Research map / Graph learning / Knowledge and temporal graphs
Recursive Social Behavior Graph for Trajectory Prediction
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 group-supervised RSBG with context features rather than STGAT (1V-1) | Trajectory prediction · ETH and UCY average | ADE: 11.1% | Helped |
| use group-supervised RSBG with context features rather than STGAT (1V-1) | Trajectory prediction · ETH and UCY average | FDE: 10.8% | Helped |
| remove human context features (ablation) | RSBG · ETH and UCY subsets | ADE: ETH: 0.79 -> 0.80; HOTEL: 0.35 -> 0.33; UNIV: 0.68 -> 0.59; ZARA1: 0.42 -> 0.40; ZARA2: 0.35 -> 0.30 | Mixed |
| replace the LSTM historical trajectory encoder with a BiLSTM (ablation) | RSBG · ETH & UCY average | ADE: 0.51 -> 0.48; 5.9% | Helped |
| replace the LSTM historical trajectory encoder with a BiLSTM (ablation) | RSBG · ETH & UCY average | FDE: 1.04 -> 0.99; 4.8% | Helped |
| use exponential L2 loss with γ=20 instead of standard L2 loss (ablation) | RSBG · ETH & UCY average | ADE: 0.50 -> 0.48; 4.0% | Helped |
| use exponential L2 loss with γ=20 instead of standard L2 loss (ablation) | RSBG · ETH & UCY average | FDE: 1.04 -> 0.99; 4.8% | Helped |
| set exponential loss γ=50 instead of standard L2 loss (ablation) | RSBG · ETH & UCY average | ADE: 0.50 -> 0.49 | Helped |
| set exponential loss γ=50 instead of standard L2 loss (ablation) | RSBG · ETH & UCY average | FDE: 1.04 -> 1.01 | Helped |
| overemphasize FDE by setting exponential loss γ=5 (ablation) | RSBG · ETH & UCY average | ADE: 0.50 -> 0.52 | Hurt |
| overemphasize FDE by setting exponential loss γ=5 (ablation) | RSBG · ETH & UCY average | FDE: 1.04 -> 1.06 | Hurt |
| add human context features (ablation) | RSBG · ETH | ADE: 0.80 -> 0.79 | Helped |
Closest papers by meaning
- Higher-order Relational Reasoning for Pedestrian Trajectory Prediction (CVPR 2024 · similarity 0.89)
- SocialCircle: Learning the Angle-based Social Interaction Representation for Pedestrian Trajectory Prediction (CVPR 2024 · similarity 0.87)
- SocialCVAE: Predicting Pedestrian Trajectory via Interaction Conditioned Latents (AAAI 2024 · similarity 0.87)
- Social-Transmotion: Promptable Human Trajectory Prediction (ICLR 2024 · similarity 0.86)
- SocialMOIF: Multi-Order Intention Fusion for Pedestrian Trajectory Prediction (CVPR 2025 · similarity 0.86)
- A Unified Environmental Network for Pedestrian Trajectory Prediction (AAAI 2024 · 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.