Research map / Probabilistic and scientific modeling / Dynamical systems
Online Learning and Control of Complex Dynamical Systems from Sensory Input
Oumayma Bounou, Jean Ponce, Justin Carpentier
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Never update the latent dynamics matrix A online (fixed after initial estimate) (ablation) | latent linear dynamics model · unforced pendulum and cartpole, 15 s horizon | per-pixel RMSE | Hurt |
| Update A at fixed intervals (750 ms pendulum, 50 ms cartpole) instead of never (ablation) | latent linear dynamics model · unforced pendulum and cartpole | per-pixel RMSE | Improved |
| Compute A by least squares per trajectory instead of learning a fixed A as a parameter (ablation) | latent linear dynamics model · pendulums of fixed length 0.6 m and varying length 0.3-0.8 m | per-pixel RMSE | Improved |
| Compute control matrix B by least squares per trajectory instead of a learned static B (ablation) | actuated pendulum and cartpole, 20 s horizon · actuated systems | per-pixel RMSE | Improved |
| Use partial online updates for forced dynamics (ablation) | latent linear dynamics model · actuated pendulum and cartpole | per-pixel RMSE | Improved |
| Use a history of one image instead of two (time-delay embedding) (ablation) | pendulum control from video · LQR control from images | feasibility of control | Hurt |
Closest papers by meaning
- Identifying latent state transitions in non-linear dynamical systems (ICLR 2025 · similarity 0.87)
- Deep-Koopman-KANDy: Dictionary Discovery for Deep-Koopman Operators with Kolmogorov-Arnold Networks for Dynamics (NeurIPS 2026 · similarity 0.86)
- ResKoopNet: Learning Koopman Representations for Complex Dynamics with Spectral Residuals (ICML 2025 · similarity 0.85)
- Efficient Dynamics Modeling in Interactive Environments with Koopman Theory (ICLR 2024 · similarity 0.85)
- Course Correcting Koopman Representations (ICLR 2024 · similarity 0.85)
- Learning Dynamics under Environmental Constraints via Measurement-Induced Bundle Structures (ICML 2025 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
