Research map / Probabilistic and scientific modeling / Molecules and proteins
Human Trajectory Prediction via Neural Social Physics
Jiangbei Yue, Dinesh Manocha, He Wang
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 collision-network and environment repulsion forces to goal force (ablation) | NSP-SFM · SDD | ADE: 6.57 -> 6.52 | Improved |
| add collision and environment forces (FDE) (ablation) | NSP-SFM · SDD | FDE: 10.68 -> 10.61 | Improved |
| add CVAE dynamics stochasticity (ablation) | NSP-SFM · SDD | ADE: 6.52 -> 1.78 | Improved |
| add CVAE dynamics stochasticity (ablation) | NSP-SFM · SDD | FDE: 10.61 -> 3.44 | Improved |
| add LSTM temporal modeling (full MLP) (ablation) | Goal/Collision networks · SDD | ADE: 6.61 -> 6.52 | Improved |
| add LSTM temporal modeling (two-layer MLP) (ablation) | Goal/Collision networks · SDD | ADE: 6.83 -> 6.66 | Improved |
| use full MLP instead of two-layer MLP (with LSTM) (ablation) | Goal/Collision networks · SDD | ADE: 6.66 -> 6.52 | Improved |
| learn tau and k_nj instead of hand-tuning (k_nj=25) (ablation) | NSP-SFM · SDD | ADE: 8.32 -> 6.53 | Improved |
| learn tau and k_nj instead of hand-tuning (k_nj=50) (ablation) | NSP-SFM · SDD | ADE: 7.54 -> 6.53 | Improved |
| learn tau and k_nj instead of hand-tuning (k_nj=25) (ablation) | NSP-SFM · SDD | FDE: 10.97 -> 10.61 | Improved |
| add remaining components to learned tau and k_nj (ablation) | NSP-SFM · SDD | ADE: 6.53 -> 6.52 | No clear effect |
| NSP-SFM instead of Y-Net | NSP-SFM · SDD (standard sampling) | ADE: 7.85 -> 6.52 | Improved |
Closest papers by meaning
- Social-STGCNN: A Social Spatio-Temporal Graph Convolutional Neural Network for Human Trajectory Prediction (CVPR 2020 · similarity 0.88)
- SocialCVAE: Predicting Pedestrian Trajectory via Interaction Conditioned Latents (AAAI 2024 · similarity 0.88)
- Social-Transmotion: Promptable Human Trajectory Prediction (ICLR 2024 · similarity 0.87)
- Stochastic Trajectory Prediction via Motion Indeterminacy Diffusion (CVPR 2022 · similarity 0.87)
- Recursive Social Behavior Graph for Trajectory Prediction (CVPR 2020 · similarity 0.87)
- Human Motion Prediction Under Unexpected Perturbation (CVPR 2024 · similarity 0.86)
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.
