Research map / Label-efficient and robust learning / Recognition and long-tail
FEND: A Future Enhanced Distribution-Aware Contrastive Learning Framework for Long-Tail Trajectory Prediction
Yuning Wang, Pu Zhang, Lei Bai, Jianru Xue
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 FEND to Traj++ EWTA | Traj++ EWTA · ETH-UCY Top 1% | minADE/minFDE: 0.92/2.33 -> 0.84/2.13 | Improved |
| Add FEND to Traj++ EWTA | Traj++ EWTA · ETH-UCY All | minADE/minFDE: 0.17/0.32 -> 0.17/0.32 | No clear effect |
| Add FEND without resampling | Traj++ EWTA · Nuscenes Top 1% | minADE/minFDE: 1.33/3.09 -> 1.21/2.50 | Improved |
| Add FEND to Y-Net | Y-Net · SDD Top 1% | minADE/minFDE: 65.82/134.01 -> 57.58/108.51 | Improved |
| Add FEND to Y-Net | Y-Net · SDD All | minADE/minFDE: 7.93/11.88 -> 7.87/11.68 | Improved |
| Add PCL only (ablation) | Traj++ EWTA · ETH-UCY Top 1% | minADE/minFDE: 0.98/2.54 -> 0.96/2.41 | Improved |
| Add future-enhanced clusters to PCL (ablation) | Traj++ EWTA · ETH-UCY Top 1% | minADE/minFDE: 0.96/2.41 -> 0.89/2.23 | Improved |
| Add hypernetwork to PCL (ablation) | Traj++ EWTA · ETH-UCY Top 1% | minADE/minFDE: 0.96/2.41 -> 0.90/2.28 | Improved |
| Set initial auxiliary loss weight a to 1 instead of 50 (ablation) | FEND · ETH-UCY Top 1% | minADE/minFDE: 0.84/2.13 -> 0.97/2.48 | Hurt |
| Set initial auxiliary loss weight a to 20 instead of 50 (ablation) | FEND · ETH-UCY Top 1% | minADE/minFDE: 0.84/2.13 -> 0.85/2.15 | Hurt |
| Set initial auxiliary loss weight a to 100 instead of 50 (ablation) | FEND · ETH-UCY Top 1% | minADE/minFDE: 0.84/2.13 -> 0.85/2.14 | Hurt |
| Disable the hard-sample threshold (ablation) | FEND · ETH-UCY Top 3% | minADE/minFDE: 0.61/1.46 -> 0.62/1.48 | Hurt |
Closest papers by meaning
- AMD: Adaptive Momentum and Decoupled Contrastive Learning Framework for Robust Long-Tail Trajectory Prediction (ICCV 2025 · similarity 0.92)
- LaKD: Length-agnostic Knowledge Distillation for Trajectory Prediction with Any Length Observations (NeurIPS 2024 · similarity 0.85)
- C2F-TP: A Coarse-to-Fine Denoising Framework for Uncertainty-Aware Trajectory Prediction (AAAI 2025 · similarity 0.85)
- STraj: Self-training for Bridging the Cross-Geography Gap in Trajectory Prediction (AAAI 2025 · similarity 0.84)
- Certified Human Trajectory Prediction (CVPR 2025 · similarity 0.83)
- TrajCLIP: Pedestrian trajectory prediction method using contrastive learning and idempotent networks (NeurIPS 2024 · similarity 0.83)
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.
