Gap metrics for stationary point processes and quantitative convexity of the free energy

Bastian M\"uller; Martin Huesmann

REVIEW 2 major objections 3 minor 57 references

Gap metrics for stationary point processes and quantitative convexity of the free energy

T0 review · 2 major / 3 minor · reviewed 2026-08-04 · deepseek-v4-flash

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

Pith's one-line read The paper proves that in a metric comparing gap sequences of stationary point processes on the line, the free energy of a large class of interactions is strictly convex, yielding unique minimizers and exponentially fast gradient-flow conver

desk verdict New gap metric and strict convexity results are real, but the λ-convexity step rests on a false inequality, so the exponential gradient-flow claims are not yet established. read the letter →

arxiv 2509.08659 v1 pith:YXGAUVKQ submitted 2025-09-10 math.PR

Martin Huesmann , Bastian Müller This is my paper · ORCID

classification math.PR MSC 49Q2260G5560K3582B21

keywords stationarypointprocessesgapmetricfreeenergygeodesicconvexityRieszgasgradientflowPalmmeasurespecificentropy

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 introduces a metric between stationary point processes on the real line that compares the ordered gaps between consecutive points rather than absolute positions. Its main theorem states that for a broad class of pairwise interaction potentials—those whose truncated, volume-weighted interaction has a uniform convexity lower bound—the free energy is strictly geodesically convex along this gap metric, with an explicit positive curvature gain. Strict convexity immediately forces uniqueness of the free-energy minimizer, giving a variational characterisation of the corresponding stationary Gibbs point processes; for hypersingular Riesz and Yukawa interactions, finite-moment subspaces additionally support a gradient flow that converges exponentially fast to that minimizer. The same convexity mechanism, applied to a renormalized electric energy, yields uniqueness for one-dimensional log- and Riesz gases with exponent 0 < s < 1.

What carries the argument

The gap coordinate system and the gap Wasserstein metric W_gap,p: every stationary simple point process with a point at 0 is coded by its bi-infinite sequence of gaps, so Palm measures become stationary probability measures on sequences. W_gap,p is the infimum over stationary couplings of |s1 − s1'|^p, and geodesics are linear interpolations of coupled gap sequences. The strict convexity engine is the second-derivative condition g_n''(x) ≥ n f(|x|) on the truncated interaction g_n = φ h_n; combined with the displacement convexity of relative entropy in these coordinates, it converts any two configurations coupled monotonically into a convexity estimate with an explicit quadratic gain.

What would settle it

Take two i.i.d. gap distributions with first-gap laws μ0 = 1/2 δ_{0.5} + 1/2 δ_{1.5} and μ1 = 1/2 δ_{0.75} + 1/2 δ_{1.25}, and set h(s) = s^2, viewed as f^{p/(p−2)} for an allowed f. For the increasing-rearrangement (monotone optimal) coupling, the left side of (5.23), ∫ sup_{z∈[s1,s1']} h(z) dU, equals (h(0.75) + h(1.5))/2 = 1.40625, while max(∫h dμ0, ∫h dμ1) = max(1.25, 1.0625) = 1.25. The stated inequality fails, so the lower bound producing λ > 0 is not justified by the proof as written.

Watch

Extended reading notes

Core claim

The central claim is Theorem 5.13: for superstable regular potentials with g_n''(x) ≥ n f(|x|), any two stationary unit-intensity point processes with finite free energy admit a stationary monotone coupling Q0 of their Palm measures such that the interpolated process P_t satisfies Fβ(P_t) ≤ (1−t)Fβ(P0) + tFβ(P1) − β(1−t)t/2 ∫ (x−y)^2 inf_{z∈[x,y]} f(|z|) Q0(dξ̄). The gain is strictly positive unless P0 = P1, so the free energy has at most one minimizer. Under an additional convexity and finite-moment assumption on the gaps, the gain is bounded below by λ t(1−t)/2 W_gap,p^2 with λ = β a^{−(2−p)/p}; this weak λ-geodesic convexity feeds the metric gradient-flow machinery and yields an energy id

Load-bearing premise

The load-bearing premise is the bound in (5.23): for the optimal stationary coupling of gap laws, the expected worst-case value of the interaction function over the coupled first-gap interval is no larger than the larger of the two marginal expectations. If that bound fails, the conversion from strict convexity to a positive multiple of the squared gap distance falls apart, and with it the exponential convergence rate.

Editorial extensions

If this is right

Unique minimizer of the free energy for hypersingular Riesz (|x|^{−s}, s > 1) and Yukawa potentials, hence a variational characterisation of the corresponding stationary Gibbs point processes.
Existence of curves of maximal slope for Fβ with the local slope as strong upper gradient and a metric energy identity.
Exponential convergence of every such curve to the unique minimizer at an explicit rate depending on β, a, and p.
Talagrand-type and log-Sobolev-type inequalities linking squared gap distance to the free-energy gap and to the local slope.
For 0 < s < 1 Riesz and log-gas interactions, the electric free energy has a unique minimizer, extending the known log-gas uniqueness to long-range Riesz interactions.

Reading between the lines

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

The gap-coordinate interpolation preserves hyperuniformity, so for long-range interactions the convexity argument may extend directly to a larger class of hyperuniform point processes than the bounded-moment spaces used here.
If the optimal-coupling stability problem the authors flag were solved, the same proof would most likely supply λ-convexity and gradient-flow convergence for log- and Riesz gases with 0 < s < 1, not just uniqueness.
For p = 2, the free (entropy-only) case suggests an interacting gap diffusion as the stochastic representation of the gradient flow; the authors remark on the difficulty of translating it back to a Markovian point process.

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

The core of this paper is genuinely new and mostly solid. The gap-based Wasserstein metric W_gap,p is a clean construction that compares relative spacings rather than absolute positions, and the strict geodesic convexity of the free energy with explicit gain (Theorem 5.13) is an important advance. The uniqueness consequences — for hypersingular Riesz/Yukawa and for long-range Riesz/log interactions, recovering and extending EHL21 — are carefully argued and appear correct. The approximation arguments through Georgii's framework and Palm calculus are competently handled.

The soft spot is exactly where the reader and stress-test put it. Inequality (5.23) is false: for convex f^{p/(p-2)}, taking f(x)=x with i.i.d. gap measures μ0=1/2δ_{0.5}+1/2δ_{1.5}, μ1=1/2δ_{0.75}+1/2δ_{1.25}, the monotone optimal coupling gives ∫sup f dU = 1.125, while each marginal integral is 1. This inequality is the step that converts strict convexity into λ-geodesic convexity with λ=βa^{-(2-p)/p}. As a result, Proposition 5.19, Theorem 5.25, Corollary 5.27, and the abstract's claim B are not proven as stated. This is a real, load-bearing flaw, not a cosmetic one.

That said, the defect looks repairable. Since max(a,b) ≤ a+b, the same Hölder argument yields a positive λ with a slightly worse constant (e.g. β(2a)^{(p-2)/p}). The qualitative conclusions — existence of gradient flow curves and exponential convergence — should survive with adjusted constants. So the reader's REJECT is a bit strong; this is a major-revision situation rather than a desk reject. The long-range section is more preliminary, and the authors candidly point out that optimality of the coupling is unknown there, so no gradient flow is claimed.

This paper is for people working in mathematical statistical mechanics and optimal transport of point processes. It deserves a serious referee: the strict convexity and uniqueness results are valuable and likely correct, and the gradient-flow gap is identifiable and plausibly fixable. I would send it to peer review, with the expectation of substantial revision.

Referee Report

2 major / 3 minor

Summary. The paper introduces gap metrics W_gap,p and W_p for stationary simple point processes on R, obtained by coupling the bi-infinite gap sequences associated with Palm measures. It proves that these are isometric geodesic extended metric spaces and that the specific relative entropy is geodesically convex. The main positive results are: (i) Theorem 5.13, a quantitative strict convexity estimate for the free energy for superstable regular potentials satisfying g_n''(x) >= n f(|x|), yielding uniqueness of minimizers; (ii) an extension to logarithmic and long-range Riesz interactions via renormalized electric energy (Theorem 6.3, Corollary 6.6); and (iii) a claimed lambda-geodesic convexity and gradient-flow/exponential-convergence theory in Section 5.1 for a class X_{p,p'}^{a,b}. The paper is clearly written and the construction is novel, but the quantitative lambda-convexity program rests on an inequality in Section 5.1 that is false as stated.

Significance. If the quantitative convexity results were established, the paper would be a substantial contribution: it provides a natural optimal-transport geometry for stationary point processes in one dimension, gives an explicit strict convexity gain for a rich class of interactions, and would yield uniqueness, variational characterizations, and exponential convergence. The strict convexity/uniqueness results (Theorem 5.13, Corollary 5.14, Theorem 6.3, Corollary 6.6) appear carefully argued and are significant on their own. The long-range extension to log and Riesz 0<s<1 is valuable. The main advertised lambda-convexity and exponential rate results, however, are not supported by the current proof because of the false inequality (5.23). The defect is local and repairable, but the constants and the statements of Corollary 5.16, Proposition 5.19, Corollary 5.27, and Corollary 5.29 need revision.

major comments (2)

[Section 5.1, Eq. (5.23)] Inequality (5.23) is false. For 1<p<2, take f(x)=x^{(p-2)/p}, so g=f^{p/(p-2)} is g(x)=x, which is convex. Let the two gap distributions be i.i.d. with common one-dimensional marginals mu0 = 1/2 delta_{0.5}+1/2 delta_{1.5} and mu1 = 1/2 delta_{0.75}+1/2 delta_{1.25}. The monotone quantile coupling, which is the W_gap,p-optimal coupling, pairs 0.5 with 0.75 and 1.5 with 1.25. Then the left side of (5.23) equals (0.75+1.5)/2 = 1.125, whereas each marginal integral of g is 1 and the right side is max(1,1)=1. This is exactly the class of couplings used in the subsequent Hölder argument. Consequently Corollary 5.16, Proposition 5.19, and the exponential rates in Corollary 5.27 are not established with the stated lambda = beta a^{-(2-p)/p}.
[Section 5.1, Corollary 5.16 and Proposition 5.19] The assertion that the displayed calculation yields W_p(P0,P1)<infinity and the claimed lambda constant depends entirely on (5.23). With the false inequality removed, the argument still yields a positive lambda through the bound max(a,b) <= a+b, giving integral sup g dU <= 2a and hence a smaller positive constant. Thus the qualitative lambda-convexity and exponential convergence may survive, but the class X_{p,p'}^{a,b}, the definition of lambda, and all quantitative bounds in Corollary 5.16, Proposition 5.19, Corollary 5.27, and Corollary 5.29 must be re-derived. As written, the central quantitative claims of Section 5.1 are unproved.

minor comments (3)

[Corollary 5.29, Eq. (5.29)] In the displayed inequality, f(s_0) should presumably be f(s_1) (or f(|s_1|)), matching the moment condition used to define a.
[Introduction, Theorem 1.1] The notation bar-xi in the gain term is used before the coupling formalism is introduced; a forward reference to Definition 2.4 or the paragraph after (1.1) would help the reader.
[Section 5.1, line before (5.23)] The phrase 'if the function f^{p/(p-2)} is convex' is not a harmless remark: the displayed inequality is exactly what is being asserted, and it is false. Consider replacing it with the valid sum bound and adjusting the constants accordingly.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the convexity theorem is derived from explicit assumptions and classical tools; the false inequality (5.23) is a correctness error, not a circular step.

full rationale

The paper's central derivation chain is self-contained rather than circular. The free energy F_beta = beta W_int + E is defined independently, and the main strict-convexity bound (Theorem 5.13) is obtained by combining (i) the geodesic convexity of the specific entropy (Theorem 3.12), proved from the classical displacement convexity of relative entropy, and (ii) Lemma 5.5, which derives the explicit quadratic gain from the standing assumption g''_n(x) >= n f(|x|). No fitted parameter is renamed as a prediction, and the gain term is computed directly from the coupling, not posited as the conclusion. The later λ-convexity result (Corollary 5.16, Proposition 5.19) does rely on inequality (5.23), but that inequality is presented as a mathematical estimate to be proved; the reader's counterexample shows it is false. A false estimate is a correctness gap, not circularity: the target result is not used as an input. The paper also contains honest limitations -- Remark 5.26 disclaims knowledge of a stochastic representation and Remark 6.4 explicitly says the long-range coupling cannot yet be shown optimal, so no λ-convexity is claimed there. Self-citations to the authors' earlier work [EHJM25] appear, but they are used for the overall strategy and for a technical entropy-uniform-integrability lemma, while the metric construction itself is re-proved in Theorem 3.5, and the cited lemma is published external mathematical support rather than a restatement of the paper's own conclusion. Thus no load-bearing step reduces by definition or by self-citation to its own input. The score of 1 reflects only the presence of minor, non-load-bearing self-citations, not any circular derivation.

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

The paper introduces a new mathematical metric (gap metric) and a restriction space X^{a,b}_{p,p'}, but no physical entities such as particles, forces, or dimensions.

assumptions (4)

standard math Displacement convexity of relative entropy along ell^p-Wasserstein geodesics
Used in Theorem 3.12 to prove geodesic convexity of the specific entropy; cited from [Vil03, Remark 5.16].
standard math Georgii's variational representation and lower semicontinuity of the free energy [Geo94, Theorem 1 and Lemma 3.4]
Invoked in Theorem 5.10 to define W_int and obtain F_beta lower semicontinuity; the paper does not reprove these.
standard math Refined Campbell theorem and Mecke inversion [LP18, Theorem 9.1; Mec67]
Basis for Palm-gap correspondence and for representation (5.3) of interaction energy.
standard math Metric gradient flow theory of Ambrosio-Gigli-Savare [AGS08, Chapter 2]
Used in Theorem 5.25 and Corollary 5.27 for existence of curves of maximal slope and exponential decay; requires lambda-convexity which is not properly established.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Gap metrics for stationary point processes and quantitative convexity of the free energy." pith.science (2026). https://pith.science/paper/YXGAUVKQ

@misc{pith2026250908659,
  author       = {Pith},
  title        = {Pith review of: Gap metrics for stationary point processes and quantitative convexity of the free energy},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/YXGAUVKQ}},
  note         = {Machine review of arXiv:2509.08659}
}

abstract

In this article, we are interested in convexity properties of the free energy for stationary point processes on $\mathbb R$ w.r.t.\ a new geometry inspired by optimal transport. We will show for a rich class of pairwise interaction energies A) quantified strict convexity of the free energy implying uniqueness of minimizers B) existence of a gradient flow curve of the free energy w.r.t. the new metric converging exponentially fast to the unique minimizer. Examples for energies for which A holds include logarithmic or Riesz interactions with parameter $0<s<1$, examples for which A and B hold are hypersingular Riesz or Yukawa interactions.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

57 extracted references · 4 linked inside Pith

[1]

urich. Birkh\

L. Ambrosio, N. Gigli, and G. Savar \'e . Gradient flows in metric spaces and in the space of probability measures . Lectures in Mathematics ETH Z\"urich. Birkh\"auser Verlag, Basel, second edition, 2008

2008
[2]

Albeverio, Y

S. Albeverio, Y. G. Kondratiev, and M. R \"o ckner. Analysis and geometry on configuration spaces. J. Funct. Anal. , 154(2):444--500, 1998

1998
[3]

Albeverio, Y

S. Albeverio, Y. G. Kondratiev, and M. R \"o ckner. Analysis and geometry on configuration spaces: The G ibbsian case. J. Funct. Anal. , 157(1):242--291, 1998

1998
[4]

Ajtai, J

M. Ajtai, J. Koml \'o s, and G \'a bor Tusn \'a dy. On optimal matchings. Combinatorica , 4:259--264, 1984

1984
[5]

Ambrosio, F

L. Ambrosio, F. Stra, and D. Trevisan. A PDE approach to a 2-dimensional matching problem. Probab. Theory Relat. Fields , 173(1-2):433--477, 2019

2019
[6]

Butez, S

R. Butez, S. Dallaporta, and D. Garc \' a-Zelada. On the Wasserstein distance between a hyperuniform point process and its mean. Preprint, arXiv :2404.09549 (2024), 2024

arXiv 2024
[7]

Boursier

J. Boursier. Decay of correlations and thermodynamic limit for the circular Riesz gas. Preprint, arXiv :2209.00396 (2022), 2022

arXiv 2022
[8]

Br \'e maud

P. Br \'e maud. Point process calculus in time and space. An introduction with applications , volume 98 of Probab. Theory Stoch. Model. Cham: Springer, 2020

2020

Show all 57 references

[9]

Caracciolo, C

S. Caracciolo, C. Lucibello, G. Parisi, and G. Sicuro. Scaling hypothesis for the E uclidean bipartite matching problem. Physical Review E , 90(1), 2014

2014
[10]

Clozeau and F

N. Clozeau and F. Mattesini. Annealed quantitative estimates for the quadratic 2d-discrete random matching problem. Probab. Theory Relat. Fields , 190(1-2):485--541, 2024

2024
[11]

Chatterjee, R

S. Chatterjee, R. Peled, Y. Peres, and D. Romik. Gravitational allocation to P oisson points. Ann. of Math. (2) , 172(1):617--671, 2010

2010
[12]

Dereudre

D. Dereudre. Variational principle for Gibbs point processes with finite range interaction. Electron. Commun. Probab. , 21:11, 2016. Id/No 10

2016
[13]

Dereudre

D. Dereudre. Introduction to the theory of G ibbs point processes. Stochastic Geometry: Modern Research Frontiers , pages 181--229, 2019

2019
[14]

Dereudre, D

D. Dereudre, D. Flimmel, M. Huesmann, and T. Lebl \'e . ( N on)-hyperuniformity of perturbed lattices. Preprint, arXiv :2405.19881 (2025), 2025

2025 arXiv
[15]

Dereudre and H.-O

D. Dereudre and H.-O. Georgii. Variational characterisation of Gibbs measures with Delaunay triangle interaction. Electron. J. Probab. , 14:2438--2462, 2009

2009
[16]

Daneri and G

S. Daneri and G. Savar \' e . Eulerian calculus for the displacement convexity in the wasserstein distance. SIAM Journal on Mathematical Analysis , 40(3):1104--1122, jan 2008

2008
[17]

Wasserstein geometry and Ricci curvature bounds for Poisson spaces

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

2024
[18]

D. J. Daley and D. Vere-Jones. An Introduction to the Theory of Point Processes. Volume II: General Theory and Structure . Springer, 2008

2008
[19]

G. Edgar. Measurability in a B anach space. Indiana Univ. Math. J. , 26:663--677, 1977

1977
[20]

Erbar, M

M. Erbar, M. Huesmann, J. Jalowy, and B. M \"u ller. Optimal transport of stationary point processes: metric structure, gradient flow and convexity of the specific entropy. J. Funct. Anal. , 289(4):67, 2025. Id/No 110974

2025
[21]

Erbar, M

M. Erbar, M. Huesmann, and T. Lebl \'e . The one-dimensional log-gas free energy has a unique minimizer. Commun. Pure Appl. Math. , 74(3):615--675, 2021

2021
[22]

Figalli and F

A. Figalli and F. Glaudo. An invitation to optimal transport, Wasserstein distances, and gradient flows . EMS Textb. Math. Berlin: European Mathematical Society (EMS), 2021

2021
[23]

Friedli and Y

S. Friedli and Y. Velenik. Statistical mechanics of lattice systems. A concrete mathematical introduction . Cambridge: Cambridge University Press, 2018

2018
[24]

H.-O. Georgii. Large deviations and the equivalence of ensembles for Gibbsian particle systems with superstable interaction. Probab. Theory Relat. Fields , 99(2):171--195, 1994

1994
[25]

H.-O. Georgii. Gibbs measures and phase transitions , volume 9. Walter de Gruyter, 2011

2011
[26]

Goldman and M

M. Goldman and M. Huesmann. A fluctuation result for the displacement in the optimal matching problem. Ann. Probab. , 50(4):1446--1477, 2022

2022
[27]

Goldman and D

M. Goldman and D. Trevisan. Convergence of asymptotic costs for random Euclidean matching problems. Probab. Math. Phys. , 2(2):121--142, 2021

2021
[28]

Georgii and H

H.-O. Georgii and H. Zessin. Large deviations and the maximum entropy principle for marked point random fields. Probab. Theory Relat. Fields , 96(2):177--204, 1993

1993
[29]

A. E. Holroyd, S. Janson, and J. W\" a stlund. Minimal matchings of point processes. Probab. Theory Related Fields , 184(1-2):571--611, 2022

2022
[30]

Huesmann and T

M. Huesmann and T. Lebl \'e . The link between hyperuniformity, Coulomb energy, and Wasserstein distance to Lebesgue for two-dimensional point processes. Preprint, arXiv :2404.18588 (2025), 2025

2025
[31]

Huesmann and B

M. Huesmann and B. M \"u ller. A Benamou - Brenier formula for transport distances between stationary random measures. Stochastic Processes Appl. , 185:21, 2025. Id/No 104633

2025
[32]

A. E. Holroyd and Y. Peres. Extra heads and invariant allocations. Ann. Probab. , 33(1):31--52, 2005

2005
[33]

A. E. Holroyd, R. Pemantle, Y. Peres, and O. Schramm. P oisson matching. Ann. Inst. Henri Poincar\'e Probab. Stat. , 45(1):266--287, 2009

2009
[34]

Huesmann and K.-T

M. Huesmann and K.-T. Sturm. Optimal transport from Lebesgue to Poisson . Ann. Probab. , 41(4):2426--2478, 2013

2013
[35]

Huesmann and H

M. Huesmann and H. Stange. Non-local Wasserstein Geometry , Gradient Flows , and Functional Inequalities for Stationary Point Processes . Preprint, arXiv :2504.12047 (2025), 2025

2025 arXiv
[36]

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

2018
[37]

The variational principle for a marked Gibbs point process with infinite-range multibody interactions

B.\ Jahnel, J.\ K \"o ppl, Y.\ Steenbeck, and A.\ Zass. The variational principle for a marked Gibbs point process with infinite-range multibody interactions. Preprint, arXiv :2408.17170 (2024), 2024

2024
[38]

Kallenberg

O. Kallenberg. Foundations of modern probability , volume 2. Springer, 1997

1997
[39]

Lebl \'e

T. Lebl \'e . Logarithmic, Coulomb and Riesz energy of point processes. J. Stat. Phys. , 162(4):887--923, 2016

2016
[40]

M. Ledoux. On optimal matching of G aussian samples. Zap. Nauchn. Sem. S.-Peterburg. Otdel. Mat. Inst. Steklov. (POMI) , 457(Veroyatnost' \ i Statistika. 25):226--264, 2017

2017
[41]

Last and M

G. Last and M. Penrose . Lectures on the Poisson process. , volume 7. Cambridge: Cambridge University Press, 2018

2018
[42]

Lachi \`e ze-Rey and D

R. Lachi \`e ze-Rey and D. Yogeshwaran. Hyperuniformity and optimal transport of point processes. arXiv :2402.13705 (2024), 2024

2024
[43]

Lebl \'e and S

T. Lebl \'e and S. Serfaty. Large deviation principle for empirical fields of L og and R iesz gases. Invent. math. , 210(3):645--757, 2017

2017
[44]

Last and H

G. Last and H. Thorisson . Invariant transports of stationary random measures and mass-stationarity. Ann. Probab. , 37(2):790--813, 2009

2009
[45]

a re zuf \

J. Mecke. Station \"a re zuf \"a llige Ma e auf lokalkompakten Abelschen Gruppen . Z. Wahrscheinlichkeitstheor. Verw. Geb. , 9:36--58, 1967

1967
[46]

H. Osada. Infinite-dimensional stochastic differential equations related to random matrices. Probab. Theory Related Fields , 153(3-4):471--509, 2012

2012
[47]

H. Osada. Interacting B rownian motions in infinite dimensions with logarithmic interaction potentials. Ann. Probab. , 41(1):1--49, 2013

2013
[48]

Pilipenko

A. Pilipenko. An Introduction to S tochastic D ifferential E quations with Reflection . Lectures in pure and applied mathematics. University Press Potsdam, 2014

2014
[49]

Petrache and S

M. Petrache and S. Serfaty. Next order asymptotics and renormalized energy for Riesz interactions. J. Inst. Math. Jussieu , 16(3):501--569, 2017

2017
[50]

Rassoul-Agha and T

F. Rassoul-Agha and T. Sepp \"a l \"a inen. A course on large deviations with an introduction to Gibbs measures , volume 162. American Mathematical Soc., 2015

2015
[51]

Santambrogio

F. Santambrogio. Optimal transport for applied mathematicians. Birk \"a user, NY , 55(58-63):94, 2015

2015
[52]

S. Serfaty. Microscopic description of log and Coulomb gases. In Random matrices , pages 341--387. Providence, RI: American Mathematical Society (AMS); Princeton, NJ: Institute for Advanced Study (IAS), 2019

2019
[53]

Sandier and S

\'E . Sandier and S. Serfaty. From the G inzburg- L andau model to vortex lattice problems. Comm. Math. Phys. , 313(3):635--743, 2012

2012
[54]

K. Suzuki. Curvature bound of Dyson Brownian motion. Commun. Math. Phys. , 406(7):56, 2025. Id/No 154

2025
[55]

L.-C. Tsai. Infinite dimensional stochastic differential equations for D yson's model. Probab. Theory Related Fields , 166(3-4):801--850, 2016

2016
[56]

C. Villani. Topics in Optimal Transportation . Graduate studies in mathematics. American Mathematical Society, 2003

2003
[57]

C. Villani. Optimal transport: old and new , volume 338. Springer, 2009

2009

Pith tools

Reviewed August 4, 2026 · model on record in the stance chip above.

[1] [1]

urich. Birkh\

L. Ambrosio, N. Gigli, and G. Savar \'e . Gradient flows in metric spaces and in the space of probability measures . Lectures in Mathematics ETH Z\"urich. Birkh\"auser Verlag, Basel, second edition, 2008

2008

[2] [2]

Albeverio, Y

S. Albeverio, Y. G. Kondratiev, and M. R \"o ckner. Analysis and geometry on configuration spaces. J. Funct. Anal. , 154(2):444--500, 1998

1998

[3] [3]

Albeverio, Y

S. Albeverio, Y. G. Kondratiev, and M. R \"o ckner. Analysis and geometry on configuration spaces: The G ibbsian case. J. Funct. Anal. , 157(1):242--291, 1998

1998

[4] [4]

Ajtai, J

M. Ajtai, J. Koml \'o s, and G \'a bor Tusn \'a dy. On optimal matchings. Combinatorica , 4:259--264, 1984

1984

[5] [5]

Ambrosio, F

L. Ambrosio, F. Stra, and D. Trevisan. A PDE approach to a 2-dimensional matching problem. Probab. Theory Relat. Fields , 173(1-2):433--477, 2019

2019

[6] [6]

Butez, S

R. Butez, S. Dallaporta, and D. Garc \' a-Zelada. On the Wasserstein distance between a hyperuniform point process and its mean. Preprint, arXiv :2404.09549 (2024), 2024

arXiv 2024

[7] [7]

Boursier

J. Boursier. Decay of correlations and thermodynamic limit for the circular Riesz gas. Preprint, arXiv :2209.00396 (2022), 2022

arXiv 2022

[8] [8]

Br \'e maud

P. Br \'e maud. Point process calculus in time and space. An introduction with applications , volume 98 of Probab. Theory Stoch. Model. Cham: Springer, 2020

2020