pith. sign in

arxiv: 2505.19787 · v3 · submitted 2025-05-26 · 🧮 math.PR

Entropy-Cost Inequalities for McKean-Vlasov SDEs with Singular Interactions

Xing Huang , Panpan Ren , Feng-Yu Wang This is my paper

Pith reviewed 2026-05-19 13:51 UTC · model grok-4.3

classification 🧮 math.PR
keywords McKean-Vlasov SDEssingular interactionsentropy-cost inequalitywell-posednessCoulomb kernelWasserstein distancediffusion semigroupsregularity estimates
0
0 comments X

The pith

Entropy-cost inequalities establish well-posedness and regularity for McKean-Vlasov SDEs with singular interactions such as Coulomb kernels.

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

The paper shows that entropy-cost inequalities can be established to prove well-posedness and regularity estimates for McKean-Vlasov stochastic differential equations whose interaction terms are singular, including Coulomb, Riesz and Biot-Savart kernels as examples. A reader would care because these equations arise in mean-field models of particles under long-range or infinite forces at close range. The authors measure the singularity by a new probability distance built from local integrable functions and bound this distance between time-marginal laws using the Wasserstein distance of the initial distributions. They further describe the admissible paths of the marginal distributions through local hyperbound estimates on the underlying diffusion semigroups.

Core claim

By establishing entropy-cost inequalities, the paper derives well-posedness and regularity estimates for McKean-Vlasov SDEs with singular interactions, including the Coulomb, Riesz, and Biot-Savart kernels. The singularity is measured using a new probability distance induced by local integrable functions, which allows estimating the distance between time-marginal laws from the Wasserstein distance of initial distributions. The path space of time-marginal distributions is characterized via local hyperbound estimates on diffusion semigroups.

What carries the argument

The entropy-cost inequality that bounds the relative entropy of the solution law by a cost functional incorporating the singular interaction, together with a new probability distance induced by local integrable functions to quantify the singularity.

If this is right

  • Unique strong solutions exist for the McKean-Vlasov SDEs with the listed singular kernels.
  • Regularity estimates hold for the time-marginal laws of the solutions.
  • Discrepancies between time-marginal laws are controlled by the initial Wasserstein distance through the new probability distance.
  • The path space of the marginal distributions satisfies local hyperbound estimates derived from diffusion semigroups.

Where Pith is reading between the lines

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

  • The entropy-cost approach could extend to deriving quantitative propagation-of-chaos rates for finite-particle approximations with the same singular forces.
  • Similar distance constructions may apply to other nonlocal mean-field equations outside the McKean-Vlasov setting.
  • Long-time convergence and ergodicity results might follow from iterating the entropy-cost bounds under suitable Lyapunov conditions.

Load-bearing premise

The singularity of the interaction kernels can be controlled by a new probability distance induced by local integrable functions.

What would settle it

A concrete singular kernel for which the entropy-cost inequality fails to produce a finite bound would indicate that well-posedness or regularity estimates break down for the associated McKean-Vlasov equation.

read the original abstract

For a class of McKean-Vlasov stochastic differential equations with singular interactions, which include the Coulomb/Riesz/Biot-Savart kernels as typical examples (Examples 2.1 and 2.2), we derive the well-posedness and regularity estimates by establishing the entropy-cost inequality. To measure the singularity of interactions, we introduce a new probability distance induced by local integrable functions, and estimate this distance for the time-marginal laws of solutions by using the Wasserstein distance of initial distributions. A key point of the study is to characterize the path space of time-marginal distributions for the solutions, by using local hyperbound estimates on diffusion semigroups.

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

2 major / 2 minor

Summary. The manuscript derives well-posedness and regularity estimates for McKean-Vlasov SDEs with singular interactions (Coulomb, Riesz, Biot-Savart kernels) by establishing entropy-cost inequalities. It introduces a new probability distance induced by local integrable functions to quantify kernel singularity, bounds the time-marginal laws of solutions via the Wasserstein distance of initial distributions, and characterizes the path space using local hyperbound estimates on diffusion semigroups.

Significance. If the entropy-cost inequality closes with constants independent of the approximation parameter, the work supplies a new metric and hyperbound framework for singular mean-field interactions. This would extend existing results on Lipschitz or mildly singular kernels and offer a concrete route to well-posedness for physically relevant models in statistical mechanics and fluid dynamics.

major comments (2)
  1. [§3] §3 (Entropy-cost inequality): the proof does not supply explicit verification that the local integrability modulus of the approximated kernels K_ε remains uniform as ε→0 in the hyperbound estimates on the diffusion semigroup; without this, the constant in the entropy-cost inequality may deteriorate and prevent closure of the fixed-point argument for the singular limit.
  2. [§4.1] §4.1 (Time-marginal estimates): the claim that d_μ(μ_t, ν_t) ≤ C W_2(μ_0, ν_0) controls the interaction term assumes C is finite and independent of the local L^1 norm of K near the origin, but no quantitative bound or uniformity statement is given to justify bootstrapping back to the singular kernel.
minor comments (2)
  1. [Abstract] The abstract refers to Examples 2.1 and 2.2 but does not state the precise assumptions on the kernels; adding a one-sentence summary of the admissible singularity class would improve readability.
  2. [Introduction] Notation for the new distance d_μ is introduced without an explicit formula in the introduction; a brief display equation would clarify the construction before the technical sections.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading and constructive comments on our manuscript. We address the major comments point by point below and will revise the manuscript to incorporate the requested clarifications on uniformity.

read point-by-point responses

  1. Referee: [§3] §3 (Entropy-cost inequality): the proof does not supply explicit verification that the local integrability modulus of the approximated kernels K_ε remains uniform as ε→0 in the hyperbound estimates on the diffusion semigroup; without this, the constant in the entropy-cost inequality may deteriorate and prevent closure of the fixed-point argument for the singular limit.

    Authors: We thank the referee for this observation. The local integrability modulus of K_ε is preserved uniformly in ε by our choice of standard mollification, which does not worsen the singularity and keeps the L^1_loc norm bounded by that of the original kernel K (as stated in Examples 2.1 and 2.2). The hyperbound estimates then inherit this uniformity. However, we agree that an explicit verification is not written out in the current proof of the entropy-cost inequality. We will add a short lemma in the revised §3 establishing that the relevant constants remain independent of ε, thereby confirming that the entropy-cost inequality closes uniformly and the fixed-point argument applies to the singular limit. revision: yes

  2. Referee: [§4.1] §4.1 (Time-marginal estimates): the claim that d_μ(μ_t, ν_t) ≤ C W_2(μ_0, ν_0) controls the interaction term assumes C is finite and independent of the local L^1 norm of K near the origin, but no quantitative bound or uniformity statement is given to justify bootstrapping back to the singular kernel.

    Authors: We agree that an explicit quantitative bound on C is needed for the singular case. The constant C is controlled by the entropy-cost inequality together with the definition of the new distance d_μ, which is constructed precisely so that it is comparable to the Wasserstein distance under the local integrability assumption on K. Because the approximating kernels satisfy a uniform local L^1 bound, the same C works in the limit. We will insert a quantitative remark or short proposition in the revised §4.1 that records the dependence of C on the local L^1 norm and states its uniformity under our approximation scheme, thereby justifying the passage to the singular kernel. revision: yes

Circularity Check

0 steps flagged

No significant circularity detected in derivation chain

full rationale

The paper introduces an independent new probability distance induced by local integrable functions to quantify interaction singularity (distinct from Wasserstein), then derives estimates of this distance for time-marginals from initial Wasserstein distance and characterizes path space via local hyperbound estimates on semigroups. The entropy-cost inequality is established within the paper to obtain well-posedness and regularity for the McKean-Vlasov SDEs, with no quoted reduction showing that any central claim (e.g., the inequality or distance bound) is equivalent to its inputs by definition, fitted parameter, or self-citation load-bearing premise. The construction uses standard tools like Gronwall-type closures and semigroup estimates without evidence of tautological renaming or ansatz smuggling.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 1 invented entities

The claim rests on the domain assumption that singular kernels admit a well-defined distance via local integrable functions and that hyperbound estimates apply to the associated diffusion semigroups; no free parameters or new physical entities are introduced in the abstract.

axioms (2)
  • domain assumption Singular interactions (Coulomb/Riesz/Biot-Savart) can be measured by a probability distance induced by local integrable functions.
    Stated in the abstract as the device used to quantify singularity and to obtain the marginal estimates.
  • domain assumption Local hyperbound estimates hold for the diffusion semigroups generated by the McKean-Vlasov dynamics.
    Invoked to characterize the path space of time-marginal distributions.
invented entities (1)
  • New probability distance induced by local integrable functions no independent evidence
    purpose: To measure the singularity of interactions and to bound the distance between time-marginal laws.
    Introduced explicitly in the abstract as the key tool for handling singular kernels.

pith-pipeline@v0.9.0 · 5642 in / 1597 out tokens · 80285 ms · 2026-05-19T13:51:25.747949+00:00 · methodology

discussion (0)

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

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Forward citations

Cited by 2 Pith papers

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

  1. Bismut Formula for Intrinsic Derivative of DDSDEs with Singular Interactions

    math.PR 2026-04 unverdicted novelty 6.0

    Derives Bismut formulas for intrinsic derivatives of DDSDEs with singular interactions by extending the differentiable-drift case.

  2. Regularity Estimates for Singular Density Dependent SDEs

    math.PR 2026-02 unverdicted novelty 6.0

    Relative and Renyi entropy estimates for singular density-dependent SDEs are bounded by initial Wasserstein distances and coincide with classical entropy-cost inequalities in one dimension with t^{1/2+} decay at zero.

Reference graph

Works this paper leans on

30 extracted references · 30 canonical work pages · cited by 2 Pith papers

  1. [1]

    Barbu, M

    V. Barbu, M. R¨ ockner.:From nonlinear Fokker-Planck equations to solutions of distribution dependent SDE, Annals of Probability,48(2020), 1902-1920

    work page 2020

  2. [2]

    Barbu, M

    V. Barbu, M. R¨ ockner, and D. Zhang.:Uniqueness of distributional solu- tions to the 2D vorticity Navier-Stokes equation and its associated nonlinear Markov process, Journal of the European Mathematical Society, available at https://ems.press/journals/jems/articles/14298872

    work page arXiv

  3. [3]

    R. F. Bass, E. Perkins.:A new technique for proving uniqueness for martingale prob- lems, Ast´ erisque,327(2009), 47-53

    work page 2009

  4. [4]

    Ben-Artzi.:Global solutions of two-dimensional Navier-Stokes and Euler equa- tions, Archive for Rational Mechanics and Analysis,128(1994), 329-358

    M. Ben-Artzi.:Global solutions of two-dimensional Navier-Stokes and Euler equa- tions, Archive for Rational Mechanics and Analysis,128(1994), 329-358. 49

    work page 1994

  5. [5]

    Carmona, F

    R. Carmona, F. Delarue.:Probabilistic Theory of Mean Field Games with Applica- tions I, Springer, 2019

    work page 2019

  6. [6]

    G.-Q. G. Chen, Z. Qian,New approach for vorticity estimates of solutions of the Navier-Stokes equations, arXiv:2210.04129

    work page arXiv

  7. [7]

    Gallagher, T

    I. Gallagher, T. Gallay.:Uniqueness for the two-dimensional Navier-Stokes equation with a measure as initial vorticity, Mathematische Annalen,332(2005), 287-327

    work page 2005

  8. [8]

    Y. Giga, T. Miyakawa, and H. Osada.:Two-dimensional Navier-Stokes flow with measures as initial vorticity, Archive for Rational Mechanics and Analysis, 104(1988), 223-250

    work page 1988

  9. [9]

    Huang, F.-Y

    X. Huang, F.-Y. Wang.:Log-Harnack inequality and Bismut formula for McKean- Vlasov SDEs with singularities in all Variables, Mathematische Annalen,available at https://doi.org/10.1007/s00208-025-03255-x

    work page doi:10.1007/s00208-025-03255-x

  10. [10]

    Huang, P

    X. Huang, P. Ren, and F.-Y. Wang.:Probability distance estimates between diffusion processes and applications to singular McKean-Vlasov SDEs, Journal of Differential Equations,420(2025), 376-399

    work page 2025

  11. [11]

    Jabin, Z

    P.-E. Jabin, Z. Wang.:Quantitative estimates of propagation of chaos for stochastic systems withW −1,∞ kernels, Inventions Mathematicae,214(2018), 523-591

    work page 2018

  12. [12]

    Kato.:The Navier-Stokes equation for an incompressible fluid inR 2 with a measure as the initial vorticity, Differential Integration Equation,7(1994), 949-966

    T. Kato.:The Navier-Stokes equation for an incompressible fluid inR 2 with a measure as the initial vorticity, Differential Integration Equation,7(1994), 949-966

    work page 1994

  13. [13]

    Lewin.:Coulomb and Riesz gases: the known and the unknown, Journal of Math- ematical Physics,63,061101(2022), available at https://doi.org/10.1063/5.0086835

    M. Lewin.:Coulomb and Riesz gases: the known and the unknown, Journal of Math- ematical Physics,63,061101(2022), available at https://doi.org/10.1063/5.0086835

    work page doi:10.1063/5.0086835 2022

  14. [14]

    J. Li, Z. Qian.:Mean field equations arising from random vortex dynamics, arXiv:2404.05054

    work page arXiv

  15. [15]

    Menozzi, A

    S. Menozzi, A. Pesce, and X. Zhang.:Density and gradient estimates for non de- generate Brownian SDEs with unbounded measurable drift, Journal of Differential Equations,272(2021), 330-369

    work page 2021

  16. [16]

    Ren.:Singular McKean-Vlasov SDEs: well-posedness, regularities and Wang’s Harnack inequality, Stochastic Processes and their Applications,156(2023), 291- 311

    P. Ren.:Singular McKean-Vlasov SDEs: well-posedness, regularities and Wang’s Harnack inequality, Stochastic Processes and their Applications,156(2023), 291- 311

    work page 2023

  17. [17]

    Ren, F.-Y.Wang.:Exponential convergence in entropy and Wasserstein for McKean-Vlasov SDEs, Nonlinear Analysis,206(2021), available at https://doi.org/10.1016/j.na.2021.112259

    P. Ren, F.-Y.Wang.:Exponential convergence in entropy and Wasserstein for McKean-Vlasov SDEs, Nonlinear Analysis,206(2021), available at https://doi.org/10.1016/j.na.2021.112259. 50

    work page doi:10.1016/j.na.2021.112259 2021

  18. [18]

    Ren, F.-Y

    P. Ren, F.-Y. Wang.:Bi-coupling method and applications, Probability Theory and Related Fields, available at https://doi.org/10.1007/s00440-025-01394-5

    work page doi:10.1007/s00440-025-01394-5

  19. [19]

    R¨ ockner, X

    M. R¨ ockner, X. Zhang.:Well-posedness of distribution dependent SDEs with singular drifts, Bernoulli,27(2021), 1131-1158

    work page 2021

  20. [20]

    Rosenzweig, S

    M. Rosenzweig, S. Serfaty.:Global-in-time mean-field convergence for singular Riesz type diffusive flows, The Annnals of Applied Probability,33(2023), 954-998

    work page 2023

  21. [21]

    Serfaty.:Lectures on Coulomb and Riesz Gases, arXiv:2407.21194

    S. Serfaty.:Lectures on Coulomb and Riesz Gases, arXiv:2407.21194

    work page arXiv

  22. [22]

    Trevisan.:Well-posedness of multidimensional diffusion processes, Electronic Journal of Probability,21(2016), 1-41

    D. Trevisan.:Well-posedness of multidimensional diffusion processes, Electronic Journal of Probability,21(2016), 1-41

    work page 2016

  23. [23]

    Wang.:Harnack Inequality and Applications for Stochastic Partial Differential Equations, Springer, 2013

    F.-Y. Wang.:Harnack Inequality and Applications for Stochastic Partial Differential Equations, Springer, 2013

    work page 2013

  24. [24]

    Wang.:Distribution dependent reflecting stochastic differential equations, Sci- ence China Mathematics,66(2023), 2411-2456

    F.-Y. Wang.:Distribution dependent reflecting stochastic differential equations, Sci- ence China Mathematics,66(2023), 2411-2456

    work page 2023

  25. [25]

    Wang.:Derivative formula for singular McKean-Vlasov SDEs, Communica- tion on Pure and Applied Analysis,22(2023), 1866-1898

    F.-Y. Wang.:Derivative formula for singular McKean-Vlasov SDEs, Communica- tion on Pure and Applied Analysis,22(2023), 1866-1898

    work page 2023

  26. [26]

    Wang.:Singular density dependent stochastic differential equations, Journal of Differential Equations,361(2023), 562-589

    F.-Y. Wang.:Singular density dependent stochastic differential equations, Journal of Differential Equations,361(2023), 562-589

    work page 2023

  27. [27]

    F.-Y. Wang, P. Ren.:Distribution Dependent Stochastic Differential Equations, World Scientific,2025

    work page 2025

  28. [28]

    P. Xia, L. Xie, X. Zhang, and G. Zhao.:L q(Lp)-theory of stochastic differential equa- tions, Stochastic Process and Their Applications,130(2020), 5188-5211

    work page 2020

  29. [29]

    Zhang, C

    S.-Q. Zhang, C. Yuan.:A Zvonkin’s transformation for stochastic differential equa- tions with singular drift and applications, Journal of Differential Equations, 297(2021), 277-319

    work page 2021

  30. [30]

    Zhao.:Existence and uniqueness for McKean-Vlasov equations with singular in- teractions, Potential Analysis,62(2025), 625-653

    G. Zhao.:Existence and uniqueness for McKean-Vlasov equations with singular in- teractions, Potential Analysis,62(2025), 625-653. 51

    work page 2025