Pith. sign in

REVIEW 4 major objections 3 minor 44 references

Tractable description of hydrodynamic limits of a class of interacting jump processes on sparse graphs

T0 review · 4 major / 3 minor · reviewed 2026-08-08 · deepseek-v4-flash

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

Pith's one-line read The paper claims that for interacting jump processes on sparse graphs with acyclic single-particle transitions, the law of a typical vertex neighborhood at every time is exactly the solution of a finite coupled ODE system.

desk verdict The paper genuinely extends the Markov local-field ODE reduction to acyclic transition graphs, but the proof of the key time-marginal 2-MRF property has a real gap that needs fixing. read the letter →

arxiv 2502.05156 v1 pith:XOPLMFUU submitted 2025-02-07 math.PR

Juniper Cocomello , Michel Davydov , Kavita Ramanan This is my paper · ORCID
classification math.PR MSC 60K3560J7460J27
keywords interactingparticlesystemshydrodynamiclimitslocal-fieldequationsMarkovianprojectionMarkovrandomfieldssparsegraphsGalton-Watsontreesjumpprocesses
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 aims to make the hydrodynamic limit of interacting jump processes on sparse random graphs tractable. It claims that when the single-particle transition graph is acyclic, the time-marginal law of the root neighborhood on a unimodular Galton-Watson tree—the local limit of configuration-model networks—is governed exactly by a finite coupled system of ordinary differential equations, the Kolmogorov forward equations of a Markov local-field process. The payoff is that one can compute the macroscopic empirical distribution of vertex states without simulating the full non-Markovian dynamics, even when interactions are nonlinear and not pairwise additive. The proof combines a local-field equation characterization, a Markovian projection that matches time marginals, and a newly established time-marginal second-order Markov random field property. If the theorem is right, a class of previously intractable spreading processes reduces to an ODE initial-value problem.

What carries the argument

The central object is the Markov local-field equation (MLFE), an SDE whose jump rates are conditional expectations of the local-field rates given only the present root-neighborhood state rather than the full trajectory. Its forward Kolmogorov equations become the ODE system (2.11)–(2.12). The passage from the non-Markovian local-field equation to the MLFE rests on two ingredients: a Markovian projection theorem for pure jump processes, which constructs a Markov process with the same time marginals, and a time-marginal second-order Markov random field property, which says that conditional on the present state of the double boundary of a set, the inside and outside of that set are independent. The acyclic transition graph assumption is what allows the authors to derive this marginal 2-MRF property from a trajectorial version up to stopping times.

What would settle it

Construct a concrete finite example satisfying Assumptions A through D in which $X_A(\bar\tau)$ and $X_B(\bar\tau)$ are conditionally independent given both $(X_{\partial^2 A}(\bar\tau-), X_{\partial^2 A}(\bar\tau))$ but not given $X_{\partial^2 A}(\bar\tau)$ alone; such a counterexample would invalidate the key implication around equation (5.20) and with it the claimed time-marginal 2-MRF property.

Watch

Extended reading notes

Core claim

Under Assumptions A, B, C and D, the time marginals of the root neighborhood are characterized by the Markov local-field equation, whose jump rates depend only on the current configuration and its one-time law. The paper proves that the corresponding forward Kolmogorov ODE system (2.11)–(2.12) has a unique solution, and that this solution equals $\mathrm{L}(X_{V_1^\theta}(t))$ for every $t \ge 0$. Consequently, for configuration-model graphs converging to a unimodular Galton-Watson tree, the empirical distribution of vertex states and root-neighborhood configurations converges in probability to this ODE solution. The key structural assumption is that the directed transition graph of a single particle is acyclic, which forces the double boundary of any set to jump only finitely many times; this is what lets the authors pass from a trajectorial 2-MRF property to a time-marginal 2-MRF property, the ingredient that makes the Markovian projection autonomous.

Load-bearing premise

The argument depends on the claim that once one knows the boundary's state right after a jump, the inside and outside of any separated region are independent—the paper asserts this follows from a two-sided conditioning but supplies no proof, and if it fails the main theorem does not follow.

Editorial extensions

If this is right

  • For configuration-model graphs with a fixed finite-support degree sequence, the empirical distribution of vertex states converges in probability to the ODE solution, so the hydrodynamic limit is explicitly computable.
  • The result applies to nonlinear, non-pairwise-linear interactions: the seizure-propagation model, the entrenched majority voter model, and thresholded multivariate Hawkes processes all satisfy the acyclicity assumption and are covered.
  • The trajectory laws of the local-field and Markov local-field processes can differ; only the time marginals are asserted to coincide.
  • On a $\kappa$-regular tree, the ODE reduces from $m^{\kappa+1}$ to $m\binom{\kappa+m-1}{m-1}$ equations, giving a dimension reduction.
  • Finite graphs can be approximated by solving the ODE with the empirical degree distribution substituted for $\theta$, as in Section 3.1.

Reading between the lines

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

  • Beyond the paper: the time-marginal 2-MRF mechanism, if made fully rigorous, would likely extend to any finite-state process in which the double boundary of every set makes only finitely many jumps, not necessarily under a globally acyclic transition graph.
  • Beyond the paper: the Markovian projection theorem stated for general bounded trajectory-dependent jump rates is a standalone tool that could be applied to other history-dependent intensity models outside the local-field setting.
  • Beyond the paper: the ODE approximation algorithm of Section 3.1 invites a quantitative finite-size error analysis, since the simulations in the paper suggest accuracy at modest graph sizes but provide no rigorous convergence-rate guarantee.
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

4 major / 3 minor

Summary. The paper studies Markov interacting pure jump processes on unimodular Galton-Watson trees, assuming a finite state space and an acyclic directed transition graph for single-particle transitions (Assumption D). The main result, Theorem 2.8, claims that the time-t law of the root together with its neighbors is exactly the solution of a finite coupled ODE system, equations (2.11)-(2.12). The proof combines a local-field equation characterization from a companion paper, a new time-marginal second-order Markov random field property, a Markovian projection theorem for pure jump processes, and a Lipschitz well-posedness analysis for the ODE. The paper also derives a hydrodynamic-limit corollary on configuration-model graphs and illustrates the approximation with simulations for a seizure-spread model, an entrenched majority voter model, and a thresholded Hawkes model.

Significance. If the main theorem is correct, this is a significant advance: it gives a finite-dimensional autonomous ODE description of the hydrodynamic limit for a class of interacting jump processes with nonlinear, non-pairwise-linear rates, going beyond the SIR/SIER results that rely on pairwise linear structure. The ODE is derived from the given rates without fitted parameters. The claimed time-marginal 2-MRF property and the general Markovian projection result are potentially of independent interest. The paper is well organized and the simulation studies are a useful sanity check. However, the proof of the time-marginal 2-MRF property has a load-bearing gap, and the main theorem depends on results from an unpublished companion paper, so the central claim is not fully established in the present manuscript.

major comments (4)
  1. [Section 5.2, after Eq. (5.20)] The assertion 'This in turn implies' is not justified. Lemma 5.12 gives conditional independence of X_A^-(tau-bar) and X_B^-(tau-bar) given both X_{d2A}^-(tau-bar) and X_{d2A}(tau-bar), but equation (5.20) only defines X_{d2A}(tau-bar) as a function of X_{d2A}^-(tau-bar), X_A^-(tau-bar), and X_B^-(tau-bar). Conditional independence given a pair does not imply conditional independence given one component when that component is a function of the other variables. For example, with independent Bernoulli variables C, A, B and D = (A XOR C, B XOR C), A is independent of B given (C,D) but not given D alone. This gap affects the equality of the rates gamma and rho-hat in the proof of Theorem 2.8, and hence Theorem 4.7 is not established by the given argument.
  2. [Section 5.2, Lemma 5.12] The proof states that conditioning on the stopped augmented process at time t is the same as conditioning on the full stopped trajectory because there are no jumps on the double boundary before tau-bar. For t < tau-bar this is correct, but at t = tau-bar the stopped trajectory includes the time of the first jump, which is not a measurable function of the pre- and post-jump states alone. Thus the sigma-algebra generated by Y_{d2A}^{tau-bar}(tau-bar) is generally smaller than that generated by Y_{d2A}^{tau-bar}[tau-bar], and conditional independence does not automatically transfer from the larger to the smaller sigma-algebra. An additional argument is required at the jump time.
  3. [Section 7, Eq. (7.1)] In the proof of Proposition 2.6, the denominator D in the definition of Psi-bar has indicators {b_empty = a1, b1 = a0}, while delta(q) immediately below is defined using {c_empty = a0, c1 = a1}. Unless a0 = a1, the inequalities delta(q) <= D <= d_max delta(q) do not follow. The Lipschitz bound therefore is not derived as written. Please correct the definition of delta(q) or the indicators in (7.1) and redo the argument.
  4. [Sections 4.1.1 and 4.4] The proof of Theorem 2.8 relies on Theorem 4.3 and Proposition 4.10, which are deferred to the companion paper [22], listed as 'in preparation.' Theorem 4.8 also invokes well-posedness 'from considerations equivalent to Proposition 4.10.' Hence the main theorem is not self-contained and cannot be fully verified from the present manuscript. The authors should provide the missing proofs or clarify exactly which results are assumed and make the companion paper available.
minor comments (3)
  1. [Section 4.4, Eq. (4.6)] In the definition of rho-hat for v not equal to the root, the conditioning is written as Xhat_empty(t-) = y_v and Xhat_1(t-) = y_empty, which appears swapped relative to the conditional expectation in (4.9), where the conditioning is on X_empty[t] = y_empty[t] and X_v[t] = y_v[t]. Please clarify the intended convention or correct the typo.
  2. [Section 3.2.1] In the seizure-propagation rate (3.3), the rate for state y(s-) = 1 contains a division by d - (1 + sum alpha^-_w 1{x_w=0}); please clarify that the denominator is positive under the intended graph and parameter assumptions, or state how the zero case is handled.
  3. [Section 7] The exposition of the Lipschitz proof would benefit from explicitly defining delta(q) with the same ordering of the root and neighbor states used in the denominator D, and from stating upfront that the permutation symmetry in Remark 2.9 is not being used to swap the root and a neighbor.

Circularity Check

2 steps flagged · score 4.0 of 10

Self-citation chain supplies the two load-bearing uniqueness/equivalence steps, but the ODE derivation is not a fit; the flagged (5.20) implication is a non-circular proof gap.

  1. self citation load bearing [Section 4.1.1, Theorem 4.3 (proof), applied at start of the proof of Theorem 2.8 in Section 4.4]
    "Theorem 4.3. Suppose that Assumptions A, B and C hold. Then, we have L(XVθ1) = L(X̃). Proof. This is established in [22]. For the special case of a regular tree, see [19]."

    This equality is the first link in the main derivation: the proof of Theorem 2.8 begins 'By Theorem 4.3, L(XVθ1) = L(X̃), where X̃ is the solution of the local-field equations (4.1).' The paper does not prove Theorem 4.3; it delegates it to [22], an in-preparation preprint by Ganguly and Ramanan, and Ramanan is a coauthor of the present paper. Everything after this point transports that equality through the Markovian projection and the forward Kolmogorov equations. So the central premise of the claimed tractable ODE description is load-bearing on an unpublished same-author citation rather than on a derivation contained in this manuscript. The cited statement does not by itself contain the ODE conclusion, so this is self-citation load-bearing rather than definitional circularity.

  2. uniqueness imported from authors [Section 4.3, Proposition 4.10, invoked in the proof of Theorem 2.8 in Section 4.4]
    "Proposition 4.10. Under Assumptions A, B and C, the Markov local-field equation (4.5) is well-posed. Proof. The well-posedness stems from considerations identical to the well-posedness of the local-field equation from Definition 4.1, for which we refer to [22]."

    The proof of the main theorem uses this imported uniqueness to close the argument: 'By well-posedness of the Markov local-field equations given in Proposition 4.10, it is enough to show that the rates γ and ρ̂ coincide.' The uniqueness that converts 'the rates agree' into 'the laws agree' is not derived here; it is referred to the same in-preparation, same-author preprint [22]. Thus the main theorem's equality of time marginals is forced through an author-supplied uniqueness theorem rather than through an external, fully verified mathematical fact. The ODE system itself is independently derived from the Markov local-field generator, so the circularity is partial, not total.

full rationale

The core derivation of Theorem 2.8 is not a fitted-input or definitional circularity: no parameter is calibrated to the conclusion, and equations (2.11)-(2.12) are obtained by writing the forward Kolmogorov equations of the Markov local-field process. The circularity burden is instead concentrated in the self-citation chain. Both the initial identification of the original dynamics with the local-field equation (Theorem 4.3) and the well-posedness/uniqueness of the Markov local-field equation (Proposition 4.10) are attributed to [22], an unpublished preprint by the same research group, and the proof of Theorem 2.8 explicitly reduces the main conclusion to these imported results. The time-marginal 2-MRF step contains a separate mathematical defect that is not circular: after equation (5.20), the paper asserts that conditioning on the post-jump boundary state alone preserves the conditional independence, but this does not follow from the stated independence given both pre- and post-jump boundary states; an additional property of f∂2A would be needed. That is an omitted proof or correctness gap in Theorem 4.7, not an input-output equivalence, so it does not by itself raise the circularity score.

Assumptions & free parameters 0 free parameters · 7 assumptions · 0 invented entities

No free parameters are fitted to data; the only inputs are the model's transition rates, degree distribution, and initial law. No new physical particles, forces, or dimensions are introduced. The Markov local-field equation is a mathematical construction, not a postulated entity requiring independent evidence.

assumptions (7)
  • domain assumption The graph is a unimodular Galton-Watson tree with finite-support offspring distribution (Assumption A).
    The ODE and local-field framework are formulated on this limit graph; finite support keeps the configuration space finite and the ODE finite-dimensional.
  • domain assumption Initial conditions form a 2-MRF and the marked tree is unimodular (Assumption B).
    These conditions are needed for the pathwise 2-MRF property and for mass-transport arguments used in the proof.
  • domain assumption Transition rates are cadlag and satisfy a sub-exponential degree-dependent bound (Assumption C).
    This ensures strong well-posedness of the interacting jump SDE via the cited result from [21].
  • ad hoc to paper The directed transition graph G_rho is finite and acyclic (Assumption D).
    This is the novel structural restriction. It guarantees only finitely many jumps on a finite boundary, which is used to prove the time-marginal 2-MRF property in Theorem 4.7.
  • domain assumption The local-field equation characterization L(X_{V_1^theta}) = L(tilde X) holds (Theorem 4.3, cited from [22]).
    The paper takes as input the same-authors result, currently listed as 'in preparation', that the root neighborhood process is characterized by the path-dependent local-field SDE.
  • domain assumption Trajectories form a 2-MRF (Theorem 4.5, cited from [20]).
    This posted-but-not-yet-peer-reviewed result is used to extend conditional independence from trajectories to stopped trajectories.
  • standard math Existence of a Markovian projection follows from Larsson-Long [32], and uniqueness follows from SDE well-posedness via Kurtz equivalence.
    External published results are adopted to construct and identify the mimicking Markov process in Theorem 6.3.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Tractable description of hydrodynamic limits of a class of interacting jump processes on sparse graphs." pith.science (2026). https://pith.science/paper/XOPLMFUU

@misc{pith2026250205156,
  author       = {Pith},
  title        = {Pith review of: Tractable description of hydrodynamic limits of a class of interacting jump processes on sparse graphs},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/XOPLMFUU}},
  note         = {Machine review of arXiv:2502.05156}
}
read the original abstract

We consider dynamics of the empirical measure of vertex neighborhood states of Markov interacting jump processes on sparse random graphs, in a suitable asymptotic limit as the graph size goes to infinity. Under the assumption of a certain acyclic structure on single-particle transitions, we provide a tractable autonomous description of the evolution of this hydrodynamic limit in terms of a finite coupled system of ordinary differential equations. Key ingredients of the proof include a characterization of the hydrodynamic limit of the neighborhood empirical measure in terms of a certain local-field equation, well-posedness of its Markovian projection, and a Markov random field property of the time-marginals, which may be of independent interest. We also show how our results lead to principled approximations for classes of interacting jump processes and illustrate its efficacy via simulations on several examples, including an idealized model of seizure spread in the brain.

Figures

Figures reproduced from arXiv: 2502.05156 by the authors.

Figure 1
Figure 1. State space representation for some example dynamics Note that the configuration space of the neighborhood of the root is then C θ := ∪n∈Θ(X n+1 × {⋆} dmax−n ) ⊂ X Vθ 1 ⋆ . For ⃗a ∈ X Vθ 1 ⋆ , we define k(⃗a) := 1{⃗a∈Cθ} (max{v : av ̸= ⋆} − 1), (2.9) that is, k(⃗a) is the degree of ∅ when XVθ 1 = ⃗a. We let P θ := n p ∈ P(X Vθ 1 ⋆ ) : p(C θ ) = 1o . In the following, having identified X with a subset of N, we let (e… view at source ↗
Figure 2
Figure 2. Fraction of individuals in states S and I in the seizure propagation model from Section 3.2.1 on the random 3-regular graph with the number of vertices equal to n = 50 (a), 200 (b), 400 (c). We compare our ODE (dashed lines) with simulations (dotted lines) averaged over 500 runs. where α − w = −αw ∨ 0. This model falls within the class of IPS dynamics (2.2) and satisfies Assumptions A, B, C and D. Unlike classical S… view at source ↗
Figure 3
Figure 3. Fraction of individuals in each state of the entrenched majority voter model from Section 3.2.2 on the random 2−regular graph with 200 vertices. We compare our ODE, simulations (500 realizations), and the mean-field approxima￾tion. 3.2.3. Multivariate Markov Hawkes processes with threshold. Hawkes processes, originally de￾veloped to model earthquakes, have in recent years been extensively used in computational neuro… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

44 extracted references · 41 canonical work pages

  1. [22]

    in preparation

    , Local Field Equations for jump processes on Unimodular Galton-Watson Trees, 2025. in preparation

    work page 2025

  2. [1]

    Agathe-Nerine, Multivariate hawkes processes on inhomogeneous random graphs, Stochastic Processes and their Applications 152 (2022), 86–148

    Z. Agathe-Nerine, Multivariate hawkes processes on inhomogeneous random graphs, Stochastic Processes and their Applications 152 (2022), 86–148

    work page 2022

  3. [2]

    Reza Aghajani and Kavita Ramanan, The hydrodynamic limit of a randomized load balancing network , Ann. Appl. Probab. 29 (2019), no. 4, 2114–2174. MR3984253

    work page 2019

  4. [3]

    Aldous and R

    D. Aldous and R. Lyons, Processes on unimodular random networks , Electronic Journal of Probability 12 (2007), 1454–1508. paper no. 54

    work page 2007

  5. [4]

    Panagiotis Andreou and Mariana Olvera-Cravioto, Opinion dynamics on non-sparse networks with commu- nity structure, 2024

    work page 2024

  6. [5]

    Baccelli and T

    F. Baccelli and T. Taillefumier, Replica-mean-field limits for intensity-based neural networks , SIAM Journal on Applied Dynamical Systems 18 (20191), no. 4, 1756–1797

    work page

  7. [6]

    Simone Baldassarri, Peter Braunsteins, Frank den Hollander, and Michel Mandjes, Opinion dynamics on dense dynamic random graphs , 2024

    work page 2024

  8. [7]

    Benjamini and O

    I. Benjamini and O. Schramm, Recurrence of Distributional Limits of Finite Planar Graphs , Electronic Journal of Probability 6 (2001), no. none, 1 –13

    work page 2001

Show all 44 references
  1. [8]

    Bentata and R

    A. Bentata and R. Cont, Mimicking the marginal distributions of a semimartingale , 2012. preprint arXiv:0910.3992

    2012 arXiv

  2. [9]

    Bordenave, Lecture notes on random graphs and probabilistic combinatorial optimization , 2016

    C. Bordenave, Lecture notes on random graphs and probabilistic combinatorial optimization , 2016

    2016

  3. [10]

    Brunick and S

    G. Brunick and S. Shreve, Mimicking an Itˆ o process by a solution of a stochastic differential equation , The Annals of Applied Probability 23 (2013), no. 4, 1584 –1628

    2013

  4. [11]

    Cocomello, C

    J. Cocomello, C. Li, and K Ramanan, Exact description of limiting generalized epidemic dynamics on locally tree-like graphs (2025). in preparation

    2025

  5. [12]

    Cocomello and K

    J. Cocomello and K. Ramanan, Exact description of limiting sir and seir dynamics on locally tree-like graphs , 2023

    2023

  6. [13]

    Davydov, Propagation of chaos and Poisson hypothesis for replica mean-field models of intensity-based neural networks, The Annals of Applied Probability 34 (2024), no

    M. Davydov, Propagation of chaos and Poisson hypothesis for replica mean-field models of intensity-based neural networks, The Annals of Applied Probability 34 (2024), no. 2, 2107 –2135

    2024

  7. [14]

    Ditlevsen and E

    S. Ditlevsen and E. L¨ ocherbach,Multi-class oscillating systems of interacting neurons , Stochastic Processes and their Applications 127 (2017), no. 6, 1840–1869

    2017

  8. [15]

    R. L. Dobrushin, Vlasov equations, Functional Analysis and Its Applications 13 (1979Apr), no. 2, 115–123

  9. [16]

    Durrett, Probability: Theory and examples , 5th ed., Cambridge Series in Statistical and Probabilistic Mathematics, Cambridge University Press, 2019

    R. Durrett, Probability: Theory and examples , 5th ed., Cambridge Series in Statistical and Probabilistic Mathematics, Cambridge University Press, 2019

    2019

  10. [17]

    Fournier and E

    N. Fournier and E. L¨ ocherbach,On a toy model of interacting neurons , Annales de l’Institut Henri Poincar´ e, Probabilit´ es et Statistiques52 (2016), no. 4, 1844 –1876

    2016

  11. [18]

    TRACTABLE DESCRIPTION OF HYDRODYNAMIC LIMITS OF IPS ON SPARSE GRAPHS 29

    Nicolas Fraiman, Tzu-Chi Lin, and Mariana Olvera-Cravioto, Opinion dynamics on directed complex net- works, 2024. TRACTABLE DESCRIPTION OF HYDRODYNAMIC LIMITS OF IPS ON SPARSE GRAPHS 29

    2024

  12. [19]

    Ganguly, Non-Markovian interacting particle systems on large sparse graphs: Hydrodynamic limits and marginal characterizations, Ph.D

    A. Ganguly, Non-Markovian interacting particle systems on large sparse graphs: Hydrodynamic limits and marginal characterizations, Ph.D. Thesis, 2022. Brown Digital Repository, Brown University Library

    2022

  13. [20]

    Ganguly and K

    A. Ganguly and K. Ramanan, Interacting jump processes preserve semi-global Markov random fields on path space, 2022. Preprint arXiv:2210.09253

    2022 arXiv

  14. [21]

    185, 1 –63

    , Hydrodynamic limits of non-markovian interacting particle systems on sparse graphs , Electronic Journal of Probability 29 (2024), no. 185, 1 –63

    2024

  15. [23]

    preprint

    , Marginal dynamics of interacting jump processes on regular trees: stationarity properties and Markov local-field equations, 2025. preprint

    2025

  16. [24]

    Gy¨ ongy, Mimicking the one-dimensional marginal distributions of processes having an ito differential , Probability Theory and Related Fields 71 (1986), no

    I. Gy¨ ongy, Mimicking the one-dimensional marginal distributions of processes having an ito differential , Probability Theory and Related Fields 71 (1986), no. 4, 501–516

    1986

  17. [25]

    Remco van der Hofstad, Random graphs and complex networks , Cambridge Series in Statistical and Proba- bilistic Mathematics, Cambridge University Press, 2016

    2016

  18. [26]

    Jabin, D

    P. Jabin, D. Poyato, and J. Soler, Mean-field limit of non-exchangeable systems , 2022. 2112.15406

    2022

  19. [27]

    I: Theory, Wiley Interscience, 1975

    Leonard Kleinrock, Queueing Systems , Vol. I: Theory, Wiley Interscience, 1975

    1975

  20. [28]

    T. G. Kurtz, Equivalence of stochastic equations and martingale problems (Dan Crisan, ed.), Springer Berlin Heidelberg, Berlin, Heidelberg, 2011

    2011

  21. [29]

    K¨ opfer and L

    B. K¨ opfer and L. R¨ uschendorf,Markov projection of semimartingales — application to comparison results , Stochastic Processes and their Applications 162 (2023), 361–386

    2023

  22. [30]

    Lacker, K

    D. Lacker, K. Ramanan, and R. Wu, Local weak convergence for sparse networks of interacting processes , The Annals of Applied Probability 33 (2023), no. 2, 843 –888

    2023

  23. [31]

    , Marginal dynamics of interacting diffusions on unimodular Galton–Watson trees, Probability Theory and Related Fields 187 (2023), 817–884

    2023

  24. [32]

    Larsson and S

    M. Larsson and S. Long, Markovian projections for itˆ o semimartingales with jumps, 2024

    2024

  25. [33]

    H. P. McKean, A class of markov processes associated with nonlinear parabolic equations , Proceedings of the National Academy of Sciences 56 (1966), no. 6, 1907–1911, available at https://www.pnas.org/doi/pdf/ 10.1073/pnas.56.6.1907

    1966 doi

  26. [34]

    10, 1094–1104

    Michael Mitzenmacher, The power of two choices in randomized load balancing, IEEE Transactions on Parallel and Distributed Systems 12 (2001), no. 10, 1094–1104

    2001

  27. [35]

    Moosavi and W

    S. Moosavi and W. Truccolo, Criticality in probabilistic models of spreading dynamics in brain networks: Epileptic seizures, PLOS Computational Biology 19 (202302), no. 2, 1–41

  28. [36]

    Oakes, The markovian self-exciting process , Journal of Applied Probability 12 (1975), no

    D. Oakes, The markovian self-exciting process , Journal of Applied Probability 12 (1975), no. 1, 69–77

    1975

  29. [37]

    Pastor-Satorras, C

    R. Pastor-Satorras, C. Castellano, P. Van Mieghem, and A. Vespignani, Epidemic processes in complex networks, Reviews of Modern Physics 87 (20158), no. 3, 925–979

  30. [38]

    Shiryaev, Optimal stopping rules , 1st ed

    A. Shiryaev, Optimal stopping rules , 1st ed. (B Rozovskii and G Grimmett, eds.), Stochastic Modelling and Applied Probability, Springer, Berlin, Germany, 2007 (en)

    2007

  31. [39]

    Truccolo, U

    W. Truccolo, U. T. Eden, M. R. Fellows, J. P. Donoghue, and E. N. Brown, A point process framework for relating neural spiking activity to spiking history, neural ensemble, and extrinsic covariate effects , Journal of Neurophysiology 93 (2005), 1074–1089

    2005

  32. [40]

    van der Hofstad, Random graphs and complex networks , Cambridge Series in Statistical and Probabilistic Mathematics, Cambridge University Press, 2024

    R. van der Hofstad, Random graphs and complex networks , Cambridge Series in Statistical and Probabilistic Mathematics, Cambridge University Press, 2024

    2024

  33. [41]

    Remco van der Hofstad and Seva Shneer, It is hard to kill fake news , 2023

    2023

  34. [42]

    Vladimirov, Serge A

    Alexander A. Vladimirov, Serge A. Pirogov, Alexander N. Rybko, and Senya B. Shlosman, Propagation of chaos and Poisson hypothesis , Problems of Information Transmission 54 (20187), no. 3, 290–299

  35. [43]

    N. D. Vvedenskaya, R. L. Dobrushin, and F. I. Karpelevich, A queueing system with a choice of the shorter of two queues—an asymptotic approach , Problemy Peredachi Informatsii 32 (1996), no. 1, 20–34. MR1384927

    1996

  36. [44]

    Xi and C

    F. Xi and C. Zhu, Jump type stochastic differential equations with non-lipschitz coefficients: Non-confluence, feller and strong feller properties, and exponential ergodicity , Journal of Differential Equations 266 (2019), no. 8, 4668–4711. Division of Applied Mathematics, Bro...

    2019

Pith tools

Reviewed August 8, 2026 · model on record in the stance chip above.