JAX for HPC/AI Hybridization Hackathon
Workshop talk, JAX for HPC/AI Hybridization Hackathon, IDRIS, Paris-Saclay, France
Advanced participant at the JAX for HPC/AI Hybridization Hackathon, September 7–9, 2026.
Talks and presentations
Upcoming Talks
Talk at IAS Frontiers Conference on Geometry, Dynamics, and Learning (title TBA)
Conference talk, IAS Frontiers Conference on Geometry, Dynamics, and Learning (GDL2026), Nanyang Technological University, Singapore
Talk at GDL2026, an international conference at the intersection of geometric analysis, dynamical systems, and machine learning.
Talk at STOP 2026: Structured and Large Scale Optimization (title TBA)
Workshop talk, STOP 2026 Workshop — Structured and Large Scale Optimization, ENS Lyon, Lyon, France
Talk at STOP 2026, a workshop on structured and large scale optimization covering inverse problems, multilevel and stochastic optimization, scientific machine learning, and matrix decomposition.
Conferences
Talk at IAS Frontiers Conference on Geometry, Dynamics, and Learning (title TBA)
Conference talk, IAS Frontiers Conference on Geometry, Dynamics, and Learning (GDL2026), Nanyang Technological University, Singapore
Talk at GDL2026, an international conference at the intersection of geometric analysis, dynamical systems, and machine learning.
Adaptive Sampling in PINNs through the lens of Kernel Theory
Conference talk, 47e Congrès National d'Analyse Numérique (CANUM), Saint-Jacut-de-la-Mer, France
Talk presenting joint work with Sheng Wan, Cyriaque Rousselot, Shorouk El-Hassanieh, Alena Shilova, and Cyril Furtlehner on adaptive sampling strategies for Physics-Informed Neural Networks (PINNs) through the lens of kernel theory.
Toward a Unified Framework for PINNs and FEMs: A Petrov–Galerkin perspective through the lens of kernel theory
Conference talk, 47e Congrès National d'Analyse Numérique (CANUM), Saint-Jacut-de-la-Mer, France
Talk presenting joint work with Roland Maier (Institute for Applied and Numerical Mathematics, KIT) on a unified framework connecting Physics-Informed Neural Networks (PINNs) and Finite Element Methods (FEMs) through a Petrov–Galerkin perspective and kernel theory.
Unifying PINNs and FEMs through the notion of Natural Neural Tangent Kernel
Conference talk, European Conference on Numerical Mathematics and Advanced Applications (ENUMATH 2025), Heidelberg, Germany
Talk presenting joint work with Roland Maier (KIT) on a unifying perspective on Physics-Informed Neural Networks (PINNs) and Finite Element Methods (FEMs) through the notion of Natural Neural Tangent Kernel.
ANaGRAM: A Natural Gradient Relative to Adapted Model for efficient PINNs learning
Conference talk, 12ème Biennale de la SMAI, Carcans-Plage, France
Talk presenting ANaGRAM, a theoretically founded and algorithmically efficient natural gradient method for training Physics-Informed Neural Networks (PINNs), at the 12th Biennale of the Société de Mathématiques Appliquées et Industrielles (SMAI).
ANaGRAM: A Natural Gradient Relative to Adapted Model for efficient PINNs learning
Conference talk, International Conference on Learning Representations (ICLR 2025), Singapore EXPO, Singapore
Presentation of ANaGRAM at ICLR 2025. ANaGRAM is a theoretically founded and algorithmically efficient natural gradient method for training Physics-Informed Neural Networks (PINNs), connecting natural gradient descent to the Neural Tangent Kernel and the Green function of the PDE operator.
Workshops & Schools
Talk at STOP 2026: Structured and Large Scale Optimization (title TBA)
Workshop talk, STOP 2026 Workshop — Structured and Large Scale Optimization, ENS Lyon, Lyon, France
Talk at STOP 2026, a workshop on structured and large scale optimization covering inverse problems, multilevel and stochastic optimization, scientific machine learning, and matrix decomposition.
JAX for HPC/AI Hybridization Hackathon
Workshop talk, JAX for HPC/AI Hybridization Hackathon, IDRIS, Paris-Saclay, France
Advanced participant at the JAX for HPC/AI Hybridization Hackathon, September 7–9, 2026.
Efficient PINNs and Natural Gradients in JAX
Workshop talk, ML & AD in JAX for Scientific Computing Workshop, Strasbourg, France
Talk presenting an efficient JAX implementation of natural gradient methods for Physics-Informed Neural Networks (PINNs).
Natural Gradients and Kernel Methods for PINNs
Workshop talk, HACE Workshop — HPC/AI Hybridization, AISSAI, Toulouse, France
Talk presenting natural gradient and kernel methods for Physics-Informed Neural Networks (PINNs).
ANaGRAM: A Natural Gradient Relative to Adapted Model for efficient PINNs learning
Workshop talk, Numerical Analysis School, EDF, Palaiseau, France
Poster presentation of ANaGRAM at the EDF Numerical Analysis school.
Geometrical perspectives on Physics-Informed Neural Networks
Workshop talk, Méca–AAC day, LISN, Paris-Saclay University – CNRS, Gif-sur-Yvette, France
Talk given at the Méca–AAC day, presenting geometrical perspectives on the training of Physics-Informed Neural Networks (PINNs).
Addressing Spectral Bias in PINNs
Workshop talk, SCALP Kick-off, LISN, Paris-Saclay University – CNRS, Gif-sur-Yvette, France
Talk given at the kick-off meeting of the SCALP project, presenting work on addressing spectral bias in Physics-Informed Neural Networks (PINNs).
Kernelizing natural gradient, with applications to PINNs
Workshop talk, Infinite-dimensional Geometry: Theory and Applications, Erwin Schrödinger International Institute for Mathematics and Physics (ESI), Vienna, Austria
Talk given at the ESI workshop on Infinite-dimensional Geometry, presenting a kernel-based perspective on natural gradient methods with applications to Physics-Informed Neural Networks (PINNs).
Seminars
Kernelization of Natural Gradient Methods for Physics Informed Neural Networks (PINNs)
Seminar, Thalès CortAIx Lab Seminar, Saclay, France
Presentation of a kernelization framework for natural gradient methods applied to Physics-Informed Neural Networks (PINNs).
Kernelization of Natural Gradient Methods for Physics Informed Neural Networks (PINNs)
Seminar, MIA Paris-Saclay Seminar, AgroParisTech-INRAE, Paris-Saclay, France
Presentation of a kernelization framework for natural gradient methods applied to Physics-Informed Neural Networks (PINNs).
Kernelization of Natural Gradient Methods for Physics Informed Neural Networks (PINNs)
Seminar, Machine Learning and Signal Processing Seminar, ENS Lyon, Lyon, France
Presentation of a kernelization framework for natural gradient methods applied to Physics-Informed Neural Networks (PINNs).
Kernelization of Natural Gradient Methods for Physics Informed Neural Networks (PINNs) with connections to Galerkin methods
Seminar, MACARON Seminar, Université de Strasbourg, Strasbourg, France
Presentation of a kernelization framework for natural gradient methods applied to Physics-Informed Neural Networks (PINNs), highlighting connections to Galerkin methods.
Kernelization of Natural Gradient Methods for Physics Informed Neural Networks
Seminar, MILES Seminar, Université Paris-Dauphine, Paris, France
Presentation of a kernelization framework for natural gradient methods applied to Physics-Informed Neural Networks (PINNs).
Kernelization of Natural Gradient Methods for Physics Informed Neural Networks
Seminar, CRUNCH Seminar, Brown University, Providence, USA
Presentation of a kernelization framework for natural gradient methods applied to Physics-Informed Neural Networks (PINNs).
ANaGRAM: A Natural Gradient Relative to Adapted Model for efficient PINNs learning
Seminar, ACSIOM Seminar, Institut Montpelliérain Alexander Grothendieck, Montpellier, France
Presentation of ANaGRAM, a theoretically founded and algorithmically efficient natural gradient method for training Physics-Informed Neural Networks (PINNs). The talk covers the functional analysis perspective on PINNs training, the connection to the Neural Tangent Kernel, and an efficient SVD-based implementation that reduces the complexity of natural gradient descent significantly.
ANaGRAM: A Natural Gradient Relative to Adapted Model for efficient PINNs learning
Seminar, SCOOL Seminar, INRIA Lille – Nord Europe, Lille, France
Presentation of ANaGRAM, a theoretically founded and algorithmically efficient natural gradient method for training Physics-Informed Neural Networks (PINNs). The talk covers the functional analysis perspective on PINNs training, the connection to the Neural Tangent Kernel, and an efficient SVD-based implementation that reduces the complexity of natural gradient descent significantly.
ANaGRAM: A Natural Gradient Relative to Adapted Model for efficient PINNs learning
Seminar, TAU Seminar, LISN – Paris-Saclay University, CNRS, INRIA, Gif-sur-Yvette, France
Presentation of ANaGRAM, a theoretically founded and algorithmically efficient natural gradient method for training Physics-Informed Neural Networks (PINNs). The talk covers the functional analysis perspective on PINNs training, the connection to the Neural Tangent Kernel, and an efficient SVD-based implementation that reduces the complexity of natural gradient descent significantly.
ANaGRAM: A Natural Gradient Relative to Adapted Metrics for efficient PINNs learning
Seminar, Institute for Applied and Numerical Mathematics, Karlsruhe Institute of Technology (KIT), Karlsruhe, Germany
Presentation of ANaGRAM, a theoretically founded and algorithmically efficient natural gradient method for training Physics-Informed Neural Networks (PINNs). The talk covers the functional analysis perspective on PINNs training, the connection to the Neural Tangent Kernel, and an efficient SVD-based implementation that reduces the complexity of natural gradient descent significantly.
PhD Defense
Natural gradients and kernel methods for Physics Informed Neural Networks (PINNs)
PhD Defense, Université Paris-Saclay – LISN, Gif-sur-Yvette, France
PhD defense in Computer Science at Université Paris-Saclay, under the supervision of Cyril Furtlehner (TAU Team, INRIA Saclay – A&O–LISN–Paris-Saclay University–CNRS).