Research map / Reinforcement learning / Visual and representation RL
DexMove: Learning Tactile-Guided Non-Prehensile Manipulation with Dexterous Hands
Pei Lin, Yuzhe Huang, Wanlin Li, Chenxi Xiao, Ziyuan Jiao
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 finger control (full dexterous hand) instead of wrist-only (ablation) | DexMove · Ablation, Large Can | success rate (%): 0.0 -> 63.3 | Improved |
| add cross-finger attention in TaFo-Net (ablation) | TaFo-Net · Book | success rate (%): 63.3 -> 90.0 | Improved |
| add cross-finger attention in TaFo-Net (ablation) | TaFo-Net · Keyboard | success rate (%): 50.0 -> 90.0 | Improved |
| add shear force component to tactile vector field (ablation) | TaFo-Net · Large Can | success rate (%): 0.0 -> 63.3 | Improved |
| add shear force component to tactile vector field (ablation) | TaFo-Net · Lego | success rate (%): 70.0 -> 66.7 | Hurt |
| replace heuristic force with learned TaFo-Net force (ablation) | DexMove · Keyboard | success rate (%): 0.0 -> 90.0 | Improved |
| teleoperated wrist instead of policy wrist | Wrist-Only · Mouse | success rate (%): 0.0 -> 73.3 | Improved |
| unseen friction Fric. B vs Fric. A | DexMove · yaw error 0-30 | success rate (%): 86.7 -> 86.7 | No clear effect |
| unseen friction Fric. B vs Fric. A | DyWA · yaw error 0-30 | success rate (%): 50.0 -> 36.7 | Hurt |
| increase tactile noise sigma from 0 to 0.1 (ablation) | DexMove · Book | success rate (%): 90.0% -> 80.0% | Hurt |
| increase tactile noise sigma from 0 to 0.5 (ablation) | DexMove · Large can | success rate (%): 63.3% -> 3.3% | Hurt |
| multi-finger contact instead of gripper | DexMove vs DyWA · pushing 0-15 cm | execution time (s): 36.1 -> 8.3 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- DyWA — DyWA: Dynamics-adaptive World Action Model for Generalizable Non-prehensile Manipulation (ICCV 2025) · better on 9 of 9 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Touch in the Wild: Learning Fine-Grained Manipulation with a Portable Visuo-Tactile Gripper (NeurIPS 2025 · similarity 0.90)
- ManipTrans: Efficient Dexterous Bimanual Manipulation Transfer via Residual Learning (CVPR 2025 · similarity 0.88)
- Scaffolding Dexterous Manipulation with Vision-Language Models (NeurIPS 2025 · similarity 0.88)
- DexHandDiff: Interaction-aware Diffusion Planning for Adaptive Dexterous Manipulation (CVPR 2025 · similarity 0.88)
- EgoTac: In-the-wild Tactile Prediction from Egocentric Vision (NeurIPS 2026 · similarity 0.87)
- Enhancing Tactile-based Reinforcement Learning for Robotic Control (NeurIPS 2025 · 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.
