"AI and Materials" workshop from M4S: feedback on Alexandre Daby-Seerasam's presentation from Laboratoire de Mécanique et de ses Interfaces (20.01.2026)
On January 20,2026, Alexandre Daby-Seesaram, lecturer at Laboratoire de Mécanique et de ses Interfaces (ENSTA), led the second edition of M4S's "AI and Materials" workshop. His presentation was entitled "Finite element neural network interpolation: Hybridization with the proper generalized decomposition for non-linear surrogate modeling" and focused on ongoing research at the laboratory.
The mechanical stresses exerted within the lungs are known to be one of the determining factors in the development of fibrosis. Building numerical models to simulate the mechanical behavior of the lungs would therefore help doctors diagnose this disease. However, the use of these models only makes sense if they can be customized for each patient. It is therefore necessary to develop calculation methodologies that allow rapid adaptation to changes in the various descriptive parameters of the pulmonary alveoli, which is complex with conventional methodologies, such as finite element calculations, which are often prohibitively expensive.
Combining traditional order reduction methods with tools commonly used in machine learning algorithms (neural networks, automatic differentiation, stochastic optimization, etc.) makes it possible to overcome the problems of high computational costs as the dimensionality of the problem increases, while maintaining explainability (often lacking when artificial intelligence methods are used). Similar methodologies are also being developed to take into account variations in the overall geometry of the lung.
Following the presentation, numerous discussions allowed those present to share their expertise in materials and AI.
About the speaker
Alexandre Daby-Seesaram holds a PhD in mechanics from Paris-Saclay University. After completing his thesis at the Laboratoire de Mécanique Paris-Saclay (CNRS, ENS Paris-Saclay), he spent two years as a postdoctoral researcher at Laboratoire de Mécanique du Solide (École polytechnique, CNRS), where he began his current work on biomechanics and digital organ twins. He joined the Laboratoire de Mécaniques et de ses Interfaces (ENSTA) in 2025 as part of a Tenure Track position for the center Engineering for Health (E4H) at the Institut Polytechnique de Paris.
About "AI and Materials" workshops
To learn more about the “AI and Materials” seminars and to register for the next meeting: