Lionel Mathelin

Centre National de la Recherche Scientifique

Papers

Total Citations

H-Index

About

Lionel Mathelin is a leading figure in data-driven modeling and machine learning for dynamical systems, with a particular focus on curriculum learning and reduced-order modeling. His most-cited work, "Curriculum learning for data-driven modeling of dynamical systems" (2023, 13 citations), introduces a novel training strategy that progressively exposes models to increasingly complex data, significantly improving their generalization and accuracy in predicting system behavior. This contribution bridges the gap between machine learning and physics-based simulation, enabling more efficient and robust modeling of complex, time-dependent phenomena. Mathelin's research has profound implications for fields such as fluid dynamics, climate science, and engineering design, where accurate and computationally tractable models are essential. His work is recognized for its innovative approach to training neural networks, making them more adept at capturing the underlying dynamics of physical systems. With a growing citation impact, Mathelin continues to shape the intersection of artificial intelligence and scientific computing, offering tools that empower researchers to tackle high-dimensional, nonlinear problems with unprecedented efficiency.