Research map / Image generation / Autoregressive and latent generation
Scenario Diffusion: Controllable Driving Scenario Generation With Diffusion
Read the paperCheck 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 |
|---|---|---|---|
| Add map conditioning (ablation) | Scenario Diffusion · Internal Dataset | MMD² Positions: 0.134 (0.006) -> 0.068 (0.005) | Helped |
| Add map conditioning (ablation) | Scenario Diffusion · Internal Dataset | MMD² Headings: 0.379 (0.012) -> 0.081 (0.004) | Helped |
| Add map conditioning (ablation) | Scenario Diffusion · Internal Dataset | Traj Off Drivable: 0.549 (0.004) -> 0.023 (0.003) | Helped |
| Add map conditioning (ablation) | Scenario Diffusion · Argoverse 2 | MMD² Positions: 0.168 (0.012) -> 0.114 (0.006) | Helped |
| Increase agent-token test masking from 0% to 100% (ablation) | Scenario Diffusion with agent tokens · Argoverse 2 | MMD² Positions: 0.009 (0.0003) -> 0.133 (0.0096) | Hurt |
| Increase agent-token test masking from 0% to 100% (ablation) | Scenario Diffusion with agent tokens · Argoverse 2 | MMD² Heading: 0.011 (0.0005) -> 0.162 (0.0110) | Hurt |
| Increase autoencoder Resnet blocks from 1 to 3 (ablation) | Scenario Diffusion autoencoder · Internal Dataset | MMD² Positions: 0.075 (0.004) -> 0.071 (0.004) | No clear effect |
| Increase autoencoder latent dimension from 32 to 256 (ablation) | Scenario Diffusion autoencoder · Internal Dataset | Traj Off Drivable: 0.022 (0.002) -> 0.156 (0.051) | Hurt |
| Increase diffusion Resnet blocks from 1 to 3 (ablation) | Scenario Diffusion · Internal Dataset | MMD² Positions: 0.076 (0.009) -> 0.074 (0.004) | No clear effect |
| Train on the full dataset instead of one region (ablation) | Scenario Diffusion · Internal Dataset, SLAC scenes | MMD² Positions: 0.053 (0.002) -> 0.060 (0.009) | Mixed |
Closest papers by meaning
- Scenario Dreamer: Vectorized Latent Diffusion for Generating Driving Simulation Environments (CVPR 2025 · similarity 0.90)
- DiffScene: Diffusion-Based Safety-Critical Scenario Generation for Autonomous Vehicles (AAAI 2025 · similarity 0.89)
- Generating Traffic Scenarios via In-Context Learning to Learn Better Motion Planner (AAAI 2025 · similarity 0.88)
- Causal Composition Diffusion Model for Closed-loop Traffic Generation (CVPR 2025 · similarity 0.87)
- SceneDiffuser: Efficient and Controllable Driving Simulation Initialization and Rollout (NeurIPS 2024 · similarity 0.86)
- Future-Aware End-to-End Driving: Bidirectional Modeling of Trajectory Planning and Scene Evolution (NeurIPS 2025 · 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.