Research map / Image generation / Autoregressive and latent generation
Optimization-Guided Diffusion for Interactive Scene Generation
Shihao Li, Naisheng Ye, Tianyu Li, Kashyap Chitta, Tuo An, Peng Su, Boyang Wang, Haiou Liu, Chen Lv, Hongyang Li
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 |
|---|---|---|---|
| apply OMEGA optimization-guided sampling to Nexus | Nexus · nuPlan free exploration | scene valid rate (%): 32.35 -> 72.27 | Improved |
| apply OMEGA to Nexus zero-shot on Waymo | Nexus · Waymo free exploration | scene valid rate (%): 61.71 -> 81.95 | Improved |
| apply OMEGA with goal conditioning | Nexus · nuPlan goal-conditioned | scene valid rate (%): 11.00 -> 80.00 | Improved |
| apply OMEGA to VBD | VBD · Waymo Open Sim Agents val | Realism Meta: 0.723 -> 0.730 | Improved |
| apply OMEGA to Nexus zero-shot | Nexus · Waymo Open Sim Agents val | Realism Meta: 0.609 -> 0.700 | Improved |
| add Rolling-Zero phase to Warm-up (ablation) | OMEGA · nuPlan free exploration | scene valid rate (%): 60.49 -> 72.27 | Improved |
| add Warm-up phase before Rolling-Zero (ablation) | OMEGA · nuPlan free exploration | scene valid rate (%): 0.00 -> 72.27 | Improved |
| add guidance objectives to two-phase noise schedule (ablation) | OMEGA · nuPlan free exploration | scene valid rate (%): 33.47 -> 72.27 | Improved |
| add phase-aligned guidance (ablation) | OMEGA · nuPlan free exploration | scene valid rate (%): 52.05 -> 72.27 | Improved |
| add phase-aligned guidance (inference time) (ablation) | OMEGA · nuPlan free exploration | inference time (s): 38.94 -> 4.96 | Improved |
| add KL trust-region constraint (ablation) | OMEGA · nuPlan free exploration | scene valid rate (%): 18.21 -> 72.27 | Improved |
| replace CTG guidance with OMEGA | Nexus · nuPlan free exploration | scene valid rate (%): 50.41 -> 72.27 | Improved |
Closest papers by meaning
- SceneDiffuser: Efficient and Controllable Driving Simulation Initialization and Rollout (NeurIPS 2024 · similarity 0.90)
- Scenario Dreamer: Vectorized Latent Diffusion for Generating Driving Simulation Environments (CVPR 2025 · similarity 0.90)
- DiffusionDrive: Truncated Diffusion Model for End-to-End Autonomous Driving (CVPR 2025 · similarity 0.88)
- Diffusion-ES: Gradient-free Planning with Diffusion for Autonomous and Instruction-guided Driving (CVPR 2024 · similarity 0.88)
- X-Scene: Large-Scale Driving Scene Generation with High Fidelity and Flexible Controllability (NeurIPS 2025 · similarity 0.88)
- Causal Composition Diffusion Model for Closed-loop Traffic Generation (CVPR 2025 · similarity 0.88)
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.
