Research map / Image generation / Autoregressive and latent generation
GraspLDP: Towards Generalizable Grasping Policy via Latent Diffusion
Enda Xiang, Haoxiang Ma, Xinzhu Ma, Zicheng Liu, Di Huang
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 graspness cue with auxiliary reconstruction (ablation) | GraspLDP · LIBERO grasping, Visual Generalization (sim) | success rate (%): 57.5 -> 64.6 | Improved |
| add graspness cue with auxiliary reconstruction (ablation) | GraspLDP · LIBERO grasping, In Domain (sim) | success rate (%): 77.4 -> 80.3 | Improved |
| add latent grasp-pose guidance instead of condition guidance (ablation) | GraspLDP · LIBERO grasping, In Domain (sim) | success rate (%): 73.5 -> 80.3 | Improved |
| add latent grasp-pose guidance instead of condition guidance (ablation) | GraspLDP · LIBERO grasping, In Domain (sim) | GFE: 1.58 -> 1.33 | Improved |
| add latent guidance (vs none) (ablation) | GraspLDP · LIBERO grasping, Visual Generalization (sim) | success rate (%): 19.4 -> 64.6 | Improved |
| add latent guidance (vs none) (ablation) | GraspLDP · LIBERO grasping, Object Generalization (sim) | success rate (%): 21.2 -> 58.2 | Improved |
| use Heuristic Pose Selector instead of highest-score selection (ablation) | GraspLDP · LIBERO grasping, In Domain (sim) | success rate (%): 72.0 -> 80.3 | Improved |
| use HPS instead of random grasp selection (ablation) | GraspLDP · LIBERO grasping, Visual Generalization (sim) | success rate (%): 54.5 -> 64.6 | Improved |
| use HPS instead of nearest selection (ablation) | GraspLDP · LIBERO grasping, In Domain (sim) | success rate (%): 69.5 -> 80.3 | Improved |
| graspness inference and latent decode overhead | GraspLDP vs Diffusion Policy · inference latency | latency: about 15% slower | Mixed |
| GraspLDP vs AnyGrasp on novel objects (real) | GraspLDP vs AnyGrasp · Real-world grasping, OG | success rate (%): 78.0 -> 75.0 | Hurt |
| GraspLDP vs diffusion policy on object generalization (sim) | GraspLDP vs Diffusion Policy · LIBERO grasping, Object Generalization (sim) | success rate (%): 11.4 -> 58.2 | Improved |
Closest papers by meaning
- Latent Diffusion Planning for Imitation Learning (ICML 2025 · similarity 0.91)
- AffordDP: Generalizable Diffusion Policy with Transferable Affordance (CVPR 2025 · similarity 0.89)
- Grasp2Grasp: Vision-Based Dexterous Grasp Translation via Schrödinger Bridges (NeurIPS 2025 · similarity 0.88)
- Scaffolding Dexterous Manipulation with Vision-Language Models (NeurIPS 2025 · similarity 0.88)
- Contact Map Transfer with Conditional Diffusion Model for Generalizable Dexterous Grasp Generation (NeurIPS 2025 · similarity 0.88)
- Cross-Embodiment Dexterous Grasping with Reinforcement Learning (ICLR 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.
