Pith. sign in

REVIEW 3 major objections 5 minor 43 references

Free energy dissipation and a decomposition of general jump diffusions on $\mathbb{R}^n$ without detailed balance

T0 review · 3 major / 5 minor · reviewed 2026-08-03 · 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 that for jump diffusions without detailed balance, the free energy dissipation is carried entirely by the reversible part of the generator and equals a nonlocal Fisher information, while the antisymmetric part generates di

desk verdict Section 3 is a real extension, but the generator decomposition at the center of Section 4 doesn't hold—Prop 4.2's proof is wrong. read the letter →

arxiv 2512.06839 v2 pith:EDBX6VRU submitted 2025-12-07 cond-mat.stat-mech math.PR

Shuyuan Fan , Qi Zhang This is my paper · ORCID
classification cond-mat.stat-mechmath.PR MSC 60J7560J6082C3135Q84 PACS 05.40.-a05.70.Ln
keywords freeenergydissipationjumpdiffusionsLévyprocessesFisherinformationentropyproductionhousekeepingheatdetailedbalancegeneratordecomposition
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

Jump diffusions—stochastic processes that combine smooth Brownian motion with discontinuous Poisson jumps—are used to model everything from intracellular cargo transport and animal foraging to financial returns and generative models. This paper proves that, whenever such a process has a smooth stationary density, its generator splits into two pieces in the Hilbert space weighted by that density: a symmetric piece that is a reversible jump diffusion, and an antisymmetric piece that circulates probability along closed loops without changing the stationary state. The free energy, defined as the relative entropy of the running distribution with respect to the stationary measure, always decreases; its entire dissipation rate equals the nonlocal Fisher information of the symmetric piece, while the antisymmetric circulation contributes zero dissipation. The same calculation yields a decomposition of the dissipation rate into housekeeping heat and entropy production, each nonnegative, and a Clausius inequality. This gives the nonlocal, jump-driven analogue of the familiar Helmholtz–Hodge decomposition of drift in Langevin dynamics: the irreversible part relaxes, the rotational part spins forever at no free-energy cost.

What carries the argument

The central object is the ρ_ss-weighted generator decomposition. In the Hilbert space L^2(R^n, ρ_ss dx), take the adjoint L^† of the generator with respect to the stationary density. Then L_s = (L+L^†)/2 is the symmetric, reversible part; its Dirichlet form E(f, log f) defines a nonlocal Fisher information I[ρ] that carries the full free-energy decay. L_a = (L−L^†)/2 is the antisymmetric part, with drift v_a = b − β^{-1}A∇logρ_ss and jump kernel k_a(x,y) = (1/2)(k(x,y)/ρ_ss(y) − k(y,x)/ρ_ss(x))ρ_ss(y). This kernel integrates to zero and the drift is divergence-free against ρ_ss, so L_a preserves the stationary density and generates pure circulation along its level sets.

What would settle it

Construct a well-posed jump diffusion with an asymmetric heavy-tailed kernel, such as k(x,y)=|x−y|^{-(1+α)}e^{y−x}, and solve for its stationary density. If the nonlocal part of e_p(t) diverges because k log(k(x,y)/k(y,x)) is unbounded, while the process itself has smooth finite-density marginals, then the dissipation identity dF/dt = Q_hk − β^{-1} e_p fails exactly when Assumption (E) is violated, marking the boundary of the theorem.

Watch

Extended reading notes

Core claim

The central discovery is an operator-theoretic split of the generator of a general jump diffusion into L = L_s + L_a, taken with respect to the inner product weighted by the stationary density ρ_ss. L_s, defined as (L+L^†)/2, is self-adjoint and generates a time-reversible jump diffusion, so it satisfies detailed balance by construction; its Dirichlet form defines a nonlocal Fisher information I[ρ], and the free energy dissipation is exactly dF/dt = −I[ρ(t)]. L_a, defined as (L−L^†)/2, generates the canonical conservative dynamics: its drift v_a = b − β^{-1}A∇logρ_ss and jump kernel k_a are constructed so that (L_a)^*ρ_ss = 0 and F_a'(ρ) vanishes identically. The antisymmetric part produces

Load-bearing premise

The whole construction leans on the assumption that the process has a smooth positive stationary density and that the jump kernel's log-ratio k(x,y)log(k(x,y)/k(y,x)) is bounded along the time-marginal distributions; if either fails, the integral defining entropy production can be infinite and the L^2 decomposition is not established.

Editorial extensions

If this is right

  • The free energy of any such jump diffusion is a Lyapunov function: F(ρ(t)) is nonincreasing, and the Clausius inequality θ dS/dt − dQ/dt = θ e_p(t) ≥ 0 holds.
  • Housekeeping heat Q_hk(t) equals the mechanical power of external driving, both local and nonlocal, and vanishes exactly when detailed balance holds.
  • Free energy dissipates only through L_s: dF/dt = −I[ρ(t)], where I is the nonlocal Fisher information built from the symmetric Dirichlet form and decomposes into a diffusion part and a jump part involving the logarithmic mean.
  • If the symmetric generator satisfies a modified logarithmic Sobolev inequality, F(t) ≤ F(0) e^{−t/(β C_I)}, so the relaxation rate is controlled by the spectral gap of L_s, not by the circulation L_a.
  • In a nonequilibrium steady state, dF/dt → 0 forces the stationary entropy production rate to equal β times the stationary housekeeping heat, so the steady state sustains positive entropy production without further free energy loss.

Reading between the lines

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

  • A practical design consequence, not drawn by the paper: adding an antisymmetric (curl-like) driving force to a jump diffusion can be expected to change the stationary circulation without changing the relaxation rate, since the decay rate is fixed solely by L_s; this could be tested by comparing equilibration times of two jump diffusions with the same L_s but different L_a.
  • The logarithmic mean appearing in the nonlocal Fisher information suggests a bridge between discrete-state Markov chains and continuum diffusions; in the small-jump limit I_nl should converge to the classical Fisher information, so the same functional family may interpolate across levels of coarse-graining.
  • The canonical conservative flow generated by L_a could be made explicit through a continuous-space cycle decomposition, analogous to the loop decomposition of master equations; expressing the antisymmetric nonlocal current as a superposition of closed loops x → y → z → x would give a graphic way to measure how far a jump diffusion is from detailed balance.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

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 studies the stochastic thermodynamics of jump diffusions on R^n combining Brownian and Poisson noise. It derives a free-energy dissipation formula dF/dt = Q_hk - β^{-1}e_p, proves (or claims) nonnegativity of Q_hk and e_p, and then decomposes the generator L = L_s + L_a in L^2(R^n, ρ_ss dx), with L_s self-adjoint and L_a antisymmetric. The central claim is that the antisymmetric part generates a 'canonical conservative dynamics' with zero free-energy dissipation, F_a'(ρ)=0, so that the total dissipation is carried by the symmetric part and equals minus a nonlocal Fisher information I[ρ]: dF/dt = -I[ρ]. Two numerical examples illustrate the intended decomposition.

Significance. The paper contains a useful algebraic decomposition of the generator into self-adjoint and antisymmetric parts (Thm. 4.1) and proposes a nonlocal Fisher information via the Dirichlet form of the symmetric part. If the dissipation-free property of L_a were true, it would provide a clean geometric picture extending Helmholtz-Hodge decompositions to jump diffusions. The numerical examples illustrate the intended phenomenology. However, the central result is not correct as stated, and the supporting calculations contain systematic factor errors; the claimed physical interpretation therefore does not follow.

major comments (3)
  1. [§4.1, Prop. 4.2] The claim F_a'(ρ)=0 is false. In the proof, the identity I2 = (1/(2β))∫∫ρ k_a (log f(y)-log f(x)) is missing a factor 1/β; the correct identity from the preceding line is I2 = (1/β)∫∫ρ k_a(φ(y)-φ(x)). Consequently the two symmetrized expressions are each (1/2)I2 with opposite signs, not I2 itself, so they cannot imply I2=0. Concrete counterexample within the hypotheses: on a three-point space with uniform ρ_ss and k(1,2)=k(2,3)=k(3,1)=2, k(1,3)=k(2,1)=k(3,2)=1, take f=(0.5,0.3,0.2). Then (4.3) and (4.4) hold but βF_a'=0.015≠0. Hence L_a is not dissipation-free and Prop. 4.3 is unsupported.
  2. [§3, proof of Thm. 3.1, Eqs. (3.11)-(3.12)] The step replacing (1/(2β))∫∫ j_nl log C, with C=kρ_ss/(k'ρ_ss'), by (1/(4β))∫∫ j_nl [log C + log(1/C)] = 0 is algebraically invalid: the two terms in the bracket cancel pointwise for the same (x,y), yielding zero, while the original integral is generically nonzero. A correct proof must use, e.g., the known relative-entropy contraction or a proper entropy-production inequality. As written, the proof of dF/dt≤0 does not follow.
  3. [§2, Def. 2.1] Detailed balance is defined by j_loc(t,x)=0 and j_nl(t,x,y)=0 for all t>0 and all x,y. This is inconsistent with relaxation: a reversible jump diffusion starting from a non-stationary density has nonzero currents at finite times. The definition should be stated for the stationary currents (j_loc^ss=0, j_nl^ss=0), or equivalently in terms of the invariant measure/reversibility, matching Prop. 2.2.
minor comments (5)
  1. [Eq. (3.10)] The stationary Fokker-Planck equation as written is missing the factor β^{-1} multiplying A(x)∇ρ_ss. The correct equation is ∇·(bρ_ss - β^{-1}A∇ρ_ss) + ∫(ρ_ss(y)k(y,x)-ρ_ss(x)k(x,y))dy = 0. This appears to be a typo, since the prior line uses the β^{-1} factor.
  2. [Prop. 4.3] Since F = β^{-1}H, the free-energy dissipation should read dF/dt = β^{-1}⟨f, L_s log f⟩_{ss} = -β^{-1}I, not -I. The factor β^{-1} is dropped in the statement.
  3. [Prop. 4.2 proof, I1] The local contribution I1 is written with the sign ∇logρ_ss - ∇logρ, whereas the adjoint calculation gives the opposite sign. Both signs yield zero by (4.3), so this is harmless but confusing.
  4. [Assumption (E)] Assumption (E) is quite strong: it requires bounded ∇logρ and k̄∈L∞, excluding many heavy-tailed asymmetric kernels for which the entropy production integrand may be infinite. The paper should state this limitation explicitly in the introduction or abstract.
  5. [General] There are several typos: 'detial', 'serveed', 'extendion', 'bablance', and inconsistent capitalization of 'Lévy'. Also, Figure 1 and Figure 2 have no descriptive captions in the text.

Circularity Check

1 steps flagged · score 6.0 of 10

Prop 4.2's proof replaces f(y) by f(x) in the nonlocal integral, so the central F_a'(ρ)=0 and the resulting dF/dt=-I[ρ] are assumed rather than derived.

  1. other [Prop 4.2 proof, §4.1, second display after Eq. (4.8)]
    "By (4.4), we also have I2 = 1/β ∫∫ ρ(t,y)ka(y,x) log(ρ(t,x)/ρss(x))dydx = 1/(2β) ∫∫ ρss(y)ka(y,x) ρ(t,x)/ρss(x)[log(ρ(t,x)/ρss(x))−log(ρ(t,y)/ρss(y))]dydx"

    Eq. (4.4) only states ∫ k_a(x,y)dy=0 (equivalently ∫ρss(y)ka(y,x)dy=0). It does not justify replacing the y-weight ρ(t,y) in the first integral by ρss(y)ρ(t,x)/ρss(x) in the second. That replacement is valid only if ∫ρss(y)ka(y,x)(f(y)-f(x))dy=0, which is precisely the claim that the nonlocal part of L_a produces no free-energy dissipation. The proof then combines this with the preceding symmetrized expression to obtain I2=-I2, hence F_a'(ρ)=0. The conclusion is thus inserted into the algebra rather than obtained from the generator decomposition, and Proposition 4.3's identification dF/dt=-I[ρ] inherits this assumed vanishing.

full rationale

The bulk of the paper is a standard algebraic decomposition: L=L_s+L_a is the symmetric/antisymmetric splitting in L^2(ρss), and the free-energy accounting dF/dt=Q_hk-β^{-1}e_p is a rearrangement of the nonlocal Fokker-Planck equation together with the definitions of Q_hk and e_p. Those definitions are legitimate and not circular by themselves. The real circular/definitional step is in Proposition 4.2. The proof's second symmetrization uses (4.4) to turn an integral containing ρ(t,y) into one containing ρ(t,x), i.e. to replace f(y) by f(x). This is the vanishing statement that the proof is supposed to establish. Once F_a'(ρ)=0 is imposed in this way, Proposition 4.3's claim that all free-energy dissipation is carried by L_s and equals -I[ρ] is not an independent consequence; it is equivalent to the assumed no-dissipation property of L_a. The self-citations to the authors' earlier [43] are present (e.g. Proposition 2.2 and the reversibility claim in §4.2), but they are not the load-bearing source of circularity: reversibility of L_s follows from its self-adjointness and the cited [43] theorem is background. For that reason I do not score the paper higher. The score of 6 reflects that the central no-dissipation/Fisher-information claim reduces, in its proof, to the very equality it purports to derive.

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

The central machinery rests on the existence of a smooth invariant density and on the integrability of log-ratio jump terms; no new physical entities are introduced. The Lévy index α in Example 5.1 is an illustrative model parameter, not a fitted constant.

assumptions (4)
  • domain assumption The jump diffusion admits a unique strong solution, smooth transition density, and a C^2 positive invariant density ρ_ss (Assumptions A-E, Theorem 2.1, Section 2).
    The entire L^2(ρ_ss) machinery and the Fokker-Planck calculations require a smooth positive stationary density and regular enough coefficients; without it, ρ_ss and the nonlocal currents are undefined.
  • domain assumption Assumption (E), Eq. (2.2): k(x,y) log(k(x,y)/k(y,x)) ∈ L^∞(μ_{X_t}(dx)dy) and ∇logρ(t,·) is bounded μ_{X_t}-a.s.
    Needed so the nonlocal entropy production and the free-energy dissipation integrals are finite; this is a genuine restriction on jump kernels, excluding e.g. strongly asymmetric heavy-tailed kernels.
  • domain assumption Proposition 2.2 / [43, Theorem 5.1]: reversibility, detailed balance, gradient drift/kernel structure, and zero steady entropy production are equivalent.
    Used to identify L_s as reversible and to connect the decomposition to detailed balance; imported from the authors' companion paper without proof here.
  • standard math Boundary terms vanish in integration by parts over R^n, and the infinitesimal generator L acts on a core where L^† is well-defined.
    The proofs of Theorem 3.1 and Propositions 4.2-4.3 integrate by parts with no boundary contributions; this is standard but unstated.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Free energy dissipation and a decomposition of general jump diffusions on $\mathbb{R}^n$ without detailed balance." pith.science (2026). https://pith.science/paper/EDBX6VRU

@misc{pith2026251206839,
  author       = {Pith},
  title        = {Pith review of: Free energy dissipation and a decomposition of general jump diffusions on $\mathbbR^n$ without detailed balance},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/EDBX6VRU}},
  note         = {Machine review of arXiv:2512.06839}
}
abstract

We analyze the thermodynamic structure of jump diffusions combining Brownian and Poisson noise, a class of stochastic dynamics relevant to non-equilibrium statistical physics. For such nonlocal dynamics, the free energy admits a full dissipation formula that decomposes into entropy production and housekeeping heat. A central result is a decomposition of the generator into symmetric and anti-symmetric parts with respect to the invariant measure $\rho_\mathrm{ss}$. The symmetric sector corresponds to a reversible dynamics and yields a nonlocal Fisher information governing free-energy decay, whereas the anti-symmetric sector generates a canonical conservative flow that produces circulation but no dissipation. Several numerical examples motivated by intracellular particle transports demonstrate how this decomposition clarifies the structure of non-equilibrium stationary states in jump-driven systems.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

43 extracted references

  1. [1]

    Applebaum.L´ evy Processes and Stochastic Calculus

    D. Applebaum.L´ evy Processes and Stochastic Calculus. Cambridge University Press, Cam- bridge, 2004

    2004

  2. [2]

    P. Ao. Potential in stochastic differential equations: Novel construction,J. Phys. A, 37, L25–L30, 2004

    2004

  3. [3]

    Ao, T.-Q

    P. Ao, T.-Q. Chen, and J.-H. Shi.Dynamical Decomposition of Markov Processes without Detailed Balance.Chinese Physics Letters, 30(7):070201, 2013

    2013

  4. [4]

    Bartumeus, J

    F. Bartumeus, J. Catalan, U. L. Fulco, M. L. Lyra, and G. M. Viswanathan. Animal search strategies: a quantitative random-walk analysis.Ecology, 86(11): 3078-3087, 2005

    2005

  5. [5]

    J. M. Beggs and D. Plenz. Neuronal avalanches in neocortical circuits.Journal of Neuro- science, 23(35): 11167-11177, 2003

    2003

  6. [6]

    N. M. Boffi and E. Vanden-Eijnden. Deep learning probability flows and entropy produc- tion rates in active matter.Proceedings of the National Academy of Sciences, 121(25), e2318106121, 2024

    2024

  7. [7]

    Chafa ¨ ı

    D. Chafa ¨ ı. Entropies, convexity, and functional inequalities, On Φ-entropies and Φ-Sobolev inequalities.Journal of Mathematics of Kyoto University, 44(2):325–363, 2004. 24

    2004

  8. [8]

    P. D. Ditlevsen and S. J. Johnsen. Tipping points: Early warning and wishful thinking. Geophysical Research Letters, 37: L19703, 2010

    2010

Show all 43 references
  1. [9]

    S. R. De Groot and P. Mazur.Non-Equilibrium Thermodynamics. North-Holland Publishing Company, Amsterdam, 1962

    1962

  2. [10]

    Delong and C

    L. Delong and C. Kl¨ uppelberg. Optimal investment and consumption in a Black–Scholes market with L´ evy-driven stochastic coefficients.Annals of Applied Probability, 18: 879–908, 2008

    2008

  3. [11]

    Da Costa and G

    L. Da Costa and G. A. Pavliotis. The entropy production of stationary diffusions.Journal of Physics A: Mathematical and Theoretical, 56(36): 365001, 2023

    2023

  4. [12]

    Esposito and C

    M. Esposito and C. Van den Broeck. Three faces of the second law. I. Master equation formulation.Phys. Rev. E, 82, 011143, 2010

    2010

  5. [13]

    Fukushima and D

    M. Fukushima and D. Stroock. Reversibility of solutions to martingale problems.Adv. Math. Suppl. Stud.9, 107–123, Academic Press, Orlando, 1986

    1986

  6. [14]

    Holderrieth, M

    P. Holderrieth, M. Havasi, J. Yim, N. Shaul, I. Gat, T. Jaakkola, B. Karrer, R. T. Q. Chen, and Y. Lipman. Generator Matching: Generative modeling with arbitrary Markov processes. The Thirteenth International Conference on Learning Representations, 2025

    2025

  7. [15]

    Ge and H

    H. Ge and H. Qian. Physical origins of entropy production, free energy dissipation, and their mathematical representations.Phys. Rev. E, 81, 051133, 2010

    2010

  8. [16]

    Gentil and C

    I. Gentil and C. Imbert. The L´ evy–Fokker–Planck equation: Φ–entropies and convergence to equilibrium.Asymptotic Analysis, 59(3-4):125-138, 2008

    2008

  9. [17]

    R. Z. Hasminskii.Stochastic Stability of Differential Equations. Sijthoff Noordhoff, Rockville, 1980

    1980

  10. [18]

    Huang, C

    Y. Huang, C. Liu, B. Miao, and X. Zhou.Entropy production in non-Gaussian active matter: A unified fluctuation theorem and deep learning framework. arXiv preprint arXiv:2504.06628, 2025

    2025

  11. [19]

    Hatano and S

    T. Hatano and S. Sasa. Steady-state thermodynamics of Langevin systems.Phys. Rev. Lett., 86: 3463, 2001

    2001

  12. [20]

    Jiang, M

    D.-Q. Jiang, M. Qian, and M.-P. Qian.Mathematical theory of nonequilibrium steady states, volume 1833 of Lecture Notes in Mathematics. Springer-Verlag, Berlin, 2004

    2004

  13. [21]

    Kunita.Stochastic flows and jump-diffusions

    H. Kunita.Stochastic flows and jump-diffusions. Springer, Singapore, 2019

    2019

  14. [22]

    Klages, G

    R. Klages, G. Radons, and I. M. Sokolov.Anomalous transport: Foundations and applica- tions. Wiley-VCH, 2008

    2008

  15. [23]

    Lisowski, D

    B. Lisowski, D. Valenti, B. Spagnolo, M. Bier, and E. Gudowska-Nowak. Stepping molecular motor amid L´ evy white noise.Phys. Rev. E, 91, 042713, 2015

    2015

  16. [24]

    Maes and K

    C. Maes and K. Netoˇ cn´ y. Time-reversal and entropy.Journal of Statistical Physics, 110(1-2): 269–310, 2003

    2003

  17. [25]

    C. Maes, K. Netoˇ cn´ y, and B. Wynants. On and beyond entropy production: The case of Markov jump processes.Markov Proc. Relat. Fields, 14, 445, 2008. 25

    2008

  18. [26]

    R. C. Merton. Option pricing when underlying stock returns are discontinuous.Journal of Financial Economics, 3: 125-144, 1976

    1976

  19. [27]

    Nicolis and I

    G. Nicolis and I. Prigogine.Self-organization in nonequilibrium systems: from dissipative structures to order through fluctuations. Wiley, New York, 1977

    1977

  20. [28]

    Oono and M

    Y. Oono and M. Paniconi. Steady state thermodynamics.Prog. Theor. Phys. Suppl, 130: 29-44, 1998

    1998

  21. [29]

    J. Picard. On the existence of smooth densities for jump processes.Probability Theory and Related Fields, 105: 481-511, 1996

    1996

  22. [30]

    H. Qian, M. Qian, and X. Tang. Thermodynamics of the general diffusion process: time- reversibility and entropy production.J. Stat. Phys., 107(5/6): 1129–1141, 2002

    2002

  23. [31]

    H. Qian. A decomposition of irreversible diffusion processes without detailed balance.J. Math. Phys., 54(5), 2013

    2013

  24. [32]

    U. Seifert. Entropy production along a stochastic trajectory and an integral fluctuation theorem.Phys. Rev. Lett., 95, 040602, 2005

    2005

  25. [33]

    U. Seifert. Stochastic thermodynamics, fluctuation theorems and molecular machines.Re- ports on Progress in Physics, 75: 126001, 2012

    2012

  26. [34]

    Sato.L´ evy processes and infinitely divisible distributions

    K. Sato.L´ evy processes and infinitely divisible distributions. Cambridge University Press, Cambridge, 1999

    1999

  27. [35]

    Sekimoto.Stochastic Energetics

    K. Sekimoto.Stochastic Energetics. Lecture Notes in Physics, Vol. 799, Springer, Berlin, Heidelberg, 2010

    2010

  28. [36]

    Schnakenberg

    J. Schnakenberg. Network theory of microscopic and macroscopic behavior of master equa- tion systems.Rev. Mod. Phys., 48: 571-585, 1976

    1976

  29. [37]

    M. F. Shlesinger, B. J. West, and J. Klafter. L´ evy dynamics of enhanced diffusion: Appli- cation to turbulence.Phys. Rev. Lett., 58: 1100, 1987

    1987

  30. [38]

    Tom´ e and M

    T. Tom´ e and M. J. de Oliveira. Entropy production in nonequilibrium systems at stationary states.Physical Review Letters, 108(2): 020601, 2012

    2012

  31. [39]

    G. M. Viswanathan, S. V. Buldyrev, S. Havlin, M. G. E. da Luz, E. P. Raposo, and H. E. Stanley. Optimizing the success of random searches.Nature, 401: 911–914, 1999

    1999

  32. [40]

    Van den Broeck and M

    C. Van den Broeck and M. Esposito. Three faces of the second law. II. Fokker-Planck for- mulation.Physical Review E, 82(1): 011144, 2010

    2010

  33. [41]

    L. Wu. A new modified logarithmic Sobolev inequality for Poisson point processes and several applications.Probability Theory and Related Fields, 118(3), 427-438, 2000

    2000

  34. [42]

    E. B. Yoon, K. Park, S. Kim and S. Lim. Score-based generative models with L´ evy processes. Advances in Neural Information Processing Systems, 36: 40694-40707, 2023

    2023

  35. [43]

    Zhang and Y

    Q. Zhang and Y. Lu. Entropy production rate and time-reversibility for general jump diffu- sions onR n.Chaos, 2025. 26

    2025

Pith tools

Reviewed August 3, 2026 · model on record in the stance chip above.