Skip to Main content Skip to Navigation
Preprints, Working Papers, ...

LEADS: Learning Dynamical Systems that Generalize Across Environments

Abstract : When modeling dynamical systems from real-world data samples, the distribution of data often changes according to the environment in which they are captured, and the dynamics of the system itself vary from one environment to another. Generalizing across environments thus challenges the conventional frameworks. The classical settings suggest either considering data as i.i.d. and learning a single model to cover all situations or learning environment-specific models. Both are sub-optimal: the former disregards the discrepancies between environments leading to biased solutions, while the latter does not exploit their potential commonalities and is prone to scarcity problems. We propose LEADS, a novel framework that leverages the commonalities and discrepancies among known environments to improve model generalization. This is achieved with a tailored training formulation aiming at capturing common dynamics within a shared model while additional terms capture environment-specific dynamics. We ground our approach in theory, exhibiting a decrease in sample complexity with our approach and corroborate these results empirically, instantiating it for linear dynamics. Moreover, we concretize this framework for neural networks and evaluate it experimentally on representative families of nonlinear dynamics. We show that this new setting can exploit knowledge extracted from environment-dependent data and improves generalization for both known and novel environments.
Complete list of metadata

https://hal.archives-ouvertes.fr/hal-03261055
Contributor : Yuan Yin Connect in order to contact the contributor
Submitted on : Tuesday, June 15, 2021 - 1:27:13 PM
Last modification on : Tuesday, July 13, 2021 - 3:27:39 AM
Long-term archiving on: : Thursday, September 16, 2021 - 6:41:33 PM

Identifiers

  • HAL Id : hal-03261055, version 1
  • ARXIV : 2106.04546

Citation

Yuan Yin, Ibrahim Ayed, Emmanuel de Bézenac, Nicolas Baskiotis, Patrick Gallinari. LEADS: Learning Dynamical Systems that Generalize Across Environments. 2021. ⟨hal-03261055⟩

Share

Metrics

Record views

37

Files downloads

28