Research map / Vision-language and multimodal / Robot action and navigation
Motion Dynamics Learning for Few-Shot Embodied Adaptation
Sibo He, Weiying Xie, Daixun Li, Junhao Zhong, Jiayun Tian, Yunke Wang, Leyuan Fang, Gang He, Yunsong 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 DCM (Dynamics-Constrained Modeling) (ablation) | DynVLA · 13 real-world tasks (Small data) | Average success rate (%): 58.5 -> 73.1 | Improved |
| add MDM contrastive grounding (ablation) | DynVLA · 13 real-world tasks (Small data) | Average success rate (%): 51.5 -> 73.1 | Improved |
| add MDM and DCM to the baseline (ablation) | Baseline policy · 13 real-world tasks (Small data) | Average success rate (%): 45.0 -> 73.1 | Improved |
| add DCM on long-horizon tasks (ablation) | DynVLA · 13 real-world tasks (Small data) | Long-horizon success rate (%): 48.3 -> 68.3 | Improved |
| increase codebook size from 512 to 1024 (ablation) | Dynamics Bank · 13 real-world tasks (Small data) | Average success rate (%): 66.5 -> 73.1 | Improved |
| increase codebook size from 1024 to 2048 (ablation) | Dynamics Bank · 13 real-world tasks (Small data) | Average success rate (%): 73.1 -> 72.3 | No clear effect |
| increase codebook size from 1024 to 4096 (ablation) | Dynamics Bank · 13 real-world tasks (Small data) | Average success rate (%): 73.1 -> 69.2 | Hurt |
| set constraint depth N to 6 instead of 2 (ablation) | DCM · 13 real-world tasks (Small data) | Average success rate (%): 46.5 -> 73.1 | Improved |
| set constraint depth N to 6 instead of 4 (ablation) | DCM · 13 real-world tasks (Small data) | Average success rate (%): 61.9 -> 73.1 | Improved |
| set constraint depth N to 6 instead of 8 (ablation) | DCM · 13 real-world tasks (Small data) | Average success rate (%): 55.8 -> 73.1 | Improved |
| set constraint depth N to 6 instead of 10 (ablation) | DCM · 13 real-world tasks (Small data) | Average success rate (%): 43.5 -> 73.1 | Improved |
| use DynVLA instead of GR00T-N1.5 with Small-scale data (about 20 demos) | few-shot VLA · 13 real-world tasks (Small data) | Average success rate (%): 57.3 -> 73.1 | Improved |
Closest papers by meaning
- PriorVLA: Prior-Preserving Adaptation for Vision-Language-Action Models (NeurIPS 2026 · similarity 0.91)
- Beyond World-Frame Action Heads: Motion-Centric Action Frames for Vision-Language-Action Models (NeurIPS 2026 · similarity 0.90)
- DiT4DiT: Jointly Modeling Video Dynamics and Actions for Generalizable Robot Control (NeurIPS 2026 · similarity 0.90)
- ElegantVLA: Learning When to Think for Efficient Vision-Language-Action Models (NeurIPS 2026 · similarity 0.90)
- RotVLA: Rotational Latent Action for Vision-Language-Action Model (NeurIPS 2026 · similarity 0.90)
- ActionUNet: Improving Robustness of VLA Models with Efficient Multi-scale Fine-tuning (NeurIPS 2026 · similarity 0.90)
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.
