Research map / Image generation / Autoregressive and latent generation
Generating Useful Accident-Prone Driving Scenarios via a Learned Traffic Prior
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 a learned traffic prior to jointly optimize scene-agent trajectories instead of optimizing one bicycle-model adversary | Replay planner scenario generation · nuScenes colliding scenarios | adversary acceleration: 2.00 -> 0.98 | Helped |
| Use a learned traffic prior to jointly optimize scene-agent trajectories instead of optimizing one bicycle-model adversary | Replay planner scenario generation · nuScenes colliding scenarios | environment collisions: 16.5 -> 10.8 | Helped |
| Use a learned traffic prior to jointly optimize scene-agent trajectories instead of optimizing one bicycle-model adversary | Replay planner scenario generation · nuScenes colliding scenarios | solution rate: 73.4 -> 83.5 | Helped |
| Generate challenging scenarios rather than replay regular scenes | Rule-based planner · nuScenes scenarios | collision rate: 1.2 -> 27.4 | Mixed |
| Generate challenging scenarios rather than replay regular scenes | Rule-based planner · nuScenes scenarios | planner acceleration: 1.63 -> 1.91 | Mixed |
| Tune the planner on regular and challenging scenarios together | Rule-based planner · Regular scenarios | collision rate: 4.6 -> 6.0 | Hurt |
| Tune the planner on regular and challenging scenarios together | Rule-based planner · Challenging scenarios | collision rate: 68.6 -> 51.4 | Helped |
| Add a learned accident mode with a second parameter set | Rule-based planner · Challenging scenarios | collision rate: 68.6 -> 54.3 | Helped |
| Add a learned accident mode with a second parameter set | Rule-based planner · Regular scenarios | collision rate: 4.6 -> 4.6 | No clear effect |
| Add an oracle accident mode with a second parameter set | Rule-based planner · Challenging scenarios | collision rate: 68.6 -> 54.3 | Helped |
Closest papers by meaning
- Generating Traffic Scenarios via In-Context Learning to Learn Better Motion Planner (AAAI 2025 · similarity 0.91)
- Scenario Dreamer: Vectorized Latent Diffusion for Generating Driving Simulation Environments (CVPR 2025 · similarity 0.88)
- ChatScene: Knowledge-Enabled Safety-Critical Scenario Generation for Autonomous Vehicles (CVPR 2024 · similarity 0.88)
- DiffScene: Diffusion-Based Safety-Critical Scenario Generation for Autonomous Vehicles (AAAI 2025 · similarity 0.87)
- Hybrid-Driving: An Autonomous Driving Decision Framework Integrating Large Language Models, Knowledge Graphs and Driving Rules (AAAI 2025 · similarity 0.87)
- AlignDrive: Aligned Lateral-Longitudinal Planning for End-to-End Autonomous Driving (NeurIPS 2026 · 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.