Research map / Image generation / Fast diffusion sampling
Optimizing Diffusion Models for Joint Trajectory Prediction and Controllable Generation
Yixiao Wang, Chen Tang, Lingfeng Sun, Simone Rossi, Yichen Xie, Chensheng Peng, Thomas Hannagan, Stefano Sabatini, Nicola Poerio, Masayoshi TOMIZUKA, Wei Zhan
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 Optimal Gaussian Diffusion instead of vanilla diffusion with the same training time T=100 (ablation) | diffusion trajectory predictor · Argoverse 2 validation | avgMinFDE_128 vs reverse steps | Improved |
| use OGD with only 40 diffusion steps versus vanilla diffusion T=500 (ablation) | OGD · Argoverse 2 validation | avgMinFDE_128 | Improved |
| replace NNM and SF guidance with ECM guidance on OGD with 10 DDIM steps | controllable generation · route U + Deceleration | minJFDE: 0.538 -> 0.072 | Improved |
| add reference joint trajectories (ECMR) to ECM (ablation) | ECM guidance · route U + Deceleration | meanJFDE: 0.237 -> 0.146 | Improved |
| use deterministic instead of stochastic injected noise in ECM (ablation) | ECM guidance · GT + Acceleration | minJFDE: 0.444 -> 0.117 | Improved |
| use learned linear mapping instead of PCA latent (ablation) | OGD latent diffusion · Argoverse 2 validation | avgMinADE_128: 0.46 -> 0.43 (ADE); 0.80 -> 0.81 (FDE) | Mixed |
| increase the number of generated samples (ablation) | OGD · Argoverse 2 leaderboard | all metrics: saturates beyond 1024 samples | Improved |
Closest papers by meaning
- Diffusion-ES: Gradient-free Planning with Diffusion for Autonomous and Instruction-guided Driving (CVPR 2024 · similarity 0.89)
- Model-based Diffusion for Trajectory Optimization (NeurIPS 2024 · similarity 0.87)
- DiffusionDrive: Truncated Diffusion Model for End-to-End Autonomous Driving (CVPR 2025 · similarity 0.87)
- Causal Composition Diffusion Model for Closed-loop Traffic Generation (CVPR 2025 · similarity 0.86)
- Gradient Guidance for Diffusion Models: An Optimization Perspective (NeurIPS 2024 · similarity 0.86)
- Variational Control for Guidance in Diffusion Models (ICML 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.
