Research map / Vision-language and multimodal / Robot action and navigation
SD2Actor: Continuous State Decomposition via Diffusion Embeddings for Robotic Manipulation
Jiayi Li
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 learnable state embeddings (state representation) (ablation) | 3D Diffuser Actor · ARNOLD novel state set | success rate (%): 4.85 -> 5.73 | Improved |
| add learnable state embeddings (state representation) (ablation) | 3D Diffuser Actor · ARNOLD any state set | success rate (%): 21.52 -> 22.61 | Improved |
| add state decomposition and combination on top of embeddings (ablation) | SD2Actor · ARNOLD novel state set | success rate (%): 5.73 -> 10.84 | Improved |
| add state decomposition and combination on top of embeddings (ablation) | SD2Actor · ARNOLD any state set | success rate (%): 22.61 -> 25.19 | Improved |
| add orthogonalization loss to full method (ablation) | SD2Actor · ARNOLD novel state set | success rate (%): 10.84 -> 12.58 | Improved |
| add orthogonalization loss to full method (ablation) | SD2Actor · ARNOLD any state set | success rate (%): 25.19 -> 27.89 | Improved |
| add orthogonalization loss without decomposition (ablation) | SD2Actor · ARNOLD novel state set | success rate (%): 5.73 -> 6.26 | No clear effect |
| combine 2 embeddings instead of 1 (ablation) | SD2Actor · ARNOLD novel state set | success rate (%): 6.74 -> 12.58 | Improved |
| combine 3 embeddings instead of 2 (ablation) | SD2Actor · ARNOLD any state set | success rate (%): 27.89 -> 26.44 | Hurt |
| use 1 closest embedding instead of none (plain 3DA) (ablation) | SD2Actor · ARNOLD novel state set | success rate (%): 4.85 -> 6.74 | Improved |
| replace 3D Diffuser Actor with SD2Actor | 3D Diffuser Actor · ARNOLD test set | success rate (%): 42.21 -> 50.23 | Improved |
| replace 3D Diffuser Actor with SD2Actor | 3D Diffuser Actor · ARNOLD novel state set | success rate (%): 4.85 -> 12.58 | Improved |
Closest papers by meaning
- D²PPO: Diffusion Policy Policy Optimization with Dispersive Loss (AAAI 2026 · similarity 0.91)
- Robots Pre-train Robots: Manipulation-Centric Robotic Representation from Large-Scale Robot Datasets (ICLR 2025 · similarity 0.89)
- Zero-Shot Robotic Manipulation with Pre-Trained Image-Editing Diffusion Models (ICLR 2024 · similarity 0.89)
- SAM2Act: Integrating Visual Foundation Model with A Memory Architecture for Robotic Manipulation (ICML 2025 · similarity 0.89)
- Exploring the Limits of Vision-Language-Action Manipulation in Cross-task Generalization (NeurIPS 2025 · similarity 0.89)
- RDT-1B: a Diffusion Foundation Model for Bimanual Manipulation (ICLR 2025 · similarity 0.89)
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.
