François Delarue
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
François Delarue has a total of 54 co-authors. Here are the top co-authors by number of shared publications:
- René Carmona
Princeton University
5 shared publications - Etienne Tanré
Université Côte d'Azur
3 shared publications - Jean-François Chassagneux
Department of Mathematics Imperial College London
3 shared publications - Aime Lachapelle
ORFE, Bendheim Center for Finance, Princeton University, Princeton, USA
2 shared publications - Dan Crisan
Department of Mathematics, Imperial College London 1 , London SW7 2AZ
2 shared publications - Franco Flandoli
Scuola Normale Superiore of Pisa
2 shared publications
Latest Publications
- 21 April 2026
- Published by Springer Nature in Applied Mathematics & Optimization
- Vol. 93 (3)
Intrinsic Regularization by Noise for 1d Mean Field Games
- 21 April 2026
- Published by Springer Nature in Applied Mathematics & Optimization
- Vol. 93 (3)
Intrinsic Regularization by Noise for 1d Mean Field Games
- 9 April 2026
- Published by Springer Nature in Journal of Dynamics and Differential Equations
- Vol. 38 (S1) , 1-54
Zero Noise Limit for Multidimensional SDEs Driven by a Pointy Gradient
- 9 April 2026
- Published by Springer Nature in Journal of Dynamics and Differential Equations
- Vol. 38 (S1) , 1-54
Zero Noise Limit for Multidimensional SDEs Driven by a Pointy Gradient
- Open Access
- 1 January 2026
- Published by American Institute of Mathematical Sciences (AIMS) in Mathematical Control and Related Fields
- Vol. 19, 240-284
Robust mean-field games under entropy-based uncertainty
In this article, we introduce a new class of entropy-penalized robust mean-field game problems in which the representative agent is opposed to Nature. The agent's objective is formulated as a min–max stochastic control problem, in which Nature distorts the reference probability measure at an entropic cost. As a consequence, the distribution of the continuum of agents represented by the player is given by the effective measure induced by Nature. Existence of a mean-field game equilibrium is established via a Schauder fixed point argument. To ensure uniqueness, we introduce a signed flat-displacement monotonicity property that combines a form of flat monotonicity in Nature's direction with a form of displacement anti-monotonicity in the player's direction. This property extends the classical Lasry–Lions monotonicity framework. Finally, we present two classes of $ N $-player games for which the mean-field game limit yields $ \varepsilon $-Nash equilibria.- 1 January 2026
- Published by American Institute of Mathematical Sciences (AIMS) in Mathematical Control and Related Fields
- Vol. 19, 240-284
Open AccessRobust mean-field games under entropy-based uncertainty
In this article, we introduce a new class of entropy-penalized robust mean-field game problems in which the representative agent is opposed to Nature. The agent's objective is formulated as a min–max stochastic control problem, in which Nature distorts the reference probability measure at an entropic cost. As a consequence, the distribution of the continuum of agents represented by the player is given by the effective measure induced by Nature. Existence of a mean-field game equilibrium is established via a Schauder fixed point argument. To ensure uniqueness, we introduce a signed flat-displacement monotonicity property that combines a form of flat monotonicity in Nature's direction with a form of displacement anti-monotonicity in the player's direction. This property extends the classical Lasry–Lions monotonicity framework. Finally, we present two classes of $ N $-player games for which the mean-field game limit yields $ \varepsilon $-Nash equilibria. - Open Access
HJB equations driven by the Dirichlet-Ferguson Laplacian in Wasserstein-Sobolev spaces
We study linear and nonlinear PDEs defined on the space of $\mathcal{P}(\mathbb{T}^d)$ over the flat torus $\mathbb{T}^d$, equipped with the Dirichlet-Ferguson measure $\mathcal{D}$. We first develop an analytic framework based on the Wasserstein-Sobolev space $H^{1,2}(\mathcal{P}(\mathbb{T}^d), W_2, \mathcal{D})$ associated with the Dirichlet form induced by the infinite-dimensional Laplacian acting on functions of measures. Within this setting, we establish existence and uniqueness results for transport-diffusion and Hamilton-Jacobi equations in the Wasserstein space. Our analysis connects the PDE approach with a corresponding interacting particles system providing a probabilistic (Kolmogorov-type) representation of strong solutions. Finally, we extend the theory to semilinear equations and mean-field optimal control problems, together with consistent finite-dimensional approximations.Open AccessHJB equations driven by the Dirichlet-Ferguson Laplacian in Wasserstein-Sobolev spaces
We study linear and nonlinear PDEs defined on the space of $\mathcal{P}(\mathbb{T}^d)$ over the flat torus $\mathbb{T}^d$, equipped with the Dirichlet-Ferguson measure $\mathcal{D}$. We first develop an analytic framework based on the Wasserstein-Sobolev space $H^{1,2}(\mathcal{P}(\mathbb{T}^d), W_2, \mathcal{D})$ associated with the Dirichlet form induced by the infinite-dimensional Laplacian acting on functions of measures. Within this setting, we establish existence and uniqueness results for transport-diffusion and Hamilton-Jacobi equations in the Wasserstein space. Our analysis connects the PDE approach with a corresponding interacting particles system providing a probabilistic (Kolmogorov-type) representation of strong solutions. Finally, we extend the theory to semilinear equations and mean-field optimal control problems, together with consistent finite-dimensional approximations. - Top Cited Papers
- 1 August 2025
- Published by Institute for Operations Research and the Management Sciences (INFORMS) in Mathematics of Operations Research
- Vol. 50 (3) , 1762-1831
Exploration Noise for Learning Linear-Quadratic Mean Field Games
The goal of this paper is to demonstrate that common noise may serve as an exploration noise for learning the solution of a mean field game. This concept is here exemplified through a toy linear-quadratic model, for which a suitable form of common noise has already been proven to restore existence and uniqueness. We here go one step further and prove that the same form of common noise may force the convergence of the learning algorithm called fictitious play, and this without any further potential or monotone structure. Several numerical examples are provided to support our theoretical analysis. Funding: F. Delarue acknowledges the financial support of the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme [AdG ELISA project, Grant 101054746]. A. Vasileiadis acknowledge the financial support of French ANR project ANR-19-P3IA-0002-3IA Côte d'Azur-Nice-Interdisciplinary Institute for Artificial Intelligence.
- 1 August 2025
- Published by Institute for Operations Research and the Management Sciences (INFORMS) in Mathematics of Operations Research
- Vol. 50 (3) , 1762-1831
Top Cited PapersExploration Noise for Learning Linear-Quadratic Mean Field Games
The goal of this paper is to demonstrate that common noise may serve as an exploration noise for learning the solution of a mean field game. This concept is here exemplified through a toy linear-quadratic model, for which a suitable form of common noise has already been proven to restore existence and uniqueness. We here go one step further and prove that the same form of common noise may force the convergence of the learning algorithm called fictitious play, and this without any further potential or monotone structure. Several numerical examples are provided to support our theoretical analysis. Funding: F. Delarue acknowledges the financial support of the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme [AdG ELISA project, Grant 101054746]. A. Vasileiadis acknowledge the financial support of French ANR project ANR-19-P3IA-0002-3IA Côte d'Azur-Nice-Interdisciplinary Institute for Artificial Intelligence.
- Open Access
Genericity of Polyak-Lojasiewicz Inequalities for Entropic Mean-Field Neural ODEs
We address the behavior of idealized deep residual neural networks (ResNets), modeled via an optimal control problem set over continuity (or adjoint transport) equations. The continuity equations describe the statistical evolution of the features in the asymptotic regime where the layers of the network form a continuum. The velocity field is expressed through the network activation function, which is itself viewed as a function of the statistical distribution of the network parameters (weights and biases). From a mathematical standpoint, the control is interpreted in a relaxed sense, taking values in the space of probability measures over the set of parameters. We investigate the optimal behavior of the network when the cost functional arises from a regression problem and includes an additional entropic regularization term on the distribution of the parameters. In this framework, we focus in particular on the existence of stable optimizers --that is, optimizers at which the Hessian of the cost is non-degenerate. We show that, for an open and dense set of initial data, understood here as probability distributions over features and associated labels, there exists a unique stable global minimizer of the control problem. Moreover, we show that such minimizers satisfy a local Polyak--Lojasiewicz inequality, which can lead to exponential convergence of the corresponding gradient descent when the initialization lies sufficiently close to the optimal parameters. This result thus demonstrates the genericity (with respect to the distribution of features and labels) of the Polyak--Lojasiewicz condition in ResNets with a continuum of layers and under entropic penalization.
Open AccessGenericity of Polyak-Lojasiewicz Inequalities for Entropic Mean-Field Neural ODEs
We address the behavior of idealized deep residual neural networks (ResNets), modeled via an optimal control problem set over continuity (or adjoint transport) equations. The continuity equations describe the statistical evolution of the features in the asymptotic regime where the layers of the network form a continuum. The velocity field is expressed through the network activation function, which is itself viewed as a function of the statistical distribution of the network parameters (weights and biases). From a mathematical standpoint, the control is interpreted in a relaxed sense, taking values in the space of probability measures over the set of parameters. We investigate the optimal behavior of the network when the cost functional arises from a regression problem and includes an additional entropic regularization term on the distribution of the parameters. In this framework, we focus in particular on the existence of stable optimizers --that is, optimizers at which the Hessian of the cost is non-degenerate. We show that, for an open and dense set of initial data, understood here as probability distributions over features and associated labels, there exists a unique stable global minimizer of the control problem. Moreover, we show that such minimizers satisfy a local Polyak--Lojasiewicz inequality, which can lead to exponential convergence of the corresponding gradient descent when the initialization lies sufficiently close to the optimal parameters. This result thus demonstrates the genericity (with respect to the distribution of features and labels) of the Polyak--Lojasiewicz condition in ResNets with a continuum of layers and under entropic penalization.
- Open Access
- 5 June 2025
- Published by American Mathematical Society (AMS) in Proceedings of the American Mathematical Society, Series B
- Vol. 12 (11) , 134-149
Itô’s formula for the Rearranged Stochastic Heat Equation
The purpose of this short note is to prove a convenient version of Itô’s formula for the Rearranged Stochastic Heat Equation (RSHE) introduced by the two authors in a previous contribution. This equation is a penalized version of the standard Stochastic Heat Equation (SHE) on the circle subject to a colored noise, whose solution is constrained to stay within the set of symmetric quantile functions by means of a reflection term. Here, we identity the generator of the solution when it is acting on functions defined on the space(of one-dimensional probability measures with a finite second moment) that are assumed to be smooth in Lions’s sense. In particular, we prove that the reflection term in the RSHE is orthogonal to the Lions (or Wasserstein) derivative of smooth functions defined on . The proof relies on nontrivial bounds for the gradient of the solution to the RSHE. - 5 June 2025
- Published by American Mathematical Society (AMS) in Proceedings of the American Mathematical Society, Series B
- Vol. 12 (11) , 134-149
Open AccessItô’s formula for the Rearranged Stochastic Heat Equation
The purpose of this short note is to prove a convenient version of Itô’s formula for the Rearranged Stochastic Heat Equation (RSHE) introduced by the two authors in a previous contribution. This equation is a penalized version of the standard Stochastic Heat Equation (SHE) on the circle subject to a colored noise, whose solution is constrained to stay within the set of symmetric quantile functions by means of a reflection term. Here, we identity the generator of the solution when it is acting on functions defined on the space(of one-dimensional probability measures with a finite second moment) that are assumed to be smooth in Lions’s sense. In particular, we prove that the reflection term in the RSHE is orthogonal to the Lions (or Wasserstein) derivative of smooth functions defined on . The proof relies on nontrivial bounds for the gradient of the solution to the RSHE. - Open Access
- 1 May 2025
- Published by Institute of Mathematical Statistics in Annales de l'Institut Henri Poincaré, Probabilités et Statistiques
- Vol. 61 (2) , 1021-1074
Uniform in time weak propagation of chaos on the torus
Nous étudions le comportement en temps long de systèmes de particules en interaction faible à valeurs dans le tore de dimension d. Le résultat principal montre que, sous des hypothèses de régularité raisonnables, l’erreur faible entre la distribution empirique marginale des particules et la loi limite théorique (gouvernée par une équation de Fokker–Planck) est de l’ordre de O(1/N), uniformément en temps sur [0,∞), où N est le nombre de particules. Ce résultat s’applique en particulier aux équations de Fokker–Planck possédant une mesure invariante globalement attractive dont le linéarisé en la mesure invariante satisfait des propriétés ergodiques appropriées. Notre approche s’appuie sur une étude systématique du comportement en temps long des dérivées du semi-groupe généré par l’équation de Fokker–Planck. Cette démarche est suffisamment flexible pour couvrir des situations variées, dont le modèle sur-critique de Kuramoto, pour lequel l’équation de Fokker–Planck correspondante admet plusieurs mesures invariantes.
- 1 May 2025
- Published by Institute of Mathematical Statistics in Annales de l'Institut Henri Poincaré, Probabilités et Statistiques
- Vol. 61 (2) , 1021-1074
Open AccessUniform in time weak propagation of chaos on the torus
Nous étudions le comportement en temps long de systèmes de particules en interaction faible à valeurs dans le tore de dimension d. Le résultat principal montre que, sous des hypothèses de régularité raisonnables, l’erreur faible entre la distribution empirique marginale des particules et la loi limite théorique (gouvernée par une équation de Fokker–Planck) est de l’ordre de O(1/N), uniformément en temps sur [0,∞), où N est le nombre de particules. Ce résultat s’applique en particulier aux équations de Fokker–Planck possédant une mesure invariante globalement attractive dont le linéarisé en la mesure invariante satisfait des propriétés ergodiques appropriées. Notre approche s’appuie sur une étude systématique du comportement en temps long des dérivées du semi-groupe généré par l’équation de Fokker–Planck. Cette démarche est suffisamment flexible pour couvrir des situations variées, dont le modèle sur-critique de Kuramoto, pour lequel l’équation de Fokker–Planck correspondante admet plusieurs mesures invariantes.
- 11 March 2025
- Published by Springer Nature
- p. 27-119
An Example Driven Introduction to Probabilistic Rough Paths
- 11 March 2025
- Published by Springer Nature
- p. 27-119
An Example Driven Introduction to Probabilistic Rough Paths
- Open Access
- 1 January 2025
- Published by American Mathematical Society (AMS) in Proceedings of the American Mathematical Society, Series B
- Vol. 12 (11) , 134-149
Itô’s formula for the Rearranged Stochastic Heat Equation
The purpose of this short note is to prove a convenient version of Itô’s formula for the Rearranged Stochastic Heat Equation (RSHE) introduced by the two authors in a previous contribution. This equation is a penalized version of the standard Stochastic Heat Equation (SHE) on the circle subject to a colored noise, whose solution is constrained to stay within the set of symmetric quantile functions by means of a reflection term. Here, we identity the generator of the solution when it is acting on functions defined on the space ${\mathcal P}_2({\mathbb R})$ (of one-dimensional probability measures with a finite second moment) that are assumed to be smooth in Lions’s sense. In particular, we prove that the reflection term in the RSHE is orthogonal to the Lions (or Wasserstein) derivative of smooth functions defined on ${\mathcal P}_2({\mathbb R})$. The proof relies on nontrivial bounds for the gradient of the solution to the RSHE.- 1 January 2025
- Published by American Mathematical Society (AMS) in Proceedings of the American Mathematical Society, Series B
- Vol. 12 (11) , 134-149
Open AccessItô’s formula for the Rearranged Stochastic Heat Equation
The purpose of this short note is to prove a convenient version of Itô’s formula for the Rearranged Stochastic Heat Equation (RSHE) introduced by the two authors in a previous contribution. This equation is a penalized version of the standard Stochastic Heat Equation (SHE) on the circle subject to a colored noise, whose solution is constrained to stay within the set of symmetric quantile functions by means of a reflection term. Here, we identity the generator of the solution when it is acting on functions defined on the space ${\mathcal P}_2({\mathbb R})$ (of one-dimensional probability measures with a finite second moment) that are assumed to be smooth in Lions’s sense. In particular, we prove that the reflection term in the RSHE is orthogonal to the Lions (or Wasserstein) derivative of smooth functions defined on ${\mathcal P}_2({\mathbb R})$. The proof relies on nontrivial bounds for the gradient of the solution to the RSHE.