Research Seminar on AI: Recognition models to learn dynamics from partial observations with neural ODEs
- + 33 6 43 50 48 75
- Send Email
Tuesday, October 04, 2022, 4:00pm
Identifying dynamical systems from experimental data is a notably difficult task. Prior knowledge generally helps, but the extent of this knowledge varies with the application, and customized models are often needed. Neural ODEs can be written as a flexible framework for system identification and can incorporate a broad spectrum of physical insight, giving physical interpretability to the resulting latent space. In the case of partial observations however, the data points cannot directly be mapped to the latent state of the ODE. Hence, Mona Buisson-Fenet and her team propose to design recognition models, in particular inspired by nonlinear observer theory, to link the partial observations to the latent state. Mona Buisson-Fenet and her team demonstrate the performance of the proposed approach on numerical simulations and on an experimental dataset from a robotic exoskeleton.
Mona is a PhD student in the Control and Systems Center at Mines Paris supervised by Florent Di Meglio, and co-supervised by Pr. Trimpe in the Institute for Data Science in Mechanical Engineering at RWTH Aachen. Before her PhD, she received a diploma in engineering from Mines Paris with a focus on applied mathematics and robotics. She works on combining state estimation and dynamics learning, aiming to learn reduced-order models of physical systems from experimental data, for example for industrial applications on the topic of digital twins.