Youssef Mesri
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
Youssef Mesri has a total of 46 co-authors. Here are the top co-authors by number of shared publications:
- Elie Hachem
Computing and Fluids Research Group (CFL), Center for Material Forming (CEMEF), Mines Paris, PSL University, UMR7635 CNRS, Sophia Antipolis, 06904, France
3 shared publications Alban Bazile
CEMEF - Centre for Material Forming, CNRS UMR 7635
2 shared publications- Augustin Parret-Fréaud
Digital Sciences and Technologies, Safran Tech, Chateaufort, France
2 shared publications - Claude Tadonki
Centre de Recherche en Informatique (CRI) Mines Paris -PSL University Fontainebleau France
2 shared publications E. Hachem
MINES ParisTech, PSL – Research University, CEMEF – Center for material forming, CNRS UMR 7635, CS 10207 rue Claude Daunesse 06904 Sophia Antipolis Cedex, France
2 shared publicationsElie Hachem
CEMEF - Centre for Material Forming, CNRS UMR 7635
2 shared publications
Latest Publications
- Open Access
- 11 May 2026
- Published by The Open Journal in Journal of Open Source Software
- Vol. 11 (121) , 9729
SOFIA: A Python Library for High-Quality 2D Triangular Mesh Adaptation
- 11 May 2026
- Published by The Open Journal in Journal of Open Source Software
- Vol. 11 (121) , 9729
Open AccessSOFIA: A Python Library for High-Quality 2D Triangular Mesh Adaptation
- Open Access
- 23 April 2026
- Published by Elsevier in Acta Materialia
- Vol. 312
InfinityEBSD : Metrics-guided infinite-size EBSD map generation with diffusion models
- 23 April 2026
- Published by Elsevier in Acta Materialia
- Vol. 312
Open AccessInfinityEBSD : Metrics-guided infinite-size EBSD map generation with diffusion models
- Open Access
- 17 January 2026
- Published by MDPI AG in Applied Sciences
- Vol. 16 (2) , 972
CompNO: A Novel Foundation Model Approach for Solving Partial Differential Equations
Partial differential equations (PDEs) govern a wide range of physical phenomena, but their numerical solution remains computationally demanding, especially when repeated simulations are required across many parameter settings. Recent Scientific Foundation Models (SFMs) aim to alleviate this cost by learning universal surrogates from large collections of simulated systems, yet they typically rely on monolithic architectures with limited interpretability and high pretraining expense. In this work, we introduce Compositional Neural Operators (CompNO), a compositional neural operator framework for parametric PDEs. Instead of pretraining a single large model on heterogeneous data, CompNO first learns a library of Foundation Blocks, where each block is a parametric Fourier neural operator specialized to a fundamental differential operator (e.g., convection, diffusion, nonlinear convection). These blocks are then assembled, via lightweight Adaptation Blocks, into task-specific solvers that approximate the temporal evolution operator for target PDEs. A dedicated boundary-condition operator further enforces Dirichlet constraints exactly at inference time. We validate CompNO on one-dimensional convection, diffusion, convection–diffusion and Burgers’ equations from the PDEBench suite. The proposed framework achieves lower relative error than strong baselines (PFNO, PDEFormer and in-context learning-based models) on linear parametric systems, while remaining competitive on nonlinear Burgers’ flows. The model maintains exact boundary satisfaction with zero loss at domain boundaries, and exhibits robust generalization across a broad range of Péclet and Reynolds numbers. These results demonstrate that Compositional Neural Operators provide a scalable and physically interpretable pathway towards foundation models for PDEs.- 17 January 2026
- Published by MDPI AG in Applied Sciences
- Vol. 16 (2) , 972
Open AccessCompNO: A Novel Foundation Model Approach for Solving Partial Differential Equations
Partial differential equations (PDEs) govern a wide range of physical phenomena, but their numerical solution remains computationally demanding, especially when repeated simulations are required across many parameter settings. Recent Scientific Foundation Models (SFMs) aim to alleviate this cost by learning universal surrogates from large collections of simulated systems, yet they typically rely on monolithic architectures with limited interpretability and high pretraining expense. In this work, we introduce Compositional Neural Operators (CompNO), a compositional neural operator framework for parametric PDEs. Instead of pretraining a single large model on heterogeneous data, CompNO first learns a library of Foundation Blocks, where each block is a parametric Fourier neural operator specialized to a fundamental differential operator (e.g., convection, diffusion, nonlinear convection). These blocks are then assembled, via lightweight Adaptation Blocks, into task-specific solvers that approximate the temporal evolution operator for target PDEs. A dedicated boundary-condition operator further enforces Dirichlet constraints exactly at inference time. We validate CompNO on one-dimensional convection, diffusion, convection–diffusion and Burgers’ equations from the PDEBench suite. The proposed framework achieves lower relative error than strong baselines (PFNO, PDEFormer and in-context learning-based models) on linear parametric systems, while remaining competitive on nonlinear Burgers’ flows. The model maintains exact boundary satisfaction with zero loss at domain boundaries, and exhibits robust generalization across a broad range of Péclet and Reynolds numbers. These results demonstrate that Compositional Neural Operators provide a scalable and physically interpretable pathway towards foundation models for PDEs. - Open Access
CompNO: A Novel Foundation Model approach for solving Partial Differential Equations
Partial differential equations (PDEs) govern a wide range of physical phenomena, but their numerical solution remains computationally demanding, especially when repeated simulations are required across many parameter settings. Recent Scientific Foundation Models (SFMs) aim to alleviate this cost by learning universal surrogates from large collections of simulated systems, yet they typically rely on monolithic architectures with limited interpretability and high pretraining expense. In this work we introduce Compositional Neural Operators (CompNO), a compositional neural operator framework for parametric PDEs. Instead of pretraining a single large model on heterogeneous data, CompNO first learns a library of Foundation Blocks, where each block is a parametric Fourier neural operator specialized to a fundamental differential operator (e.g. convection, diffusion, nonlinear convection). These blocks are then assembled, via lightweight Adaptation Blocks, into task-specific solvers that approximate the temporal evolution operator for target PDEs. A dedicated boundary-condition operator further enforces Dirichlet constraints exactly at inference time. We validate CompNO on one-dimensional convection, diffusion, convection--diffusion and Burgers' equations from the PDEBench suite. The proposed framework achieves lower relative L2 error than strong baselines (PFNO, PDEFormer and in-context learning based models) on linear parametric systems, while remaining competitive on nonlinear Burgers' flows. The model maintains exact boundary satisfaction with zero loss at domain boundaries, and exhibits robust generalization across a broad range of Peclet and Reynolds numbers. These results demonstrate that compositional neural operators provide a scalable and physically interpretable pathway towards foundation models for PDEs.
Open AccessCompNO: A Novel Foundation Model approach for solving Partial Differential Equations
Partial differential equations (PDEs) govern a wide range of physical phenomena, but their numerical solution remains computationally demanding, especially when repeated simulations are required across many parameter settings. Recent Scientific Foundation Models (SFMs) aim to alleviate this cost by learning universal surrogates from large collections of simulated systems, yet they typically rely on monolithic architectures with limited interpretability and high pretraining expense. In this work we introduce Compositional Neural Operators (CompNO), a compositional neural operator framework for parametric PDEs. Instead of pretraining a single large model on heterogeneous data, CompNO first learns a library of Foundation Blocks, where each block is a parametric Fourier neural operator specialized to a fundamental differential operator (e.g. convection, diffusion, nonlinear convection). These blocks are then assembled, via lightweight Adaptation Blocks, into task-specific solvers that approximate the temporal evolution operator for target PDEs. A dedicated boundary-condition operator further enforces Dirichlet constraints exactly at inference time. We validate CompNO on one-dimensional convection, diffusion, convection--diffusion and Burgers' equations from the PDEBench suite. The proposed framework achieves lower relative L2 error than strong baselines (PFNO, PDEFormer and in-context learning based models) on linear parametric systems, while remaining competitive on nonlinear Burgers' flows. The model maintains exact boundary satisfaction with zero loss at domain boundaries, and exhibits robust generalization across a broad range of Peclet and Reynolds numbers. These results demonstrate that compositional neural operators provide a scalable and physically interpretable pathway towards foundation models for PDEs.
- Open Access
Benchmark-based Study of CPU/GPU Power-Related Features through JAX and TensorFlow
Power management has become a crucial focus in the modern computing landscape, considering that {\em energy} is increasingly recognized as a critical resource. This increased the importance of all topics related to {\em energy-aware computing}. This paper presents an experimental study of three prevalent power management techniques that are {\em power limitation, frequency limitation}, and {\em ACPI/P-State governor modes} (OS states related to power consumption). Through a benchmark approach with a set of six computing kernels, we investigate {\em power/performance} trade-off with various hardware units and software frameworks (mainly TensorFlow and JAX). Our experimental results show that {\em frequency limitation} is the most effective technique to improve {\em Energy-Delay Product (EDP)}, which is a convolution of energy and running time. We also observe that running at the highest frequency compared to a reduced one could lead to a reduction of factor $\frac{1}{10}$ in EDP. Another noticeable fact is that frequency management shows a consistent behavior with different CPUs, whereas opposite effects sometimes occur between TensorFlow (TF) and JAX with the same power management settings.Open AccessBenchmark-based Study of CPU/GPU Power-Related Features through JAX and TensorFlow
Power management has become a crucial focus in the modern computing landscape, considering that {\em energy} is increasingly recognized as a critical resource. This increased the importance of all topics related to {\em energy-aware computing}. This paper presents an experimental study of three prevalent power management techniques that are {\em power limitation, frequency limitation}, and {\em ACPI/P-State governor modes} (OS states related to power consumption). Through a benchmark approach with a set of six computing kernels, we investigate {\em power/performance} trade-off with various hardware units and software frameworks (mainly TensorFlow and JAX). Our experimental results show that {\em frequency limitation} is the most effective technique to improve {\em Energy-Delay Product (EDP)}, which is a convolution of energy and running time. We also observe that running at the highest frequency compared to a reduced one could lead to a reduction of factor $\frac{1}{10}$ in EDP. Another noticeable fact is that frequency management shows a consistent behavior with different CPUs, whereas opposite effects sometimes occur between TensorFlow (TF) and JAX with the same power management settings. - Open Access
- 20 September 2024
- Published by MDPI AG in Mathematics
- Vol. 12 (18) , 2933
Graph Neural Networks for Mesh Generation and Adaptation in Structural and Fluid Mechanics
The finite element discretization of computational physics problems frequently involves the manual generation of an initial mesh and the application of adaptive mesh refinement (AMR). This approach is employed to selectively enhance the accuracy of resolution in regions that encompass significant features throughout the simulation process. In this paper, we introduce Adaptnet, a Graph Neural Networks (GNNs) framework for learning mesh generation and adaptation. The model is composed of two GNNs: the first one, Meshnet, learns mesh parameters commonly used in open-source mesh generators, to generate an initial mesh from a Computer Aided Design (CAD) file; while the second one, Graphnet, learns mesh-based simulations to predict the components of an Hessian-based metric to perform anisotropic mesh adaptation. Our approach is tested on structural (Deforming plate–Linear elasticity) and fluid mechanics (Flow around cylinders–steady-state Stokes) problems. Our findings demonstrate the model’s ability to precisely predict the dynamics of the system and adapt the mesh as needed. The adaptability of the model enables learning resolution-independent mesh-based simulations during training, allowing it to scale effectively to more intricate state spaces during inference.
- 20 September 2024
- Published by MDPI AG in Mathematics
- Vol. 12 (18) , 2933
Open AccessGraph Neural Networks for Mesh Generation and Adaptation in Structural and Fluid Mechanics
The finite element discretization of computational physics problems frequently involves the manual generation of an initial mesh and the application of adaptive mesh refinement (AMR). This approach is employed to selectively enhance the accuracy of resolution in regions that encompass significant features throughout the simulation process. In this paper, we introduce Adaptnet, a Graph Neural Networks (GNNs) framework for learning mesh generation and adaptation. The model is composed of two GNNs: the first one, Meshnet, learns mesh parameters commonly used in open-source mesh generators, to generate an initial mesh from a Computer Aided Design (CAD) file; while the second one, Graphnet, learns mesh-based simulations to predict the components of an Hessian-based metric to perform anisotropic mesh adaptation. Our approach is tested on structural (Deforming plate–Linear elasticity) and fluid mechanics (Flow around cylinders–steady-state Stokes) problems. Our findings demonstrate the model’s ability to precisely predict the dynamics of the system and adapt the mesh as needed. The adaptability of the model enables learning resolution-independent mesh-based simulations during training, allowing it to scale effectively to more intricate state spaces during inference.
- Open Access
- 27 August 2024
- Published by MDPI AG in Preprints.org
Graph Neural Networks for Mesh Generation and Adaptation in Structural and Fluid Mechanics
Finite element discretizations of problems in computational physics often rely on hand-generated initial mesh and adaptive mesh refinement (AMR) to preferentially resolve regions containing important features during simulation. We propose Adaptnet, a Graph Neural Networks (GNNs) framework for learning mesh generation and adaptation. The model is composed of two GNNs: the first one, Meshnet, learns mesh parameters commonly used in open-source mesh generators, to generate an initial mesh from a Computer Aid Design (CAD) file; while the second one, Graphnet, learns mesh-based simulations to predict the components of an Hessian-based metric to perform anisotropic mesh adaptation. Our approach is tested on structural (Deforming plate - Linear elasticity) and fluid mechanics (Flow around cylinders - steady-state Stokes) problems. Our results show it can accurately predict the dynamics of the system and adapt the mesh accordingly. The adaptivity of the model supports learning resolution-independent dynamics and can scale to more complex state spaces at test time.
- 27 August 2024
- Published by MDPI AG in Preprints.org
Open AccessGraph Neural Networks for Mesh Generation and Adaptation in Structural and Fluid Mechanics
Finite element discretizations of problems in computational physics often rely on hand-generated initial mesh and adaptive mesh refinement (AMR) to preferentially resolve regions containing important features during simulation. We propose Adaptnet, a Graph Neural Networks (GNNs) framework for learning mesh generation and adaptation. The model is composed of two GNNs: the first one, Meshnet, learns mesh parameters commonly used in open-source mesh generators, to generate an initial mesh from a Computer Aid Design (CAD) file; while the second one, Graphnet, learns mesh-based simulations to predict the components of an Hessian-based metric to perform anisotropic mesh adaptation. Our approach is tested on structural (Deforming plate - Linear elasticity) and fluid mechanics (Flow around cylinders - steady-state Stokes) problems. Our results show it can accurately predict the dynamics of the system and adapt the mesh accordingly. The adaptivity of the model supports learning resolution-independent dynamics and can scale to more complex state spaces at test time.
- Open Access
- 1 July 2023
- Published by AIP Publishing in Physics of Fluids
- Vol. 35 (7)
A graph neural network-based framework to identify flow phenomena on unstructured meshes
Driven by the abundant data generated from computational fluid dynamics (CFD) simulations, machine learning (ML) methods surpass the deterministic criteria on flow phenomena identification in the way that they are independent of a case-by-case threshold by combining the flow field properties and the topological distribution of the phenomena. The current most popular and successful ML models based on convolutional neural networks are limited to structured meshes and unable to directly digest the data generated from unstructured meshes, which are more widely used in real industrial CFD simulations. We proposed a framework based on graph neural networks with the proposed fast Gaussian mixture model as the convolution kernel and U-Net architecture to detect flow phenomena based on a graph hierarchy generated by the algebraic multigrid method embedded in the open-source CFD solver, code_saturne. We demonstrate the superiority of the proposed kernel and U-Net architecture, along with the generality of the framework to unstructured mesh and unseen case, on detecting the vortices once trained on a single backward-facing step case. Our proposed framework can be trivially extended to detect more flow phenomena in three dimensional cases, which is ongoing work.
- 1 July 2023
- Published by AIP Publishing in Physics of Fluids
- Vol. 35 (7)
Open AccessA graph neural network-based framework to identify flow phenomena on unstructured meshes
Driven by the abundant data generated from computational fluid dynamics (CFD) simulations, machine learning (ML) methods surpass the deterministic criteria on flow phenomena identification in the way that they are independent of a case-by-case threshold by combining the flow field properties and the topological distribution of the phenomena. The current most popular and successful ML models based on convolutional neural networks are limited to structured meshes and unable to directly digest the data generated from unstructured meshes, which are more widely used in real industrial CFD simulations. We proposed a framework based on graph neural networks with the proposed fast Gaussian mixture model as the convolution kernel and U-Net architecture to detect flow phenomena based on a graph hierarchy generated by the algebraic multigrid method embedded in the open-source CFD solver, code_saturne. We demonstrate the superiority of the proposed kernel and U-Net architecture, along with the generality of the framework to unstructured mesh and unseen case, on detecting the vortices once trained on a single backward-facing step case. Our proposed framework can be trivially extended to detect more flow phenomena in three dimensional cases, which is ongoing work.
- Open Access
- 23 March 2022
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Parallel Computing
- Vol. 9 (1) , 1-23
High-performance 3D Unstructured Mesh Deformation Using Rank Structured Matrix Computations
The Radial Basis Function (RBF) technique is an interpolation method that produces high-quality unstructured adaptive meshes. However, the RBF-based boundary problem necessitates solving a large dense linear system with cubic arithmetic complexity that is computationally expensive and prohibitive in terms of memory footprint. In this article, we accelerate the computations of 3D unstructured mesh deformation based on RBF interpolations by exploiting the rank structured property of the matrix operator. The main idea consists in approximating the matrix off-diagonal tiles up to an application-dependent accuracy threshold. We highlight the robustness of our multiscale solver by assessing its numerical accuracy using realistic 3D geometries. In particular, we model the 3D mesh deformation on a population of the novel coronaviruses. We report and compare performance results on various parallel systems against existing state-of-the-art matrix solvers.
- 23 March 2022
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Parallel Computing
- Vol. 9 (1) , 1-23
Open AccessHigh-performance 3D Unstructured Mesh Deformation Using Rank Structured Matrix Computations
The Radial Basis Function (RBF) technique is an interpolation method that produces high-quality unstructured adaptive meshes. However, the RBF-based boundary problem necessitates solving a large dense linear system with cubic arithmetic complexity that is computationally expensive and prohibitive in terms of memory footprint. In this article, we accelerate the computations of 3D unstructured mesh deformation based on RBF interpolations by exploiting the rank structured property of the matrix operator. The main idea consists in approximating the matrix off-diagonal tiles up to an application-dependent accuracy threshold. We highlight the robustness of our multiscale solver by assessing its numerical accuracy using realistic 3D geometries. In particular, we model the 3D mesh deformation on a population of the novel coronaviruses. We report and compare performance results on various parallel systems against existing state-of-the-art matrix solvers.