Research map / Probabilistic and scientific modeling / Dynamical systems
Forecasting Continuous Non-Conservative Dynamical Systems in SO(3)
Lennart Bastian, Mohammad Rashed, Nassir Navab, Tolga Birdal
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 |
|---|---|---|---|
| use SO(3) Savitzky-Golay control path instead of Hermite spline for the neural CDE (ablation) | SO3-nCDE · Freely rotating, t=1.2s | rotational geodesic error (deg): 3.62 -> 1.91 | Improved |
| add second-order control signal to SG-nCDE (ablation) | SG-nCDE · Freely rotating, t=1.2s | rotational geodesic error (deg): 1.91 -> 1.31 | Improved |
| add learnable weight matrix W without second-order control (ablation) | SG-nCDE · Freely rotating, t=1.2s | rotational geodesic error (deg): 1.91 -> 1.98 | No clear effect |
| add learnable weight matrix W to second-order SG-nCDE (ablation) | SG-nCDE · Freely rotating, t=1.2s | rotational geodesic error (deg): 1.31 -> 1.28 | Improved |
| use full SG-nCDE instead of SO3-nCDE baseline runtime (ablation) | SG-nCDE · Runtime | runtime (ms): 74.6 -> 76.8 | No clear effect |
| use Savitzky-Golay control path to reduce function evaluations | SG-nCDE · training NFE | number of function evaluations | Improved |
| use SG-nCDE instead of SO3-GRU baseline | SG-nCDE · Oxford Motion Dataset | rotation prediction error (deg): 3.04 -> 2.32 (translational) | Improved |
| increase GRU latent dimension (ablation) | SO3-GRU · latent dimension ablation | extrapolation error | Hurt |
| increase LEAP latent dimension (ablation) | LEAP · latent dimension ablation | extrapolation error | No clear effect |
| use SG-nCDE vs energy-conserving method with ground-truth momentum, freely rotating at long horizon | Conservational · Freely rotating, t=1.2s | rotational geodesic error (deg): 1.15 vs 1.28 | Mixed |
| increase input noise level (ablation) | all methods · velocity damping | rotational geodesic error | Hurt |
Closest papers by meaning
- Markovian Dynamics Enforcer: Feasibility Preserving Correction on Learned Dyanmics Manifolds (NeurIPS 2026 · similarity 0.85)
- ODEWorld: A Continuous Predictive Architecture via Physical-Time Flow (NeurIPS 2026 · similarity 0.85)
- Learning Physics From Video: Unsupervised Physical Parameter Estimation for Continuous Dynamical Systems (CVPR 2025 · similarity 0.84)
- Optimal-state Dynamics Estimation for Physics-based Human Motion Capture from Videos (NeurIPS 2024 · similarity 0.84)
- The Dynamic-Probabilistic Consistency Gap in Chaotic Surrogate Modeling (NeurIPS 2026 · similarity 0.84)
- Degrees of Freedom Matter: Inferring Dynamics from Point Trajectories (CVPR 2024 · similarity 0.84)
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.
