Pith. sign in

REVIEW 2 major objections 6 minor 1 cited by

Coarse graining of stochastic differential equations: averaging and projection method

T0 review · 2 major / 6 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read Under the assumptions of Theorem 5.3, the projection-method reduced dynamics has the same long-time statistics as the slow variables of the full SDE, and its equality with averaging holds only under explicit conditions.

desk verdict A useful, honest clarification of when projection-type coarse graining and averaging disagree, with sufficient-condition results that are checkable in principle but hard to verify in practice; the explicit errors in the counterexample section are sloppy but not load-bearing. read the letter →

arxiv 2506.14939 v1 pith:K44RJYJA submitted 2025-06-17 math.PR math-phmath.DSmath.MP

Manh Hong Duong , Carsten Hartmann , Michela Ottobre This is my paper · ORCID
classification math.PRmath-phmath.DSmath.MP MSC 60H1034F0560J6034K3335B4082C31
keywords coarsegrainingstochasticdifferentialequationsprojectionmethodaveragingGyöngymimickingmarginalsequilibriumconditionaldensityslow-fastsystemstwo-parameterMarkovsemigroups
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

Two standard ways to reduce a system of stochastic differential equations to the variables that matter are classical averaging, which exploits a separation of time scales, and the projection method, which replaces the eliminated variables by their equilibrium conditional distribution given the remaining ones. This paper's first claim is that, under explicit conditions, the projection-method reduced dynamics has a unique invariant measure, equal to the marginal of the full system's invariant measure on the slow variables; long-time averages of the reduced dynamics therefore reproduce the slow-variable statistics of the original system. The proof works by connecting the projection method to the Gyöngy mimicking-marginals construction, an exact but hard-to-simulate rewriting of the slow dynamics, and then showing the two have the same long-time behaviour. The paper's second claim is that, contrary to a common belief, the projection method and averaging do not in general coincide under time-scale separation; they coincide exactly under a kernel condition on the generators, and a degenerate-noise example shows the projection method can even produce a constant dynamics where averaging gives a nontrivial Ornstein-Uhlenbeck process.

What carries the argument

The load-bearing object is the Gyöngy SDE (8), $$dX^G(t)=f_G(t,X^G(t))\,dt+\$alpha_G^{{1/2}}$(t,X^G(t))\,dU_t,$$ whose drift and diffusion are obtained by integrating the original coefficients against the conditional density $\rho_t(y|x)$ of the full system. Because $\rho_t(y|x)$ evolves in time, (8) is a non-autonomous SDE; the projection method $X^P$ is exactly the autonomous limit obtained by replacing $\rho_t(y|x)$ with the equilibrium conditional density $\rho(y|x)$. The argument runs through the theory of evolution systems of measures: condition [C3]---the obtuse-angle condition on the commutator $[\sigma^{[i]}\cdot\nabla, B\cdot\nabla]$ of the drift and diffusion vector fields---ensures uniqueness of the evolution system of measures for the non-autonomous dynamics, so its long-time limit is the invariant measure of the limit autonomous equation, which is the projection dynamics. A second structural identity, Proposition 4.1, characterizes when the invariant measure of the frozen fast dynamics equals the equilibrium conditional density in terms of the kernels of the dual generators $L'_x$ and $L'_y$.

What would settle it

Take the uniform-elliptic modification of Example 2.1, i.e. add an independent noise term $\sqrt{2}\,dU_t$ to the slow equation (66), so the joint law is Gaussian with an explicit invariant measure. Compute the projection-method coefficients and the Gyöngy coefficients from the Gaussian conditional densities, check the obtuse-angle inequality of condition [C3] directly, and compare the invariant measure of the projected dynamics with the marginal $\bar\rho=N(0,1/2)$; a mismatch while [C3] holds would refute Theorem 5.3, and a match in a regime where [C3] fails would show the condition is too strong.

Watch

Extended reading notes

Core claim

The central discovery is that the autonomous SDE produced by the projection method, $$dX^P(t)=f_P(X^P(t))\,dt+\$alpha_P^{{1/2}}$(X^P(t))\,dU_t, \quad f_P(x)=\int f(x,y)\rho(y|x)\,dy,\quad \alpha_P(x)=\int \$\alpha$\$\alpha$^T(x,y)\rho(y|x)\,dy,$$ inherits its long-time behaviour from the Gyöngy SDE whose coefficients use the time-dependent conditional density $\rho_t(y|x)$ instead of the equilibrium conditional density $\rho(y|x)$. Theorem 5.3 states that if the original system has smooth, uniformly elliptic, ergodic coefficients, the Gyöngy SDE is well posed, and its non-autonomous coefficients satisfy Assumption A1---uniform ellipticity, a Lyapunov condition, the obtuse-angle commutator condition [C3], and convergence to the autonomous coefficients---then $X^P$ admits a unique invariant measure, equal to the marginal $\bar\rho$ of the invariant measure $\rho$ of the full system, so $\lim_{t\to\infty}\mathbb{E}h(X^P(t))=\int h(x)\bar\rho(x)\,dx$ for bounded continuous $h$. The proof does not rely on reversibility or an explicit formula for $\rho$, and it makes rigorous the relation between the projection method and Gyöngy's exact marginal-mimicking method.

Load-bearing premise

The time-dependent Gyöngy coefficients must satisfy the obtuse-angle commutator condition [C3], and because these coefficients are defined through the conditional density $\rho_t(y|x)$, which is usually not known explicitly, the condition is hard to check; the paper states this difficulty explicitly in Note 5.4 and leaves it to future work.

Editorial extensions

If this is right

  • If the conditions of Theorem 5.3 hold, practitioners can use the projection-method SDE to compute long-time averages of the slow variables without simulating the fast variables, with the guarantee that the limiting statistics coincide with the marginal $\bar\rho$ of the full invariant measure.
  • In the scale-separation limit, the equality of projection and averaging holds precisely when $\lim_{\varepsilon\to0}\rho_\varepsilon(y|x)\to\rho^{(x)}(y)$, which the paper shows is equivalent to the marginal of the invariant measure remaining strictly positive in the limit and to the kernel condition of Proposition 4.1 being satisfied.
  • When the slow equation has degenerate noise, as in the Ornstein-Uhlenbeck example (66)-(67), the projection method can produce a zero-drift, zero-noise dynamics with a continuum of invariant measures, whereas averaging produces a nontrivial OU process; the two methods are then not only different, the projection method fails to capture the correct long-time behaviour.
  • For reversible (gradient) systems the two methods coincide because the frozen invariant measure equals the equilibrium conditional law; the paper's examples show that non-reversibility alone does not determine whether they coincide.

Reading between the lines

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

  • Editorial inference: the commutator condition [C3] reduces in one dimension to the monotonicity test $\partial_x f_G(t,x)\le -\lambda_0$, suggesting a cheap practical diagnostic---compute the conditional-drift derivative and check for contraction before trusting the projected reduced model.
  • Editorial inference: the formal expansion in Section 4.2 ties the validity of the $\varepsilon\to0$ limit to the strict positivity of the limiting marginal; when the slow marginal degenerates (as in Example 4.5), the equilibrium conditional density cannot converge to the frozen measure, so a practical rule emerges: expect projection and averaging to disagree whenever the slow-variable invariant me
  • Editorial inference: the same two-parameter-semigroup framework could be applied to other conditional-expectation reductions, such as nonlinear reaction coordinates, by substituting the reaction-coordinate conditional law for $\rho_t(y|x)$.
  • Editorial inference: the degenerate-noise failure of the projection method indicates that the method should be applied only when the resolved variables carry their own noise; adding a small ellipticity to the slow equation and letting it vanish may provide a selection criterion among the many invariant measures, a possibility the paper does not explore.
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

2 major / 6 minor

Summary. The paper studies coarse-graining of stochastic differential equations by comparing the classical averaging principle with the projection method (PM), which constructs an effective dynamics by conditional expectation with respect to the equilibrium conditional distribution of the unresolved variables. The authors link the PM to Gyöngy's mimicking-marginals method, prove that the PM generator leaves the marginal of the original invariant measure infinitesimally invariant (Proposition 3.1), give necessary and sufficient conditions for the equilibrium conditional distribution to coincide with the frozen-process invariant measure (Proposition 4.1), and state a theorem (Theorem 5.3) under which the Gyöngy dynamics and the PM dynamics have the same long-time behavior, so that the PM samples the correct marginal. The comparison between PM and averaging is illustrated by examples and counterexamples, including a formal asymptotic expansion of the conditional distribution in the slow-fast limit. The main theorem relies on a non-autonomous SDE result (Theorem 5.1) adapted from the authors' prior work [CCDO21] and [ALL13].

Significance. If the main results hold, the paper provides a reversibility-independent justification of the projection method and clarifies its relationship to Gyöngy's exact mimicking construction. Proposition 3.1 and Proposition 4.1 are clean and useful, and the counterexample in Section 5.3 is instructive. The formal expansion in Section 4.2 is also a helpful heuristic. However, the central Theorem 5.3 is conditional on the obtuse angle condition [C3] imposed on the implicitly defined Gyöngy coefficients, and the paper does not exhibit any non-trivial system for which this condition is verified; Note 5.4 explicitly defers this check to future work. This substantially limits the practical applicability of the main theorem. In addition, several displayed formulas in the examples are incorrect. With corrections and a more honest framing of the theorem's scope, the paper could be a valuable contribution to the coarse-graining literature.

major comments (2)
  1. [§5.2, Theorem 5.3 and Assumption A1 [C3], with Note 5.4] The central theorem is gated by the obtuse angle condition [C3], which is imposed on the Stratonovich vector fields of the Gyöngy SDE (8). Since those fields are defined through the time-dependent conditional density ρ_t(y|x), which is explicit only in Gaussian or linear settings, [C3] is not checkable for typical systems. Note 5.4 concedes that checking Assumption iii) 'might be difficult' and defers it to future work. The paper provides no non-trivial class of systems for which [C3] is certified; the only sufficient condition mentioned (the two-dimensional case with α=σ, where [C3] reduces to ∂_x f_G ≤ −λ_0) is not applied to any example. The introduction's claim that the PM works for 'a large class of SDEs' is therefore not substantiated by the results. The authors should either add a non-trivial class of systems where [C3] is verified, or explicitly state in the abstract and introduction that Theorem 5.3 is conditional and that no instance of its hypotheses is provided.
  2. [§4.2, formal asymptotic expansion and the limit (29)] The paper claims to provide sufficient conditions under which the PM and averaging coincide in the limit of time-scale separation, but the derivation of the limit (29) is formal. The text states: 'To turn the above formal expansion into an actual proof one needs to combine these results on Poisson equations together with a more careful study of the difference ρε − ρ(x)(y) − ερ1(x,y), which needs to be shown to converge to zero as ε → 0. While not too difficult, this is not within the scope of this paper.' Thus the rigorous comparison of PM and averaging in the slow-fast regime is not established. Since this comparison is one of the paper's two advertised contributions, the formal status of this part should be clearly flagged in the abstract and introduction; as written, the reader may overestimate the strength of the results.
minor comments (6)
  1. [Example 2.1, equation (26)] The conditional variance in equation (26) is incorrect. For the joint Gaussian invariant measure with covariance matrix (25), the conditional variance of Y given X=x is 1 − (ε/(1+ε))^2 / (ε/(1+ε)) = 1/(1+ε), not (1−ε)/(1+ε). For ε=1, the stated formula gives variance zero, contradicting the direct computation in Section 5.3 where the correct value is 1/2.
  2. [Example 4.6] The conditional variance formula is incorrect. Given the stated covariance matrix Σε, the conditional variance is 1 − ε^2/((1+ε)(1+2ε)) = (1+3ε+ε^2)/(1+3ε+2ε^2), not (1−ε^2)/(1+3ε+4ε^2). The qualitative conclusion that ρε(·|x) converges to N(0,1) as ε→0 remains correct with the corrected formula, but the displayed expression should be fixed.
  3. [Example 4.4] The 'for instance' construction of the vector field A is incorrect: the condition is ∇_x V = ∇_y A, so the correct formula is A(x,y) = ∫_0^y ∂_x V(x,w) dw, not ∫_0^y ∂_y V(x,w) dw as written.
  4. [Note 4.2] The bullet list in Note 4.2 is malformed: bullets appear mid-sentence ('• As a direct consequence of Proposition 4.1, • if α12(x,y)=0 ... • states that L′ρ=0'). Please reformat so that the text reads as complete, coherent sentences.
  5. [Title/Abstract] The abstract displays 'A VERAGING' instead of 'AVERAGING' in the title; this typo should be corrected.
  6. [Example 4.6, pathwise convergence claim] The claim that 'it follows by Itô's formula and a standard Gronwall estimate that ... the projected dynamics converges pathwise' is not proved or even sketched. A brief indication of the estimate would strengthen the example.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: Theorem 5.3 chains external Gyöngy results and published non-autonomous SDE theory; its restrictive hypothesis [C3] is an unverified condition, not an assumed conclusion.

full rationale

The paper's central claim, Theorem 5.3, is conditional and its proof is a legitimate chaining of external results. Proposition 3.2, which identifies the law of the Gyöngy SDE (8) with the x-marginal of the full system, is an adaptation of Gyöngy's published theorem, not a restatement of the target conclusion. Theorem 5.1, which transfers the long-time behaviour of the non-autonomous Gyöngy dynamics to its autonomous limit (the projection dynamics XP), is taken from and proved via [ALL13] and [CCDO21]; those are published, parameter-free theorems whose assumptions do not include the conclusion that XP has invariant measure \bar\rho. The target invariant measure is not assumed in Assumption A1: condition [C4] only assumes convergence of the time-dependent coefficients to their limits, and the identification of that limit with the invariant measure is then derived. The difficulty in checking [C3] (the obtuse angle condition for the implicitly defined Gyöngy coefficients) is explicitly acknowledged in Note 5.4 and deferred to future work; this makes the theorem conditional but does not make it circular. Proposition 3.1, showing that \bar\rho is infinitesimally invariant for the projected generator, is a direct computation from the definition of the conditional-expectation coefficients, not a predictive claim obtained by fitting. No parameter fitting, no renaming of known results, and no self-citation chain substituting for proof appear. The paper is self-contained in the sense that every claimed implication is either proved in the text or supported by independently published theorems with stated assumptions.

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

The paper introduces no new entities or fitted parameters. Its results are conditional on standard existence, uniqueness, regularity, and ergodicity hypotheses, plus the harder-to-check Assumption A1 on the implicit Gyongy coefficients.

assumptions (6)
  • domain assumption The full system (1) is uniformly elliptic (equation (3)) and has smooth coefficients.
    Standing assumption in Section 2; used to guarantee strictly positive densities and smoothness of conditional distributions.
  • domain assumption System (1) admits a unique weak solution and a unique invariant measure rho with density.
    Stated as a standing assumption in Section 2; used throughout for the marginal and conditional densities.
  • domain assumption The Gyongy SDE (8) admits a unique weak solution with absolutely continuous law.
    Assumption ii) of Theorem 5.3; needed to apply Proposition 3.2. The authors note this is hard to verify from conditions on (1) alone.
  • domain assumption The coefficients of (8) satisfy Assumption A1, including the obtuse angle condition [C3] and the convergence condition [C4].
    Assumption iii) of Theorem 5.3; it is the load-bearing technical condition that lets the non-autonomous Gyongy dynamics be compared with the autonomous projected dynamics.
  • domain assumption The frozen process (19) admits a unique invariant measure rho^(x) and converges to it.
    Standard averaging assumption, Section 2.2, needed to define the averaged dynamics.
  • standard math Boundary terms in the integration by parts for L'_y vanish (Proposition 3.2 assumption ii).
    Requires functions g(x,.)rho_t(x,.) and derivatives of beta beta^T rho_t to vanish at infinity; stated and argued as standard.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Coarse graining of stochastic differential equations: averaging and projection method." pith.science (2026). https://pith.science/paper/K44RJYJA

@misc{pith2026250614939,
  author       = {Pith},
  title        = {Pith review of: Coarse graining of stochastic differential equations: averaging and projection method},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/K44RJYJA}},
  note         = {Machine review of arXiv:2506.14939}
}
read the original abstract

We study coarse-graining methods for stochastic differential equations. In particular we consider averaging and a type of projection operator method, sometimes referred to as effective dynamic via conditional expectations. The projection method (PM) we consider is related to the ``mimicking marginals'' coarse graining approach proposed by Gy\"ongy. The first contribution of this paper is to provide further theoretical background for the PM and a rigorous link to the Gy\"ongy method. Moreover, we compare PM and averaging. While averaging applies to systems with time scale separation, the PM can in principle be applied irrespective of this. However it is often assumed that the two methods coincide in presence of scale separation. The second contribution of this paper is to make this statement precise, provide sufficient conditions under which these two methods coincide and then show -- via examples and counterexamples -- that this needs not be the case.

Figures

Figures reproduced from arXiv: 2506.14939 by the authors.

Figure 1
Figure 1. Relations and dependencies between Gy¨ongy method (GM), projection method (PM) and averaging of slow-fast systems. separations, i.e. whether the following limit holds, limε→0 ρ ε (y|x) = ρ (x) (y), (29) in some appropriate sense. Again, Example 2.1 shows that this needs not be the case. In Section 4.2 we look at this limit through a formal asymptotic expansion in ε and give sufficient conditions under which this lim… view at source ↗
Figure 2
Figure 2. Behaviour of Gy¨ongy approximation (79) of the 2-dimensional system (66) for Gaussian initial conditions X0 ∼ N (−1, 0.1) and Y0 ∼ N (5, 1). The plotted mean and variance have been computed by averaging over trajectories starting from N = 105 independent initial conditions X0. If Σxx 0 > 0 the function φ has the property that φ(t) → 1 as t → ∞, φ(t) → 0 as t → 0, and it is smooth and bounded. When Σxx 0 = 0 then the… view at source ↗
Figure 3
Figure 3. Long term dynamics of the Gy¨ongy approximation (79): The histogram was computed by binning N = 105 trajectories starting from independent random initial data X0 ∼ ρ¯0 6= N (0, 1/2) and evaluated at t = 20; the red curve shows ¯ρ = N (0, 1/2). 0 5 10 15 20 t -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5 3 x projection method Gyöngy method 0 5 10 15 20 t -4 -3 -2 -1 0 1 2 3 4 5 x projection method Gyöngy method [PITH_FULL_IMAGE:fi… view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Solution of (78) and (79) for 10 representative initial condi￾tions X0 ∼ N (−1, 0.1) (left panel) and X0 ∼ N (−1, 10) (right panel). Both figures show convergence of the Gy¨ongy trajectories (solid lines) to an ensemble with the correct asymptotic mean and variance on …

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

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

  1. Model Reduction of Multivariate Geometric Brownian Motions and Localization in a Two-State Quantum System

    math-ph 2025-07 conditional novelty 6.0 of 10

    A hybrid invariant-manifold and fluctuation-dissipation reduction is shown to produce a one-dimensional additive-noise surrogate that preserves the localization dynamics of a two-state quantum geometric Brownian motion.

Reference graph

Works this paper leans on

53 extracted references · 49 canonical work pages · cited by 1 Pith paper

  1. [1]

    Hypercontractivity and asymptotic behavior in nonautonomous kolmogorov equations

    Luciana Angiuli, Luca Lorenzi, and Alessandra Lunardi. Hypercontractivity and asymptotic behavior in nonautonomous kolmogorov equations. Communications in Partial Differential Equations , 38(12):2049--2080, 2013

    work page 2013

  2. [2]

    Linear algebra done right

    Sheldon Axler. Linear algebra done right . Springer Science & Business Media, 1997

    work page 1997

  3. [3]

    A non-linear kinetic model of self-propelled particles with multiple equilibria

    Paolo Butt \`a , Franco Flandoli, Michela Ottobre, and Boguslaw Zegarlinski. A non-linear kinetic model of self-propelled particles with multiple equilibria. arXiv preprint arXiv:1804.01247 , 2018

    work page Pith review arXiv 2018

  4. [4]

    Global regularity and bounds for solutions of parabolic equations for probability measures

    Vladimir I Bogachev, MICHAEL Rockner, and Stanislav V Shaposhnikov. Global regularity and bounds for solutions of parabolic equations for probability measures. Theory of Probability & Its Applications , 50(4):561--581, 2006

    work page 2006

  5. [5]

    Mimicking an it\^ o process by a solution of a stochastic differential equation

    Gerard Brunick and Steven Shreve. Mimicking an it\^ o process by a solution of a stochastic differential equation. The Annals of Applied Probability , 23:1584--1628, 2010

    work page 2010

  6. [6]

    Long-time behaviour of degenerate diffusions: Ufg-type sdes and time-inhomogeneous hypoelliptic processes

    Thomas Cass, Dan Crisan, Paul Dobson, and Michela Ottobre. Long-time behaviour of degenerate diffusions: Ufg-type sdes and time-inhomogeneous hypoelliptic processes. Electron. J. Probab. , 26:1--72, 2021

    work page 2021

  7. [7]

    Poisson equations with locally-lipschitz coefficients and uniform in time averaging for stochastic differential equations via strong exponential stability

    Dan Crisan, Paul Dobson, Ben Goddard, Michela Ottobre, and Iain Souttar. Poisson equations with locally-lipschitz coefficients and uniform in time averaging for stochastic differential equations via strong exponential stability. arXiv preprint arXiv:2204.02679 , 2022

    arXiv 2022

  8. [8]

    Uniform in time estimates for the weak error of the euler method for sdes and a pathwise approach to derivative estimates for diffusion semigroups

    Dan Crisan, Paul Dobson, and Michela Ottobre. Uniform in time estimates for the weak error of the euler method for sdes and a pathwise approach to derivative estimates for diffusion semigroups. Transactions of the American Mathematical Society , 374(5):3289--3330, 2021

    work page 2021

Show all 53 references
  1. [9]

    Optimal prediction and the mori--zwanzig representation of irreversible processes

    Alexandre J Chorin, Ole H Hald, and Raz Kupferman. Optimal prediction and the mori--zwanzig representation of irreversible processes. Proceedings of the National Academy of Sciences , 97(7):2968--2973, 2000

    2000

  2. [10]

    Optimal prediction with memory

    Alexandre J Chorin, Ole H Hald, and Raz Kupferman. Optimal prediction with memory. Physica D: Nonlinear Phenomena , 166(3-4):239--257, 2002

    2002

  3. [11]

    Prediction from partial data, renormalization, and averaging

    Alexandre J Chorin, Ole H Hald, and Raz Kupferman. Prediction from partial data, renormalization, and averaging. Journal of Scientific Computing , 28(2-3):245--261, 2006

    2006

  4. [12]

    Optimal prediction for hamiltonian partial differential equations

    Alexandre J Chorin, Raz Kupferman, and Doron Levy. Optimal prediction for hamiltonian partial differential equations. Journal of Computational Physics , 162(1):267--297, 2000

    2000

  5. [13]

    Pointwise gradient bounds for degenerate semigroups (of ufg type)

    Dan Crisan and Michela Ottobre. Pointwise gradient bounds for degenerate semigroups (of ufg type). Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences , 472(2195):20160442, 2016

    2016

  6. [14]

    Global existence of smooth solutions for the vlasov-fokker-planck equation in 1 and 2 space dimensions

    Pierre Degond. Global existence of smooth solutions for the vlasov-fokker-planck equation in 1 and 2 space dimensions. Annales scientifiques de l' \'E cole Normale Sup \'e rieure , 19(4):519--542, 1986

    1986

  7. [15]

    Quantification of coarse-graining error in langevin and overdamped langevin dynamics

    Manh Hong Duong, Agnes Lamacz, Mark A Peletier, Andr \'e Schlichting, and Upanshu Sharma. Quantification of coarse-graining error in langevin and overdamped langevin dynamics. Nonlinearity , 31(10):4517, 2018

    2018

  8. [16]

    Ornstein--uhlenbeck operators with time periodic coefficients

    Giuseppe Da Prato and Alessandra Lunardi. Ornstein--uhlenbeck operators with time periodic coefficients. Journal of Evolution Equations , 7(4):587--614, 2007

    2007

  9. [17]

    Analysis of multiscale methods for stochastic differential equations

    Weinan E, Di Liu, and Eric Vanden-Eijnden. Analysis of multiscale methods for stochastic differential equations. Communications on Pure and Applied Mathematics , 58(11):1544--1585, 2005

    2005

  10. [18]

    Extracting macroscopic dynamics: model problems and algorithms

    Dror Givon, Raz Kupferman, and Andrew Stuart. Extracting macroscopic dynamics: model problems and algorithms. Nonlinearity , 17(6):R55, 2004

    2004

  11. [19]

    Mimicking the one-dimensional marginal distributions of processes having an It \^o differential

    Istv \'a n Gy \"o ngy. Mimicking the one-dimensional marginal distributions of processes having an It \^o differential. Probability theory and related fields , 71(4):501--516, 1986

    1986

  12. [20]

    Accelerating diffusions

    Chii-Ruey Hwang, Shu-Yin Hwang-Ma, and Shuenn-Jyi Sheu. Accelerating diffusions. The Annals of Applied Probability , 15(2), 2005

    2005

  13. [21]

    Coarse graining of nonreversible stochastic differential equations: Quantitative results and connections to averaging

    Carsten Hartmann, Lara Neureither, and Upanshu Sharma. Coarse graining of nonreversible stochastic differential equations: Quantitative results and connections to averaging. SIAM Journal on Mathematical Analysis , 52(3):2689--2733, 2020

    2020

  14. [22]

    Adaptive dimensionality reduction of stochastic differential equations for protein dynamics

    Jes \'u s A Izaguirre and Christopher R Sweet. Adaptive dimensionality reduction of stochastic differential equations for protein dynamics. In Proc. second international workshop on model reduction in reacting flows , 2009

    2009

  15. [23]

    Stochastic modelling: replacing fast degrees of freedom by noise

    Wolfram Just, Holger Kantz, Christian R \"o denbeck, and Mario Helm. Stochastic modelling: replacing fast degrees of freedom by noise. Journal of Physics A: Mathematical and General , 34(15):3199, 2001

    2001

  16. [24]

    R. Z. Khasminskij. On the principle of averaging the it \^o 's stochastic differential equations. Kybernetika , 4:260--279, 1968

    1968

  17. [25]

    Statistical mechanics of fluid mixtures

    John G Kirkwood. Statistical mechanics of fluid mixtures. The Journal of chemical physics , 3(5):300--313, 1935

    1935

  18. [26]

    Fast chaos versus white noise: entropy analysis and a fokker--planck model for the slow dynamics

    Holger Kantz, Wolfram Just, Nil \"u fer Baba, Katrin Gelfert, and Anja Riegert. Fast chaos versus white noise: entropy analysis and a fokker--planck model for the slow dynamics. Physica D: Nonlinear Phenomena , 187(1-4):200--213, 2004

    2004

  19. [27]

    Nonautonomous kolmogorov parabolic equations with unbounded coefficients

    Markus Kunze, Luca Lorenzi, and Alessandra Lunardi. Nonautonomous kolmogorov parabolic equations with unbounded coefficients. Transactions of the american mathematical society , 362(1):169--198, 2010

    2010

  20. [28]

    Brownian motion and stochastic calculus , volume 113

    Ioannis Karatzas and Steven Shreve. Brownian motion and stochastic calculus , volume 113. springer, 2014

    2014

  21. [29]

    Fractional kinetics in kac--zwanzig heat bath models

    Raz Kupferman. Fractional kinetics in kac--zwanzig heat bath models. Journal of statistical physics , 114:291--326, 2004

    2004

  22. [30]

    Analytical methods for Markov semigroups

    Luca Lorenzi and Marcello Bertoldi. Analytical methods for Markov semigroups . Chapman and Hall/CRC, 2006

    2006

  23. [31]

    Effective dynamics using conditional expectations

    Fr \'e d \'e ric Legoll and Tony Lelievre. Effective dynamics using conditional expectations. Nonlinearity , 23(9):2131, 2010

    2010

  24. [32]

    Some remarks on free energy and coarse-graining

    Fr \'e d \'e ric Legoll and Tony Lelievre. Some remarks on free energy and coarse-graining. In Numerical Analysis of Multiscale Computations: Proceedings of a Winter Workshop at the Banff International Research Station 2009 , pages 279--329. Springer, 2011

    2009

  25. [33]

    Pathwise estimates for an effective dynamics

    Fr \'e d \'e ric Legoll, Tony Lelievre, and Stefano Olla. Pathwise estimates for an effective dynamics. Stochastic Processes and their Applications , 127(9):2841--2863, 2017

    2017

  26. [34]

    Effective dynamics for non-reversible stochastic differential equations: a quantitative study

    Frederic Legoll, Tony Lelievre, and Upanshu Sharma. Effective dynamics for non-reversible stochastic differential equations: a quantitative study. Nonlinearity , 32(12):4779--4816, 2019

    2019

  27. [35]

    A mathematical theory of optimal milestoning (with a detour via exact milestoning)

    Ling Lin, Jianfeng Lu, and Eric Vanden-Eijnden. A mathematical theory of optimal milestoning (with a detour via exact milestoning). Commun Pure Appl. Math. , 71(6):1149--1177, 2018

    2018

  28. [36]

    Optimal non-reversible linear drift for the convergence to equilibrium of a diffusion

    Tony Lelievre, Francis Nier, and Grigorios A Pavliotis. Optimal non-reversible linear drift for the convergence to equilibrium of a diffusion. Journal of Statistical Physics , 152(2):237--274, 2013

    2013

  29. [37]

    Pathwise estimates for effective dynamics: the case of nonlinear vectorial reaction coordinates

    Tony Leli \`e vre and Wei Zhang. Pathwise estimates for effective dynamics: the case of nonlinear vectorial reaction coordinates. Multiscale Modeling & Simulation , 17(3):1019--1051, 2019

    2019

  30. [38]

    G. M. Moy, J. J. Hope, and C. M. Savage. Born and markov approximations for atom lasers. Phys. Rev. A , 59:667--675, Jan 1999

    1999

  31. [39]

    A continued-fraction representation of the time-correlation functions

    Hazime Mori. A continued-fraction representation of the time-correlation functions. Progress of Theoretical Physics , 34(3):399--416, 1965

    1965

  32. [40]

    A mathematical framework for stochastic climate models

    Andrew J Majda, Ilya Timofeyev, and Eric Vanden Eijnden. A mathematical framework for stochastic climate models. Communications on Pure and Applied Mathematics: A Journal Issued by the Courant Institute of Mathematical Sciences , 54(8):891--974, 2001

    2001

  33. [41]

    Spectral properties of effective dynamics from conditional expectations

    Feliks N \"u ske, P \'e ter Koltai, Lorenzo Boninsegna, and Cecilia Clementi. Spectral properties of effective dynamics from conditional expectations. Entropy , 23(2):134, 2021

    2021

  34. [42]

    Asymptotic analysis for the generalized langevin equation

    Michela Ottobre and Grigorios A Pavliotis. Asymptotic analysis for the generalized langevin equation. Nonlinearity , 24(5):1629, 2011

    2011

  35. [43]

    Stochastic processes and applications

    Grigorios A Pavliotis. Stochastic processes and applications. Texts in applied mathematics , 60, 2014

    2014

  36. [44]

    K. B. Petersen and M. S. Pedersen. The matrix cookbook, 2012. Version 20121115

    2012

  37. [45]

    Multiscale methods: averaging and homogenization , volume 53

    Grigorios A Pavliotis and Andrew Stuart. Multiscale methods: averaging and homogenization , volume 53. Springer Science & Business Media, 2008

    2008

  38. [46]

    The time-dependent born-oppenheimer approximation

    Gianluca Panati, Herbert Spohn, and Stefan Teufel. The time-dependent born-oppenheimer approximation. ESAIM: Mathematical Modelling and Numerical Analysis , 41(2):297--314, 2007

    2007

  39. [47]

    On poisson equation and diffusion approximation 2

    E Pardoux and A Yu Veretennikov. On poisson equation and diffusion approximation 2. The Annals of Probability , 31(3):1166--1192, 2003

    2003

  40. [48]

    Conditions for uniform in time convergence: applications to averaging, numerical discretisations and mean-field systems

    Katharina Schuh and Iain Souttar. Conditions for uniform in time convergence: applications to averaging, numerical discretisations and mean-field systems. arXiv preprint arXiv:2412.05239 , 2024

    2024 arXiv

  41. [49]

    Longtime convergence of the temperature-accelerated molecular dynamics method

    Gabriel Stoltz and Eric Vanden-Eijnden. Longtime convergence of the temperature-accelerated molecular dynamics method. Nonlinearity , 31(8):3748, 2018

    2018

  42. [50]

    Topics in propagation of chaos

    Alain-Sol Sznitman. Topics in propagation of chaos. Ecole d? \'e t \'e de probabilit \'e s de Saint-Flour XIX?1989 , 1464:165--251, 1991

    1989

  43. [51]

    A. Yu. Veretennikov. On the averaging principle of averaging for systems of stochastic differential equations. Mathematics of the USSR-Sbornik , 69(1):271, 1991

    1991

  44. [52]

    Numerically optimal runge--kutta pairs with interpolants

    James H Verner. Numerically optimal runge--kutta pairs with interpolants. Numerical Algorithms , 53(2):383--396, 2010

    2010

  45. [53]

    Effective dynamics along given reaction coordinates, and reaction rate theory

    Wei Zhang, Carsten Hartmann, and Christof Sch \"u tte. Effective dynamics along given reaction coordinates, and reaction rate theory. Faraday discussions , 195:365--394, 2016

    2016

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.