NYU's Steertuning teaches robots chores from 5 human demos in under 5 minutes, 71% success
CyberRobooo · x · 2026-11-11
NYU's Lerrel Pinto group released Steertuning: a system that lets robots learn new tasks from just 5 human demonstrations in under 5 minutes, achieving 71% success across 12 tasks in 12 new environments including real home kitchens.
- Two-stage approach: (1) pretrain a 400M-parameter diffusion policy on 100 hours of densely annotated egocentric human interactions for robustness and generalization; (2) finetune for only 2 minutes (2,000 updates) and steer actions during reverse sampling using the demo closest to the deployment observation.
- Runs on a wheeled semi-humanoid (YOR mobile base + two Wuji dexterous hands). Dataset, model weights, and code are open.
Related event: NYU's Steertuning Teaches Robots New Tasks in 5 Minutes from 5 Demos(3 posts)→
More from Embodied
- Clean point clouds: Tersus GNSS MVP S1 scans turned into high-quality splats in LichtFeld Studio — janusch_patas · 2026-11-11
- Flashing cheap ESP32 IoT devices with ESPHome for a smart-home upgrade — haydendevs · 2026-11-11
- MessyMem on Compact Long-Term Memory for Agentic Robots Hits CoRL 2026 Main Poster — leto__jean · 2026-11-11
- NYU's Steertuning Lets a Robot Learn a Chore From 5 Minutes of Human Demo — CyberRobooo · 2026-11-11
- Boston Dynamics Atlas demoed with Blender 5.3 — TomLikesRobots · 2026-11-11
- Lamport Robotics posts robot 'wake-up test' demo video — R4nsack · 2026-11-11