Research map / Multi-agent systems and games / Human-AI and goal-conditioned
Long-term Traffic Simulation with Interleaved Autoregressive Motion and Scenario Generation
Xiuyu Yang, Shuhan Tan, Philipp Krähenbühl
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 |
|---|---|---|---|
| use the control token design (Cont. token) (ablation) | Ours · WOMD long-term ablation | Composite: 0.6328 -> 0.6674 | Improved |
| use the position token design (Pos. token) (ablation) | Ours · WOMD long-term ablation | Composite: 0.6564 -> 0.6674 | Improved |
| use the heading token design (Head. token) (ablation) | Ours · WOMD long-term ablation | Composite: 0.6509 -> 0.6674 | Improved |
| use Pos. and Head. tokens together without Cont. token (ablation) | Ours · WOMD long-term ablation | Composite: 0.6297 -> 0.6674 | Improved |
| use Cont. token (ablation) | Ours · WOMD long-term ablation | Interactive: 0.7043 -> 0.7688 | Improved |
| use Head. token (ablation) | Ours · WOMD long-term ablation | Interactive: 0.7276 -> 0.7688 | Improved |
| use Pos. token (ablation) | Ours · WOMD long-term ablation | Kinematic: 0.5580 -> 0.5921 | Improved |
| add interleaved scene generation versus motion-only rollout (SMART) (ablation) | Ours · WOMD long-term motion-only | Composite: 0.7428 -> 0.7432 | No clear effect |
| interleaved scene generation vs CatK | Ours vs CatK · WOMD long-term | Placement-based overall: 0.4424 -> 0.4542 | Improved |
| interleaved scene generation vs CatK | Ours vs CatK · WOMD long-term | Map-based: 0.8186 -> 0.8087 | Hurt |
| interleaved scene generation vs CatK | Ours vs CatK · WOMD long-term | Mean ACE: 12.2 -> 8.1 | Improved |
| interleaved scene generation vs CatK | Ours vs CatK · WOMD long-term | ACE slope: 0.32 -> 0.15 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- SMART-7M — SMART: Scalable Multi-agent Real-time Motion Generation via Next-token Prediction (NeurIPS 2024) · better on 8 of 12 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
- Trajeglish: Traffic Modeling as Next-Token Prediction (ICLR 2024 · similarity 0.85)
- SceneDiffuser++: City-Scale Traffic Simulation via a Generative World Model (CVPR 2025 · similarity 0.85)
- Overcoming Challenges of Long-Horizon Prediction in Driving World Models (NeurIPS 2025 · similarity 0.85)
- Is Ego Status All You Need for Open-Loop End-to-End Autonomous Driving? (CVPR 2024 · similarity 0.85)
- BehaviorGPT: Smart Agent Simulation for Autonomous Driving with Next-Patch Prediction (NeurIPS 2024 · similarity 0.85)
- Causal Composition Diffusion Model for Closed-loop Traffic Generation (CVPR 2025 · 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.
