# LEADS: Learning Dynamical Systems that Generalize Across Environments Source: [LEADS: Learning Dynamical Systems that Generalize Across Environments]() Collection index: [llms.txt]() Data updated: 2026-01-04T05:07:39.622Z Authors: Yuan Yin, Ibrahim Ayed, Emmanuel de B\\'ezenac, Nicolas Baskiotis, Patrick Gallinari Published: 2021-06-08T17:28:19.000Z ## 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. Code is available at https://github.com/yuan-yin/LEADS. ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [Learning in novel environments on LV novel environments n=1K (test)](): Learning in novel environments; LV novel environments n=1K (test); Test MSE: 0.00111. - [Learning in novel environments on GS novel environments n=1K (test)](): Learning in novel environments; GS novel environments n=1K (test); Test MSE: 0.00127. - [Learning in novel environments on NS novel environments n=8K (test)](): Learning in novel environments; NS novel environments n=8K (test); Test MSE: 0.00714. - [Dynamical system modeling on Lotka-Volterra (LV) m=10, n=1 (train)](): Dynamical system modeling; Lotka-Volterra (LV) m=10, n=1 (train); MSE: 0.00000328. - [Dynamical system modeling on Navier-Stokes (NS) m=4, n=8 (train)](): Dynamical system modeling; Navier-Stokes (NS) m=4, n=8 (train); MSE: 0.000103. - [Dynamical system modeling on Gray-Scott (GS) m=3, n=1 (train)](): Dynamical system modeling; Gray-Scott (GS) m=3, n=1 (train); MSE: 0.0000575. ## Original sources and implementation [Original paper]() [PDF]() [Implementation]()