Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Variational Hierarchical Mixtures for Probabilistic Learning of Inverse Dynamics
Hany Abdulsamad, Peter Nickl, Pascal Klink, Jan Peters
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 ILR instead of LWPR | LWPR · SARCOS | MSE: 2.600 -> 0.480 | Improved |
| use ILR instead of LGR | LGR · SARCOS | MSE: 8.600 -> 0.480 | Improved |
| use ILR instead of LWPR | LWPR · Barrett-WAM | MSE: 1.000 -> 0.290 | Improved |
| use hierarchical HILR instead of ILR (number of experts) | ILR · SARCOS | number of experts: 1700 -> 1450 | Improved |
| use hierarchical HILR instead of ILR (accuracy) | ILR · SARCOS | MSE: 0.480 -> 0.530 | Hurt |
| use hierarchical HILR instead of ILR (number of experts) | ILR · Barrett-WAM | number of experts: 1350 -> 1110 | Improved |
| use hierarchical HILR instead of ILR (accuracy) | ILR · Barrett-WAM | MSE: 0.290 -> 0.310 | Hurt |
| compare ILR with rBCM product of experts | ILR · SARCOS | MSE: 0.480 vs 0.452 | Mixed |
| compare ILR with full GPR | ILR · Barrett-WAM | MSE: 0.100 vs 0.290 | Hurt |
| add ILR feed-forward torque to PD controller (T1) | PD controller · Barrett-WAM real robot | tracking MSE: 2.33e-2 -> 1.03e-3 | Improved |
| add ILR feed-forward torque instead of analytic model (T1) | Analytic+PD · Barrett-WAM real robot | tracking MSE: 2.16e-2 -> 1.03e-3 | Improved |
| use ILR+PD instead of PD, PD torque contribution (T1) | PD controller · Barrett-WAM real robot | PD torque: 8.25 -> 1.40 | Improved |
Closest papers by meaning
- Invariance-based Learning of Latent Dynamics (ICLR 2024 · similarity 0.87)
- The Dynamic-Probabilistic Consistency Gap in Chaotic Surrogate Modeling (NeurIPS 2026 · similarity 0.86)
- Online Variational Sequential Monte Carlo (ICML 2024 · similarity 0.86)
- Predictive Inverse Dynamics Models are Scalable Learners for Robotic Manipulation (ICLR 2025 · similarity 0.86)
- Support-Safe Variational Hybrid Filtering for Contact-Mode and Sparse-Law Recovery (NeurIPS 2026 · similarity 0.86)
- Safely Learning Controlled Stochastic Dynamics (NeurIPS 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.
