LEADS: Learning Dynamical Systems That Generalize Across Environments
Dec 6, 2021
Speakers
Ibrahim Ayed
Speaker · 0 followers
Emmanuel de Bézenac
Speaker · 0 followers
Nicolas Baskiotis
Speaker · 0 followers
Patrick Gallinari
Speaker · 0 followers
About
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.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 forme…
Organizer
NeurIPS 2021
Account · 1.9k followers
About NeurIPS 2021
Neural Information Processing Systems (NeurIPS) is a multi-track machine learning and computational neuroscience conference that includes invited talks, demonstrations, symposia and oral and poster presentations of refereed papers. Following the conference, there are workshops which provide a less formal setting.
Like the format? Trust SlidesLive to capture your next event!
Professional recording and live streaming, delivered globally.
Sharing
Recommended Videos
Presentations on similar topic, category or speaker
Linear Convergence of Gradient Methods for Estimating Structured Transition Matrices in High-dimensional Vector Autoregressive Models
Watch later
Xiao Lv, …
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
Best Papers and Closing Remarks
Watch later
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
Variational Automatic Curriculum Learning for Sparse-Reward Cooperative Multi-Agent Problems
Watch later
Jiayu Chen, …
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
Understanding Post-hoc Adaptation for Improving Subgroup Robustness
Watch later
David Madras, …
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
Human-Adversarial Visual Question Answering
Watch later
Sasha Sheng, …
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
Datasets for Online Controlled Experiments
Watch later
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%