Pith. sign in

REVIEW 3 major objections 5 minor 2 cited by

Space-Time Log-Sobolev Inequality and Hypocoercive Hypercontractivity for Underdamped Langevin Dynamics

T0 review · 3 major / 5 minor · reviewed 2026-08-02 · deepseek-v4-flash

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

Pith's one-line read For underdamped Langevin dynamics with a convex potential whose spatial Gibbs marginal satisfies log-Sobolev with constant ρ and friction γ=Γ√ρ, the semigroup boosts L^p integrability exponentially on the kinetic time scale t∼ρ^{-1/2}, enfo

desk verdict Genuinely new entropic space-time LSI that yields kinetic-scale hypercontractivity and √ρ Rényi decay — but the load-bearing analytic inputs are delegated to an unpublished companion preprint, so the headline theorem is conditional until that companion is verified. read the letter →

arxiv 2605.25083 v2 pith:5K7FF466 submitted 2026-05-24 math.AP math.FAmath.PR

Bowen Li , Jianfeng Lu This is my paper · ORCID
classification math.APmath.FAmath.PR MSC 35Q8447D0760J60
keywords underdampedLangevindynamicshypocoercivityhypercontractivityspace-timelogarithmicSobolevinequalityRényidivergencekineticFokker–Planckequationoptimaltransportlog-concavesampling
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 tries to prove that underdamped Langevin dynamics—where noise only touches velocities—still enjoys classical hypercontractivity, despite having no instantaneous log-Sobolev inequality in phase space. It claims that when the position marginal has log-Sobolev constant ρ and friction scales as Γ√ρ, the semigroup maps L^p to L^q with q(t)-1 growing like e^{λ√ρ t}. That means Hamiltonian transport plus mild velocity noise regularizes the whole system on the fast ballistic time scale t∼ρ^{-1/2}. As a corollary, all Rényi divergences decay exponentially at the sharp hypocoercive rate √ρ. The engine is a space-time log-Sobolev inequality that transfers velocity dissipation into position information, plus a forward/backward interpolation using the adjoint semigroup.

What carries the argument

The space-time log-Sobolev inequality (Theorem 3.4): for a path of densities g_t on [0,T], Ent_{time-averaged}(g) ≤ C[averaged velocity Fisher information + (1/ρ) average of ||(∂_t+L_a)g_t||^2_{-1,g_t}]. The negative norm is a velocity-fibre Wasserstein tangent norm; it measures how much the path fails to be a solution, while the velocity Fisher information measures dissipation. The inequality is proved from a controlled version of hypocoercive entropy decay using a modified entropy with an optimal-transport corrector coupling position and velocity, and the hypercontractivity step follows by applying it to a forward/backward interpolation path G_s=φ_s ψ_s/Z with φ_s=P_s φ, ψ_s=P^*_{T-s} ψ, s

What would settle it

Find a convex confining potential with spatial LSI and tame growth for which the forward/backward interpolation path G_s fails the assumed regularity—for example, the centered velocity-stress tensor is not C^1 or the spatial-cutoff limits do not hold—and exhibit a density path on which the space-time log-Sobolev inequality (3.25) is violated. A simpler numerical check: compute the Rényi-decay rate for a strongly convex potential and compare with the predicted λ_{Γ,τ}√ρ; a rate slower than √ρ would falsify the hypocoercive claim.

Watch

Extended reading notes

Core claim

The central claim is a kinetic analogue of the classical equivalence between log-Sobolev inequalities and hypercontractivity. For underdamped Langevin dynamics, assume the spatial marginal of the invariant Gibbs measure has LSI constant ρ and the potential satisfies tame confining growth. If γ=Γ√ρ, then for every p>1, ||P_t f||_{L^{q(t)}} ≤ ||f||_{L^p} with q(t)-1=α^{⌊√ρ t/τ⌋}(p-1) and constants independent of dimension and ρ. The admissible integrability exponent therefore grows exponentially at rate λ√ρ on the ballistic time scale, matching the sharp hypocoercive entropy-decay rate. The proof route is a space-time LSI for finite-action kinetic paths, a duality argument built on a non-rever

Load-bearing premise

The argument assumes that the optimal-transport displacement and velocity-stress estimates imported from the companion entropy-decay work remain valid for the forward/backward interpolation path; the paper inherits those regularity and cutoff conditions instead of verifying them, and if they fail, the space-time inequality and the hypercontractivity result collapse.

Editorial extensions

If this is right

  • Integrability improves on the kinetic time scale: q(t)-1 grows like exp(λ_{Γ,τ}√ρ t), so the gain is ballistic (ρ^{-1/2}) rather than diffusive (ρ^{-1}).
  • For every q>1 and every initial law, Rényi divergence decays as R_q(ν_t|μ) ≤ q α_{Γ,τ} e^{-λ_{Γ,τ}√ρ t} R_q(ν_0|μ), with the same sharp hypocoercive rate as entropy decay.
  • All constants α, λ depend only on the dimensionless friction Γ and the chosen period τ, not on dimension d or ρ.
  • The one-step estimate iterates cleanly: P_{nT} maps L^p into L^{1+α^n(p-1)}, and the full-time statement follows by Markov contraction.
  • Optimizing in friction selects Γ≈1 (γ≈√ρ); at extreme friction the Rényi decay rate degrades cubically in max{Γ,Γ^{-1}}.

Reading between the lines

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

  • If the theory holds, sampling algorithms built on underdamped Langevin dynamics can expect Rényi warm-start bounds at the faster √ρ rate rather than the diffusive ρ rate, potentially improving complexity estimates; the paper flags this as future work.
  • The same space-time-LSI plus non-reversible interpolation scheme may extend to other degenerate kinetic equations admitting a controlled entropy decay, such as kinetic Fokker–Planck equations with velocity-dependent friction or non-convex potentials with local log-Sobolev structure.
  • An explicit test on the harmonic oscillator, where the spectrum is known, could reveal how far the constants α and λ are from optimal; the constants here are explicit but likely not sharp.
  • The dimension-free constants suggest that concentration or tail inequalities for underdamped Langevin paths could be derived from this hypercontractivity, by analogy with reversible diffusions.
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. The paper proves a space-time logarithmic Sobolev inequality for the underdamped Langevin dynamics with a convex confining potential whose spatial marginal satisfies an LSI with constant rho, under the additional tame Hérau–Nier hypotheses imported from the same-author preprint [28]. Using this inequality, the authors establish a Gross-type hypercontractivity estimate: for friction gamma = Gamma sqrt(rho) and T = tau / sqrt(rho), the semigroup satisfies ||P_t f||_{L^q} <= ||f||_{L^p} for q <= 1 + alpha_{Gamma,tau}^{floor(sqrt(rho) t / tau)} (p-1), so the admissible integrability exponent grows exponentially on the kinetic time scale t ~ rho^{-1/2}. The proof combines a controlled entropy-corrector differential inequality with a duality argument based on a forward/backward interpolation of the non-reversible semigroup. As a corollary, exponential decay of Rényi divergences at the sharp hypocoercive rate O(sqrt(rho)) is derived.

Significance. If the technical gaps identified below are closed, this would be a significant contribution: it provides the first Gross-type hypercontractivity mechanism for the low-friction underdamped Langevin dynamics, with the correct kinetic time scale, and it introduces a space-time entropic inequality of independent interest. The central algebraic structure is largely original and carefully executed: the corrector computation in Lemma 3.2, the constant bookkeeping in Proposition 3.3, the endpoint estimate in Theorem 3.4, the cancellation identities in Lemmas 4.1 and 4.3, and the Riesz–Thorin interpolation in Theorem 5.2 all check out. The paper is also honest about what is imported from [28]. However, the current manuscript delegates one load-bearing regularity step to a citation rather than a proof, and does not verify the hypotheses of that step for the interpolation path used in Section 4.

major comments (3)
  1. [Section 3, Lemma 3.6] Lemma 3.6 is the only bridge from the regular-path inequality (3.25) to the non-regular data used in the rest of the paper. Its 'proof' is a citation to [28, Definition 2.5 and Section 7], asserting convergence of entropy, Fisher information, and action while preserving the controlled equation. But the regular pairs produced by the asserted regularization must satisfy Definition 2.2(A1)-(A3), the characteristics condition (R3), and the C^1 stress-tensor plus boundary-limit hypotheses of Lemma 2.5. None of these are checked in the present manuscript. The sentence that the Brenier maps, moment equations, stress pairings, and lower-semicontinuity inputs are 'precisely those of [28]' is an assertion, not a proof. Since Theorem 3.4 and hence Proposition 4.2 rest on this lemma, this is a load-bearing gap. The authors should either prove Lemma 3.6 in full or avoid it by direct regularity estima
  2. [Section 4, Proposition 4.2] The space-time LSI (Theorem 3.4) is applied to the forward/backward interpolation density G_s defined in (4.1). Lemma 4.1 verifies the velocity continuity equation (4.2) and the Fisher/action bounds (4.4), but Theorem 3.4 requires the path to satisfy the regularity assumptions of Lemma 3.2, in particular that (q_s, j_s) = (Pi_v G_s, Pi_v(v G_s)) is a regular density-current pair in the sense of Definition 2.2, including the cutoff condition (A3) and the characteristics condition (A2), and that the centered stress tensor Theta_{G_s} is C^1_loc with the boundary limits required by Lemma 2.5. None of these are verified for G_s = phi_s psi_s / Z. This is not a cosmetic issue: Lemma 3.2 and Proposition 3.3 use the stress estimate (2.11) with exactly those hypotheses. Thus the application of the central theorem in Proposition 4.2 is not justified as written.
  3. [Section 2 and references to [28]] The paper adopts 'the standing hypotheses of [28, Assumptions 2.1 and 2.2]' and imports several lemmas from [28] (Lemmas 2.1, 2.4, 2.5 and the regularization scheme of Lemma 3.6) as black boxes. The tame Hérau–Nier bounds and the precise regularity classes are not reproduced. Since [28] is a same-author arXiv preprint appearing concurrently, this dependency is hard for a reader to verify. Please state explicitly all hypotheses and either prove or precisely state the imported results, or provide a version of [28] that is publicly available and stable. This is especially important because Lemma 3.6's correctness is asserted to be 'precisely those of [28]' without a proof.
minor comments (5)
  1. [Throughout] The symbol sqrt(rho) is frequently rendered as 'p rho' (e.g., in the abstract and introduction). Please fix the typography in the final version.
  2. [Theorem 3.4 proof] The proof selects a measurable velocity field A_t with the near-minimal action bound (3.27). The measurable selection step is not addressed. A standard measurable selection or a careful regularization argument should be included.
  3. [Section 4] The hypotheses on phi and psi needed to ensure E_T(phi,psi) < infinity should be stated. The proof implicitly assumes finiteness of the two-sided Fisher information; if E_T = infinity the bounds are vacuous but the interpolation argument should still be justified for the approximating sequence.
  4. [Definition 2.2(A3)] The condition 'locally uniformly for t' is not given a precise topology. Please specify the mode of convergence in t (e.g., uniform on compact subsets of (0,T)).
  5. [Corollary 5.5] In part (ii), the monotonicity of Rényi divergence in the order is used but not stated. Adding one sentence would make the argument self-contained.

Circularity Check

1 steps flagged · score 4.0 of 10

Hypercontractivity result is genuinely derived, but the space-time LSI's limiting step is delegated by self-citation to the same-author preprint [28], making the core proof partially load-bearing on unverified imported regularity.

  1. self citation load bearing [Lemma 3.6 (proof), Section 3]
    "This is the regularization step from [28, Definition 2.5 and Section 7], applied to the linear finite-action continuity equation with velocity current. ... the Brenier maps, moment equations, stress pairings, and lower-semicontinuity inputs invoked are precisely those of [28, Definition 2.5 and Sections 4–7], so no additional Brenier regularity argument is needed beyond the one already carried out there."

    Theorem 3.4 is the single engine of the paper, and its proof covers only regular controlled paths. The passage from those regular pairs to the smooth positive finite-action densities actually used for the Section 4 interpolation G_s is Lemma 3.6, whose proof is not carried out: it invokes [28, Definition 2.5 and Sections 4–7] and asserts convergence of entropy, Fisher information, and action by the cited regularization. The paper never checks (A1)–(A3) or stress-tensor regularity for G_s. Thus the central space-time LSI reduces, at its limiting step, to a same-author preprint rather than to a proof; this is load-bearing self-citation. It is not full circularity because the final hypercontractivity estimate is not a restatement of [28]'s entropy decay and is obtained by a new duality/interp

full rationale

The main derivation chain (controlled entropy inequality -> space-time LSI -> forward/backward interpolation -> Gross-type hypercontractivity -> Rényi decay) is not equivalent to its inputs. The target theorem is a new Lp-to-Lq estimate at the kinetic time scale; it is not present in [28], and the equations in Section 5 do not merely re-derive [28]'s entropy decay. The constants alpha_T, eta_T, lambda are computed explicitly from CST and tau, not fitted. However, the proof of Theorem 3.4 has an unproved limiting step (Lemma 3.6) delegated wholly to the same-author preprint [28], and the Section 4 path's satisfaction of (A1)-(A3) is asserted rather than verified. This is a load-bearing self-citation and an omitted proof, which I weigh in the score as partial self-citation dependence, but not as construction-level circularity. If [28]'s regularization and stress estimates are taken as valid premises, the derivation chain is self-contained. Thus score 4 rather than 6.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The paper's net new contribution is the entropic space-time LSI and its use in a non-reversible duality argument; everything else — convexity + LSI + Hérau–Nier regularity, the Brenier/OT corrector estimates, and the regularization scheme — is imported from [28] (same author, contemporaneous preprint). No parameters are fitted to data. No new physical entities are postulated; the new mathematical objects (space-time LSI, the velocity negative norm (3.1), the modified entropy H_θ) are proof constructions, not independent postulates.

free parameters (3)
  • Γ (friction scaling, γ = Γ√ρ)
    Theorem parameter controlling the friction regime; not fitted to data. All constants depend on M = max{Γ,Γ^{-1}}; Remark 5.6 shows the rate degrades cubically at extreme friction, so Γ≈1 is the regime where the estimates are sharpest.
  • θ = (1/8)min{Γ,Γ^{-1}} = 1/(8M)
    Hand-chosen weight in the modified entropy H_θ (3.14), selected so that the coefficient inequality (3.23) holds. It is a proof-design constant, not an empirical fit.
  • τ (time-window parameter, T = τρ^{-1/2}) = ≥ τ_ST(Γ) = 32M+4
    Chosen large enough to make the coefficient (θ/2)√ρ − (1+θ)/T positive in Theorem 3.4. The final Rényi rate λ_{Γ,τ} = (log α_{Γ,τ})/τ depends on this choice; any τ above threshold is admissible.
assumptions (5)
  • domain assumption U ∈ C^∞(R^d) is convex and confining; the spatial marginal μ_x satisfies an LSI with constant ρ, (2.1).
    Standing hypothesis of the paper (Section 2). The abstract's 'convex confining potential' glosses over the additional requirements listed below.
  • domain assumption Tame Hérau–Nier confining bounds [23] hold, inherited via [28, Assumptions 2.1–2.2].
    Not stated in the paper; invoked at Section 2 (p.4) and used 'only through the regularization and semigroup-smoothing arguments of [28, Lemmas 5.3 and 7.1]', i.e., in Lemma 3.6 and the stress estimates.
  • domain assumption Validity of the imported corrector machinery of [28]: Lemma 2.1 ([28, Lemmas 3.1–3.2]), Lemma 2.4 ([28, Lemma 4.1], Wasserstein acceleration), Lemma 2.5 ([28, Lemma 5.3], Brenier stress).
    These carry the hard analysis (optimal transport, second-order Wasserstein expansion, cutoff limits). They are by the same author and are neither proved nor fully stated here.
  • domain assumption Regular density-current regularity conditions (A1)–(A3) plus [28, Definition 2.5 (R3)] are satisfiable by the paths considered, after regularization.
    Definition 2.2 and Lemma 3.6. The forward/backward interpolation path G_s of Section 4 is not explicitly verified against these conditions in the text.
  • standard math Standard background: Brenier's theorem, Talagrand's inequality following from LSI, Gross's Gaussian LSI, Riesz–Thorin interpolation, L^p Markov contraction of P_t.
    Used throughout, e.g., (2.6), (2.4), and Theorem 5.2. Standard and uncontroversial.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Space-Time Log-Sobolev Inequality and Hypocoercive Hypercontractivity for Underdamped Langevin Dynamics." pith.science (2026). https://pith.science/paper/5K7FF466

@misc{pith2026260525083,
  author       = {Pith},
  title        = {Pith review of: Space-Time Log-Sobolev Inequality and Hypocoercive Hypercontractivity for Underdamped Langevin Dynamics},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5K7FF466}},
  note         = {Machine review of arXiv:2605.25083}
}
abstract

We study hypercontractivity for underdamped Langevin dynamics with a convex confining potential whose spatial Gibbs marginal satisfies a logarithmic Sobolev inequality (LSI). Unlike in the overdamped case, the noise acts only on the velocity variable, so the usual LSI-based argument does not apply. Nevertheless, under an additional tame Hessian-growth assumption on the potential, we prove that, when the spatial LSI constant is $\rho$ and the friction parameter is of order $\sqrt{\rho}$, the semigroup satisfies a Gross-type $L^p$-to-$L^q$ estimate in which the integrability exponent grows exponentially on the kinetic time scale $t \sim \rho^{-1/2}$. The central ingredient, of independent interest, is a space-time LSI, valid for controlled kinetic paths of finite action, which quantifies how dissipation in the velocity variable is transferred to the position variable. We derive it from a controlled version of the hypocoercive entropy-decay estimate, and we then restore the Gross hypercontractivity mechanism through a duality argument built on a non-reversible forward/backward interpolation of the underdamped Langevin semigroup. As a corollary, R\'{e}nyi divergences of every order decay exponentially at the sharp hypocoercive rate $\mathcal{O}(\sqrt{\rho})$.

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

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

  1. Log-Sobolev inequalities for boundary-driven anharmonic chains

    math-ph 2026-07 conditional novelty 7.0 of 10

    A weakly anharmonic boundary-driven oscillator chain obeys a dimension-free full-gradient logarithmic Sobolev inequality and an O(N^3)-time boundary space-time logarithmic Sobolev inequality.

  2. Windowed thinning and query complexity for the bouncy particle and Zigzag samplers

    math.NA 2026-07 accept novelty 6.0 of 10

    Windowed thinning yields exact BPS and Zigzag simulation with cold-start query complexities O(κ^{1/2}d(d log κ + log 1/ε)) and O(κ d^{1/4}(d log κ + log 1/ε)) full-gradient equivalents.

Reference graph

Works this paper leans on

38 extracted references · 2 linked inside Pith · cited by 2 Pith papers

  1. [28]

    Jianfeng Lu,A sharp hypocoercive entropy decay estimate for underdamped Langevin dynamics, arXiv:2605.01933 (2026)

    arXiv 2026

  2. [1]

    6, 1953–2010

    Dallas Albritton, Scott Armstrong, Jean-Christophe Mourrat, and Matthew Novack,Variational methods for the kinetic Fokker–Planck equation, Analysis & PDE17(2024), no. 6, 1953–2010

    2024

  3. [2]

    Altschuler and Sinho Chewi,Faster high-accuracy log-concave sampling via algorithmic warm starts, Journal of the ACM71(2024), no

    Jason M. Altschuler and Sinho Chewi,Faster high-accuracy log-concave sampling via algorithmic warm starts, Journal of the ACM71(2024), no. 3, 24:1–24:55

    2024

  4. [3]

    Altschuler, Sinho Chewi, and Matthew S

    Jason M. Altschuler, Sinho Chewi, and Matthew S. Zhang,Shifted composition IV: Toward ballistic acceleration for log-concave sampling, arXiv:2506.23062 (2025)

    arXiv 2025

  5. [4]

    Luigi Ambrosio, Nicola Gigli, and Giuseppe Savaré,Gradient flows: in metric spaces and in the space of probability measures, Springer, 2005

    2005

  6. [5]

    Dominique Bakry and Michel Émery,Diffusions hypercontractives, Séminaire de probabilités, XIX, 1983/84, 1985, pp. 177–206. 26 BOWEN LI AND JIANFENG LU

    1983

  7. [6]

    348, Springer, Cham, 2014

    Dominique Bakry, Ivan Gentil, and Michel Ledoux,Analysis and geometry of Markov diffusion operators, Grundlehren der Mathematischen Wissenschaften, vol. 348, Springer, Cham, 2014

    2014

  8. [7]

    3, 993–1067

    Franck Barthe, Patrick Cattiaux, and Cyril Roberto,Interpolated inequalities between exponential and Gaussian, Orlicz hypercontractivity and isoperimetry, Revista Matemática Iberoamericana22(2006), no. 3, 993–1067

    2006

Show all 38 references
  1. [8]

    7, 2275–2291

    Fabrice Baudoin,Bakry–émery meets Villani, Journal of Functional Analysis273(2017), no. 7, 2275–2291

    2017

  2. [9]

    Étienne Bernard, Max Fathi, Antoine Levitt, and Gabriel Stoltz,Hypocoercivity with Schur complements, Annales Henri Lebesgue5(2022), 523–557

    2022

  3. [10]

    4, 375–417

    Yann Brenier,Polar factorization and monotone rearrangement of vector-valued functions, Communications on Pure and Applied Mathematics44(1991), no. 4, 375–417

    1991

  4. [11]

    4, 3587–3622

    Giovanni Brigati and Gabriel Stoltz,How to construct explicit decay rates for kinetic Fokker–Planck equations?, SIAM Journal on Mathematical Analysis57(2025), no. 4, 3587–3622

    2025

  5. [12]

    5, 1172–1184

    Yu Cao, Jianfeng Lu, and Yulong Lu,Exponential decay of Rényi divergence under Fokker–Planck equations, Journal of Statistical Physics176(2019), no. 5, 1172–1184

    2019

  6. [13]

    Yu Cao, Jianfeng Lu, and Lihan Wang,On explicit L2-convergence rate estimate for underdamped Langevin dynamics, Archive for Rational Mechanics and Analysis247(2023), no. 5, 90

    2023

  7. [14]

    Yuansi Chen and Khashayar Gatmiry,When does Metropolized Hamiltonian Monte Carlo provably outperform Metropolis-adjusted Langevin algorithm?, arXiv:2304.04724 (2023)

    2023

  8. [15]

    Laurent Desvillettes and Cédric Villani,On the trend to global equilibrium in spatially inhomogeneous entropy- dissipating systems: The linear Fokker–Planck equation, Communications on Pure and Applied Mathematics54 (2001), no. 1, 1–42

    2001

  9. [16]

    6, 3807–3828

    Jean Dolbeault, Clément Mouhot, and Christian Schmeiser,Hypocoercivity for linear kinetic equations conserving mass, Transactions of the American Mathematical Society367(2015), no. 6, 3807–3828

    2015

  10. [17]

    Zexi Fan, Bowen Li, and Jianfeng Lu,Sharp hypocoercive convergence estimates for underdamped Langevin dynamics via the modified L2 method, arXiv:2604.10068 (2026)

    2026 arXiv

  11. [18]

    Vasseur,Harnack inequality for kinetic Fokker– Planck equations with rough coefficients and application to the Landau equation, Annali della Scuola Normale Superiore di Pisa

    François Golse, Cyril Imbert, Clément Mouhot, and Alexis F . Vasseur,Harnack inequality for kinetic Fokker– Planck equations with rough coefficients and application to the Landau equation, Annali della Scuola Normale Superiore di Pisa. Classe di Scienze. Serie V19(2019), no. 1...

    2019

  12. [19]

    Leonard Gross,Logarithmic Sobolev inequalities, American Journal of Mathematics97(1975), 1061–1083

    1975

  13. [20]

    Alice Guionnet and Bogusław Zegarli ´ nski,Lectures on logarithmic Sobolev inequalities, Séminaire de probabilités xxxvi, 2004, pp. 1–134

    2004

  14. [21]

    1862, Springer-Verlag, Berlin, 2005

    Bernard Helffer and Francis Nier,Hypoelliptic estimates and spectral theory for Fokker–Planck operators and Witten Laplacians, Lecture Notes in Mathematics, vol. 1862, Springer-Verlag, Berlin, 2005

    2005

  15. [22]

    1, 95–118

    Frédéric Hérau,Short and long time behavior of the Fokker–Planck equation in a confining potential and applications, Journal of Functional Analysis244(2007), no. 1, 95–118

    2007

  16. [23]

    2, 151–218

    Frédéric Hérau and Francis Nier,Isotropic hypoellipticity and trend to equilibrium for the Fokker–Planck equation with a high-degree potential, Archive for Rational Mechanics and Analysis171(2004), no. 2, 151–218

    2004

  17. [24]

    Lars Hörmander,Hypoelliptic second order differential equations, Acta Mathematica119(1967), 147–171

    1967

  18. [25]

    Richard Jordan, David Kinderlehrer, and Felix Otto,The variational formulation of the Fokker–Planck equation, SIAM Journal on Mathematical Analysis29(1998), no. 1, 1–17

    1998

  19. [26]

    1, 116–117

    Andrey Kolmogoroff,Zufällige Bewegungen (Zur Theorie der Brownschen Bewegung), Annals of Mathematics35 (1934), no. 1, 116–117

    1934

  20. [27]

    Michel Ledoux,Concentration of measure and logarithmic Sobolev inequalities, Seminaire de probabilites xxxiii, 2006, pp. 120–216

    2006

  21. [29]

    SPACE-TIME LSI AND HYPOCOERCIVE HYPERCONTRACTIVITY FOR UNDERDAMPED LANGEVIN DYNAMICS 27

    Edward Nelson,The free Markoff field, Journal of Functional Analysis12(1973), 211–227. SPACE-TIME LSI AND HYPOCOERCIVE HYPERCONTRACTIVITY FOR UNDERDAMPED LANGEVIN DYNAMICS 27

    1973

  22. [30]

    Section B

    Jacques Neveu,Sur l’espérance conditionnelle par rapport à un mouvement brownien, Annales de l’Institut Henri Poincaré. Section B. Calcul des probabilités et statistiques12(1976), no. 2, 105–109

    1976

  23. [31]

    1–2, 101–174

    Felix Otto,The geometry of dissipative evolution equations: the porous medium equation, Communications in Partial Differential Equations26(2001), no. 1–2, 101–174

    2001

  24. [32]

    Felix Otto and Cédric Villani,Generalization of an inequality by Talagrand and links with the logarithmic Sobolev inequality, Journal of Functional Analysis173(2000), 361–400

    2000

  25. [33]

    Vempala and Andre Wibisono,Rapid convergence of the unadjusted Langevin algorithm: isoperimetry suffices, Advances in neural information processing systems, 2019, pp

    Santosh S. Vempala and Andre Wibisono,Rapid convergence of the unadjusted Langevin algorithm: isoperimetry suffices, Advances in neural information processing systems, 2019, pp. 8092–8104

    2019

  26. [34]

    58, American Mathematical Soc., 2003

    Cédric Villani,Topics in optimal transportation, Vol. 58, American Mathematical Soc., 2003

    2003

  27. [35]

    ,Hypocoercivity, Memoirs of the American Mathematical Society202(2009), no. 950

    2009

  28. [36]

    Feng-Yu Wang,Logarithmic Sobolev inequalities on noncompact Riemannian manifolds, Probability Theory and Related Fields109(1997), 417–424

    1997

  29. [37]

    12, 5360–5383

    ,Hypercontractivity and applications for stochastic Hamiltonian systems, Journal of Functional Analysis 272(2017), no. 12, 5360–5383

    2017

  30. [38]

    Zhang, Jason M

    Matthew S. Zhang, Jason M. Altschuler, and Sinho Chewi,Algorithmic warm starts for Hamiltonian Monte Carlo, arXiv:2603.22741 (2026). DEPARTMENT OFMATHEMATICS, CITYUNIVERSITY OFHONGKONG Email address:boweli4@cityu.edu.hk DEPARTMENT OFMATHEMATICS, DEPARTMENT OFPHYSICS,ANDDEPARTM...

    2026

Pith tools

Reviewed August 2, 2026 · model on record in the stance chip above.