Research map / Probabilistic and scientific modeling / Dynamical systems
Variational Dependent Multi-output Gaussian Process Dynamical Systems
Jing Zhao, Shiliang Sun
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 |
|---|---|---|---|
| add explicit output dependency (multi-output GP) to the GP dynamical system (VDM-GPDS vs VGPDS) | VGPDS · Synthetic output-dependent, prediction y1-y4 | RMSE (%): 1.51 -> 1.43 | Improved |
| add explicit output dependency to VGPDS when outputs are independent | VGPDS · Synthetic output-independent, prediction | RMSE (%): 2.18 -> 2.21 | No clear effect |
| use VDM-GPDS instead of VGPDS for reconstructing missing y4 | VGPDS · Synthetic, reconstruction of missing y4 | RMSE (%): 6.79 -> 5.56 | Improved |
| use VDM-GPDS instead of VGPDS for reconstructing missing y1 | VGPDS · Synthetic, reconstruction of missing y1 | RMSE (%): 1.49 -> 0.98 | Improved |
| use VDM-GPDS instead of VGPDS on mocap leg reconstruction (subject 35) | VGPDS · CMU mocap subject 35, LA | RMSE: 5.54 -> 5.30 | Improved |
| use VDM-GPDS instead of VGPDS on mocap upper body reconstruction (subject 35) | VGPDS · CMU mocap subject 35, BA | RMSE: 2.81 -> 2.60 | Improved |
| use VDM-GPDS instead of VGPDS on unseen subject 16 upper body angle (BA) | VGPDS · CMU mocap subject 16, BA | RMSE: 10.64 -> 6.69 | Improved |
| use VDM-GPDS instead of VGPDS for traffic forecasting at link Gd | VGPDS · Traffic flow forecasting, Gd | RMSE: 162.81 -> 147.14 | Improved |
| use VDM-GPDS instead of VGPDS for traffic forecasting at link Ka | VGPDS · Traffic flow forecasting, Ka | RMSE: 88.23 -> 75.23 | Improved |
| use VDM-GPDS instead of VGPDS on SARCOS 2nd joint, 100 training points | VGPDS · SARCOS 2nd joint (N=100) | RMSE: 4.89 -> 4.55 | Improved |
| use VDM-GPDS instead of VGPDS on SARCOS 7th joint, 100 training points | VGPDS · SARCOS 7th joint (N=100) | RMSE: 0.65 -> 0.66 | No clear effect |
| use VDM-GPDS instead of GPDM for synthetic prediction MSLL | GPDM · Synthetic output-dependent, prediction y1 | MSLL: 2.21×104 -> -2.79 | Improved |
Closest papers by meaning
- Free-Form Variational Inference for Gaussian Process State-Space Models (ICML 2023 · similarity 0.88)
- Heterogeneous Multi-output Gaussian Process Prediction (NeurIPS 2018 · similarity 0.87)
- Physics-Informed Variational State-Space Gaussian Processes (NeurIPS 2024 · similarity 0.86)
- Exact Inference for Continuous-Time Gaussian Process Dynamics (AAAI 2024 · similarity 0.85)
- Sparse Gaussian Processes: Structured Approximations and Power-EP Revisited (NeurIPS 2025 · similarity 0.85)
- Bayesian Optimization via Continual Variational Last Layer Training (ICLR 2025 · 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.
