Research map / Probabilistic and scientific modeling / Molecules and proteins
W2W: Language-Model-Based Trajectory Prediction with Reinforcement Learning
Zirui Xu, Biao Yang, Rongrong Ni, Zhongkai Zhou, Shaobo Shen
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 interaction cues (companion/following/obstacle) to the prompt (ablation) | W2W-SFT · ETH/UCY avg | ADE/FDE: 0.22/0.32 -> 0.21/0.30 | Improved |
| add RL (PPO+LoRA) after SFT (ablation) | W2W-SFT · ETH/UCY avg | ADE/FDE: 0.21/0.30 -> 0.21/0.29 | Improved |
| add RL with off-road penalty (ablation) | W2W-SFT · ETH/UCY | off-road rate: 11.30% -> 8.85% | Improved |
| SFT of pretrained T5-Small (ablation) | T5-Small · trajectory text generation | format execution rate: FER ~0 -> ~100% | Improved |
| remove lengthy scene-description text from the prompt (ablation) | W2W-SFT+ · ETH/UCY avg | ADE/FDE: 0.22/0.32 -> 0.22/0.31 | Improved |
| replace scene text with concise interaction semantics (ablation) | W2W-SFT* · ETH/UCY avg | ADE/FDE: 0.22/0.31 -> 0.21/0.30 | Improved |
| weight accuracy reward above off-road reward (RW-C vs RW-A) (ablation) | W2W RL · ETH/UCY avg | ADE/FDE | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- MoFlow — MoFlow: One-Step Flow Matching for Human Trajectory Forecasting via Implicit Maximum Likelihood Estimation based Distillation (CVPR 2025) · better on 1 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Social Reasoning-Aware Trajectory Prediction via Multimodal Language Model (TPAMI 2025 · similarity 0.90)
- AutoTraces: Autoregressive Trajectory Forecasting via Multimodal Large Language Models (CVPR 2026 · similarity 0.89)
- Social-Transmotion: Promptable Human Trajectory Prediction (ICLR 2024 · similarity 0.89)
- Can Language Beat Numerical Regression? Language-Based Multimodal Trajectory Prediction (CVPR 2024 · similarity 0.89)
- Multi-modal Knowledge Distillation-based Human Trajectory Forecasting (CVPR 2025 · similarity 0.87)
- Trajeglish: Traffic Modeling as Next-Token Prediction (ICLR 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.
