Rare events of small-noise Doob conditioned processes

arXiv:2605.24537 · cond-mat.stat-mech, math-ph, math.MP, math.PR · Submitted 2026-05-23 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Condensed Matter. The day in condensed matter: materials, phases and the measurements under them.

Kai: We're looking at the paper: "Rare events of small-noise Doob conditioned processes".

Paper summary: Kai: So, looking at the paper "Rare events of small-noise Doob conditioned processes," it really boils down to showing how you can encode conditioning through a terminal boundary term <ref:2605.24537#pg2>. They derive a generalized Feynman-Kac equation for the unnormalised generating function from this setup <ref:2605.24537#pg1>.

Mira: And the crucial part is that they found a Lagrangian variational representation for the action, which means we can treat these fluctuations as an optimal control problem <ref:2605.24537#pg1>. This method is useful because it lets us tackle systems where the explicit Doob drift isn't available analytically <ref:2605.24537#pg1>.

Lev: For hardware, this implies we might not need to solve the difficult backward Kolmogorov equation for every system; instead, we can use these optimization schemes directly when characterizing fluctuations of conditioned processes <ref:2605.24537#pg1>. It moves the complexity from solving an intractable differential equation to setting up an optimization problem <ref:2605.24537#pg1>.

Kai: The authors used this framework on path observables, like heat dissipated by a Doob-conditioned process <ref:2605.24537#pg1>. They found that for the Ornstein-Uhlenbeck bridge, the rare trajectories that connect the bottom of the potential to a value far away in fixed time T will dissipate heat <ref:2605.24537#pg1>.

Mira: That connects to how we think about thermodynamics and rare events in these systems <ref:2605.24537#pg1>. It shows that the larger the parameter s, which relates to the conditioning, the greater the heat dissipated under this Doob protocol <ref:2605.24537#pg1>.

Lev: But I have to mention what they didn't do here; they focus on small-noise regimes and weak-noise large deviation regime for their statistics <ref:2605.24537#pg1>. So, applying this directly to very high noise or highly non-linear systems might need more work <ref:2605.24537#pg1>.

Kai: Right, so the paper proves we can use post-selection to find a way around the difficulty of calculating Doob's h-function and instead get an action that comes from an optimal control formulation <ref:2605.24537#pg1>. That’s what they set up in this paper.

Conclusion: Kai: So, this paper is about finding rare paths in Markov processes by using a technique called Doob conditioning to make sure you hit a specific target at exactly one time.

Mira: It’s looking at how to find these rare paths when the noise is small, which is kind of what we actually expect in many physical systems.

Kai: The authors are focusing on modifying the drift term in a way that guarantees the process reaches a certain spot by a specific time, and they're using this to study the statistics of those conditioned paths.

Mira: It’s really about taking these complicated paths and turning them into something we can actually calculate using tools like moment generating functions.

Kai: They show that you can derive a kind of action principle from this whole conditioning setup, which turns the problem into an optimal control problem for characterising fluctuations.

Mira: That means they aren't just looking at one path; they’re looking at the whole ensemble of paths and finding a way to quantify how much those rare events contribute to the overall statistics.

Kai: It seems like this method could be useful for understanding things in quantum hardware where you want to know about very specific, rare error events that might otherwise be too hard to measure directly.

Mira: Exactly, because it gives us a way to encode that desired outcome into the math itself rather than just hoping the path happens.

Kai: We'll see how this variational approach plays out when we apply it to things like heat dissipation in these conditioned processes next.

Iago N. Mamede, Francesco Coghi

Instituto de Física da Universidade de Sãao Paulo · UHasselt Faculty of Sciences, Theory Lab · School of Physics and Astronomy University of Nottingham · Centre for the Mathematics and Theoretical Physics of Quantum Non-Equilibrium Systems University of Nottingham

cond-mat.stat-mech, math-ph, math.MP, math.PR

Submitted: 2026-05-23

Updated: 2026-11-16

Comments: 13 pages + Appendices, 4 figures (v2: fixed typos, added clarifications and more explanations, and updated references prior to publication in JSP)

License: http://creativecommons.org/licenses/by/4.0/

Importance score: 80/100

The gist: Doob fixed-time conditioning enables the sampling of rare trajectories of Markov processes by modifying the drift so that reaching a prescribed target at a given time is guaranteed The paper studies…

Key concepts

Doob's h-function
This function, ht(x), represents the probability that a Markov process starting at state x will reach a specific target state within a given time T. It is used to dynamically modify the process drift to ensure it hits the target at time T.
Modified diffusion (X~ϵ t)
This is the Doob conditioned process. Its dynamics are altered by adding a drift term proportional to the gradient of Doob's h-function and scaled by $\sqrt{\epsilon}$. This modification makes the resulting path measure optimal in terms of minimizing relative entropy compared to the original process.
Action (M~τ,x(s))
The action is a leading exponential term extracted from the moment generating function. It can be represented as a variational principle involving a Lagrangian. This allows researchers to characterize fluctuations of conditioned processes by treating them as optimal control problems.
Weak-noise limit ($\epsilon \to 0$)
When the noise parameter $\epsilon$ approaches zero, the generalized Feynman–Kac equation simplifies into a Hamilton–Jacobi problem. The solution to this problem yields the leading exponential contribution to the moment generating function, which is crucial for finding rare event statistics.

Terminology

Summary

Doob fixed-time conditioning enables the sampling of rare trajectories of Markov processes by modifying the drift so that reaching a prescribed target at a given time is guaranteed The paper studies the statistics of this conditioned path ensemble through the moment generating function in the weaknoise large deviation regime

Doob Conditioning Framework

The framework involves modifying the drift to realize a final-conditioned path law dynamically, where the additional drift is written as a function of Doob’s h-function This setup is formally defined by the Doob’s h-function as ht(x):= P(Xϵ T ∈ RXϵ t = x) The modified diffusion (X˜ϵ t) represents the Doob conditioned process and is a solution of dX˜ϵ t = F(X˜ϵ t) + ϵD ∇ ln ht(X˜ϵ t) dt + √ϵ σ ◦ dWt The Doob’s process infinitesimal generator L˜ϵ is obtained by Doob transforming the original process infinitesimal generator L ϵ via [23] This transformed process is optimal in the sense that its path measure minimizes the relative entropy with respect to the path measure of the original process

Moment Generating Function and Action

The moment generating function satisfies a generalized Feynman–Kac equation (for details on Feynman–Kac theory see [26, 27]) Its leading exponential behaviour is called the action, which is extracted from a Hamilton–Jacobi equation supplemented by a new boundary term The unnormalised moment generating function Zˆϵ τ,x(s) satisfies the generalized Feynman–Kac relation ∂τ + Lˆϵ s Zˆϵ τ,x(s) = 0 In the weak-noise limit ϵ → 0, this equation reduces to a Hamilton–Jacobi problem whose solution gives the leading exponential contribution to the generating function The action is represented by a variational principle where Mˆ τ,x(s) = sup xτ ∈ AC([τ,T]) xτ = k(xT) − Z T τ (L(xt, x˙ t)) dt

Application to Physical Observables

The framework is applied to path observables of the Doob process with time-additive form A T τ = Z T τ f(X˜ϵ t) dt + Z T τ g(X˜ϵ t) ◦ dX˜ϵ t The heat dissipated by a Doob-conditioned process is considered as an observable for f = 0 and g(x) = 2D−1(F(x) + D∇Mˆ t,x(0)) Fluctuations of this object can be studied by solving Eq. (32) with the specific Lagrangian L(xt, x˙ t) = 1/2 x˙ t − F(xt)2D − 2s(F(xt) + D∇Mˆ t,x(0)) · D−1x˙ t For the Brownian bridge, the area observable provides a simple benchmark for the variational formalism

Illustrative Results

For the one-dimensional Brownian bridge, the Euler–Lagrange equation associated with Eq. (42) is x¨t + s = 0 Solving this boundary value problem yields the optimal trajectory x∗ t = s squared t(T − t) + t(xf − x) T + x The mean rate at which area accumulates in a Brownian bridge process is given by A T 0:= AT 0/T = s 1/2 T squared + x + xf/2 For the Ornstein–Uhlenbeck bridge, the mean rate of dissipated heat over these instanton trajectories can be calculated using the instanton (53) The rare trajectories that connect the bottom of the potential with a value far away from it in a fixed time T will typically dissipate heat—the larger the value of s, the greater the heat dissipated under the Doob protocol

Conclusion and Outlook

The work demonstrates how to encode conditioning through a terminal boundary term and derive a generalized Feynman–Kac equation for the unnormalised generating function The paper shows that the action admits a Lagrangian variational representation, turning the problem of characterising fluctuations of conditioned processes into an optimal-control problem The method can be used numerically in higher-dimensional systems where analytical expressions for the Doob drift are not available Future work includes extending the formalism to genuinely timedependent drifts and studying many-particle systems The variational representation developed here suggests direct optimisation schemes for conditioned fluctuations in systems where the h-function is inaccessible Finally, it would be interesting to connect this weak-noise bridge formalism more closely with thermodynamic questions

Acknowledgments

IM acknowledges financial support from the Brazilian agency FAPESP under grant 2023/17704-2 and from the Special Research Fund (BOF) of Hasselt University under Grant No. BOF25BL12 FC is supported by a Leverhulme Early Career Fellowship No. p. 9

Where the paper says it

Conversely, atypical trajectories with negative s tend to reduce the heat dissipated by staying far away from the bottom of the OU potential for longer. For s < sc, a critical value that we identify in the following, such a behaviour is supported by absorbing heat from the environment for values of the conditioning-time T larger than a certain threshold (which depends on the target value xf ).

ECF-2025-482 The authors are grateful to Juan P. Garrahan and Kay Brandner for insightful discussions and to Nordita (Stockholm) for hosting the workshop ‘Fluctuations in self-interacting and learning processes’ in July 2025, during which this project was initiated

Appendix A Derivation

The derivation of the generalized Feynman–Kac relation via Itˆo’s calculus shows how to derive the Feynman–Kac Eqs. (20)–(21) for time-additive observables that include both state and current dependent terms The next step is to calculate the differential of Zˆϵ u,Xϵ t(s) defined in Eq. p. 3

Where the paper says it

The typical Brownian bridge trajectory is linear in time, whereas an OU bridge tends to get closer to the bottom of the quadratic potential quickly and stalls there until the final target is reached. SMALL-NOISE LARGE DEVIATIONS: HAMILTON–JACOBI FORMALISM We focus on path observables of the Doob process with time-additive form A T τ = Z T τ f(X˜ϵ t ) dt + Z T τ g(X˜ϵ t ) ◦ dX˜ϵ t , (15) which accumulate, linearly with time, state- and current-dependent features of the dynamics [19–21, 23].

(18), which will be useful in the following This is readily done by using Itˆo’s lemma and takes the form dZˆϵ u,Xϵ t = (∂uZˆϵ u,Xϵ u + L ϵZˆϵ u,Xϵ u)du + √ϵ∇(Zˆϵ u,Xϵ t)⊤σ(X˜u)dW

Appendix B Instanton Calculation

The solution of the instanton equation (53) is given by x∗ τ = A1P 1νs(coth[γ(T − τ)]) + A2Q 1νs(coth[γT]) + xf cosh[γ(T − τ)] The contribution to the action of Eq. (57) is therefore dominated by Mˆ 0,x(s) ∝ Z T 0 (T − t) 2νs dt This integral is finite if and only if νs > −1/2, which implies s 1/8, when the instanton develops complex behaviour

References

[1] S. Albeverio, V. Jentsch, and H. Kantz, eds., Extreme Events in Nature and Society (Springer, 2006) [2] M. p. 0

Where the paper says it

Rare events of small-noise Doob conditioned processes Iago N. Mamede1, 2, ∗ and Francesco Coghi3, 4, † 1 Instituto de F´ısica da Universidade de S˜ao Paulo, 05314-970 S˜ao Paulo, Brazil 2UHasselt, Faculty of Sciences, Theory Lab, Agoralaan, 3590 Diepenbeek, Belgium 3School of Physics and Astronomy, University of Nottingham, Nottingham, NG7 2RD, United Kingdom 4Centre for the Mathematics and Theoretical Physics of Quantum Non-Equilibrium Systems, University of Nottingham, Nottingham, NG7 2RD, United Kingdom Doob fixed-time conditioning enables the sampling of rare trajectories of Markov processes by modifying the drift so that reaching a prescribed target at a given time is guaranteed.

Scheffer, Critical Transitions in Nature and Society (Princeton University Press, 2009) [3] G. M. Wang, E. M. Sevick, E. Mittag, D. J. Searles, and D. J. Evans, Experimental demonstration of violations of the second law of thermodynamics for small systems and short time scales (Physical Review Letters 89, 050601 (2002)) [4] S. p. 0
Where the paper says it

However important they are, ensembles of rare trajectories are in general hard to access, both experimentally and computationally. Experiments are typically confined to typical behaviour and typical fluctuations of noisy systems, although carefully designed small-system experiments in stochastic thermodynamics have made it possible to access rare fluctuations at the trajectory level [3, 4].

Ciliberto, Experiments in stochastic thermodynamics: Short history and perspectives (Physical Review X 7, 021051 (2017)) [5] J. p. 0
Where the paper says it

However important they are, ensembles of rare trajectories are in general hard to access, both experimentally and computationally. Experiments are typically confined to typical behaviour and typical fluctuations of noisy systems, although carefully designed small-system experiments in stochastic thermodynamics have made it possible to access rare fluctuations at the trajectory level [3, 4].

A. Bucklew, Introduction to Rare Event Simulation (Springer, 2004) [6] C. p. 0
Where the paper says it

INTRODUCTION Having access to specific behaviours and dynamical regimes of processes influenced by noise is a central objective in the study of stochastic dynamics. This objective becomes increasingly important when the specific behaviour or event we are interested in is far from typical as, although rarely seen, atypical events may lead to drastic changes in the dynamics that generate long-time correlations potentially leading to cascading effects in the future [1, 2].

Giardin´a, J. Kurchan, V. Lecomte, and J. Tailleur, Simulating rare events in dynamical processes (Journal of Statistical Physics 145, 787 (2011)) [7] C. p. 0
Where the paper says it

Rare events of small-noise Doob conditioned processes Iago N. Mamede1, 2, ∗ and Francesco Coghi3, 4, † 1 Instituto de F´ısica da Universidade de S˜ao Paulo, 05314-970 S˜ao Paulo, Brazil 2UHasselt, Faculty of Sciences, Theory Lab, Agoralaan, 3590 Diepenbeek, Belgium 3School of Physics and Astronomy, University of Nottingham, Nottingham, NG7 2RD, United Kingdom 4Centre for the Mathematics and Theoretical Physics of Quantum Non-Equilibrium Systems, University of Nottingham, Nottingham, NG7 2RD, United Kingdom Doob fixed-time conditioning enables the sampling of rare trajectories of Markov processes by modifying the drift so that reaching a prescribed target at a given time is guaranteed.

Hartmann, R. Banisch, M. Sarich, T. Badowski, and C. Sch¨utte, Characterization of rare events in molecular dynamics (Entropy 16, 350 (2014)) [8] F. p. 0
Where the paper says it

This objective becomes increasingly important when the specific behaviour or event we are interested in is far from typical as, although rarely seen, atypical events may lead to drastic changes in the dynamics that generate long-time correlations potentially leading to cascading effects in the future [1, 2]. However important they are, ensembles of rare trajectories are in general hard to access, both experimentally and computationally.

Bouchet, J. Rolland, and J. Wouters, Rare event sampling methods (Chaos 29, 080402 (2019)) [9] T. p. 0
Where the paper says it

Once the ensemble of these rare trajectories is identified and reconstructed via Doob’s theory, one has access to its statistics, such as the typical conditioned dynamics as well as to the fluctuations of these conditioned processes and observables associated to them [18–21]. Fluctuations are in particular important because they help clarify whether the rare event is realised via a unique mechanism—then the typical rare-event dynamics is well representative—or multiple mechanisms/metastable dynamics, leading potentially to coexistence of dynamical phases and biases in sampling the conditioned dynamics.

Grafke and E. Vanden-Eijnden, Numerical computation of rare events via large deviation theory (Chaos 29, 063118 (2019)) [10] J. p. 0
Where the paper says it

Accessing rare events by brute force requires a computational effort that grows inversely with their probability, and therefore exponentially in the large-deviation regime. This difficulty has motivated rare-event theory to branch into several directions, leading to efficient numerical methods to sample rare trajectories [5–9].

L. Doob, Conditional Brownian motion and the boundary limits of harmonic functions (Bulletin de la Soci´et´e Math´ematique de France 85, 431 (1957)) [11] J. p. 1
Where the paper says it

V with a summary and outlook for future research. DOOB CONDITIONED PROCESSES We consider the time-homogeneous diffusion process (Xϵ t ) 0≤t≤T in R d solution of the following stochastic differential equation (SDE) dXϵ t = F(Xϵ t ) dt + √ ϵ σ ◦ dWt , (1) where F ∈ R d is a smooth space-dependent drift, σ ∈ R d×m is a constant noise matrix and Wt = R t 0 dWs is a Brownian motion in R m characterised by independent Gaussian increments, i.

L. Doob, Classical Potential Theory and Its Probabilistic Counterpart (Springer, 1984) [12] H. p. 3
Where the paper says it

4 Brownian bridge The classical one-dimensional Brownian bridge is obtained by setting F(x) = 0 in Eq. (1), then by fixing X0 = 0 and conditioning the process on reaching a target at time T.

Orland, Generating transition paths by Langevin bridges (Journal of Chemical Physics 134, 174114 (2011)) [13] S. p. 0
Where the paper says it

The outstanding result of Doob is that it is possible to derive a new conditioned process via the so-called Doob’s h-transform which introduces a new drift—in terms of the so-called Doob’s h-function—but keeps the same noise amplitude as well as preserves Markovianity. Doob’s framework is the mathematical framework of stochastic/Langevin bridges [12–14] as well as Schr¨odinger bridges [15] where the process is conditioned on reaching a target distribution rather than a target point.

N. Majumdar and H. Orland, Effective Langevin equations for constrained stochastic processes (Journal of Statistical Mechanics: Theory and Experiment 2015, P06039) [14] J. p. 0
Where the paper says it

We illustrate the framework with two analytical examples and with an application to heat dissipation of a minimal model of biomolecular folding. Keywords: large deviation theory; small noise; Doob processes; Feynman–Kac; Hamilton–Jacobi; stochastic control; biomolecule folding I.

Aguilar, J. W. Baron, T. Galla, and R. Toral, Sampling rare trajectories using stochastic bridges (Physical Review E 105, 064138 (2022)) [15] C. p. 0
Where the paper says it

Experiments are typically confined to typical behaviour and typical fluctuations of noisy systems, although carefully designed small-system experiments in stochastic thermodynamics have made it possible to access rare fluctuations at the trajectory level [3, 4]. Similarly, direct numerical simulations in silico, such as na¨ıve Monte Carlo sampling, are naturally confined to typical trajectories.

L´eonard, A survey of the Schr¨odinger problem and some of its connections with optimal transport (Discrete and Continuous Dynamical Systems - A 34, 1533 (2014)] [16] V. p. 0
Where the paper says it

Doob’s framework is the mathematical framework of stochastic/Langevin bridges [12–14] as well as Schr¨odinger bridges [15] where the process is conditioned on reaching a target distribution rather than a target point. Relevant examples include studying the finite-time dynamics of noisy systems moving between multiple basins of attraction such as icehouse-greenhouse transitions in climate systems [16] and open-to-native folding in proteins [17].

Lucarini and T. B´odai, Transitions across melancholia states in a climate model: Reconciling the deterministic and stochastic points of view (Physical Review Letters 122, 158701 (2019)] [17] J. p. 0
Where the paper says it

Doob’s framework is the mathematical framework of stochastic/Langevin bridges [12–14] as well as Schr¨odinger bridges [15] where the process is conditioned on reaching a target distribution rather than a target point. Relevant examples include studying the finite-time dynamics of noisy systems moving between multiple basins of attraction such as icehouse-greenhouse transitions in climate systems [16] and open-to-native folding in proteins [17].

N. Onuchic, Z. Luthey-Schulten, and P. G. Wolynes, Theory of protein folding: The energy landscape perspective (Annual Review of Physical Chemistry 48, 545 (1997)] [18] H. p. 0
Where the paper says it

We illustrate the framework with two analytical examples and with an application to heat dissipation of a minimal model of biomolecular folding. Keywords: large deviation theory; small noise; Doob processes; Feynman–Kac; Hamilton–Jacobi; stochastic control; biomolecule folding I.

Touchette, The large deviation approach to statistical mechanics (Physics Reports 478, 1 (2009)] [19] R. p. 0
Where the paper says it

Mamede1, 2, ∗ and Francesco Coghi3, 4, † 1 Instituto de F´ısica da Universidade de S˜ao Paulo, 05314-970 S˜ao Paulo, Brazil 2UHasselt, Faculty of Sciences, Theory Lab, Agoralaan, 3590 Diepenbeek, Belgium 3School of Physics and Astronomy, University of Nottingham, Nottingham, NG7 2RD, United Kingdom 4Centre for the Mathematics and Theoretical Physics of Quantum Non-Equilibrium Systems, University of Nottingham, Nottingham, NG7 2RD, United Kingdom Doob fixed-time conditioning enables the sampling of rare trajectories of Markov processes by modifying the drift so that reaching a prescribed target at a given time is guaranteed. We study the statistics of this conditioned path ensemble through the moment generating function in the weaknoise large deviation regime.

L. Jack and P. Sollich, Large deviations and ensembles of trajectories in stochastic models (Progress of Theoretical Physics Supplement 184, 304 (2010)] [20] J. p. 0
Where the paper says it

Mamede1, 2, ∗ and Francesco Coghi3, 4, † 1 Instituto de F´ısica da Universidade de S˜ao Paulo, 05314-970 S˜ao Paulo, Brazil 2UHasselt, Faculty of Sciences, Theory Lab, Agoralaan, 3590 Diepenbeek, Belgium 3School of Physics and Astronomy, University of Nottingham, Nottingham, NG7 2RD, United Kingdom 4Centre for the Mathematics and Theoretical Physics of Quantum Non-Equilibrium Systems, University of Nottingham, Nottingham, NG7 2RD, United Kingdom Doob fixed-time conditioning enables the sampling of rare trajectories of Markov processes by modifying the drift so that reaching a prescribed target at a given time is guaranteed. We study the statistics of this conditioned path ensemble through the moment generating function in the weaknoise large deviation regime.

P. Garrahan, Classical stochastic dynamics and continuous matrix product states: Gauge transformations, conditioned and driven processes, and equivalence of trajectory ensembles (Journal of Statistical Mechanics: Theory and Experiment 2016, 73208) [2605.245371,two] [21] H. p. 0
Where the paper says it

This viewpoint then yields an optimal-control representation for the leading exponential contribution to the generating function, expressed as a variational principle with terminal boundary conditions set by the Doob end-point constraint. We illustrate the framework with two analytical examples and with an application to heat dissipation of a minimal model of biomolecular folding.

Touchette, Introduction to dynamical large deviations of Markov processes (Physica A: Statistical Mechanics and its Applications 504, 5 (2018)] [22] M. p. 0
Where the paper says it

Keywords: large deviation theory; small noise; Doob processes; Feynman–Kac; Hamilton–Jacobi; stochastic control; biomolecule folding I. INTRODUCTION Having access to specific behaviours and dynamical regimes of processes influenced by noise is a central objective in the study of stochastic dynamics.

I. Freidlin and A. D. Wentzell, Random Perturbations of Dynamical Systems (Springer, 1984) [30] B. p. 0
Where the paper says it

Among all possible ensembles of rare trajectories one might be interested in, fixed-time conditioning a noisy system at a specific position in the state space has a prominent role. If the original system is Markov, then this setup can be seen as an instance of Doob’s original conditioning problem, which dealt with exiting a certain domain via a subset of its boundary at a random (first-passage) time [10, 11].

Derrida and T. Sadhu, Large deviations conditioned on large deviations I: Markov chain and Langevin equation (Journal of Statistical Physics 176, 773 (2019)] [31] F. p. 1
Where the paper says it

As discussed in [28], rather than studying the characteristics of the Hamilton–Jacobi equation, one switches to the Lagrangian formulation and writes the action in its optimal control representation, which is accessible through iterative numerical optimisation methods [9, 29]. We also remark that our approach is close in spirit to the weak-noise variational analysis of large deviations conditioned on large deviations developed in [30].

Baudoin, Conditioned stochastic differential equations: Theory, examples and application to finance (Stochastic Processes and their Applications 100, 109 (2002)] [32] J. p. 0
Where the paper says it

We illustrate the framework with two analytical examples and with an application to heat dissipation of a minimal model of biomolecular folding. Keywords: large deviation theory; small noise; Doob processes; Feynman–Kac; Hamilton–Jacobi; stochastic control; biomolecule folding I.

Lehec, Representation formula for the entropy and functional inequalities (Annales de l’Institut Henri Poincar´e, Probabilit´es et Statistiques 49, 885 (2013)] [33] P. Del Moral, Feynman-Kac Formulae: Genealogical and Interacting Particle Systems with Applications (Springer, 2004)] [34] K. p. 0
Where the paper says it

We illustrate the framework with two analytical examples and with an application to heat dissipation of a minimal model of biomolecular folding. Keywords: large deviation theory; small noise; Doob processes; Feynman–Kac; Hamilton–Jacobi; stochastic control; biomolecule folding I.

Sekimoto, Stochastic Energetics, Lecture Notes in Physics, Vol. p. 0
Where the paper says it

Rare events of small-noise Doob conditioned processes Iago N. Mamede1, 2, ∗ and Francesco Coghi3, 4, † 1 Instituto de F´ısica da Universidade de S˜ao Paulo, 05314-970 S˜ao Paulo, Brazil 2UHasselt, Faculty of Sciences, Theory Lab, Agoralaan, 3590 Diepenbeek, Belgium 3School of Physics and Astronomy, University of Nottingham, Nottingham, NG7 2RD, United Kingdom 4Centre for the Mathematics and Theoretical Physics of Quantum Non-Equilibrium Systems, University of Nottingham, Nottingham, NG7 2RD, United Kingdom Doob fixed-time conditioning enables the sampling of rare trajectories of Markov processes by modifying the drift so that reaching a prescribed target at a given time is guaranteed.

799 (Springer, 2010)] [35] U. Seifert, Stochastic thermodynamics, fluctuation theorems and molecular machines (Reports on Progress in Physics 75, 126001 (2012)] [36] W. H. Fleming and H. M. Soner, Controlled Markov Processes and Viscosity Solutions (Springer, 2006)] [37] M. p. 0
Where the paper says it

Rare events of small-noise Doob conditioned processes Iago N. Mamede1, 2, ∗ and Francesco Coghi3, 4, † 1 Instituto de F´ısica da Universidade de S˜ao Paulo, 05314-970 S˜ao Paulo, Brazil 2UHasselt, Faculty of Sciences, Theory Lab, Agoralaan, 3590 Diepenbeek, Belgium 3School of Physics and Astronomy, University of Nottingham, Nottingham, NG7 2RD, United Kingdom 4Centre for the Mathematics and Theoretical Physics of Quantum Non-Equilibrium Systems, University of Nottingham, Nottingham, NG7 2RD, United Kingdom Doob fixed-time conditioning enables the sampling of rare trajectories of Markov processes by modifying the drift so that reaching a prescribed target at a given time is guaranteed.

Assaf and B. Meerson, WKB theory of large deviations in stochastic populations (Journal of Physics A: Mathematical and Theoretical 50, 263001 (2017)] [38] M. p. 0
Where the paper says it

Mamede1, 2, ∗ and Francesco Coghi3, 4, † 1 Instituto de F´ısica da Universidade de S˜ao Paulo, 05314-970 S˜ao Paulo, Brazil 2UHasselt, Faculty of Sciences, Theory Lab, Agoralaan, 3590 Diepenbeek, Belgium 3School of Physics and Astronomy, University of Nottingham, Nottingham, NG7 2RD, United Kingdom 4Centre for the Mathematics and Theoretical Physics of Quantum Non-Equilibrium Systems, University of Nottingham, Nottingham, NG7 2RD, United Kingdom Doob fixed-time conditioning enables the sampling of rare trajectories of Markov processes by modifying the drift so that reaching a prescribed target at a given time is guaranteed. We study the statistics of this conditioned path ensemble through the moment generating function in the weaknoise large deviation regime.

F. Weber and E. Frey, Master equations and the theory of stochastic path integrals (Reports on Progress in Physics 80, 046601 (2017)] [39] F. p. 0
Where the paper says it

Mamede1, 2, ∗ and Francesco Coghi3, 4, † 1 Instituto de F´ısica da Universidade de S˜ao Paulo, 05314-970 S˜ao Paulo, Brazil 2UHasselt, Faculty of Sciences, Theory Lab, Agoralaan, 3590 Diepenbeek, Belgium 3School of Physics and Astronomy, University of Nottingham, Nottingham, NG7 2RD, United Kingdom 4Centre for the Mathematics and Theoretical Physics of Quantum Non-Equilibrium Systems, University of Nottingham, Nottingham, NG7 2RD, United Kingdom Doob fixed-time conditioning enables the sampling of rare trajectories of Markov processes by modifying the drift so that reaching a prescribed target at a given time is guaranteed. We study the statistics of this conditioned path ensemble through the moment generating function in the weaknoise large deviation regime.

den Hollander, Large Deviations (American Mathematical Society, 2000) [41] I. p. 1
Where the paper says it

As discussed in [28], rather than studying the characteristics of the Hamilton–Jacobi equation, one switches to the Lagrangian formulation and writes the action in its optimal control representation, which is accessible through iterative numerical optimisation methods [9, 29]. We also remark that our approach is close in spirit to the weak-noise variational analysis of large deviations conditioned on large deviations developed in [30].

N. Mamede, P. Singh, A. Pal, C. E. Fiore, and K. Proesmans, Work statistics at first-passage times (New Journal of Physics 26, 033034 (2024)] [42] S. p. 0
Where the paper says it

If the original system is Markov, then this setup can be seen as an instance of Doob’s original conditioning problem, which dealt with exiting a certain domain via a subset of its boundary at a random (first-passage) time [10, 11]. The outstanding result of Doob is that it is possible to derive a new conditioned process via the so-called Doob’s h-transform which introduces a new drift—in terms of the so-called Doob’s h-function—but keeps the same noise amplitude as well as preserves Markovianity.

N. Majumdar and A Comtet, Airy distribution function: From the area under a Brownian excursion to the maximal height of fluctuating interfaces (Journal of Statistical Physics 119, 777 (2005)] [43] F. p. 5
Where the paper says it

In the statistical physics literature, the approximation in the display above is also known as WKBJ (after Wentzel, Kramers, Brillouin and Jeffreys) ansatz [37, 38]. By returning to normalised objects using Mϵ τ,x(s) = Mˆ ϵ τ,x(s)−ϵ ln hτ (x), the finite-time weak-noise scaled cumulant generating function (SCGF) takes the form Mτ,x(s) = Mˆ τ,x(s) − Mˆ τ,x(0), (28)

W. J. Olver, D. W. Lozier, R. F. Boisvert, and C. W Clark, eds., NIST Handbook of Mathematical Functions (Cambridge University Press, 2010)] [44] J Grela, S N Majumdar, and G Schehr, Non-intersecting Brownian bridges in the flat-to-flat geometry (Journal of Statistical Physics 183, 49 (2021)]

The solution of the instanton equation (53) is given by x∗ τ = A1P 1νs(coth[γ(T − τ)]) + A2Q 1νs(coth[γT]) + xf cosh[γ(T − τ)] The contribution to the action of Eq.

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Optimization of rare event sampling for complex systems using optimal control representation by reinterpreting Doob conditioning as post-selection on a terminal constraint, yielding an optimal-control representation for the leading exponential contribution to the generating function.

  2. Derivation of a generalized Feynman–Kac equation that accounts for time-additive observables, including state and current dependent terms, allowing for the study of path statistics in systems with multiplicative noise.

  3. Application of Lagrangian variational representation to characterize fluctuations of conditioned processes, transforming the problem into an optimal control problem to find the most probable transition paths connecting metastable states in biomolecular folding models.

  4. Calculation of instanton trajectories for Ornstein–Uhlenbeck bridges using a hyperbolic Pöschl–Teller equation with a source term, which reveals how the Doob protocol reshapes heat dissipation along rare conditioned trajectories.

  5. Development of a numerical framework that allows for the study of fluctuations in high-dimensional systems and interacting diffusions by utilizing the variational representation derived from the action functional.

Abstract

Doob fixed-time conditioning enables the sampling of rare trajectories of Markov processes by modifying the drift so that reaching a prescribed target at a given time is guaranteed. We study the statistics of this conditioned path ensemble through the moment generating function in the weak-noise large deviation regime. Since the Doob drift is rarely available in closed form, we reinterpret the conditioned ensemble as the original process post-selected on the terminal constraint, thereby avoiding explicit construction of the Doob transform. This viewpoint then yields an optimal-control representation for the leading exponential contribution to the generating function, expressed as a variational principle with terminal boundary conditions set by the Doob end-point constraint. We illustrate the framework with two analytical examples and with an application to heat dissipation under Doob-controlled dynamics in a minimal model of biomolecular folding.

Related papers