UHAS: Unified Hand Action Space Enables Cross-Embodiment Dexterous Manipulation
chris_j_paxton · x · 2026-08-14
RoboPapers Episode #97 discusses UHAS (Unified Hand Action Space), a sphere-based action representation that enables robots to learn dexterous manipulation skills across different hand embodiments. Presented by lfcasas7 and YuXiangIRVL, the research emphasizes the importance of action representation in cross-embodiment learning.
Related event: UHAS: Unified Hand Action Space Enables Skill Transfer Across Robot Hands(3 posts)→
More from Embodied
- FoveaVLA replaces wrist cameras with learned foveation, hitting 70.8% success — Haoyu_Xiong_ · 2026-10-16
- Musk teases two-way Grok x Optimus integration: Grok in the bot, bot in Grok — beffjezos · 2026-10-16
- SimVTA uses simulator-rendered video to post-train driving world action models with RL — OpenDriveLab · 2026-10-16
- Robotics data startups compared: Scale AI tops $1B revenue as 9 LLM vs robotics data suppliers benchmarked — FinanceYF5 · 2026-10-16
- Kroger cuts 3,000 corporate jobs (37%) for AI; humanoid robots next in the 2030s — csuwildcat · 2026-10-16
- Tesla dares its humanoid robot to do a kickflip in playful demo — yunta_tsai · 2026-10-16