Research map / Multi-agent systems and games / Human-AI and goal-conditioned
Unlimited Neighborhood Interaction for Heterogeneous Trajectory Prediction
Fang Zheng, Le Wang, Sanping Zhou, Wei Tang, Zhenxing Niu, Nanning Zheng, Gang Hua
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 hierarchical graph attention (category-to-category) instead of MLP (ablation) | UNIN · nuScenes | ADE/FDE: 1.613/1.969 -> 1.049/1.521 | Improved |
| add hierarchical graph attention instead of CNN (ablation) | UNIN · nuScenes | ADE/FDE: 1.763/1.982 -> 1.049/1.521 | Improved |
| add hierarchical graph attention instead of MLP (ablation) | UNIN · Apolloscape | WADE/WFDE: 1.460/1.794 -> 1.094/1.545 | Improved |
| add hierarchical graph attention instead of CNN (ablation) | UNIN · Apolloscape | WADE/WFDE: 1.837/2.014 -> 1.094/1.545 | Improved |
| add unlimited neighborhood interaction instead of MLP (ablation) | UNIN · nuScenes | ADE/FDE: 1.547/1.728 -> 1.049/1.521 | Improved |
| add unlimited neighborhood interaction instead of CNN (ablation) | UNIN · nuScenes | ADE/FDE: 1.701/1.934 -> 1.049/1.521 | Improved |
| add unlimited neighborhood interaction instead of MLP (ablation) | UNIN · Apolloscape | WADE/WFDE: 1.576/1.843 -> 1.094/1.545 | Improved |
| add unlimited neighborhood interaction instead of CNN (ablation) | UNIN · Apolloscape | WADE/WFDE: 1.792/1.955 -> 1.094/1.545 | Improved |
| use UNIConv kernel size 2 instead of 1 (ablation) | UNIN · Argoverse+nuScenes avg | ADE: 1.179 -> 0.921 | Improved |
| use UNIConv kernel size 2 instead of 5 (ablation) | UNIN · Argoverse+nuScenes avg | ADE: 1.247 -> 0.921 | Improved |
| use UNIConv kernel size 2 instead of 10 (ablation) | UNIN · Argoverse+nuScenes avg | ADE: 2.691 -> 0.921 | Improved |
| use UNIConv kernel size 3 instead of 2 (ablation) | UNIN · Argoverse+nuScenes avg | FDE: 1.388 -> 1.323 (FDE); 0.921 -> 0.998 (ADE) | Mixed |
Closest papers by meaning
- Density-Adaptive Model Based on Motif Matrix for Multi-Agent Trajectory Prediction (CVPR 2024 · similarity 0.89)
- TraPHic: Trajectory Prediction in Dense and Heterogeneous Traffic Using Weighted Interactions (CVPR 2019 · similarity 0.89)
- M2I: From Factored Marginal Trajectory Prediction to Interactive Prediction (CVPR 2022 · similarity 0.89)
- SWIFT: A Small-World Interaction Framework for Flow-Aware Trajectory Prediction in Autonomous Driving (TPAMI 2026 · similarity 0.89)
- Leveraging Future Relationship Reasoning for Vehicle Trajectory Prediction (ICLR 2023 · similarity 0.88)
- Higher-order Relational Reasoning for Pedestrian Trajectory Prediction (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.
