AI
RoboCoach: World Models as Active Coaches for Compositional Robot Skills
Publication: Hugging Face Daily Papers
Author: Not available in collected metadata
Source published: 29/09/2026, 20:00:00 UTC
Briefing created: 01/10/2026, 09:23:25 UTC
Based on a feed excerpt; full article content has not been acquired. Source evidence retrieved 01/10/2026, 03:18:24 UTC.
Briefings
RoboCoach is a world-model-guided framework that uses imagined failures to guide robot skill improvement. It employs a Route-Imagine-Diagnose-Improve (RIDI) loop within COACHWORLD to select demonstrations and update experts. Results show that imagined and deployed success correlates over 22 task-policy pairs with a rho of 0.840, and coached experts transfer to four held-out compositions with 35.0% average success versus 0% for a shared-policy baseline.
Key points
- The ROBOCOACH framework uses a Route-Imagine-Diagnose-Improve (RIDI) loop to execute reusable skill experts inside COACHWORLD, a shared action-conditioned world model.
- Imagined failures guide demonstration requests and expert updates, with success rates rising from 13.3% to 75.0% on Franka and from 40.0% to 83.8% on AgileX using only 150 additional subtask demonstrations.
Context and limitations
Results are limited to two simulation suites and two real-robot platforms; success rates are observed over 22 task-policy pairs rather than across all possible compositions.
Generated by Bittrees-hosted qwen35-2b sha256:aaf42c8b7c3c on 01/10/2026, 09:23:25 UTC. Based on publisher-provided feed evidence; may not cover the full original article. Check the source for details.
View archived briefing on IPFS
bafybeia27odwurslv7jui6dfztul25reulb3h4driv63muujcit5oyjgny