Pith. sign in

REVIEW 3 major objections 5 minor 66 references

Sampling and Optimization meet Enhanced Flows

T0 review · 3 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read This paper constructs deterministic shear-flow dynamics that converge to a prescribed Gibbs measure at the enhanced rate $O(\nu^{1/2})$ instead of the classical $O(\nu)$ Langevin rate, and it offers a quadrature-free way to compute the…

desk verdict New deterministic samplers with provably faster convergence, but the main enhanced-dissipation proof has a misstated lemma that needs fixing. read the letter →

arxiv 2608.07329 v1 pith:EGMAZ642 submitted 2026-08-07 math.OC math.AP

Yuan Gao , Siming He , Eitan Tadmor This is my paper · ORCID
classification math.OCmath.AP MSC 35B4035Q8435Q3537A2565C05
keywords LangevindynamicsenhanceddissipationGibbsmeasurealternatingshearflowsamplingoptimizationmixingmass-searching
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

This paper asks whether an external drift can make a diffusion converge to a prescribed Gibbs density $\Pi\propto e^{-U}$ much faster than the classical overdamped Langevin dynamics, whose convergence time scales like $O(\nu^{-1})$ for small noise $\nu$. The authors propose two deterministic transport-diffusion models: a first-order equation with a time-periodic alternating shear drift, and a second-order kinetic equation with a static momentum-dependent drift, and they prove that both converge to $\Pi$ at the enhanced exponential rate $O(\nu^{1/2})$. The engine is enhanced dissipation: the shear flow chops the density into thin filaments, so diffusion acts on fine scales and dissipates fluctuations at a rate proportional to $\sqrt{\nu}$ rather than $\nu$. The paper also supplies a quadrature-free 'mass-searching' dynamics that computes the normalization constant $Z$ at the same enhanced rate, making the schemes numerically implementable without high-dimensional integration. If correct, this gives a deterministic, numerically tractable route to accelerated sampling and optimization.

What carries the argument

The load-bearing object is the alternating shear flow (1.18): the velocity field switches between $u_1=\sum_{j=1}^{d}\sin(y_{d+j})e_j$ and $u_2=\sum_{j=1}^{d}\sin(x_j)e_{d+j}$ on time intervals of length $O(\nu^{-1/2})$, so that each coordinate direction is successively sheared and the passive scalar's fluctuations obey the enhanced dissipation estimate (1.22): $\|\eta(s+t)-\overline{\eta(s+t)}\|_{L^2}\le C_{ED}\|\eta(s)-\overline{\eta(s)}\|_{L^2}e^{-\delta_{ED}\nu^{1/2}t}$. The drift-defect alignment term $Q_{U,V}$, defined by $D=V\cdot\nabla U$ and $Q_{U,V}[\rho]=\rho(\int(D(y)-D(x))\rho(y)\,dy)$, preserves the Gibbs invariant measure while keeping the drift divergence-free, and the mass-searching dynamics (1.29) computes $\|e^{-W}\|_{L^1}$ without quadrature by running a second enhanced-dissipation equation whose solution converges to the desired mass at rate $e^{-\delta_{ED}\nu^{1/2}t}$.

What would settle it

A concrete numerical check: for the passive scalar equation (1.21) on $\mathbb{T}^4$ with the alternating sine-shear flow, measure the time $T_\nu$ at which $\|\eta(t)-\overline{\eta}\|_{L^2}$ first drops below 1% of its initial value for $\nu=10^{-2},10^{-3},10^{-4}$. If $T_\nu$ does not scale like $\nu^{-1/2}$, the central enhanced-dissipation theorem fails. A second check: simulate the first-order enhanced dynamics (1.5) on a double-well potential and verify that the $L^2$ relative error reaches a fixed tolerance in time proportional to $\nu^{-1/2}\log(1/\varepsilon)$.

Watch

Extended reading notes

Core claim

The central claim is Theorem 1.1: for the first-order enhanced dynamics (1.5) with an alternating sine-shear flow on the torus of even dimension, the relative density $\rho(t)/\Pi-1$ decays in $L^2$ with rate $\delta\nu^{1/2}$ for all $0<\nu\le\nu_0$, provided the initial data is smooth and normalized. Theorem 1.2 establishes the analogous $L^1$ and $L^\infty$ convergence for the hydrodynamic density of the second-order kinetic model (1.8), which uses a fixed drift $v(p)=(\sin p_1,\dots,\sin p_d)$ rather than a time-dependent one. Both rates improve the classical Langevin convergence $O(\nu)$ to $O(\nu^{1/2})$ in the small-noise regime. The proof rests on a new higher-dimensional enhanced-dissipation theorem (Theorem 1.3) for alternating shear flows, and the construction yields concrete velocity fields rather than a probabilistic existence statement.

Load-bearing premise

The entire rate claim depends on the higher-dimensional enhanced-dissipation estimate for alternating shear flows, whose proof relies on a spectral inequality quoted from another paper and whose explicit construction is given only for even dimensions.

Editorial extensions

If this is right

  • Sampling from a general smooth Gibbs density on the even-dimensional torus costs $O(\nu^{-1/2})$ time to reach a fixed $L^2$ error, instead of $O(\nu^{-1})$ for classical Langevin dynamics.
  • The normalization constant $Z$ can be computed without quadrature: the mass-searching dynamics (1.29) converges to $\|e^{-W}\|_{L^1}$ at the enhanced rate $O(\nu^{1/2})$.
  • The second-order kinetic model achieves the same $\nu^{1/2}$ rate for the hydrodynamic density in every dimension $d\ge1$, using a static drift rather than time-dependent controls.
  • Any divergence-free flow satisfying the enhanced dissipation condition (1.22) can be substituted for the sine flow, so the result is not tied to a single velocity field.
  • The deterministic construction avoids the rapid random switching required by earlier random-flow samplers, which makes long-time numerical simulation more practical.

Reading between the lines

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

  • The even-dimension restriction looks removable; if the promised odd-dimensional flow construction can be made explicit, the first-order rate would hold on $\mathbb{T}^d$ for all $d\ge2$, and the same argument would also give dimension-robust samplers on odd tori.
  • The D-alignment mechanism is reminiscent of birth-death and 'environmental averaging' terms in collective dynamics; one could test whether replacing the nonlocal alignment by a local or stochastic variant preserves the $\nu^{1/2}$ rate at lower computational cost.
  • The mass-searching dynamics is a standalone deterministic partition-function estimator; a natural test is to compare its variance and cost against bridge sampling and annealed importance sampling on high-dimensional Gaussian mixture targets.
  • Because the rate is only $\nu^{1/2}$ while randomized flows achieve $|\log\nu|^{-1}$, a hybrid that uses deterministic shears for most of the time and rare random kicks might interpolate between numerical simplicity and near-optimal speed.
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 / 5 minor

Summary. This paper proposes two deterministic enhanced-sampling dynamics for Gibbs measures Π∝e^{-U}: a first-order Fokker–Planck-type equation with a time-dependent divergence-free shear drift (1.5), and a second-order kinetic model with a static drift (1.8). In both, a defect-alignment term preserves Π. The main theorems assert convergence to Π at the enhanced rate O(ν^{1/2}) (Theorems 1.1 and 1.2), together with a mixing estimate for observables (1.13). The proofs are reduced to an enhanced-dissipation theorem (Theorem 1.3) for alternating sine-shear flows on T^{2d}, proved in Appendix D. The paper also gives a quadrature-free mass-searching dynamics (Theorem 1.4) and particle systems.

Significance. If the central estimate held, the paper would provide a useful deterministic and numerically tractable alternative to the randomized-flow Langevin sampler of [12], at the modest cost of replacing a |logν|^{-1} rate by ν^{1/2}. The strength of the paper is its explicit, derivation-based proof structure with no fitted parameters, concrete shear flows, and a quadrature-free normalization procedure. The numerical section illustrates the qualitative speedup. However, the proof of the key enhanced-dissipation estimate contains a misstated comparison lemma that currently leaves the main theorems unsupported, so the significance can only be assessed after that gap is repaired.

major comments (3)
  1. [Appendix D, Lemma D.2 and Eq. (D.8)] For the functional (D.3), with A=αφε^{1/2}, B=γφ³ε^{-1/2}, C=βφ², the quadratic form in (∂_{y_m}f, cos(y_m)f) is nonnegative only if 4β²≤αγ, not merely β²≤αγ as stated. The proof of Theorem D.1 sets α=β^{1/2}/4 and γ=4β^{3/2}, giving αγ=β², so the stated hypothesis of Lemma D.2 is violated. Therefore the final inequality (D.8), which invokes Lemma D.2, does not follow from the quoted comparison lemma as written. Since Corollary D.1 and Theorem 1.3 depend on (D.8), the proofs of Theorems 1.1, 1.2, and 1.4 are incomplete at this point. The likely repair is to enlarge γ (e.g., γ=16β^{3/2}) and re-optimize the constants, but this must be carried out explicitly.
  2. [Theorem 1.1, Eq. (1.6)] The theorem states that C† is independent of d, but the proof in §2.2 defines C† := C1∥ĝ(ρ0/Π)∥_{L2} + |T|^{d/2} C2, where |T|^{d/2}=(2π)^{d/2} and the L2 norms depend on d. Thus the asserted d-independence of C† is not supported and, as written, the theorem statement is false. The statement should either explicitly allow d-dependence of C† or redefine C† to absorb the prefactor after proving a d-independent bound.
  3. [Remark 1.6] The paper promises an odd-dimensional extension of the alternating shear construction and says it will be detailed in Section D, but Section D presents only the even-dimensional construction (D.1). This is an unmet promise. Since Theorems 1.1 and 1.3 are stated for even d only, this does not invalidate those theorems, but the remark should be fulfilled or removed.
minor comments (5)
  1. [Title] The title in the text contains 'OPTIMIZA TION' with an inserted space; please correct.
  2. [Section 2.2] The notation ĝ{ρ0 e^U} appears in (2.3) where the fluctuation notation ĝ{\cdot} is introduced immediately before; move the definition of the fluctuation ĝ{F}=F−̅F ahead of (2.1).
  3. [Section 4.2] The numerical experiment uses ν=0.25, which is outside the small-ν regime for which the O(ν^{1/2}) rate is proved; the figure shows a speedup but not the ν-scaling. Please add a remark that the test is illustrative or include a ν-sweep.
  4. [Appendix D, Lemma D.1] Lemma D.1 from [13] is stated with an explicit constant but no proof; because this lemma is load-bearing for the spectral inequality, a proof or precise location in [13] would improve self-containedness.
  5. [Theorem D.2] The notation σ:=1∨{d/2+ς} is introduced inside the statement; define it in the Notation section for consistency.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity: the ν^{1/2} enhanced-sampling rate is derived from an alternating-shear dissipation argument with explicit constants, not from a fitted parameter or self-referential definition; the authors' prior lemmas are used as modular, externally checkable ingredients.

full rationale

I walked the claimed derivation chain. Theorem 1.1 reduces to the nonlinear enhanced-dissipation estimate (2.1) for h=ρ/Π; the bootstrap proof uses the linear passive-scalar estimate (1.22) only as an input on reference solutions η(i), and the constants C†, δ are explicit functions of ρ0, U, d and the universal constants of Theorem 1.3. No parameter is fitted to the target quantity, and the predicted O(ν^{1/2}) rate is not inserted by ansatz. Theorem 1.2 is structured identically, using the linear estimate (D.5) for the renormalized fluctuation, with δ fixed as δ0/5. Theorem 1.3 is proved by reducing to the single-shear hypocoercivity estimate (D.4) and then assembling the alternating cutoff functions Φ1, Φ2; the final contraction argument is a direct time-discretization argument, not an assumption of the conclusion. The supporting lemmas quoted from [13] and [33] (Lemmas D.1–D.4) are general spectral/uncertainty-type and hypocoercivity comparison inequalities with stated assumptions that do not include the target rate ν^{1/2} for the d-dimensional alternating flow; although S. He is a co-author of those works, they are parameter-free and externally checkable, so their citation is legitimate modularization rather than a self-referential proof. Theorem 1.4 is a direct substitution: v:=ωe^{-W} solves the passive scalar equation, so its convergence to the mass M is literally Theorem 1.3 applied to v; this is an application, not a circular prediction. I found no step where a 'prediction' is equivalent by construction to a fitted input or to the defining equations. Proof-completeness flags that are not circularity: Lemma D.2 and Lemma D.3 have proofs omitted with citations to [33]; Lemma D.2 is stated with condition β²≤αγ but the proof of Theorem D.1 later chooses α=β^{1/2}/4 and γ=4β^{3/2}, giving αγ=β², and the claimed 1/2-equivalence for arbitrary ratios of the two quadratic components may require the sharper threshold 4β²≤αγ, so the bound (D.8) does not follow as written from the quoted lemma. Additionally, Remark 1.6 promises an odd-dimensional extension that Section D does not actually provide. These are correctness risks in the proof of Theorem 1.3, not instances of the theorems assuming their own conclusions.

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

The proofs depend on several external results (spectral inequality, 2D shear enhanced dissipation, hypocoercivity lemmas from [33]) and on an unproven local well-posedness assertion. No numerical fitting or ad hoc parameters are used in the theoretical claims.

assumptions (5)
  • standard math The uncertainty-principle spectral inequality (Lemma D.1) relating the L2 norm of a function on T^d to the L2 norm of its derivative and the cos-weighted norm, cited from [13].
    Used in the proof of the hypocoercivity estimate (D.4) in Appendix D to control low frequencies; the proof is not reproduced.
  • domain assumption The 2D shear-enhanced dissipation estimate for shear profile U=sin with finitely many non-degenerate critical points, cited from [6,65].
    Baseline result used to motivate and extend the higher-dimensional alternating shear construction.
  • domain assumption Local well-posedness and positivity (maximum principle) for the nonlinear nonlocal Fokker-Planck equations (1.5) and (1.8), asserted by 'a standard argument' in Sections 2.1 and 3.1.
    The a priori estimates in the bootstrap proofs assume existence of smooth solutions on [0,T*).
  • ad hoc to paper The hypocoercivity lemmas (Lemma D.2 and D.3) from [33] (co-authored by S. He), used without proof in the proof of Theorem D.1.
    These lemmas provide the energy estimates for the hypocoercivity functional; the paper says proofs are 'similar to [33]' and omits them.
  • domain assumption The alternating shear flow construction requires even dimension d=2d; the odd-dimensional extension promised in Remark 1.6 is not detailed in the visible text.
    The main theorems 1.1 and 1.3 are stated for even d; the general claim in the abstract is not fully supported for odd dimensions.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Sampling and Optimization meet Enhanced Flows." pith.science (2026). https://pith.science/paper/EGMAZ642

@misc{pith2026260807329,
  author       = {Pith},
  title        = {Pith review of: Sampling and Optimization meet Enhanced Flows},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/EGMAZ642}},
  note         = {Machine review of arXiv:2608.07329}
}
abstract

It is well known that the computational realization of Gibbs probability measures, $e^{-\mathbb{U}(\mathbf{x})}/Z$, plays a central role in sampling and optimization. In this paper, we introduce two types of dynamics that exhibit rapid convergence towards these Gibbs measures. The mechanism driving this rapid convergence is the enhanced dissipation associated with these transport-diffusion dynamics. Motivated by these enhanced dynamics, we design numerical algorithms for sampling from the target Gibbs measure. Finally, we provide the corresponding particle systems that may yield other effective numerical samplers.

Figures

Figures reproduced from arXiv: 2608.07329 by the authors.

Figure 1
Figure 1. The coefficient functions Φ(i) Remark 1.4. The parameter δ in Theorem 1.1 is closely connected to the parameters in Theorem 1.3: δ := δED 2 + log CED (1.23) < δED. Remark 1.5. We remark that our construction is robust in the sense that replacing the shear profile U by another function with finitely many nondegenerate critical points leaves the result unchanged [PITH_FULL_IMAGE:figures/full_fig_p007_1.png] view at source ↗
Figure 2
Figure 2. The time evolution of density f computed from scheme (4.4). The initial density f0 is taken as (4.6) and the target invariant measure is taken as (4.5). Snapshots of ft are shown at t = 0.02, 0.05, 0.2, 6. the original linear Fokker-Planck equation in the gradient flow form are shown in [PITH_FULL_IMAGE:figures/full_fig_p024_2.png] view at source ↗
Figure 3
Figure 3. The decays of L 1 and relative L 2 errors between the dynamic density and the invariant measure are shown for both the flow-accelerated Fokker-Planck equation (4.1) and the original Fokker-Planck equation in the gradient flow form (4.7). The solid red and black lines represent the decay of L 1 and L 2 errors for (4.7). The dashed red and black lines represent the decay of L 1 and L 2 errors for (4.1) after flow acce… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

66 extracted references · 54 canonical work pages

  1. [33]

    Mixing, Enhanced Dissipation and Phase Transition in the Kinetic Vicsek Model

    M. Gu and S. He. Mixing, enhanced dissipation and phase transition in the kinetic vicsek model. arXiv:2409.19859, 2024

    work page Pith review arXiv 2024

  2. [13]

    Coble and S

    D. Coble and S. He. A note on enhanced dissipation and taylor dispersion of time-dependent shear flows. arXiv:2309.15738, 2023

    arXiv 2023

  3. [12]

    Christie, Y

    A. Christie, Y. Feng, G. Iyer, and A. Novikov. Speeding up langevin dynamics by mixing.Multiscale Modeling & Simulation, 23(4):1696–1743, 2025

    work page 2025

  4. [1]

    Albritton, S

    D. Albritton, S. Armstrong, J.-C. Mourrat, and M. Novack. Variational methods for the kinetic fokker–planck equation.Analysis & PDE, 17(6):1953–2010, 2024

    work page 1953

  5. [2]

    Albritton, R

    D. Albritton, R. Beekie, and M. Novack. Enhanced dissipation and H¨ ormander’s hypoellipticity.J. Funct. Anal., 283(3):Paper No. 109522, 38, 2022

    work page 2022

  6. [3]

    J. M. Altschuler and K. Talwar. Resolving the mixing time of the langevin algorithm to its stationary distribution for log-concave sampling.arXiv preprint arXiv:2210.08448, 2022

    arXiv 2022

  7. [4]

    F. Baudoin. Bakry–´ emery meet villani.Journal of functional analysis, 273(7):2275–2291, 2017

    work page 2017

  8. [5]

    Almost-sure enhanced dissipation and uniform-in-diffusivity exponential mixing for advection-diffusion by stochastic Navier-Stokes

    J. Bedrossian, A. Blumenthal, and S. Punshon-Smith. Almost-sure enhanced dissipation and uniform-in- diffusivity exponential mixing for advection-diffusion by stochastic Navier-Stokes.arXiv:1911.01561

    work page Pith review arXiv 1911

Show all 66 references
  1. [6]

    Bedrossian and M

    J. Bedrossian and M. Coti Zelati. Enhanced dissipation, hypoellipticity, and anomalous small noise inviscid limits in shear flows.Arch. Ration. Mech. Anal., 224(3):1161–1204, 2017

    2017

  2. [7]

    Ben-Porat, J

    I. Ben-Porat, J. A. Carrillo, and P.-E. Jabin. Singular flows with time-varying weights.arXiv preprint arXiv:2503.02276, 2025

    2025 arXiv

  3. [8]

    R. N. Bhattacharya. Criteria for recurrence and existence of invariant measures for multidimensional diffusions. The Annals of Probability, pages 541–553, 1978

    1978

  4. [9]

    Blumenthal, M

    A. Blumenthal, M. Coti Zelati, and R. S. Gvalani. Exponential mixing for random dynamical systems and an example of Pierrehumbert.Ann. Probab., 51(4):1559–1601, 2023

    2023

  5. [10]

    Y. Cao, J. Lu, and L. Wang. On explicitL 2-convergence rate estimate for underdamped Langevin dynamics. Arch. Ration. Mech. Anal., 247(5):Paper No. 90, 34, 2023

    2023

  6. [11]

    Cattiaux, A

    P. Cattiaux, A. Guillin, P. Monmarch´ e, and C. Zhang. Entropic multipliers method for langevin diffusion and weighted log sobolev inequalities.Journal of Functional Analysis, 277(11):108288, 2019

    2019

  7. [14]

    Constantin, A

    P. Constantin, A. Kiselev, L. Ryzhik, and A. Zlatoˇ s. Diffusion and mixing in fluid flow.Ann. of Math. (2), 168(2):643–674, 2008

    2008

  8. [15]

    Coti Zelati, G

    M. Coti Zelati, G. Crippa, G. Iyer, and A. L. Mazzucato. Mixing in incompressible flows: transport, dissipation, and their interplay.Notice of AMS, 71(5), 2024

    2024

  9. [16]

    Coti Zelati, M

    M. Coti Zelati, M. G. Delgadino, and T. M. Elgindi. On the relation between enhanced dissipation timescales and mixing rates.Comm. Pure Appl. Math., 73(6):1205–1244, 2020

    2020

  10. [17]

    Coti Zelati, H

    M. Coti Zelati, H. Dietert, and D. G´ erard-Varet. Orientation mixing in active suspensions.Ann. PDE, 9(20), 2023

    2023

  11. [18]

    Coti-Zelati and T

    M. Coti-Zelati and T. D. Drivas. A stochastic approach to enhanced diffusion.arXiv:1911.09995v1, 2019

    1911 arXiv

  12. [19]

    Cucker and S

    F. Cucker and S. Smale. On the mathematics of emergence.Japanese Journal of Mathematics, 2(1):197–227, 2007

    2007

  13. [20]

    Dolbeault, C

    J. Dolbeault, C. Mouhot, and C. Schmeiser. Hypocoercivity for linear kinetic equations conserving mass.Trans- actions of the American Mathematical Society, 367(6):3807–3828, 2015. SAMPLING AND OPTIMIZATION MEET ENHANCED FLOWS 35

    2015

  14. [21]

    T. M. Elgindi, K. Liss, and J. C. Mattingly. Optimal enhanced dissipation and mixing for a time-periodic, lipschitz velocity field onT 2.arXiv preprint arXiv:2304.05374, 2023

    2023 arXiv

  15. [22]

    Feng and G

    Y. Feng and G. Iyer. Dissipation enhancement by mixing.Nonlinearity, 32(5):1810–1851, 2019

    2019

  16. [23]

    Y. Feng, A. L. Mazzucato, and C. Nobili. Enhanced dissipation by circularly symmetric and parallel pipe flows. Physica D: Nonlinear Phenomena, 445:133640, 2023

    2023

  17. [24]

    Frouvelle and J.-G

    A. Frouvelle and J.-G. Liu. Dynamics in a kinetic model of oriented particles with phase transition.SIAM Journal on Mathematical Analysis, 44(2):791–826, 2012

    2012

  18. [25]

    Gallay and M

    T. Gallay and M. C. Zelati. Enhanced dissipation and taylor dispersion in higher-dimensional parallel shear flows.arXiv preprint arXiv:2108.11192, 2021

    2021 arXiv

  19. [26]

    Gao, J.-G

    Y. Gao, J.-G. Liu, and N. Wu. Data-driven efficient solvers for langevin dynamics on manifold in high dimensions. Applied and Computational Harmonic Analysis, 62:261–309, 2023

    2023

  20. [27]

    Gardner, K

    V. Gardner, K. L. Liss, and J. C. Mattingly. A pathwise approach to the enhanced dissipation of passive scalars advected by shear flows.arXiv preprint arXiv:2410.05657, 2024

    2024 arXiv

  21. [28]

    Gelman and X.-L

    A. Gelman and X.-L. Meng. Simulating normalizing constants: From importance sampling to bridge sampling to path sampling.Statistical Science, 13(2):163–185, 1998

    1998

  22. [29]

    Girolami and B

    M. Girolami and B. Calderhead. Riemann manifold langevin and hamiltonian monte carlo methods.Journal of the Royal Statistical Society Series B: Statistical Methodology, 73(2):123–214, 2011

    2011

  23. [30]

    Griffin, S

    C. Griffin, S. Rajtmajer, A. Squicciarini, and A. Belmonte. Consensus and information cascades in game-theoretic imitation dynamics with static and dynamic network topologies.SIAM Journal on Applied Dynamical Systems, 18(2):597–628, 2019

    2019

  24. [31]

    Grindrod

    P. Grindrod. Models of individual aggregation or clustering in single and multi-species communities.J. Math. Biol. 26:651-660, 1988

    1988

  25. [32]

    Grothaus and P

    M. Grothaus and P. Stilgenbauer. Geometric langevin equations on submanifolds and applications to the sto- chastic melt-spinning process of nonwovens and biology.Stochastics and Dynamics, 13(04):1350001, 2013

    2013

  26. [34]

    Ha and E

    S.-Y. Ha and E. Tadmor. From particle to kinetic and hydrodynamic descriptions of flocking.Kinetic and Related Models, 1(3):415–435, 2008

    2008

  27. [35]

    S. Haber. Numerical evaluation of multiple integrals.SIAM Review, 12(4):481–526, 1970

    1970

  28. [36]

    He and A

    S. He and A. Kiselev. Stirring speeds up chemical reaction.Nonlinearity, 35(8):4599, 2022

    2022

  29. [37]

    He and E

    S. He and E. Tadmor. A game of alignment: collective behavior of multi-species.Ann. Inst. H. Poincar´ e C Anal. Non Lin´ eaire, 38(4):1031–1053, 2021

    2021

  30. [38]

    H´ erau and F

    F. H´ erau and F. Nier. Isotropic hypoellipticity and trend to equilibrium for the Fokker-Planck equation with a high-degree potential.Arch. Ration. Mech. Anal., 171(2):151–218, 2004

    2004

  31. [39]

    H¨ ormander

    L. H¨ ormander. Hypoelliptic second order differential equations.Acta Math., 119:147–171, 1967

    1967

  32. [40]

    L. Kelvin. Stability of fluid motion-rectilinear motion of viscous fluid between two parallel plates.Phil. Mag., (24):188, 1887

  33. [41]

    Kolmogoroff

    A. Kolmogoroff. Zuf¨ allige Bewegungen (zur Theorie der Brownschen Bewegung).Ann. of Math. (2), 35(1):116– 117, 1934

    1934

  34. [42]

    Y. Lu, J. Lu, and J. Nolen. Accelerating langevin sampling with birth-death.arXiv preprint arXiv:1905.09863, 2019

    1905 arXiv

  35. [43]

    J. C. Mattingly, A. M. Stuart, and D. J. Higham. Ergodicity for sdes and approximations: locally lipschitz vector fields and degenerate noise.Stochastic processes and their applications, 101(2):185–232, 2002

    2002

  36. [44]

    Motsch and E

    S. Motsch and E. Tadmor. A new model for self-organized dynamics and its flocking behavior.Journal of Statistical Physics 144(5), 2011

    2011

  37. [45]

    Motsch and E

    S. Motsch and E. Tadmor. Heterophilious dynamics enhances consensus.SIAM Review 56(4), 2014

    2014

  38. [46]

    R. M. Neal. Annealed importance sampling.Statistics and Computing, 11(2):125–139, 2001

    2001

  39. [47]

    Nier and B

    F. Nier and B. Helffer.Hypoelliptic estimates and spectral theory for Fokker-Planck operators and Witten Lapla- cians. Springer, 2005

    2005

  40. [48]

    E. Novak. Some results on the complexity of numerical integration. In R. Cools and D. Nuyens, editors,Monte Carlo and Quasi-Monte Carlo Methods: MCQMC, Leuven, Belgium, April 2014, volume 163 ofSpringer Pro- ceedings in Mathematics & Statistics, pages 161–183. Springer, Cham, 2016

    2014

  41. [49]

    G. Parisi. Correlation functions and computer simulations.Nuclear Physics B, 180(3):378–384, 1981

    1981

  42. [50]

    G. A. Pavliotis. Stochastic processes and applications.Texts in applied mathematics, 60, 2014

    2014

  43. [51]

    Pierrehumbert

    R. Pierrehumbert. Tracer microstructure in the large-eddy dominated regime.Chaos, Solitons&Fractals, 4(6):1091–1110, 1994. Special Issue: Chaos Applied to Fluid Mixing

    1994

  44. [52]

    Pouradier Duteil

    N. Pouradier Duteil. Mean-field limit of collective dynamics with time-varying weights.Netw. Heterog. Media, 17(2):129–161, 2022

    2022

  45. [53]

    C. P. Robert and G. Casella.Monte Carlo Statistical Methods. Springer, New York, second edition, 2004. 36 YUAN GAO, SIMING HE, AND EITAN TADMOR

    2004

  46. [54]

    G. O. Roberts and R. L. Tweedie. Exponential convergence of langevin distributions and their discrete approxi- mations.Bernoulli, 2(4):341–363, 1996

    1996

  47. [55]

    Shvydkoy

    R. Shvydkoy. Global hypocoercivity of kinetic Fokker-Planck-alignment equations.Kinet. Relat. Models, 15(2):213–237, 2022

    2022

  48. [56]

    Shvydkoy

    R. Shvydkoy. Environmental averaging.EMS Surv. Math. Sci., 11(2):277–413, 2024

    2024

  49. [57]

    Shvydkoy et al.Dynamics and analysis of alignment models of collective behavior

    R. Shvydkoy et al.Dynamics and analysis of alignment models of collective behavior. Springer, 2021

    2021

  50. [58]

    Shvydkoy and E

    R. Shvydkoy and E. Tadmor. Eulerian dynamics with a commutator forcing.Transactions of Mathematics and its Applications, 1(1):1–26, 2017

    2017

  51. [59]

    I. H. Sloan and S. Joe.Lattice Methods for Multiple Integration. Clarendon Press, Oxford University Press, Oxford, 1994

    1994

  52. [60]

    I. H. Sloan and H. Wo´ zniakowski. An intractability result for multiple integration.Mathematics of Computation, 66(219):1119–1124, July 1997

    1997

  53. [61]

    D. Talay. Stochastic hamiltonian systems: exponential convergence to the invariant measure, and discretization by the implicit euler scheme.Markov Process. Related Fields, 8(2):163–198, 2002

    2002

  54. [62]

    Tan and J

    L. Tan and J. Lu. Accelerate langevin sampling with birth-death process and exploration component.arXiv preprint arXiv:2305.05529, 2023

    2023 arXiv

  55. [63]

    N. G. Trillos and D. Sanz-Alonso. From optimization to sampling through gradient flows.arXiv preprint arXiv:2302.10037, 2023

    2023 arXiv

  56. [64]

    Vicsek, A

    T. Vicsek, A. Czir´ ok, E. Ben-Jacob, I. Cohen, and O. Shochet. Novel type of phase transition in a system of self-driven particles.Phys. Rev. Lett., 75(6):1226–1229, 1995

    1995

  57. [65]

    D. Wei. Diffusion and mixing in fluid flow via the resolvent estimate.Science China Mathematics, pages 1–12, 2019

    2019

  58. [66]

    L. Wu. Large and moderate deviations and exponential convergence for stochastic damping hamiltonian systems. Stochastic processes and their applications, 91(2):205–238, 2001. Department of Mathematics, Purdue University, West Lafayette, IN, USA Email address:gao662@purdue.edu ...

    2001

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.