Research map / Time series / Causal and generative time series
Flow-Based Spatio-Temporal Structured Prediction of Motion Dynamics
Mohsen Zand, Ali Etemad, Michael Greenspan
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 masked-convolution spatio-temporal conditioning (ablation) | MotionFlow · CMU Mocap basketball | MPJPE at 1000 ms (mm): 104.33 -> 84.27 | Improved |
| add masked-convolution spatio-temporal conditioning (ablation) | MotionFlow · CMU Mocap walking | MPJPE at 1000 ms (mm): 113.69 -> 34.11 | Improved |
| add dynamic latent space (temporal conditional prior) (ablation) | MotionFlow · CMU Mocap basketball | MPJPE at 1000 ms (mm): 110.17 -> 88.37 | Improved |
| add dynamic latent space (temporal conditional prior) (ablation) | MotionFlow · CMU Mocap walking | MPJPE at 1000 ms (mm): 121.58 -> 72.69 | Improved |
| add residual network on latent variables (ablation) | MotionFlow · CMU Mocap basketball | MPJPE at 1000 ms (mm): 88.37 -> 84.27 | Improved |
| add residual network on latent variables (ablation) | MotionFlow · CMU Mocap walking | MPJPE at 1000 ms (mm): 72.69 -> 34.11 | Improved |
| use random layer initialization instead of Gaussian/zero init (ablation) | MotionFlow · CMU Mocap | final MPJPE | No clear effect |
| MotionFlow vs NRI on 5-particle simulation | MotionFlow · Simulated trajectories (5 particles) | MSE at step 25: 2.05e-2 -> 1.28e-3 | Improved |
| MotionFlow vs DNRI on 3-particle simulation | MotionFlow · Simulated trajectories (3 particles) | MSE at step 25: 5.36e-4 -> 4.32e-4 | Improved |
| MotionFlow vs IMMA on NBA | MotionFlow · NBA trajectory forecasting | FDE (m): 1.438 -> 1.509 | Hurt |
| MotionFlow vs IMMA on NBA | MotionFlow · NBA trajectory forecasting | ADE (m): 0.769 -> 0.763 | Improved |
| MotionFlow vs ALEN on binary segmentation | MotionFlow · Weizmann Horses segmentation (32x32) | Mean IoU (%): 85.4 -> 89.2 | Improved |
Closest papers by meaning
- VideoFlow: A Conditional Flow-Based Model for Stochastic Video Generation (ICLR 2020 · similarity 0.88)
- Flow Matching with Gaussian Process Priors for Probabilistic Time Series Forecasting (ICLR 2025 · similarity 0.87)
- Adaptive Quantization in Generative Flow Networks for Probabilistic Sequential Prediction (NeurIPS 2025 · similarity 0.87)
- Normalizing Trajectory Models (NeurIPS 2026 · similarity 0.86)
- Stream-level Flow Matching with Gaussian Processes (ICML 2025 · similarity 0.85)
- Normalizing Flows are Capable Generative Models (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.
