RoboTryOn turns human videos into robot demos, lifting success rate from 2% to 60%
kaist-ai · hf · 2026-11-04
KAIST and RLWRLD's RoboTryOn treats human-to-robot video translation as virtual try-on: it swaps the human for a reference robot while preserving motion, scene and objects, then infers actions via an inverse dynamics model. Policies trained purely on synthetic data improve novel-task success from 2% to 60% (including bimanual manipulation), matching teleoperated-data performance at 1/3 the collection cost.
More from Embodied
- Baidu's Apollo Go wraps two-week public robotaxi test rides in Tung Chung, Hong Kong — Baidu_Inc · 2026-11-04
- imec neural probe integrates photonics directly into 130nm CMOS backend — jwt0625 · 2026-11-04
- Chris Paxton: interactive robot sim is the future but glaringly data-limited — Scobleizer · 2026-11-04
- DreamX: video diffusion models as universal latent action learners for robots — kaist-ai · 2026-11-04
- INMO Future glasses: dual 1080p displays, 125g, $799 early-bird on Kickstarter Dec 10 — Scobleizer · 2026-11-04
- DISCO distilled tree search named CoRL 2026 best-paper finalist, top 6 of 687 — yining_hong · 2026-11-04