Pith. sign in

REVIEW 2 major objections 4 minor 69 references

On the numerical integration of the Fokker-Planck equation driven by a mechanical force and the Bismut-Elworthy-Li formula

T0 review · 2 major / 4 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read The paper derives exact path-average representations for Fokker-Planck densities and value-function gradients, turning stochastic optimal control into grid-free Monte Carlo.

desk verdict The overdamped Girsanov integration and BEL gradients are solid and useful, but the underdamped Fokker-Planck algorithm samples the wrong bridge measure and the τ-scaling in (32) is off, so that part needs rework. read the letter →

arxiv 2411.08518 v2 pith:QYGHAHWY submitted 2024-11-13 math.OC cond-mat.stat-mechmath-phmath.MP

Julia Sanders , Paolo Muratore-Ginanneschi This is my paper · ORCID
classification math.OCcond-mat.stat-mechmath-phmath.MP MSC 35Q8449L2060H3065C0565C30
keywords Fokker-PlanckequationGirsanovtheoremBismut-Elworthy-LiformulaHamilton-Jacobi-BellmanSchrödingerbridgeMonteCarlomethodsunderdampedLangevindynamicsstochasticoptimalcontrol
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's aim is to turn two PDEs that arise in stochastic optimal control into quantities that can be estimated by Monte Carlo sampling, with no spatial grid. For Fokker-Planck equations whose drift is the gradient of a mechanical potential, it derives an exact representation of the probability density as an expectation over paths of an auxiliary diffusion, weighted by a Girsanov factor, for both overdamped and underdamped (inertia-plus-friction) dynamics. For the Hamilton-Jacobi-Bellman equation, it uses the Bismut-Elworthy-Li formula to express the gradient of the value function as an expectation of the terminal payoff times a stochastic integral, which is the quantity the stationarity condition needs to fix the optimal protocol. Both representations are evaluated numerically by Euler-Maruyama sampling of the underlying diffusion, so the cost scales with the number of sampled trajectories rather than with grid points per dimension. The paper demonstrates the combined scheme on Schrödinger bridge problems (steering a distribution between assigned endpoints with minimal divergence from a free diffusion) and on a neural-network-trained optimal control protocol, checking against iterative and perturbative reference solutions.

What carries the argument

The machinery has two parts. First, the factorization $p_t(q) = e^{-\beta U_t(q)} f_t(q)$ recasts the forward Fokker-Planck equation as a backward Kolmogorov equation for $f_t$ along a time-reversed diffusion; Girsanov's change of measure then replaces that backward path measure by the measure of the free Wiener process, so the solution becomes an average of the initial density times the exponential Radon-Nikodym weight, accumulated as a running cost on backward Euler-Maruyama trajectories. Second, the Bismut-Elworthy-Li formula, derived through Malliavin's integration-by-parts identity, writes a directional derivative of the value function as an expectation of the terminal payoff multiplied by the stochastic integral $\int_t^{t_f}\langle dw_u, A_u^{-1}(x_u) x^*_{u,t} e_i\rangle$, where $x^*_{u,t}$ is the derivative of the stochastic flow with respect to its initial point; the algorithms accumulate this weight along forward trajectories of the controlled diffusion. The same weight appears in the degenerate-diffusion version, with an auxiliary variation field chosen to enforce the required boundary conditions in phase space.

What would settle it

For a harmonic potential, where the transition density is Gaussian and exactly known, simulate the backward scheme (15), reconstruct the density at $t_\iota$, and compare with the analytic forward kernel; any systematic bias as the time step shrinks would indicate that the reverse-time sampler is not an unbiased surrogate for the true backward transition kernel. Alternatively, in the quartic example of Section 2.2.3, record the empirical variance of the Girsanov weight $e^{-g}$ across samples and over longer horizons; exponential growth of the variance would show that the Monte Carlo estimate is dominated by rare paths.

Watch

Extended reading notes

Core claim

The central discovery is that a Fokker-Planck equation whose drift is the gradient of a mechanical potential can be solved pointwise by a conditional expectation over paths of an auxiliary diffusion. For the overdamped equation (3), the solution is $$p_t(q) = $E^{{\mathbb{Q}}$}\Bigl[ p_{t_\iota}(q_{t_\iota}) \, $e^{{-\frac{\beta}}${2}\int_{t_\iota}^{t}\bigl(\langle dq_s,(\partial U_s)(q_s)\rangle + \frac{\mu}{2}\|(\partial U_s)(q_s)\|^2 ds\bigr)} \Bigm| q_t = q \Bigr],$$ with the expectation over paths of the backward free diffusion; equation (32) is the underdamped analogue, in which the Girsanov weight absorbs the force and friction terms. For the Hamilton-Jacobi-Bellman equation (34), the Bismut-Elworthy-Li formula expresses each component of the gradient of the value function as an expectation of the terminal data and the running cost multiplied by stochastic-integral weights, equation (46), and Section 5 extends this to degenerate diffusions such as inertia-plus-friction dynamics. These identities are exact; Algorithms 1-5 evaluate them by Euler-Maruyama sampling of the relevant diffusion, which is what makes the integration grid-free.

Load-bearing premise

The load-bearing premise is that the exponential path weight has finite variance and that the reverse-time Euler-Maruyama sampler is an unbiased representation of the backward transition kernel; if either fails, the Monte Carlo averages are inaccurate even though the analytic representations are exact.

Editorial extensions

If this is right

  • The pointwise Fokker-Planck estimate costs one reverse-time trajectory per sample, so the method has no spatial mesh and can in principle be used in high-dimensional state spaces where grid-based solvers become unaffordable.
  • The same Girsanov construction covers underdamped (inertia-plus-friction) Langevin dynamics, not just overdamped diffusion, giving a Monte Carlo route to joint position-momentum densities in stochastic thermodynamics.
  • The Bismut-Elworthy-Li gradient estimator makes the stationarity condition (19) and its underdamped analogue (31) directly evaluable from samples, so optimal control updates do not require solving a nonlinear PDE on a grid.
  • Combining the two estimators yields an iterative Schrödinger bridge solver that the paper validates against a proximal iteration in the overdamped case and against perturbative predictions in the underdamped case.
  • The neural-network protocol of Algorithm 5 is a direct corollary: at each gradient-descent step the BEL estimator supplies the gradient of the value function, and the Girsanov estimator supplies the final density used to update the Lagrange multiplier.

Reading between the lines

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

  • Beyond the paper: the reciprocal-process identity (49) suggests a way to estimate the score function itself by conditioning the Bismut-Elworthy-Li weight on both endpoints, which would give a particle scheme for probability-flow equations without training a score network.
  • Beyond the paper: the Girsanov weight is an exponential of a path action, so its variance can be expected to grow with the horizon and with dimension; importance sampling or weight clipping may be needed before the method beats spectral or tensor-grid solvers in high-dimensional problems.
  • Beyond the paper: a direct numerical stress test would apply Algorithm 1 to a harmonic potential with an analytically known Gaussian transition density and monitor the empirical variance of the Girsanov weight as $\beta$, $\mu$, or the horizon is increased.
  • Beyond the paper: the degenerate BEL construction of Section 5 could serve as the training signal for an underdamped machine-learning bridge, extending Algorithm 5 to inertial dynamics; the paper suggests this as a possible generalization but does not demonstrate it.
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 / 4 minor

Summary. The paper proposes two grid-free Monte Carlo methods for PDEs arising in stochastic optimal control: (i) a Girsanov/Feynman-Kac representation of the Fokker-Planck equation for a diffusion driven by a mechanical potential, with separate treatments for overdamped (Section 2) and underdamped (Section 3) dynamics and corresponding Algorithms 1 and 2; and (ii) Bismut-Elworthy-Li (BEL) representations for the gradient of the solution of a Hamilton-Jacobi-Bellman equation, in both non-degenerate (Section 4) and degenerate (Section 5) settings, with Algorithms 3 and 4. The methods are demonstrated on Schrödinger bridge problems and on a neural-network-trained optimal control protocol (Algorithm 5, Section 6). The overdamped results are checked against analytic examples and against the independent Caluya-Halder iterative method; the underdamped FP integration is checked against a perturbative reference from the authors' own prior work.

Significance. If correct, the proposed representations would provide attractive, dimension-robust Monte Carlo alternatives to spatial discretization for the coupled FP/HJB systems that arise in stochastic thermodynamics and control. The overdamped Girsanov representation, the BEL formulas, and the analytic and numerical verifications in Sections 2.2, 4.3, 4.4, and 5.2 are useful and appear sound. The code availability and the external comparison with Caluya-Halder in the overdamped cases are strengths. However, the underdamped FP algorithm in Section 3 is not correctly sampling the bridge measure it claims to use, and the displayed formula (32) contains a τ-scaling error. As a result, the numerical underdamped demonstration in Fig. 3 does not validate the method as presented.

major comments (2)
  1. [Section 3.1 and Algorithm 2] Algorithm 2 samples a backward Euler discretization of the forward process (27) starting from the terminal point x. This is not the conditional law of (27) given X_t = x. For a diffusion with generator b·∇ + (1/2)σσ^T:∇∇, the reverse-time bridge drift is -b + σσ^T ∇ log p_{T-t}; the divergence-free condition on the deterministic part only gives Liouville volume preservation, not time-reversal invariance of the path measure. Concretely, in the free case ∂U = 0, the update for p in Algorithm 2 has conditional mean p_t, whereas the true Brownian-bridge conditional mean at the earlier time is not p_t. The estimator therefore has a bias of order one at every fixed step size and does not converge to the conditional expectation in (32) as h → 0. This undermines the underdamped FP integration claim and the evidence in Fig. 3.
  2. [Equation (32)] The Girsanov weight in (32) has an incorrect scaling in the parameter τ. The derivation immediately preceding (32) gives, for m = 1, an exponent with β/(4τ) ||p_s||^2 and √β/(2√τ) ⟨dw_s, p_s⟩, whereas (32) states τβ/4 ||p_s||^2 and √(τβ)/2 ⟨dw_s, p_s⟩. These two expressions agree only when τ = 1. Since the numerical example in Section 3.2 uses τ = 1, the error is not visible in Fig. 3, but the formula as written is false for general τ and needs correction.
minor comments (4)
  1. [Section 3.2, Fig. 3 caption] The underdamped FP reference solution is taken from the authors' own perturbative work [55], so the agreement shown in Fig. 3 is not an independent validation of the Monte Carlo scheme. This is a limitation, particularly because the algorithm itself is in question.
  2. [Section 3.1] The phrase "divergence-less" should be "divergence-free"; the wording conflates a property of the deterministic flow (Liouville volume preservation) with a property of the path measure, which is precisely the source of the sampling error in Algorithm 2.
  3. [Section 2.2.3, Algorithm 1] The index notation "∂U_{t_{n-1-i}}" in the pseudocode is confusing; clarity would be improved by writing the potential derivative as a function of q_{t_{n-1-i}}.
  4. [General] There are several typographical errors, including "auxilliary" (Section 2.2.3), "Kullback-Liebler" (Section 2.2.4), and "divergence-less" (Section 3.1).

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity: the Girsanov and Bismut-Elworthy-Li derivations are self-contained, and the underdamped benchmark's reliance on the authors' prior work is a validation limitation, not a circular reduction.

full rationale

The central representations (7), (11), (26), (32), and (46) are derived from the stated stochastic differential equations by standard Feynman-Kac, Girsanov change-of-measure, and Malliavin integration-by-parts arguments; none of these formulas is defined in terms of the numerical quantity it is used to predict. The overdamped and BEL numerical experiments are checked against an independent external benchmark, the Caluya-Halder iterative method [9], and against explicit analytic examples in Sections 2.2.1, 4.3, and 5.2, so the main numerical claims have independent grounding. The only self-referential element is Section 3.2 and Fig. 3, where both the optimal underdamped drift used in the integration and the reference densities used for comparison come from the authors' own prior work [55]; the resulting agreement is a self-consistency check rather than an independent validation. This is a benchmark limitation caused by self-citation, not a circular reduction by construction, and it does not compromise the independence of the derivation of the representation itself. The reviewer-level concern about the missing reverse-time score correction in Algorithm 2 is a correctness objection, not an instance of a fitted parameter being renamed a prediction or of a formula being equivalent to its own input. Overall, no circular step satisfying the evidentiary standard of this review was found.

Assumptions & free parameters 4 free parameters · 6 assumptions · 0 invented entities

The mathematical core rests on standard stochastic calculus (Girsanov, Feynman-Kac, BEL) plus regularity/confining assumptions on U. The only ad hoc additions are the reverse-time sampling approximation and the machine-learning hyperparameters. No new physical entities are proposed.

free parameters (4)
  • Learning rates γ1, γ2 = γ1=0.1 to 10^-2, γ2=10^-3 to 10^-5 (Section 6)
    Chosen by hand for the prototype ML example; they affect convergence of the trained control protocol, but not the mathematical derivation.
  • Neural network architecture and hyperparameters = 4-10-1 feed-forward, swish, Glorot init, 512 points, 10-100 MC trajectories, time step 0.005 (Section 6)
    Chosen for the prototype; no sensitivity analysis. These are numerical optimization choices, not fitted physical constants.
  • Euler-Maruyama step size h and sample counts M = h=10^-3 (Fig 1), h=0.005 (Fig 2), h=0.025 (Fig 3), h=0.01 (Fig 5); M between 10^3 and 10^4
    Chosen per example; no convergence study.
  • Polynomial degree for λ = 6 (Section 6)
    Arbitrary choice for the Lagrange multiplier model.
assumptions (6)
  • domain assumption The time-dependent potential U_t is sufficiently regular and confining, ensuring integrability of the density and the martingale property of the Girsanov/Feynman-Kac weights (Section 2, before Prop 2.1; Prop 2.1 uses 'confining potential also guarantees integrability').
    Needed for the representation (7) and for the Monte Carlo averages to be well-defined; the paper does not give explicit conditions on growth or regularity beyond 'sufficiently regular and confining.'
  • standard math Backward Itô calculus with post-point integrals is used for the backward diffusion (8), and the martingale property under the backward filtration is assumed (Section 2, proof of Prop 2.1).
    Follows [36, 42] and classical stochastic calculus; accepted background.
  • standard math Girsanov theorem applies to the backward diffusion with the stated Radon-Nikodym derivative (10) (Section 2.1).
    Standard result for absolutely continuous path measures; the paper cites [33, 49].
  • domain assumption For the underdamped case, the measure I generated by (27) and P♭ are mutually absolutely continuous with the given density (28) (Section 3.1).
    The frictionless process (27) and the backward process (25) must be equivalent; this holds for the constant diffusion coefficient but requires non-degeneracy of the noise on the momentum component.
  • ad hoc to paper The reverse-time Euler-Maruyama sampling in Algorithms 1 and 2 computes the conditional expectation in (7)/(32) without a density-gradient correction (Sections 2.2.3, 3.2).
    The paper samples the backward SDE as if the reverse-time transition kernel were the time reversal of the forward drift, without justifying this or adding the Fokker-Planck density correction. This is the main methodological risk.
  • domain assumption The stationarity condition (19) ∂qV = (2/β)∂qU and the update rules (68)-(69) are used to train the neural network (Section 6).
    These are specific to the Schrödinger bridge / KL-minimization problem and assumed to hold for the approximate solutions during training.

how reviews work

0 comments
Cite this review

Pith. "Pith review of On the numerical integration of the Fokker-Planck equation driven by a mechanical force and the Bismut-Elworthy-Li formula." pith.science (2026). https://pith.science/paper/QYGHAHWY

@misc{pith2026241108518,
  author       = {Pith},
  title        = {Pith review of: On the numerical integration of the Fokker-Planck equation driven by a mechanical force and the Bismut-Elworthy-Li formula},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QYGHAHWY}},
  note         = {Machine review of arXiv:2411.08518}
}
read the original abstract

Optimal control theory aims to find an optimal protocol to steer a system between assigned boundary conditions while minimizing a given cost functional in finite time. Equations arising from these types of problems are often non-linear and difficult to solve numerically. In this note, we describe numerical methods of integration for two partial differential equations that commonly arise in optimal control theory: the Fokker-Planck equation driven by a mechanical potential for which we use Girsanov theorem; and the Hamilton-Jacobi-Bellman, or dynamic programming, equation for which we find the gradient of its solution using the Bismut-Elworthy-Li formula. The computation of the gradient is necessary to specify the optimal protocol. Finally, we give an example application of the numerical techniques to solving an optimal control problem without spacial discretization using machine learning.

Figures

Figures reproduced from arXiv: 2411.08518 by the authors.

Figure 3
Figure 3. Solution of a Fokker-Planck equation driven by a non-linear mechanical underdamped diffusion [PITH_FULL_IMAGE:figures/full_fig_p013_3.png] view at source ↗
Figure 5
Figure 5. The gradient of the optimal control potential minimizing the Kullback-Leibler divergence ( [PITH_FULL_IMAGE:figures/full_fig_p023_5.png] view at source ↗
Figure 6
Figure 6. Solution of the optimal control problem minimizing the Kullback-Leibler divergence from a [PITH_FULL_IMAGE:figures/full_fig_p025_6.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

69 extracted references · 44 canonical work pages

  1. [55]

    DifferentialEquations.jl–a performant and feature-rich ecosystem for solving differential equations in Julia

    Christopher Rackauckas and Qing Nie. “DifferentialEquations.jl–a performant and feature-rich ecosystem for solving differential equations in Julia”. In:Journal of Open Research Software5.1 (2017)

    work page 2017

  2. [1]

    Random Dynamical Systems

    Ludwig Arnold. Random Dynamical Systems. Springer Monographs in Mathematics. Springer, 1998

    work page 1998

  3. [2]

    Refined Second Law of Thermodynamics for fast random processes

    Erik Aurell, Krzysztof Gawędzki, Carlos Mejía-Monasterio, Roya Mohayaee, and Paolo Muratore- Ginanneschi. “Refined Second Law of Thermodynamics for fast random processes”. In:Journal of Statistical Physics 147.3 (Apr. 2012), pp. 487–505.doi: 10.1007/s10955- 012- 0478- x . arXiv: 1201.3207 [cond-mat.stat-mech]. url: https://doi.org/10.1007/s10955-012-0478-x

    work page Pith review arXiv 2012

  4. [3]

    The Bismut–Elworthy–Li formula for mean-field stochastic differential equations

    David Baños. “The Bismut–Elworthy–Li formula for mean-field stochastic differential equations”. In: Annales de l’Institut Henri Poincaré, Probabilités et Statistiques54.1 (2018), pp. 220–233.doi: 10.1214/16-AIHP801. url: https://doi.org/10.1214/16-AIHP801

    work page doi:10.1214/16-aihp801 2018

  5. [4]

    Control Theory for Physicists

    John Bechhoefer. Control Theory for Physicists. Cambridge: Cambridge University Press, 2021. isbn: 9781009028493. doi: 10.1017/9780511734809 . url: https://www.cambridge.org/gb/ academic/subjects/physics/mathematical-methods/control-theory-physicists?format= AR

    doi:10.1017/9780511734809 2021

  6. [5]

    Large Deviations and the Malliavin Calculus

    Jean-Michel Bismut. Large Deviations and the Malliavin Calculus. Vol. 45. Progress in Mathemat- ics. Birkhäuser Verlag, 1984, p. 232.isbn: 978-3-7643-3220-4

    work page 1984

  7. [6]

    Probability flow solution of the Fokker-Planck equa- tion

    Nicholas M. Boffi and Eric Vanden-Eijnden. “Probability flow solution of the Fokker-Planck equa- tion”. In: Machine Learning: Science and Technology(June 9, 2022). doi: 10.1088/2632-2153/ ace2aa. arXiv: 2206.04642 [cs.LG]. url: https://doi.org/10.1088/2632-2153/ace2aa

    arXiv 2022

  8. [7]

    Improving control based importance sampling strategies for metastable diffusions via adapted metadynamics

    Enric Ribera Borrell, Jannes Quer, Lorenz Richter, and Christof Schütte. “Improving control based importance sampling strategies for metastable diffusions via adapted metadynamics”. In:SIAM Journal on Scientific Computing 46 (2 June 14, 2022), S298–S323.doi: 10 . 1137/ 22M1503464. arXiv: 2206.06628 [math.OC]. url: https://epubs.siam.org/doi/10.1137/22M1503464

    work page Pith review arXiv 2022

Show all 69 references
  1. [8]

    Gradient Flow Algorithms for Density Propagation in Stochastic Systems

    Kenneth F. Caluya and Abhishek Halder. “Gradient Flow Algorithms for Density Propagation in Stochastic Systems”. In:IEEE Transactions on Automatic Control65.10 (2020), pp. 3991–4004. doi: 10.1109/TAC.2019.2951348

    2020

  2. [9]

    Wasserstein Proximal Algorithms for the Schrödinger Bridge Problem: Density Control with Nonlinear Drift

    Kenneth F. Caluya and Abhishek Halder. “Wasserstein Proximal Algorithms for the Schrödinger Bridge Problem: Density Control with Nonlinear Drift”. In:IEEE Transactions on Automatic Con- trol 67.3 (Mar. 2022). doi: 10.1109/TAC.2021.3060704 . arXiv: 1912.01244 [math.OC] . url: h...

    2022

  3. [10]

    Fronts in passive scalar turbulence

    Antonio Celani, Alessandra Lanotte, Andrea Mazzino, and Massimo Vergassola. “Fronts in passive scalar turbulence”. In:Physics of Fluids 13.6 (June 2001), pp. 1768–1783.issn: 1089-7666. doi: 10.1063/1.1367325. arXiv: nlin/0009014. url: http://dx.doi.org/10.1063/1.1367325

    2001 arXiv

  4. [11]

    Stochastic Control Liaisons: Richard Sinkhorn Meets Gaspard Monge on a Schrödinger Bridge

    Yongxin Chen, Tryphon T. Georgiou, and Michele Pavon. “Stochastic Control Liaisons: Richard Sinkhorn Meets Gaspard Monge on a Schrödinger Bridge”. In:SIAM Review 63.2 (Jan. 2021), pp. 249–313. doi: 10.1137/20m1339982. url: https://doi.org/10.1137%2F20m1339982

    2021 doi

  5. [12]

    E. Schrödinger’s 1931 paper “On the Reversal of the Laws of Nature

    Raphaël Chetrite, Paolo Muratore-Ginanneschi, and Kay Schwieger. “E. Schrödinger’s 1931 paper “On the Reversal of the Laws of Nature” [“Über die Umkehrung der Naturgesetze”, Sitzungsberichte der preussischen Akademie der Wissenschaften, physikalisch-mathematische Klasse, 8 N9 ...

    1931 arXiv

  6. [13]

    Fluctuation relations for diffusion processes

    Raphaël Chétrite and Krzysztof Gawędzki. “Fluctuation relations for diffusion processes”. In:Com- munications in Mathematical Physics282.2 (Sept. 2008), pp. 469–518.doi: 10.1007/s00220-008- 0502-9. arXiv: 0707.2725 [math-ph]

    2008 arXiv

  7. [14]

    Entropic turnpike esti- mates for the kinetic Schrödinger problem

    Alberto Chiarini, Giovanni Conforti, Giacomo Greco, and Zhenjie Ren. “Entropic turnpike esti- mates for the kinetic Schrödinger problem”. In:Electronic Journal of Probability27 (Jan. 2021), pp. 1–32. arXiv:2108.09161 [math.PR]

    2021 arXiv

  8. [15]

    Diffusion Schrödinger Bridge with Applications to Score-Based Generative Modeling

    Valentin De Bortoli, James Thornton, Jeremy Heng, and Arnaud Doucet. “Diffusion Schrödinger Bridge with Applications to Score-Based Generative Modeling”. In:NeurIPS 2021 (spotlight) and arXiv: 2106.01357 (June 1, 2021). arXiv:2106.01357 [stat.ML]. 27

    2021 arXiv

  9. [16]

    Quantum trajectory framework for general time- local master equations

    Brecht Donvil and Paolo Muratore-Ginanneschi. “Quantum trajectory framework for general time- local master equations”. In: Nature Communications 13 (June 2022), p. 4140. doi: 10 . 1038 / s41467 - 022 - 31533 - 8. arXiv: 2102 . 10355 [quant-ph]. url: https : / / doi . org / 10 ...

    2022

  10. [17]

    Deep learning-based numerical methods for high- dimensional parabolic partial differential equations and backward stochastic differential equations

    E, Weinan, Jiequn Han, and Arnulf Jentzen. “Deep learning-based numerical methods for high- dimensional parabolic partial differential equations and backward stochastic differential equations”. In: Communications in Mathematics and Statistics5 (June 2017), pp. 349–380. doi: 10...

    2017 arXiv

  11. [18]

    Multilevel Picard itera- tions for solving smooth semilinear parabolic heat equations

    E, Weinan, Martin Hutzenthaler, Arnulf Jentzen, and Thomas Kruse. “Multilevel Picard itera- tions for solving smooth semilinear parabolic heat equations”. In:Partial Differential Equations and Applications 2.80 (June 2021). doi: 10 . 1007 / s42985 - 021 - 00089 - 5. arXiv: 160...

    2021

  12. [19]

    Differentiation of heat semigroups and applications

    Kenneth David Elworthy and Xue-Mei Li. “Differentiation of heat semigroups and applications”. In: Probability Theory And Mathematical Statistics: Proceedings of the 6th Vilnius Conference 1993. Ed. by Bronius Grigelionis. V.S.P. Interanational science, Utrecht, 1994.isbn: 978-...

    1993

  13. [20]

    Formulae for the Derivatives of Heat Semigroups

    Kenneth David Elworthy and Xue-Mei Li. “Formulae for the Derivatives of Heat Semigroups”. In: Journal of Functional Analysis125.1 (Oct. 1994), pp. 252–286.issn: 0022-1236. doi: 10.1006/ jfan.1994.1124. url: https://doi.org/10.1006/jfan.1994.1124

    1994

  14. [21]

    2nd, revised

    Wendell Helms Fleming and Mete Halil Soner.Controlled Markov processes and viscosity solutions. 2nd, revised. Vol. 25. Stochastic modelling and applied probability. Springer, 2006, p. 428.isbn: 978-0-387-26045-7

    2006

  15. [22]

    Irreversibility and Biased Ensembles in Active Matter: Insights from Stochastic Thermodynamics

    Étienne Fodor, Robert L. Jack, and Michael E. Cates. “Irreversibility and Biased Ensembles in Active Matter: Insights from Stochastic Thermodynamics”. In:Annual Review of Condensed Matter Physics 13.1 (Mar. 2022), pp. 215–238. issn: 1947-5462. doi: 10.1146/annurev- conmatphys-...

    2022 arXiv

  16. [23]

    ApplicationsofMalli- avin calculus to Monte-Carlo methods in finance. II

    EricFournié,Jean-MichelLasry,JérômeLebuchoux,andPierre-LouisLions.“ApplicationsofMalli- avin calculus to Monte-Carlo methods in finance. II”. In:Finance and Stochastics5.2 (Apr. 2001), pp. 201–236. doi: 10.1007/pl00013529. url: https://doi.org/10.1007%2Fpl00013529

    2001 doi

  17. [24]

    Ap- plications of Malliavin calculus to Monte Carlo methods in finance

    Eric Fournié, Jean-Michel Lasry, Jérôme Lebuchoux, Pierre-Louis Lions, and Nizar Touzi. “Ap- plications of Malliavin calculus to Monte Carlo methods in finance”. In:Finance and Stochastics 3.4 (Aug. 1999), pp. 391–412.doi: 10.1007/s007800050068. url: https://doi.org/10.1007% 2...

    1999 doi

  18. [25]

    The geometry of physics: an introduction

    Theodore Frankel. The geometry of physics: an introduction. 3rd. Cambridge University Press, 2012, pp. LXII, 686.isbn: 978-0-521-53927-2. url: http://www.cambridge.org/9780521833301

    2012

  19. [26]

    Solvinghigh-dimensionalpartialdifferentialequations using deep learning

    JiequnHan,ArnulfJentzen,andE,Weinan.“Solvinghigh-dimensionalpartialdifferentialequations using deep learning”. In: Proceedings of the National Academy of Sciences115.34 (Aug. 2018), pp. 8505–8510. doi: 10 . 1073 / pnas . 1718942115. url: https : / / doi . org / 10 . 1073 / pna...

    2018

  20. [27]

    Horn and Charles R

    Roger A. Horn and Charles R. Johnson. Topics in Matrix Analysis. New York, NY: Cambridge University Press, Apr. 1991.doi: 10.1017/cbo9780511840371. url: https://doi.org/10.1017% 2Fcbo9780511840371

    1991 doi

  21. [28]

    Estimation of Non-Normalized Statistical Models by Score Matching

    Aapo Hyvärinen. “Estimation of Non-Normalized Statistical Models by Score Matching”. In:Jour- nal of Machine Learning Research6 (2055). Ed. by Peter Dayan, pp. 695–709.doi: 10 . 5555 / 1046920.1088696. url: https://dl.acm.org/doi/abs/10.5555/1046920.1088696

  22. [29]

    Fashionable Modelling with Flux

    Michael Innes, Elliot Saba, Keno Fischer, Dhairya Gandhi, Marco Concetto Rudilosso, Neethu Mariya Joy, Tejan Karmali, Avik Pal, and Viral Shah. “Fashionable Modelling with Flux”. In: CoRR abs/1811.01457 (2018). arXiv:1811.01457. url: https://arxiv.org/abs/1811.01457

    2018 arXiv

  23. [30]

    Flux: Elegant Machine Learning with Julia

    Mike Innes. “Flux: Elegant Machine Learning with Julia”. In: Journal of Open Source Software (2018). doi: 10.21105/joss.00602

    2018 doi

  24. [31]

    Synthetic Molecular Motors and Mechan- ical Machines

    Euan R. Kay, David A. Leigh, and Francesco Zerbetto. “Synthetic Molecular Motors and Mechan- ical Machines”. In:Angewandte Chemie International Edition46.1-2 (Jan. 2007), pp. 72–191.issn: 1521-3773. doi: 10.1002/anie.200504313. 28

    2007 doi

  25. [32]

    Adam: A Method for Stochastic Optimization

    Diederik Kingma and Jimmy Ba. “Adam: A Method for Stochastic Optimization”. In:International Conference on Learning Representations(Dec. 2014)

    2014

  26. [33]

    Klebaner

    Fima C. Klebaner. Introduction to stochastic calculus with applications. 2nd ed. Imperial College Press, London UK., 2005, p. 432. isbn: 978-1-84816-822-0. doi: 10 . 1142 / p386. url: https : //doi.org/10.1142/p386

    2005 doi

  27. [34]

    Reciprocal diffusions in flat space

    Arthur J. Krener. “Reciprocal diffusions in flat space”. In:Probability Theory and Related Fields 107 (2 1997), pp. 243–281.issn: 0178-8051. doi: 10.1007/s004400050085

    1997 doi

  28. [35]

    Lectures on Stochastic Flows And Applications

    Hiroshi Kunita. Lectures on Stochastic Flows And Applications. Vol. 78. Lectures on mathematics and physics: Mathematics. Springer, 1986.isbn: 978-3-540-12878-6

    1986

  29. [36]

    On backward stochastic differential equations

    Hiroshi Kunita. “On backward stochastic differential equations”. In:Stochastics 6.3–4 (Jan. 1982), pp. 293–313. issn: 0090-9491. doi: 10.1080/17442508208833209. url: http://dx.doi.org/10. 1080/17442508208833209

    1982 doi

  30. [37]

    Flor Langouche, Dirk Roekaerts, and Enrique Tirapegui.Functional integration and semiclassical expansions. Vol. 10. Mathematics and Its Applications. Springer Netherlands, 1982, p. 313.isbn: 978-94-017-1634-5. doi: 10.1007/978-94-017-1634-5

    1982 doi

  31. [38]

    A survey of the Schrödinger problem and some of its connections with optimal transport

    Christian Léonard. “A survey of the Schrödinger problem and some of its connections with optimal transport”. In:Discrete and Continuous Dynamical Systems - Series A34.4 (Apr. 2014), pp. 1533–

    2014

  32. [39]

    InteractingParticleSolutionsofFokker–Planck Equations Through Gradient–Log–Density Estimation

    DimitraMaoutsa,SebastianReich,andManfredOpper.“InteractingParticleSolutionsofFokker–Planck Equations Through Gradient–Log–Density Estimation”. In:Entropy 22.8 (July 2020), p. 802.doi: 10.3390/e22080802. arXiv: 2006.00702. url: https://doi.org/10.3390%2Fe22080802

    2020 arXiv

  33. [40]

    BrownianCarnotengine

    Ignacio A. Martínez, Édgar Roldán, Luis Dinis, Dmitri Petrov, Juan M. R. Parrondo, and Raúl A. Rica.“BrownianCarnotengine”.In: Nature Physics12(2016),pp.67–70. doi: 10.1038/nphys3518. arXiv: 1412.1282 [cond-mat.stat-mech]. url: https://doi.org/10.1038/nphys3518

    2016 arXiv

  34. [41]

    Passive scalar turbulence in high dimensions

    Andrea Mazzino and Paolo Muratore-Ginanneschi. “Passive scalar turbulence in high dimensions”. eng. In:Physical Review E63.1 Pt 2 (Jan. 2001), p. 015302.doi: 10.1103/PhysRevE.63.015302. eprint: nlin/0010042

    2001 arXiv

  35. [42]

    Géométriedifférentiellestochastique,II

    Paul-AndréMeyer.“Géométriedifférentiellestochastique,II”.In: Séminaire de probabilités de Stras- bourg S16 (1982), pp. 165–207.url: http://www.numdam.org/item?id=SPS_1982__S16__165_0

    1982

  36. [43]

    Diffusion processes and Riemannian geometry

    Stanislav A. Molchanov. “Diffusion processes and Riemannian geometry”. In:Uspekhi Matematich- eskikh Nauk 30.1(181) (1975), pp. 3–59. doi: 10.1070/RM1975v030n01ABEH001400 . url: http: //mi.mathnet.ru/umn4123

    1975 doi

  37. [44]

    On extremals of the entropy production by “Langevin–Kramers

    Paolo Muratore-Ginanneschi. “On extremals of the entropy production by “Langevin–Kramers” dy- namics”. In:Journal of Statistical Mechanics: Theory and Experiment2014.5 (May 2014), P05013. issn:1742-5468. doi: 10.1088/1742-5468/2014/05/p05013.arXiv: 1401.3394 [cond-mat.stat-mec...

    2014 arXiv

  38. [45]

    Efficient protocols for Stirling heat engines at the micro-scale

    Paolo Muratore-Ginanneschi and Kay Schwieger. “Efficient protocols for Stirling heat engines at the micro-scale”. In:EPL (Europhysics Letters)112 (Oct. 2015), p. 20002.doi: 10.1209/0295- 5075/112/20002. arXiv: 1503.05788 [cond-mat.stat-mech]

    2015 arXiv

  39. [46]

    How nanomechanical systems can minimize dis- sipation

    Paolo Muratore-Ginanneschi and Kay Schwieger. “How nanomechanical systems can minimize dis- sipation”. In:Physical Review E90.6 (Dec. 2014), 060102(R).doi: 10.1103/PhysRevE.90.060102. arXiv: 1408.5298 [cond-mat.stat-mech]

    2014 arXiv

  40. [47]

    Dynamical Theories of Brownian Motion

    Edward Nelson. Dynamical Theories of Brownian Motion. 2nd. Princeton University Press, 2001, p. 148. isbn: 978-0-691-07950-9. doi: 10 . 2307 / j . ctv15r57jg. url: https : / / web . math . princeton.edu/~nelson/books.html

    2001

  41. [48]

    The Malliavin calculus and related topics

    David Nualart. The Malliavin calculus and related topics. Probability and its applications. Springer, 1995, p. 266

    1995

  42. [49]

    Pavliotis

    Grigorios A. Pavliotis. Stochastic Processes and Applications: Diffusion Processes, the Fokker- Planck and Langevin Equations.NewYork:SpringerNewYork,2014,p.339. isbn:978-1-4939-1322-

    2014

  43. [50]

    Stochastic Thermodynamics

    Luca Peliti and Simone Pigolotti. Stochastic Thermodynamics. Princeton and Oxford: Princeton University Press, 2020. isbn: 9780691215525. url: https : / / press . princeton . edu / books / ebook/9780691215525/stochastic-thermodynamics

    2020

  44. [51]

    url: https://link.springer.com/book/10.1007/978- 1-4939-1323-7

    doi: 10.1007/978-1-4939-1323-7 . url: https://link.springer.com/book/10.1007/978- 1-4939-1323-7. 29

    doi

  45. [52]

    Finite-Time Landauer Principle

    Karel Proesmans, Jannik Ehrich, and John Bechhoefer. “Finite-Time Landauer Principle”. In: Physical Review Letters125.10 (Sept. 2020), p. 100602.doi: 10.1103/physrevlett.125.100602. arXiv: 2006.03242. url: https://doi.org/10.1103%2Fphysrevlett.125.100602

    2020 arXiv

  46. [53]

    Constrained Differential Optimization

    John Platt and Alan Barr. “Constrained Differential Optimization”. In:Neural Information Pro- cessing Systems. Ed. by D. Anderson. Vol. 0. American Institute of Physics, 1987.url: https:// proceedings.neurips.cc/paper_files/paper/1987/file/a87ff679a2f3e71d9181a67b7542122c- Paper.pdf

    1987

  47. [54]

    Minimal-work protocols for inertial particles in non-harmonic traps

    Julia Sanders, Marco Baldovin, and Paolo Muratore-Ginanneschi. “Minimal-work protocols for inertial particles in non-harmonic traps”. In:Eprint arXiv:2407.15678 (July 22, 2024). doi: 10. 48550/ARXIV.2407.15678. arXiv: 2407.15678 [cond-mat.stat-mech]

    2024 doi arXiv

  48. [56]

    Efficiency at maximum power: An analytically solvable model for stochastic heat engines

    Tim Schmiedl and Udo Seifert. “Efficiency at maximum power: An analytically solvable model for stochastic heat engines”. In:EPL (Europhysics Letters)81.2 (Jan. 2008), p. 20003.issn: 1286-4854. doi: 10.1209/0295-5075/81/20003. arXiv: 0710.4097 [cond-mat.stat-mech]

    2008 arXiv

  49. [57]

    Optimal Control of Under- damped Systems: An Analytic Approach

    Julia Sanders, Marco Baldovin, and Paolo Muratore-Ginanneschi. “Optimal Control of Under- damped Systems: An Analytic Approach”. In:Journal of Statistical Physics191 (Mar. 1, 2024), p. 117. doi: 10.1007/s10955- 024- 03320- w. arXiv: 2403.00679 [cond-mat.stat-mech] . url: https...

    2024 arXiv

  50. [58]

    Über die Umkehrung der Naturgesetze

    Erwin Schrödinger. “Über die Umkehrung der Naturgesetze”. In:Sitzungsberichte der preussischen Akademie der Wissenschaften, physikalische mathematische Klasse8.9 (1931), pp. 144–153.doi: 10.1002/ange.19310443014

    1931 doi

  51. [59]

    Efficiency of molecular motors at maximum power

    Tim Schmiedl and Udo Seifert. “Efficiency of molecular motors at maximum power”. In: EPL (Europhysics Letters)83.3 (July 2008), p. 30005.issn: 1286-4854. doi: 10.1209/0295-5075/83/ 30005. arXiv: 0801.3743 [cond-mat.stat-mech]

    2008 arXiv

  52. [60]

    Stochastic Energetics

    Ken Sekimoto. Stochastic Energetics. Vol. 799. Lecture Notes in Physics. Springer, 2010, p. 322. isbn: 978-3-642-05410-5. doi: 10.1007/978-3-642-05411-2

    2010 doi

  53. [61]

    Stochastic thermodynamics, fluctuation theorems and molecular machines

    Udo Seifert. “Stochastic thermodynamics, fluctuation theorems and molecular machines”. In:Re- ports on Progress in Physics75.12 (Dec. 2012), p. 126001. doi: 10 . 1088 / 0034 - 4885 / 75 / 12 / 126001

    2012

  54. [62]

    Efficient computation of optimal actions

    Emanuel Todorov. “Efficient computation of optimal actions”. In: Proceedings of the National Academy of Sciences 106.28 (July 2009), pp. 11478–11483. doi: 10 . 1073 / pnas . 0710743106. url: http://Userss.cs.washington.edu/~todorov/papers.html

    2009

  55. [63]

    Estimating functions for diffusion-type processes

    Michael Sørensen. “Estimating functions for diffusion-type processes”. In: Statistical Methods for Stochastic Differential Equations. Ed. by Mathieu Kessler. Ed. by Alexander Lindner. Ed. by MichaelSorensen.NewYork:ChapmanandHall/CRCPress,May30,2012,p.507. isbn:9781439849767. ...

    2012 doi

  56. [64]

    Solving Schrödinger Bridges via Maximum Likelihood

    Francisco Vargas, Pierre Thodoroff, Austen Lamacraft, and Neil Lawrence. “Solving Schrödinger Bridges via Maximum Likelihood”. In: Entropy 23.9 (2021). issn: 1099-4300. doi: 10 . 3390 / e23091134. url: https://www.mdpi.com/1099-4300/23/9/1134

    2021

  57. [65]

    Theoretical guarantees for sampling and inference in gen- erative models with latent diffusions

    Belinda Tzen and Maxim Raginsky. “Theoretical guarantees for sampling and inference in gen- erative models with latent diffusions”. In:32nd Annual Conference on Learning Theory. Vol. 99. Mar. 5, 2019, pp. 1–31. arXiv:1903.01608 [math.PR]

    2019 arXiv

  58. [66]

    Stochastic flows and Bismut formulas for stochastic Hamiltonian systems

    Xicheng Zhang. “Stochastic flows and Bismut formulas for stochastic Hamiltonian systems”. In: Stochastic Processes and their Applications120.10 (Sept. 2010), pp. 1929–1949.doi: 10.1016/j. spa.2010.05.015. url: https://doi.org/10.1016%2Fj.spa.2010.05.015

    2010 doi

  59. [67]

    Optimal transport: old and new

    Cédric Villani. Optimal transport: old and new. Vol. 338. Grundlehren der mathematischen Wis- senschaften. Springer, 2009, p. 973.doi: 10.1007/978- 3- 540- 71050- 9 . url: https://link. springer.com/book/10.1007/978-3-540-71050-9. 30

    2009 doi

  60. [69]

    Nonequilibrium statistical mechanics

    Robert Zwanzig. Nonequilibrium statistical mechanics. Oxford University Press, 2001, p. 240.isbn: 978-0-19-514018-7. url: http://ukcatalogue.oup.com/product/9780195140187.do. 31

    2001

  61. [1574]

    doi: 10.3934/dcds.2014.34.1533

    issn: 1078-0947. doi: 10.3934/dcds.2014.34.1533. arXiv: 1308.0215 [math.PR]

    2014 arXiv

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.