Reversible birth-and-death dynamics in continuum: free-energy dissipation and attractor properties

Alexander Zass; Benedikt Jahnel; Jonas K\"oppl; Yannic Steenbeck

REVIEW 3 major objections 5 minor 38 references

Reversible birth-and-death dynamics in continuum: free-energy dissipation and attractor properties

T0 review · 3 major / 5 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read For reversible birth-and-death dynamics in the continuum, every long-time weak limit point of a regular starting measure is a Gibbs point process for the same area interaction.

desk verdict Genuinely novel continuum entropy-dissipation machinery, but the central attractor proof (Theorem 2.4) has a load-bearing Palm normalization error that needs fixing before the result can be trusted. read the letter →

arxiv 2508.21196 v1 pith:3SJVVEVO submitted 2025-08-28 math.PR

Yannic Steenbeck , Alexander Zass , Jonas Köppl , Benedikt Jahnel This is my paper · ORCID

classification math.PR MSC 82C2182B2160K3560G5560J25

keywords Gibbspointprocessesspatialbirth-and-deathrelativeentropydissipationFisherinformationPalmmeasuresareainteractionattractorphasetransition

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 proves a Lyapunov result for interacting particle systems in continuous space. For reversible birth-and-death dynamics driven by the area interaction, the specific relative entropy—the free-energy density with respect to any equilibrium state—never increases along a trajectory. For regular starting distributions, the entropy lost is bounded below by the time-integral of a non-negative 'Palm Fisher information' that vanishes exactly at equilibrium. It follows that every long-time limit point of the trajectory is itself an equilibrium (Gibbs) state for the same interaction, and hence reversible. This is the first out-of-equilibrium convergence result for such continuum dynamics, and it works without assuming uniqueness of the equilibrium state.

What carries the argument

The Lyapunov function is the specific relative entropy density Ī(μ|ν) with respect to a reversible Gibbs measure ν. Entropy dissipation is computed through a double-layer representation: the entropy loss under the localized finite-volume dynamics equals the space integral of the relative entropy between two Palm-type measures μ∗G_x and μ∘G_x, which respectively add and remove a point at x according to the birth rate; this is re-expressed as a modified Fisher information J_Λ. Three structural ingredients carry the argument: a finite-speed-of-propagation estimate comparing the global dynamics with a finite-volume dynamics whose boundary conditions are sampled from ν; a quasi-superadditivity p

What would settle it

Take a translation-invariant regular starting measure μ that is not Gibbs, evolve the area-interaction dynamics, and measure the double-layer relative entropy I(μT_s∗G_o | μT_s∘G_o). The theorem predicts this is strictly positive for every s with μT_s∉G_θ and vanishes only at times where the state approaches G_θ. A single time s with zero double-layer entropy but a non-Gibbs state, or a τ_L-limit point outside G_θ (detected, for example, by a local function whose expectation differs from every Gibbs state), would refute Theorem 2.3(iii) or Theorem 2.4, respectively.

Watch

Extended reading notes

Core claim

The central claim is the pair of statements in Theorems 2.3 and 2.4. Theorem 2.3 establishes free-energy dissipation: for ν an infinite-volume Gibbs measure for the area interaction, the map t ↦ Ī(μT_t|ν) is non-increasing for every translation-invariant starting measure μ, and for regular μ the lost entropy is at least the time-integral of the functional ξ_μ(s), built from Palm measures; ξ_μ(s) vanishes if and only if μT_s is a Gibbs state. Theorem 2.4 converts this into an attractor property: for regular starting measures, every τ_L-weak limit point of the trajectory lies in G_θ, the set of infinite-volume Gibbs measures for the same interaction. The proof works in both the uniqueness and

Load-bearing premise

Everything rests on a bound, stated as an easy DLR calculation, that area-interaction Gibbs measures are nearly factorized across large boxes: the density of ν against the product of its inside and outside marginals is at most exp(ε|volume|); if that bound failed, as it does for longer-range interactions, the quasi-superadditivity of entropy—and with it the attractor theorem—would collapse.

Editorial extensions

If this is right

Every weak limit point of a regular trajectory is a Gibbs state, so the long-time attractor contains no non-equilibrium states; combined with reversibility of Gibbs states, the omega-limit set coincides with the set of Gibbs measures.
The entropy-monotonicity statement holds for every translation-invariant starting measure, without requiring absolute continuity with respect to the reversible measure.
The characterization ξ_μ(s)=0 if and only if μT_s∈G_θ gives a continuum analogue of the classical entropy-dissipation-to-Fisher-information relation for birth-death dynamics.
All area-interaction Gibbs measures are regular, so the attractor theorem applies to the physically natural grand-canonical starting distributions.
The results hold in both the uniqueness and the phase-transition regime, so no ergodicity assumption is needed.

Reading between the lines

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

If the conjectured equality in the Fisher-information lower bound could be proved, the specific entropy would satisfy a de Bruijn-type identity, opening the door to quantitative convergence rates via log-Sobolev or Talagrand inequalities for continuum Gibbs point processes.
The same strategy should transfer to any finite-range interaction whose Papangelou intensity is globally bounded and whose Gibbs measures satisfy an exp(ε|Λ|) factor property; the paper explicitly flags the factor property as the main bottleneck.
The finite-speed-of-propagation lemma suggests a light-cone structure that could support perfect-simulation or efficient approximate sampling algorithms for area-interaction Gibbs measures, a practical consequence the paper does not pursue.
A natural test is to run the dynamics from a high-intensity Poisson start in the phase-transition region and check that the empirical local statistics converge to a Gibbs state consistent with the dynamics rather than to a non-Gibbs metastable measure.

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.

Desk Editor's Note

read the letter

You should know two things about this paper. First, it is the first real attempt to transfer the lattice entropy-dissipation/attractor toolkit to interacting continuum birth-and-death dynamics, and the core construction—entropy dissipation via a Palm-version double-layer representation—is genuinely new. Second, the central attractor proof, Theorem 2.4, has a concrete normalization error as written, and I think the paper needs revision before it can be trusted.

What is genuinely good: the framework is clean. The finite-speed-of-propagation estimate (Lemma 4.4) is a real continuum analogue of the lattice argument. The double-layer representation (Prop. 6.3) connecting entropy loss to a relative entropy of Palm-type measures is elegant. Theorem 2.3(i) gives entropy monotonicity for all translation-invariant starts, and the conditional Fisher-information representation is plausible. These are substantial mathematical contributions, not just repackaged lattice arguments.

Now the soft spots, in proportion. The factor property Lemma 5.2 is asserted after a one-line sketch; that is a real gap, but the result is likely true for the area interaction (surface-order correction), so it is a presentation problem more than a fatal one. Proposition 6.5's local-to-global Fisher information passage is abbreviated and needs careful checking.

The more serious issue is in the proof of Theorem 2.4. They define G(η)=Σ_{x∈η∩[0,1]^d} exp(F(θ_xη\{o})) and then use ρ[G] = ρ^!_o[exp(F)]. But the reduced Palm formula gives ρ[G] = λ ρ^!_o[exp(F)], where λ is the intensity of ρ. The same missing λ appears in the Donsker–Varadhan lower bound for ξ. Unless λ=1—which is not guaranteed; the area-interaction Gibbs measure with an intensity-1 reference does not have intensity 1 in general—the displayed inequality ξ_μ(t) ≥ μT_t[b(o,·)F] − μT_t[G] omits a log λ(t) term, and the contradiction that generates infinite dissipation is not justified as written. This is load-bearing, and it is not a typo that can be fixed by inspection; the definition of G or the normalization of the Palm measure would need to be adjusted, and the rest of the argument rechecked.

This doesn't sink the whole paper. The entropy-dissipation theorem and the technical machinery likely survive. But the attractor property is one of the two headline results, and right now its proof has a hole.

I would send this to a serious referee. The paper is important and the core ideas are worth engagement, but the referee should be asked to verify the Palm normalization carefully.

Recommendation: conditional, with a request to fix the normalization issue.

Referee Report

3 major / 5 minor

Summary. The paper studies continuous-time birth-and-death dynamics on the space of locally finite point configurations in R^d, with rates derived from the area-interaction Hamiltonian, and with at least one infinite-volume Gibbs measure as a reversible equilibrium. The main results are Theorem 2.3, which states that the specific relative entropy is non-increasing along trajectories and that, for regular initial measures, the entropy drop is bounded below by an integrated Palm-type Fisher information term, and Theorem 2.4, which concludes that every tau_L-weak limit point of a trajectory starting from a regular measure is a Gibbs measure for the same interaction. The proof strategy follows Sullivan's lattice method: a local finite-volume dynamics with stochastic boundary conditions is introduced and compared with the global dynamics via a finite-speed-of-propagation estimate; quasi-superadditivity of the entropy is derived from a factor property of area-interaction Gibbs measures; and the entropy dissipation is expressed through a double-layer/Palm representation and bounded below by a DV-type functional. The paper is clearly organized and explicitly discusses limitations and possible generalizations.

Significance. If the proofs are completed, the paper would be a substantial contribution: it provides the first entropy-dissipation and attractor results for continuum birth-and-death dynamics in a regime that includes phase transitions, and it does so without fitting parameters, using a derivation that rests on classical external results. The finite-speed-of-propagation estimate and the double-layer representation are likely to be of independent interest. The paper is also honest about the model-specific assumptions and future directions. However, the proof of the central attractor theorem (Theorem 2.4) currently contains a concrete Palm-normalization error in the Donsker-Varadhan bound, and the DV bound in Proposition 6.5 omits a mass term. These are load-bearing gaps, though they appear fixable by tracking the intensity and using the correct finite-measure DV formula.

major comments (3)

[Section 6.3, proof of Theorem 2.4] The identity leading to (6.5) is incorrect. By the reduced Palm formula in Section 2.3, for a translation-invariant measure rho with intensity lambda one has rho[G] = lambda rho^!_o[e^F], not rho^!_o[e^F], where G(eta)=sum_{x in eta∩[0,1]^d} exp(F(theta_x eta \ {o})). Consequently the DV lower bound in (6.5) is missing a factor or a log lambda(t) term, and lambda(t) = intensity of mu T_t is not shown to be constant. In addition, b(o,·)rho is not in general a probability measure, so the standard DV inequality applied to I(b(o,·)rho || rho^!_o) needs normalization or the finite-measure DV formula. This gap is load-bearing: the uniform lower bound xi_mu(t) >= delta/4 and the resulting infinite-dissipation contradiction are not justified as written.
[Section 6.2, proof of Proposition 6.5] The variational approximation step claims (mu T_s^(Λ) * G_x)[F_x] - (mu T_s^(Λ) ◦ G_x)[exp(F_x)] >= I(mu T_s * G_o || mu T_s ◦ G_o) - epsilon. For positive finite measures A, B with common total mass m, the correct DV formula is I(A||B) = sup_F [A[F] - B[e^F] + m]. Here the double-layer measures have common mass E_{mu_s}[b(o,·)] + lambda(s), which is not 1 in general. The displayed inequality is therefore off by a mass term, and the asserted lower bound liminf_n n^{-d} J_{Λ_n}(mu T_s^(Λ_n)|ν) >= xi_mu(s) is not established by the proof given.
[Lemma 5.2] The factor property is the key structural input for quasi-superadditivity of the entropy (Lemma 5.3) and hence for the existence and lower semicontinuity of the relative entropy density (Proposition 5.4). Its proof is only a sketch: after writing a formula for g_Lambda, the bound g_Lambda <= exp(c|∂Lambda|) is asserted without argument. For the area interaction, the boundary-layer contribution depends on the number of particles within distance 2R of the boundary, which is not uniformly bounded over configurations; a pointwise exponential bound requires a separate argument or a precise reference. Since the attractor and dissipation claims rely on this estimate, the proof should be supplied in detail.

minor comments (5)

[Lemma 4.5] The first displayed estimate includes a factor |Λ| times e^{-dist(Λ,bΛ^c)}, but the proof yields 2||f||_∞ e^{-dist} without |Λ|. Either the statement or the derivation should be adjusted.
[Section 5] The notation for the local relative-entropy density, rendered as fI_mu(t) in the text, appears to be a typo for an ar I or ilde I functional. Please clarify the notation.
[Proposition 5.4] The condition involving M contains a misplaced fraction: it should read (I_{Λ_n}(mu|ν)/|Λ_n|) * |Δ|/|Λ_m| <= epsilon, not the displayed ratio.
[Section 6.3] The definition of G is ambiguous: write G(eta) = sum_{x in eta∩[0,1]^d} exp(F(theta_x(eta - delta_x))) or equivalently use the notation for the Palm shift explicitly, to avoid confusion about removing the origin.
[Throughout] When applying the Donsker-Varadhan inequality to finite measures, the paper should state, once, the exact convention for relative entropy of unnormalized measures and the associated variational formula. This would prevent the mass-term issues in Sections 6.2 and 6.3.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the attractor theorem is derived from parameter-free entropy dissipation, external classical characterizations, and model-defined auxiliary dynamics.

full rationale

Walking the derivation chain, Theorem 2.4 is deduced from Theorem 2.3(iii), whose zero-Fisher-information criterion is proved via Lemma 3.4, an external GNZ/Palm characterization (Georgii, Glötzl) that does not presuppose the target result. The entropy monotonicity (Proposition 5.1) is obtained by comparing the global dynamics with a local dynamics with stochastic boundary conditions; the comparison rests on the finite-speed-of-propagation estimate Lemma 4.4/4.5, based on the boundedness of the area-interaction birth rate, and on the factor property Lemma 5.2, which is derived (sketchily but not circularly) from the DLR equations. The double-layer measures in Definition 6.1 and the Fisher-information representation Proposition 6.4 are defined from the model and the propagated measure, not fitted. There is no fitted parameter renamed as a prediction, no definitional identity between input and output, and no load-bearing self-citation: citations to the authors' earlier lattice work [JK23, JK25] and to [JKSZ24] are motivational or comparative, not the justification of any step. The reader's Palm-normalization objection to the DV lower bound in Section 6.3 is a possible correctness gap, not circularity: even if the inequality is missing an intensity term, that would make the proof incomplete, not make the conclusion equivalent to the assumptions. The paper also explicitly flags its model assumptions and conjectures in Section 2.4. Accordingly the circularity burden is not met; no step reduces by construction to its own input.

Assumptions & free parameters 0 free parameters · 6 assumptions · 0 invented entities

The paper introduces no physical entities, fitted parameters, or ad hoc constants. The double-layer measures mu * G_x and mu o G_x are auxiliary mathematical objects constructed from existing Gibbs and Palm machinery, not independent postulates. The only model-level assumptions are the area interaction, finite range, globally bounded birth rate, and the factor property, all explicit.

assumptions (6)

standard math Existence, uniqueness, and Markov property of the infinite-volume birth-death process via the graphical representation of Garcia-Kurtz (Proposition 2.1, from GK06 Theorem 2.13).
Used throughout as the definition of the semigroup T_t; not re-proved in this paper.
standard math The DLR characterization of infinite-volume Gibbs measures and the non-emptiness of G_theta for the area interaction, including the phase-transition regime (Section 2.1, after Rue71, GLM95, CCK95).
Background theorem from the literature; the attractor result must hold in both regimes.
standard math Gloetzl's characterization: a translation-invariant mu is Gibbs if and only if its reduced Campbell measure has density b(x, eta) with respect to dx tensor mu (Proposition 3.2, from Gloetzl 1981).
Load-bearing equivalence used in Lemma 3.4 and in the criterion xi_mu(s)=0 iff mu T_s in G_theta.
domain assumption Factor property of area-interaction Gibbs measures: for large boxes Lambda, d nu/d(nu_Lambda tensor nu_Lambda^c) is bounded by exp(epsilon|Lambda|) (Lemma 5.2).
Interaction-specific structural input; only sketched in the paper, it drives quasi-superadditivity, the existence of the entropy density, and the comparison estimates.
standard math Donsker-Varadhan variational formula for local relative entropy and its lower semicontinuity (used in Lemma 5.5, Proposition 5.4, and Proposition 6.5).
Classical functional-analytic tool from DV76 and BLM13.
domain assumption Regular starting measures: local densities satisfy z^{-N_Lambda} <= d mu_Lambda/d pi_Lambda <= z^{N_Lambda}, and these bounds persist along the dynamics for finite time.
This is the 'large class of starting measures' in the abstract; it justifies the density evolution and derivative computations in Proposition 6.3 and the tail estimates in Theorem 2.4.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Reversible birth-and-death dynamics in continuum: free-energy dissipation and attractor properties." pith.science (2026). https://pith.science/paper/3SJVVEVO

@misc{pith2026250821196,
  author       = {Pith},
  title        = {Pith review of: Reversible birth-and-death dynamics in continuum: free-energy dissipation and attractor properties},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/3SJVVEVO}},
  note         = {Machine review of arXiv:2508.21196}
}

abstract

We consider continuous-time birth-and-death dynamics in $\mathbb{R}^d$ that admit at least one infinite-volume Gibbs point process based on area interactions as a reversible measure. For a large class of starting measures, we show that the specific relative entropy decays along trajectories, and that all possible long-time weak limit points are also Gibbs point processes with respect to the same interaction. Our proof rests on a representation of the entropy dissipation in terms of the Palm version of the propagated measure.

Figures

Figures reproduced from arXiv: 2508.21196 by the authors.

**Figure 1.** Realization of the graphical representation of Proposition 2.1. At time t = 0, an initial configuration η (in green) is given; for positive times, the driving space-time Poisson measure N proposes new points to be born at a given location and time (dots), with assigned death times (crosses); points are accepted (blue) or rejected (gray) depending on the current configuration. For any B ⋐ R d (red) and t > 0, X (η) t… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

38 extracted references · 38 canonical work pages

[1]

Boucheron, G

S. Boucheron, G. Lugosi, and P. Massart. Concentration Inequalities : A Nonasymptotic Theory of Independence . Oxford University Press, 2013

work page 2013
[2]

J. T. Chayes, L. Chayes, and R. Koteck \'y . The analysis of the Widom--Rowlinson model by stochastic geometric methods . Comm. Math. Phys. , 172(3):551 -- 569, 1995

work page 1995
[3]

Dello Schiavo, R

L. Dello Schiavo, R. Herry, and K. Suzuki. Wasserstein geometry and Ricci curvature bounds for Poisson spaces. J. \'E c. Polytech. Math. , 11:957--1010, 2024

work page 2024
[4]

M. D. Donsker and S. R. S. Varadhan. Asymptotic evaluation of certain Markov process expectations for large time -- III . Comm. Pure Appl. Math. , 29(4):389--461, 1976

work page 1976
[5]

Dereudre and T

D. Dereudre and T. Vasseur. Existence of G ibbs point processes with stable infinite range interaction. J. Appl. Probab. , 57(3):775–791, 2020

work page 2020
[6]

D. J. Daley and D. Vere-Jones. An Introduction to the Theory of Point Processes. V ol. II . Probability and its Applications (New York). Springer, New York, second edition, 2008

work page 2008
[7]

S. N. Ethier and T. G. Kurtz. Markov Processes . Wiley Series in Probability and Mathematical Statistics: Probability and Mathematical Statistics. John Wiley & Sons Inc., 1986

work page 1986
[8]

A. M. Etheridge and T. G. Kurtz. Genealogical constructions of population models. Ann. Probab. , 47(4), 2019

work page 2019

Show all 38 references

[9]

Finkelshtein, Y

D. Finkelshtein, Y. Kondratiev, O. Kutoviy, and M. J. Oliveira. Dynamical Widom -- Rowlinson model and its mesoscopic limit. J. Stat. Phys. , 158(1):57--86, 2014

2014
[10]

N. L. Garcia. Birth and death processes as projections of higher-dimensional Poisson processes. Adv. Appl. Probab. , 27(4):911--930, 1995

1995
[11]

H.-O. Georgii. Canonical and grand canonical G ibbs states for continuum systems. Comm. Math. Phys. , 48:31--51, 1976

1976
[12]

H.-O. Georgii. The equivalence of ensembles for classical systems of particles. J. Stat. Phys. , 80(5):1341--1378, 1995

1995
[13]

N. L. Garcia and T. G. Kurtz. Spatial birth and death processes as solutions of stochastic equations. Alea , 1:281--303, 2006

2006
[14]

Giacomin, J

G. Giacomin, J. L. Lebowitz, and C. Maes. Agreement percolation and phase coexistence in some Gibbs systems. J. Stat. Phys. , 80(5-6):1379--1403, 1995

1995
[15]

Gl \"o tzl

E. Gl \"o tzl. Time reversible and G ibbsian point processes. I . M arkovian spatial birth and death processes on a general phase space. Math. Nachr. , 102:217--222, 1981

1981
[16]

R. A. Holley. Free energy in a Markovian model of a lattice spin system. Comm. Math. Phys. , 23(2):87--99, 1971

1971
[17]

Higuchi and T

Y. Higuchi and T. Shiga. Some results on Markov processes of infinite lattice spin systems. Kyoto J. Math. , 15(1), 1975

1975
[18]

R. A. Holley and D. W. Stroock. In one and two dimensions, every stationary measure for a stochastic Ising model is a Gibbs state. Comm. Math. Phys. , 55(1):37--45, 1977

1977
[19]

R. A. Holley and D. W. Stroock. Nearest neighbor birth and death processes on the real line . Acta Math. , 140:103 -- 154, 1978

1978
[20]

Huesmann and H

M. Huesmann and H. Stange. Non-local W asserstein geometry, gradient flows, and functional inequalities for stationary point processes. Preprint arXiv:2504.12047 , 2025

2025 arXiv
[21]

S. Jansen. Gibbsian point processes. Lecture notes https://www.mathematik.uni-muenchen.de/ jansen/gibbspp.pdf, 2018

2018
[22]

Jahnel and Ch

B. Jahnel and Ch. Külske. Attractor properties for irreversible and reversible interacting particle systems. Comm. Math. Phys. , 366(1):139--172, 2019

2019
[23]

Jahnel and J

B. Jahnel and J. Köppl. Dynamical Gibbs variational principles for irreversible interacting particle systems with applications to attractor properties. Ann. Appl. Probab. , 33(6A), 2023

2023
[24]

Jahnel and J

B. Jahnel and J. K \"o ppl. On the long-time behaviour of reversible interacting particle systems in one and two dimensions. Probab. Math. Phys. , 6(2):479--503, 2025

2025
[25]

Jahnel, J

B. Jahnel, J. Köppl, Y. Steenbeck, and A. Zass. The variational principle for a marked G ibss point process with infinite-range multibody interactions. Preprint arXiv:2408.17170 , 2024

2024
[26]

Kallenberg

O. Kallenberg. Random Measures . Academic Press, 1983

1983
[27]

Kondratiev, O

Y. Kondratiev, O. Kutoviy, and S. Pirogov. Correlation functions and invariant measures in continuous contact model. Infin. Dimens. Anal. Quantum Probab. Relat. Top. , 11(02):231–258, 2008

2008
[28]

Kondratiev and A

Y. Kondratiev and A. Skorokhod. On contact processes in continuum. Infin. Dimens. Anal. Quantum Probab. Relat. Top. , 09(02):187--198, 2006

2006
[29]

T. G. Kurtz. Representations of Markov Processes as multiparameter time changes. Ann. Probab. , 8(4), 1980

1980
[30]

H. Künsch. Time reversal and stationary Gibbs measures. Stoch. Process. their Appl. , 17(1):159--166, 1984

1984
[31]

T. M. Liggett. Interacting Particle Systems . Classics in Mathematics . Springer Berlin Heidelberg, Berlin, Heidelberg, 2005

2005
[32]

X. X. Nguyen and H. Zessin. Integral and differential characterizations of the G ibbs process. Math. Nachr. , 88(1):105--115, 1979

1979
[33]

H. Osada. Interacting Brownian motions in infinite dimensions with logarithmic interaction potentials. Ann. Probab. , 41(1), 2013

2013
[34]

M. D. Penrose. Existence and spatial limit theorems for lattice and continuum particle systems . Probab. Surv. , 5:1 -- 36, 2008

2008
[35]

C. Preston. Spatial birth-and-death processes. Bull. Inst. Internat. Statist. , 46(2):371--391, 405--408, 1975

1975
[36]

Poghosyan and H

S. Poghosyan and H. Zessin. Penrose- Stable interactions in classical statistical mechanics. Ann. Henri Poincar \' e , 23(3):739--771, 2021

2021
[37]

D. Ruelle. Existence of a phase transition in a continuous classical system. Phys. Rev. Lett. , 27:1040--1041, 1971

1971
[38]

W. G. Sullivan. Specific information gain for interacting M arkov processes. Z. Wahrscheinlichkeitstheorie Verw. Gebiete , 37(1):77--90, 1976/77

1976

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.

[1] [1]

Boucheron, G

S. Boucheron, G. Lugosi, and P. Massart. Concentration Inequalities : A Nonasymptotic Theory of Independence . Oxford University Press, 2013

work page 2013

[2] [2]

J. T. Chayes, L. Chayes, and R. Koteck \'y . The analysis of the Widom--Rowlinson model by stochastic geometric methods . Comm. Math. Phys. , 172(3):551 -- 569, 1995

work page 1995

[3] [3]

Dello Schiavo, R

L. Dello Schiavo, R. Herry, and K. Suzuki. Wasserstein geometry and Ricci curvature bounds for Poisson spaces. J. \'E c. Polytech. Math. , 11:957--1010, 2024

work page 2024

[4] [4]

M. D. Donsker and S. R. S. Varadhan. Asymptotic evaluation of certain Markov process expectations for large time -- III . Comm. Pure Appl. Math. , 29(4):389--461, 1976

work page 1976

[5] [5]

Dereudre and T

D. Dereudre and T. Vasseur. Existence of G ibbs point processes with stable infinite range interaction. J. Appl. Probab. , 57(3):775–791, 2020

work page 2020

[6] [6]

D. J. Daley and D. Vere-Jones. An Introduction to the Theory of Point Processes. V ol. II . Probability and its Applications (New York). Springer, New York, second edition, 2008

work page 2008

[7] [7]

S. N. Ethier and T. G. Kurtz. Markov Processes . Wiley Series in Probability and Mathematical Statistics: Probability and Mathematical Statistics. John Wiley & Sons Inc., 1986

work page 1986

[8] [8]

A. M. Etheridge and T. G. Kurtz. Genealogical constructions of population models. Ann. Probab. , 47(4), 2019

work page 2019