SJTU's UVTA Unifies Vision and Touch for Robotic Dexterity

SJTU and partners introduced UVTA, a unified visual-tactile-action model that achieves 70% success in page-turning and other contact-rich tasks, trained on 1000 human tactile demonstrations. The paper, code, and datasets are open-sourced.

2026-10-10 ~ 2026-10-10 · 3 related posts

• SJTU's UVTA teaches robot dexterous manipulation from 1,000 human tactile demos per task — siyuanhuang95 · 2026-10-10
• Robot hand turns book pages: UVTA visual-tactile-action model hits 70% vs 29% baseline — CyberRobooo · 2026-10-10
• UVTA paper, code and dataset open-sourced: shared visual-tactile-action representation from human demos — CyberRobooo · 2026-10-10