Research map / Image generation / Autoregressive and latent generation
Refining Diffusion Planner for Reliable Behavior Synthesis by Automatic Detection of Infeasible Plans
Kyowoon Lee, Seongun Kim, Jaesik Choi
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 |
|---|---|---|---|
| add restoration gap guidance (RGG) to Diffuser | Diffuser · Maze2D Large | score: 123.5 -> 135.4 | Improved |
| add attribution regularizer (RGG+) over RGG (ablation) | RGG · Maze2D Large | score: 135.4 -> 143.9 | Improved |
| add RGG to Diffuser | Diffuser · Multi2D Large | score: 141.2 -> 148.3 | Improved |
| add RGG to Diffuser | Diffuser · Maze2D U-Maze | score: 108.6 -> 108.8 | No clear effect |
| add RGG to Diffuser | Diffuser · Multi2D Medium | score: 130.1 -> 130.0 | No clear effect |
| add RGG to Diffuser | Diffuser · D4RL locomotion average | normalized return: 75.3 -> 81.2 | Improved |
| add RGG to Diffuser | Diffuser · Medium Hopper | normalized return: 58.5 -> 82.5 | Improved |
| add RGG to Diffuser | Diffuser · Med-Expert Walker2d | normalized return: 108.4 -> 107.8 | No clear effect |
| add RGG to Diffuser | Diffuser · Unconditional block stacking | score: 53.3 -> 63.3 | Improved |
| add RGG+ over RGG (ablation) | RGG · Conditional block stacking | score: 53.0 -> 56.7 | Improved |
| use Saliency instead of Grad-CAM attribution in RGG+ (ablation) | RGG+ · Maze2D Large | score: 143.9 -> 142.7 | No clear effect |
| use DeepLIFT instead of Grad-CAM attribution in RGG+ (ablation) | RGG+ · Multi2D Large | score: 150.9 -> 150.9 | No clear effect |
Closest papers by meaning
- Prior-Guided Diffusion Planning for Offline Reinforcement Learning (NeurIPS 2025 · similarity 0.88)
- DiRecT: Safe Diffusion-Based Planning via Receding-Horizon Denoising (NeurIPS 2026 · similarity 0.88)
- Diffusion Subgoal Planning for Long-Horizon Offline Goal-Conditioned Reinforcement Learning (NeurIPS 2026 · similarity 0.88)
- Task-Agnostic Pre-training and Task-Guided Fine-tuning for Versatile Diffusion Planner (ICML 2025 · similarity 0.88)
- SafeDiffuser: Safe Planning with Diffusion Probabilistic Models (ICLR 2025 · similarity 0.87)
- Simple Hierarchical Planning with Diffusion (ICLR 2024 · similarity 0.87)
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.
