dissipation condition the pathwise and pointwise estimates improve to O(N^{-1/(α-1)}) times logarithmic-type factors invol

Load-bearing premise

The quantitative chaos rates rest on Theorem 2.13's assertion that the interacting particle system (2.9) has a unique strong solution on the product space; the provided proof is one sentence appealing to an external product-space theorem and omits the verification, so if the empirical-measure coupling fails the required monotonicity or growth conditions, the rates are not anchored to a well-defined particle system.

Editorial extensions

If this is right

  • If the main theorems are correct, infinite-dimensional mean-field systems with super-linear kernels—previously known only to converge qualitatively—now come with explicit rates, making the mean-field approximation quantitatively reliable.
  • In Hilbert or Euclidean spaces (α=2), the general pathwise rate is N^{-1/4}; adding dissipation lifts it to near N^{-1/2+ε} in the strong sense, close to the classical sharp N^{-1/2} rate.
  • The framework yields quantitative propagation-of-chaos estimates for stochastic Stein variational gradient descent with polynomial super-linear kernels, providing a theoretical convergence rate for this Bayesian-inference algorithm.
  • The mean-field Allen-Cahn and Lagrangian-averaged Burgers equations are covered, so the results apply directly to stochastic quantization and fluid-mechanics models.
  • The convexity-exponent dependence indicates that the geometry of the state space is not merely an analytic device but a design consideration: spaces with better convexity give faster mean-field convergence.

Reading between the lines

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

  • The geometric dependence of the rate likely extends beyond this paper: any interacting-particle analysis that uses martingale differences in a uniformly convex space should encounter the same convexity exponent, so the rate formula may serve as a general heuristic.
  • The largest unresolved step is the one-sentence proof of Theorem 2.13; until the product-space conditions on the empirical-measure coupling are verified in detail, the quantitative rates should be read as conditional on that omitted check.
  • For SVGD-type algorithms, the near-N^{-1/2+} rate under dissipation suggests that adding enough noise or dissipation is not just practically stabilizing but theoretically rate-optimal; a numerical test on a Gaussian target could check the predicted slope.
  • The ℓ^q i.i.d. example suggests the N^{-1/(α-1)} exponent may be optimal without additional smoothing, so future improvements would require exploiting the specific PDE structure rather than the martingale argument alone.
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

4 major / 5 minor

Summary. The paper develops a variational framework for mean-field stochastic PDEs with measure-dependent pseudo-monotone operators, proves existence of weak solutions (Theorem 2.9), existence of strong solutions under local monotonicity (Theorem 2.11), and uniqueness/continuous dependence under a decoupled local monotonicity condition (Theorem 2.12). The main advertised contribution is quantitative, dimension-free propagation of chaos: Theorems 2.14 and 2.16 give pathwise and pointwise rates of order N^{-1/α} and, under an added dissipativity condition, rates near N^{-1/(α-1)} in α-uniformly convex Banach spaces. Applications are given to stochastic SVGD, mean-field Allen-Cahn equations, and Lagrangian-averaged Burgers equations. The well-posedness section is lengthy and detailed, using Galerkin approximations, stochastic compactness, and a measure cut-off argument. The PoC proof uses Sznitman's synchronous coupling, martingale difference inequalities in uniformly smooth dual spaces, and stopping-time arguments.

Significance. If the main results are correct, this is the first quantitative, dimension-free propagation-of-chaos rate for general infinite-dimensional weakly interacting systems, and the dependence of the rate on the modulus of convexity is a genuinely new insight. The paper also provides a self-contained i.i.d. example (Proposition 2.15) showing that the geometry-dependent LLN rate is real, which is a valuable falsifiable check. The well-posedness framework is broad enough to cover super-linear kernels that are outside standard Lipschitz/monotone settings. However, two load-bearing points are currently not adequately proved: the existence of the interacting particle system (Theorem 2.13) is asserted with a one-sentence proof, and a key fluctuation estimate in the PoC proof (5.13)/(5.16) relies on an uncontrolled/non-verified step. These gaps must be repaired before the advertised rates are attached to a well-defined particle system.

major comments (4)
  1. [Theorem 2.13] The proof of Theorem 2.13 is one sentence: 'This result follows directly from verifying the coefficients of IPS (2.9) satisfy the assumptions in Theorem 2.6 in [65] on the product spaces. We omit details.' This is load-bearing: the coefficient of the i-th particle is A(t, X^{i,N}, \mu^N), so the product-space operator F_N(U)=(A(t,u_i,\mu^U))_{i=1}^N is not a product operator. One must verify on H^N the local monotonicity, growth, and coercivity conditions of [65, Thm 2.6], including the explicit local-monotonicity coefficient and its integrability. No such verification is supplied. Since Theorems 2.14 and 2.16 are rates for this IPS, the central PoC claims are not attached to a proven object unless this gap is filled.
  2. [§5.2, Eq. (5.13), Step 2] In bounding I3, the proof writes I3 ≲ N^{-2} Σ_i (E∫_0^T ||X^i_s - X^{i,N}_s||^α_V ds)^{1/α} (…) and then, without justification, drops the V-norm factor and concludes I3 ≲ N^{-1/α}. Theorem 2.14 does not assume dissipation that would give L^α([0,T];V) control of the difference; (A'_5) only controls H-norm increments. The factor (E∫ ||diff||^α_V)^{1/α} is not shown to be bounded uniformly in N, so (5.13) is not established. This directly affects the central claim (2.11).
  3. [§5.2, claim (5.16); §5.3, Eq. (5.26)] The assertion (5.16) — E∫||Σ_j(∫\tilde A dμ - \tilde A(·,X^j))||^{α/(α-1)} ≲ N — is used in both PoC proofs, but its derivation is only sketched. Lemma 5.5 is applied to a 'martingale difference sequence' without defining the filtration or verifying conditional centering for {∫\tilde A(t,x,y)μ_t(dy)-\tilde A(t,x,X^j_t)}_{j≠i} after conditioning on X^i_t. The independence structure makes this plausible, but the verification is nontrivial, and this estimate is the main fluctuation term used to absorb I3 and \tilde I3. The proof must be written out; the statement 'We claim' is not enough.
  4. [Theorem 2.16, (2.12)–(2.13)] The notation for the rates is ambiguous: it is not clear whether (q−β)/(q−2) and (p′−2)/p′ are multiplicative factors outside the exponent or part of the exponent. If they are outside, then taking q↓β or p′↓2 makes the claimed right-hand side O(1), contradicting the 'near-optimal' wording; if they are in the exponent, the theorem must state this unambiguously. The proof in §5.3 (5.31) appears to produce the product inside the exponent, but the final theorem statement should match the proof with an explicit exponent γ for N^{-γ}.
minor comments (5)
  1. [Global] The abstract and Section 1.2 call the rates 'near-optimal in a suitable sense', but no formal definition of near-optimality is given. Please state precisely which parameters are allowed to depend on N and what 'near' means.
  2. [Theorem 3.1] The condition p ≥ max{32k−24, 32(m+n)−40} appears without derivation; please indicate how β := max{8k−6, 8(m+n)−10} and the PoC assumptions on p are obtained.
  3. [§5.3, (5.28)] The exponent on the second term of (5.28) appears to be inconsistent with (5.21): it reads 1/ϑp where (5.21) has 2/ϑp. Please correct the typo.
  4. [References] The paper relies heavily on [42]–[44] by overlapping authors for qualitative PoC and on [65] for the IPS existence theorem. Please add a precise statement of which parts of the proof are new and which are quoted, so the reader can verify the self-containedness of the central claims.
  5. [Eq. (2.12)–(2.13)] As a typesetting issue, the superscripts in (2.12)–(2.13) are nearly unreadable in the current version; please rewrite with an explicit exponent, e.g. N^{-γ}, and give γ.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the PoC rates are derived from stated assumptions via Sznitman coupling and independent martingale estimates, not fitted or defined in terms of the conclusions.

full rationale

The quantitative propagation-of-chaos estimates in Theorems 2.14 and 2.16 are not fitted inputs renamed as predictions. The N^{-1/alpha} rate comes from applying Pisier's uniformly-smooth-space martingale-difference inequality to the V*-valued empirical-measure error in the proof of (5.16), and the improved N^{-1/(alpha-1)} rate in Theorem 2.16 comes from the same estimate together with the dissipative term delta0 ||u-v||_V^alpha in (A''_5). No assumption contains the target rate as a hidden parameter. Proposition 2.15 is an independent i.i.d. lower-bound computation illustrating optimality, not a source of the upper bounds. The well-posedness proofs are largely in-house; the use of [44, Lemma 5.9] is a self-citation, but it is a published, externally checkable theorem used only after pathwise uniqueness is proved in the paper, so it does not make the central claim equivalent to its inputs. The most notable issue is Theorem 2.13, whose one-sentence proof omits the verification of [65, Thm 2.6] for the non-product empirical-measure coupling in (2.9); that is a genuine proof gap and correctness risk, but an omitted verification is not circularity.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

The paper contains no fitted parameters and no new physical or mathematical entities such as new fields or particles. Its assumptions are structural conditions on the state space and coefficients. The main auxiliary construct, measure-dependent pseudo-monotone operators, is a definition rather than an invented entity. The proof depends on the unproved applicability of an external theorem for the interacting particle system, and this is the most significant assumption.

assumptions (5)
  • domain assumption Gelfand triple V⊂H≃H*⊂V* with V reflexive, compact and dense embedding into H.
    Used throughout Section 2 to define the variational framework and the spaces for weak and strong solutions; introduced in (1.3).
  • domain assumption V is a separable α-uniformly convex Banach space with α≥2, and V* is correspondingly α/(α-1)-uniformly smooth.
    This geometric assumption is the key driver of the PoC rates (Theorems 2.14 and 2.16); it is not derived within the paper. Proposition 2.5 supplies the duality.
  • domain assumption Coefficient conditions (A1)-(A6), (A'5), (A''5), including growth, coercivity, local/decoupled monotonicity, and the kernelization condition (A6).
    These are the theorem hypotheses defining the class of mean-field SPDEs covered. They are not consequences of more basic assumptions and are verified only for the specific examples in Section 3.
  • domain assumption Theorem 2.6 in [65] provides strong well-posedness for SPDEs with fully local monotone coefficients on product spaces.
    Invoked as the justification for Theorem 2.13, which asserts existence and uniqueness of the interacting particle system. The applicability to the empirical-measure coupling is not demonstrated.
  • standard math Standard probabilistic tools: Itô's formula, B-D-G inequality, martingale representation, Skorokhod representation, Prokhorov tightness, Rosenthal's inequality, Lenglart inequality.
    These underpin the proofs in Sections 4 and 5 and are cited or used without proof.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Mean-Field Stochastic PDEs: Well-posedness and Quantitative Dimension-Free Propagation of Chaos." pith.science (2026). https://pith.science/paper/5FEUYZG3

@misc{pith2026260717195,
  author       = {Pith},
  title        = {Pith review of: Mean-Field Stochastic PDEs: Well-posedness and Quantitative Dimension-Free Propagation of Chaos},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5FEUYZG3}},
  note         = {Machine review of arXiv:2607.17195}
}
abstract

This work investigates the mean-field stochastic PDEs involving a class of pseudo-monotone kernels. We first study the well-posedness -- in both the strong and weak sense -- within the variational framework by introducing a notion of measure-dependent pseudo-monotone operators, which generalizes the classical framework due to Br\'{e}zis. Furthermore, we establish the quantitative dimension-free propagation of chaos within a $p$-uniformly convex Banach space for general infinite-dimensional weakly interacting systems, obtaining convergence rates that are near-optimal in a suitable sense. Our results reveal a new insight: the convergence rate of the mean-field limit is intrinsically governed by the geometry of the underlying solution space, specifically its modulus of convexity. As applications, we study several finite- and infinite-dimensional interacting particle systems arising in machine learning and fluid mechanics, including stochastic Stein variational gradient descent, mean-field Allen-Cahn equations, and Lagrangian-averaged Burgers equations.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

83 extracted references · 1 linked inside Pith

  1. [65]

    and Zhang, T

    R¨ockner, M., Shang, S. and Zhang, T. (2025).Well-posedness of stochastic partial differential equations with fully local monotone coefficients. Math. Ann.390, 3419–3469

    2025

  2. [1]

    and Veraar, M

    Agresti, A. and Veraar, M. (2024).The critical variational setting for stochastic evolution equations. Probab. Theory Relat. Fields188, 957–1015

    2024

  3. [2]

    (1978).Stopping times and tightness

    Aldous, D. (1978).Stopping times and tightness. Ann. Probab.6, 335–340

    1978

  4. [3]

    Allen, S. M. and Cahn, J. W. (1979).A microscopic theory for antiphase boundary motion and its application to antiphase domain coarsening. Acta. Metall.27, 1084–1095

    1979

  5. [4]

    and Xu, F

    Bagnara, M., Maurelli, M. and Xu, F. (2025).No blow-up by nonlinear Itˆ o noise for the Euler equations. Electron. J. Probab.30, 1–29

    2025

  6. [5]

    and R ¨ockner, M

    Barbu, V. and R ¨ockner, M. (2024).Nonlinear Fokker-Planck flows and their probabilistic counterparts. Lecture Notes in Math.2353, Springer

    2024

  7. [6]

    and Slade, G

    Bauerschmidt, R. and Slade, G. (2020).Mean-field tricritical polymers. Probab. Math. Phys.1, 167–204

    2020

  8. [7]

    (2011).Introduction to Banach spaces and their geometry

    Beauzamy, B. (2011).Introduction to Banach spaces and their geometry. North-Holland Mathe- matics Studies68, Elsevier

    2011

Show all 83 references
  1. [8]

    G., Kallianpur, G., Karandikar, R

    Bhatt, A. G., Kallianpur, G., Karandikar, R. L. and Xiong, J. (1998).On interacting systems of Hilbert-space-valued diffusions. Appl. Math. Optim.37, 151–188. 71

    1998

  2. [9]

    M., Kucukelbir, A

    Blei, D. M., Kucukelbir, A. and McAuliffe, J. D. (2017).Variational inference: A review for statisticians. J. Am. Stat. Assoc.112, 859–877

    2017

  3. [10]

    Bolley, F., Canizo, J. A. and Carrillo, J. A. (2011).Stochastic mean-field limit: non-Lipschitz forces and swarming. Math. Models Methods Appl. Sci.21, 2179–2210

    2011

  4. [11]

    (1968).´Equations et in´ equations non lin´ eaires dans les espaces vectoriels en dualit´ e

    Br´ ezis, H. (1968).´Equations et in´ equations non lin´ eaires dans les espaces vectoriels en dualit´ e. Ann. Inst. Fourier (Grenoble)18, 115–175

    1968

  5. [12]

    and Rainer, C

    Buckdahn, R., Li, J., Peng, S. and Rainer, C. (2017).Mean-field stochastic differential equations and associated PDEs. Ann. Probab.45, 824–878

    2017

  6. [13]

    and Delarue, F

    Carmona, R. and Delarue, F. (2018).Probabilistic Theory of Mean Field Games with Applica- tions I/II. Probab. Theory Stoch. Model.83/84, Springer

    2018

  7. [14]

    and Diez, A

    Chaintron, L.-P. and Diez, A. (2022).Propagation of chaos: a review of models, methods and applications. II. Applications. Kinet. Relat. Models15, 1017–1173

    2022

  8. [15]

    and Chen, L

    Chen, C., Zhang, R., Wang, W., Li, B. and Chen, L. (2018).A unified particle-optimization framework for scalable Bayesian sampling. In Uncertainty in Artificial Intelligence (UAI)

    2018

  9. [16]

    S., Kallianpur, G

    Chiang, T. S., Kallianpur, G. and Sundar, P. (1991).Propagation of chaos and the McKean- Vlasov equation in duals of nuclear spaces. Appl. Math. Optim.24, 55–83

    1991

  10. [17]

    Chidume, C. E. (2009).Geometric properties of Banach spaces and nonlinear iterations. Lecture Notes in Mathematics1965, Springer

    2009

  11. [18]

    (2023).Propagation of chaos for weakly interacting mild solutions to stochastic partial differential equations

    Criens, D. (2023).Propagation of chaos for weakly interacting mild solutions to stochastic partial differential equations. J. Statist. Phys.190, 114

    2023

  12. [19]

    and Zabczyk J

    Da Prato, G. and Zabczyk J. (1992).Stochastic Equations in Infinite Dimensions. Cambridge University Press

    1992

  13. [20]

    (2013).Mean field simulation for Monte Carlo integration

    Del Moral, P. (2013).Mean field simulation for Monte Carlo integration. Monogr. Stat. Appl. Probab.126, CRC Press

    2013

  14. [21]

    and Tugaut, J

    Del Moral, P. and Tugaut, J. (2018).On the stability and the uniform propagation of chaos properties of ensemble Kalman-Bucy filters. Ann. Appl. Probab.28, 790–850

    2018

  15. [22]

    Diop, C., Sow, T. M. M., Djitte, N. and Chidume, C. E. (2015).Constructive techniques for zeros of monotone mappings in certain Banach spaces. SpringerPlus,4, 383

    2015

  16. [23]

    D., Holm, D

    Drivas, T. D., Holm, D. D. and Leahy, J.-M. (2020).Lagrangian averaged stochastic advection by Lie transport for fluids. J. Stat. Phys.179, 1304–1342

    2020

  17. [24]

    Dudley, R. M. (1969).The speed of mean Glivenko–Cantelli convergence. Ann. Math. Statist. 40, 40–50

    1969

  18. [25]

    and Szpruch, L

    Duncan, A., N ¨usken, N. and Szpruch, L. (2023).On the geometry of Stein variational gradient descent. J. Mach. Learn. Res.24, 1–39

    2023

  19. [26]

    and Shen, H

    E, W. and Shen, H. (2013).Mean field limit of a dynamical model of polymer systems. Sci. China Math.56, 2591–2598

    2013

  20. [27]

    (2022).Well-posedness and propagation of chaos for McKean-Vlasov equations with jumps and locally Lipschitz coefficients

    Erny, X. (2022).Well-posedness and propagation of chaos for McKean-Vlasov equations with jumps and locally Lipschitz coefficients. Stoch. Process. Appl.150, 192–214

    2022

  21. [28]

    Ertel, S. W. and Stannat, W. (2024).Analysis of the ensemble Kalman-Bucy filter for correlated observation noise. Ann. Appl. Probab.34, 1072–1107

    2024

  22. [29]

    Folland, G. B. (1999).Real analysis: modern techniques and their applications. John Wiley & Sons

    1999

  23. [30]

    and Guillin, A

    Fournier, N. and Guillin, A. (2015).On the rate of convergence in Wasserstein distance of the empirical measure. Probab. Theory Relat. Fields162, 707–738

    2015

  24. [31]

    Rosati, T

    Gabriel, S. Rosati, T. and Zygouras, N. (2025).The Allen-Cahn equation with weakly critical random initial datum. Probab. Theory Relat. Fields192, 1373–1446

    2025

  25. [32]

    Galeati, L., Harang, F. A. and Mayorcas, A. (2022).Distribution dependent SDEs driven by additive continuous noise. Electron. J. Probab.27, Paper No. 37, 38 pp

    2022

  26. [33]

    and Insua, D

    Gallego, V. and Insua, D. R. (2020).Stochastic gradient MCMC with repulsive forces. arXiv:1812.00071

    2020 arXiv

  27. [34]

    and Ziane, M

    Glatt-Holtz, N. and Ziane, M. (2009).Strong pathwise solutions of the stochastic Navier-Stokes system. Adv. Differ. Equ.14, 567–600

    2009

  28. [35]

    and Krylov, N

    Gy¨ongy, I. and Krylov, N. V. (2003).On the Rate of Convergence of Splitting-up Approximations for SPDEs. Progr. Probab.56, 301–321

    2003

  29. [36]

    and Rosati, T

    Hairer, M., Lˆ e, K. and Rosati, T. (2023).The Allen-Cahn equation with generic initial datum. Probab. Theory Relat. Fields186, 957–998. 72

    2023

  30. [37]

    and Szpruch L

    Hammersley, W., ˇSiˇ ska D. and Szpruch L. (2021).McKean-Vlasov SDEs under measure dependent Lyapunov conditions. Ann. Inst. Henri Poincar´ e Probab. Stat.57, 1032–1057

    2021

  31. [38]

    and Zhang, X

    Hao, Z., R¨ockner, M. and Zhang, X. (2024).Strong convergence of propagation of chaos for McKean-Vlasov SDEs with singular interactions. SIAM J. Math. Anal.56, 2661–2713

    2024

  32. [39]

    Hastings, W. K. (1970).Monte Carlo sampling methods using Markov chains and their applications. Biometrika57, 97–109

    1970

  33. [40]

    and Liu, W

    Hong, W., Li, S. and Liu, W. (2022).Strong convergence rates in averaging principle for slow-fast McKean-Vlasov SPDEs. J. Differ. Equ.316, 94–135

    2022

  34. [41]

    and Liu, W

    Hong, W., Li, S. and Liu, W. (2021).Freidlin-Wentzell type large deviation principle for multiscale locally monotone SPDEs. SIAM J. Math. Anal.53, 6517–6561

    2021

  35. [42]

    and Liu, W

    Hong, W., Hu, S. and Liu, W. (2024).McKean–Vlasov SDE and SPDE with locally monotone coefficients. Ann. Appl. Probab.342136–2189

    2024

  36. [43]

    and Liu, W

    Hong, W., Li, S. and Liu, W. (2025).McKean–Vlasov stochastic partial differential equations: existence, uniqueness and propagation of chaos. Probab. Theory Relat. Fields193, 717–793

    2025

  37. [44]

    and Liu, W

    Hong, W., Li, S. and Liu, W. (2026).Mean field stochastic partial differential equations with nonlinear kernels. Ann. Appl. Probab.36, 206–274

    2026

  38. [45]

    and Wang, Z

    Jabin, P.-E. and Wang, Z. (2018).Quantitative estimates of propagation of chaos for stochastic systems withW −1,∞ kernels. Invent. Math.214, 523–591

    2018

  39. [46]

    (1986).On the Skorokhod topology

    Jakubowski, A. (1986).On the Skorokhod topology. Ann. Inst. H. Poincar´ e Probab. Statist.22, 263–285

    1986

  40. [47]

    (1956).Foundations of Kinetic Theory

    Kac, M. (1956).Foundations of Kinetic Theory. University of California Press, Berkeley and Los Angeles,III, 171–197

    1956

  41. [48]

    and Xiong, J

    Kallianpur, G. and Xiong, J. (1994).Asymptotic behavior of a system of interacting nuclear- space-valued stochastic differential equations driven by Poisson random measures. Appl. Math. Optim.30, 175–201

    1994

  42. [49]

    Krylov, N. V. and Rozovskii, B. L. (1981).Stochastic evolution equations. Translated from Itogi Naukii Tekhniki, Seriya Sovremennye Problemy Matematiki. Plenum Publishing Corp.14, 71–146

    1981

  43. [50]

    (2018).Mean field games and interacting particle systems

    Lacker, D. (2018).Mean field games and interacting particle systems. Springer

    2018

  44. [51]

    and Tzafriri, L

    Lindenstrauss, J. and Tzafriri, L. (1996).Classical Banach spaces I and II: Sequence spaces and function spaces. Classics in Mathematics, Springer

    1996

  45. [52]

    (2017).Stein variational gradient descent as gradient flow

    Liu, Q. (2017).Stein variational gradient descent as gradient flow. Adv. Neural Inf. Process. Syst.30, 3115–3123

    2017

  46. [53]

    and Wang, D

    Liu, Q. and Wang, D. (2016).Stein variational gradient descent: A general purpose Bayesian inference algorithm. Adv. Neural Inf. Process. Syst.29, 2378–2386

    2016

  47. [54]

    and Wang, D

    Liu, Q. and Wang, D. (2018).Stein variational gradient descent as moment matching. Adv. Neural Inf. Process. Syst.31, 8868–8877

    2018

  48. [55]

    and R ¨ockner, M

    Liu, W. and R ¨ockner, M. (2015).Stochastic Partial Differential Equations: An Introduction. Universitext, Springer

    2015

  49. [56]

    McKean, H. P. (1966).A class of Markov processes associated with nonlinear parabolic equations. Proc. Nat. Acad. Sci. U.S.A.56, 1907–1911

    1966

  50. [57]

    McKean H. P. (1967).Propagation of chaos for a class of non-linear parabolic equations. Stochastic Differential Equations (Lecture Series in Differential Equations, Session 7, Catholic Univ., 1967), 41–57

    1967

  51. [58]

    and Veretennikov, A

    Mishura, Y. and Veretennikov, A. (2020).Existence and uniqueness theorems for solutions of McKean-Vlasov stochastic equations. Theory Probab. Math. Statist.103, 59–101

    2020

  52. [59]

    (2020).Coercivity condition for higher moment a priori estimates for nonlinear SPDEs and existence of a solution under local monotonicity

    Neelima and ˇSiˇ ska D. (2020).Coercivity condition for higher moment a priori estimates for nonlinear SPDEs and existence of a solution under local monotonicity. Stochastics92, 684–715

    2020

  53. [60]

    and Temam, R

    Nguyen, P., Tawri, K. and Temam, R. (2021).Nonlinear stochastic parabolic partial differential equations with a monotone operator of the Ladyzenskaya-Smagorinsky type, driven by a L´ evy noise. J. Funct. Anal.281, 109157

    2021

  54. [61]

    and Tan, C

    Nguyen, V., Shvydkoy, R. and Tan, C. (2026).Mean-field limits of deterministic and stochastic flocking models with nonlinear velocity alignment. Adv. Math.494, 45 pp

    2026

  55. [62]

    and Martin S

    Pinnau R., Totzeck C., Tse O. and Martin S. (2017).A consensus-based model for global optimization and its mean-field limit. Math. Models Methods Appl. Sci.27, 183–204. 73

    2017

  56. [63]

    (1975).Martingales with values in uniformly convex spaces

    Pisier, G. (1975).Martingales with values in uniformly convex spaces. Israel J. Math.20, 326–350

    1975

  57. [64]

    (1975).Equations aux d´ eriv´ ees partielles stochastiques non lin´ eaires monotones

    Pardoux, ´E. (1975).Equations aux d´ eriv´ ees partielles stochastiques non lin´ eaires monotones. PhD thesis, Universit´ e Paris XI

    1975

  58. [66]

    Rosenthal, H. P. (1970).On the subspaces ofL p (p >2) spanned by sequences of independent random variables. Israel J. Math.8, 273–303

    1970

  59. [67]

    and Vanden-Eijnden, E

    Rotskoff, G. and Vanden-Eijnden, E. (2018).Parameters as interacting particles: Long time convergence and asymptotic error scaling of neural networks. Adv. Neural Inf. Process. Syst.31, 7146–7155

    2018

  60. [68]

    and Vanden-Eijnden, E

    Rotskoff, G. and Vanden-Eijnden, E. (2022).Trainability and accuracy of artificial neural networks: An interacting particle system approach. Commun. Pure Appl. Math.75, 1889–1935

    2022

  61. [69]

    Rozovskii, B. L. and Lototsky, S. V. (2018).Stochastic evolution systems. Linear theory and applications to non-linear filtering. Probability Theory and Stochastic Modelling89, Springer

    2018

  62. [70]

    (1987).Uniqueness and nonuniqueness of solutions of Vlasov-McKean equations

    Scheutzow, M. (1987).Uniqueness and nonuniqueness of solutions of Vlasov-McKean equations. J. Aust. Math. Soc. Ser. A43, 246–256

    1987

  63. [71]

    (2020).Mean field limit for Coulomb-type flows

    Serfaty, S. (2020).Mean field limit for Coulomb-type flows. Duke Math. J.169, 2887–2935

    2020

  64. [72]

    and Zhu, X

    Shen, H., Smith, S., Zhu, R. and Zhu, X. (2022).Large N limit of the O(N) linear sigma model via stochastic quantization. Ann. Probab.50, 131–202

    2022

  65. [73]

    (1991).Topics in propagation of chaos

    Sznitman, A.-S. (1991).Topics in propagation of chaos. ´Ecole d’ ´Et´ ede Probabilit´ es de Saint- Flour XIX-1989, 165-251, Lecture Notes in Math. 1464, Springer

    1991

  66. [74]

    (1995).Analytic semigroups and semilinear initial boundary value problems

    Taira, K. (1995).Analytic semigroups and semilinear initial boundary value problems. Cambridge University Press

    1995

  67. [75]

    (2024).Sharp propagation of chaos for the ensemble Langevin sampler

    Vaes, U. (2024).Sharp propagation of chaos for the ensemble Langevin sampler. J. Lond. Math. Soc.110, e13008

    2024

  68. [76]

    (2015).Asymptotic couplings by reflection and applications for nonlinear monotone SPDES

    Wang, F.-Y. (2015).Asymptotic couplings by reflection and applications for nonlinear monotone SPDES. Nonlinear Anal.117, 169–188

    2015

  69. [77]

    (2018).Distribution dependent SDEs for Landau type equations

    Wang, F.-Y. (2018).Distribution dependent SDEs for Landau type equations. Stochastic Process. Appl.128, 595–621

    2018

  70. [78]

    and Ren, P

    Wang, F.-Y. and Ren, P. (2025).Distribution dependent stochastic differential equations. World Scientific Publishing Co. Pte. Ltd., Hackensack, NJ

    2025

  71. [79]

    (2010).Sharp interface limit for invariant measures of a stochastic Allen-Cahn equation

    Weber, H. (2010).Sharp interface limit for invariant measures of a stochastic Allen-Cahn equation. Comm. Pure Appl. Math.63, 1071–1109

    2010

  72. [80]

    and Bach, F

    Weed, J. and Bach, F. (2019).Sharp asymptotic and finite-sample rates of convergence of empirical measures in Wasserstein distance. Bernoulli25, 2620–2648

    2019

  73. [81]

    (1991).Inequalities in Banach spaces with applications

    Xu, H. (1991).Inequalities in Banach spaces with applications. Nonlinear Anal.16, 1127–1138

    1991

  74. [82]

    Xu, Z. B. and Roach, G. F. (1991).Characteristic inequalities of uniformly convex and uniformly smooth Banach spaces. J. Math. Anal. Appl.157, 189–210

    1991

  75. [83]

    and Chen, C

    Zhang, J., Zhang, R., Carin, L. and Chen, C. (2020).Stochastic particle-optimization sampling and the non-asymptotic convergence theory. In Proc. Int. Conf. Artif. Intell. Stat. 1877–1887, PMLR

    2020

Pith tools

Reviewed August 1, 2026 · model on record in the stance chip above.