Emmanuel Menier
Deep learning researcher focusing on reduced-order modeling
Summary
Work
Education
Projects
Writing
Interpretable learning of effective dynamics for multiscale systems
January 1, 2025Framework (iLED) motivated by Mori–Zwanzig and Koopman operator theory to learn interpretable effective dynamics for multiscale systems; demonstrated on benchmark systems.
Deep Learning for Reduced Order Modeling
January 1, 2024PhD thesis presenting methods to exploit dynamical data for reduced-order models, introducing CD-ROM and iLED approaches to hybridize physical and data-driven modeling for interpretable, efficient dynamical models.
CD-ROM (Journal article)
January 1, 2023Journal article 'CD-ROM: Complemented Deep - Reduced Order Model' published in Computer Methods in Applied Mechanics and Engineering (CMAME).
CD-ROM: Complemented Deep-Reduced Order Model
January 1, 2022Paper proposing a deep-learning-based closure modeling approach for POD-Galerkin reduced order models (CD-ROM), with a continuous memory formulation to improve stability and interpretability; demonstrated on CFD examples including Kuramoto-Sivashinsky.
Continuous Methods : Hamiltonian Domain Translation
January 1, 2022Work proposing a Hamiltonian-structured reformulation of Cycle-GAN-style domain translation models, connecting generative models and dynamical systems.